AUTHOR Proceedings of the NECC/5 National Educational REPORT NO

DOCUMENT RESUME

.ED 235 779 IR 010 852

AUTHOR Bonnette, Della, Ed. TITLE Proceedings of the NECC/5 National Educational Computing Conference 1983 (5th, Baltimore, Maryland, June 6-8, 1983). REPORT NO ISBN-0-8186-0050-0 PUB DATE Jun 83 NOTE 422p.; Published by the IEEE Computer Society Press. AVAILABLE FROMProfessor Ted Sjoerdsma, University of Iowa, Computer Science Division, Iowa City, Iowa 52242 ($15.00). PUB TYPE Collected Works - Conference Proceedings (021) -- Viewpoints (120) -- Reports - Descriptive (141) EDRS PRICE MF01 Plus Postage. PC Not Available from EDRS. DESCRIPTORS *Computer Assisted Instruction; *Computer Literacy; Computer Managed Instruction; *Computer Programs; *Computer Science Education; Instructional Materials; Material Development; Programing; *Programing Languages; Science Education; Teacher Education IDENTIFIERS *Computer Uses in Education; LOGO Programing Language ABSTRACT Recent research and current trends in the field of computers and education are reflected in this collection of reviewed papers, tutorials, panels, project presentations and other sessions. More than 80 papers are grouped in the following topic areas: administrative applications, composition and literature, computing for the learning disabled or handicapped, computer services, computer science curricula, computers in education, computing in the non-curricular support role, pre-college computer science, computer science software, LOGO, alternative approaches to providing computing facilities, computer uses in education, science, computer literacy, computer education for elementary school teachers, commerce, computer science -- teaching programming, computers in science education, computer assisted instruction, computers in education at an early education level, computer education for secondary school teachers, mathematical needs of computer sciences, computer-based education,. teacher training, pre-college instructional uses of computers, mathematics and statistics, courseware development and evaluation, and pre-college computer services. An additional 39 papers are included under the headings of tutorials, invited sessions, and special sessions. (LMM)

********************************************************************* Reproductions supplied by EDRS are the best that can be made from the original document. *********************************************************************** U.S. DEPARTMENT OF EDUCATION NATIONAL INSTITUTE OF EDUCATION EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC) 111, This document has been reproduced as received from the person or organization migaiMingin Minor changes have been made to improve reproduction quality.

Points of view or opinions stated in this document do not necessarily represent official NIE position, or policy. Proceedings of NECC/5 National Educational Computing Conference 1983

CONFERENCE: June 6-8, 1983, Baltimore, Maryland ISBN 0-8186-0050-0 EEE CATALOG NO. 83CH1888-7 LIBRARY OF CONGRESS M3.83-60814 HOST: Towson State University, Baltimore, Maryland EEE COMPUTER SOCETY NO. 490

COMPUTER SOCMETY PRESSIrs

EDITED BY Della Bonnette "PERMISSION TO REPRODUCE THIS MATERIAL INMICROFICHE ONLY University of Southwestern Louisiana HAS BEEN GRANTED BY Lafayette, Louisiana Ted Sjoerdsma

TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC)." The papers appearing in this book comprise the proceedings of the meeting mentioned on the cover and title page. They reflect the authors' opinions and are published as presentedand without change, in the interests of timely dissemination. Their inclusion in this publication does not necessarily constitute endorsement by the editors. IEEE Computer Society Press, or the Institute of Electrical and Electronics Engineers, Inc.

Published by IEEE Computer Society Press 1109 Spring Street Suite 300 Silver Spring, MD 20910

ISBN 0-8186-0050-0 (Paper) ISBN 0-8186-0051-9 (Casebound) ISBN 0-8186-0052-7 (Microfiche)

National Educational Computing Conference June 1983

Cover designed by Madeline Windauer NECC STEERING COMMITTEE

Ronald Anderson Doris Lidtke University of Minnesota Towson State University

Alfred Bork James Lubkin University of California, Irvine Michigan State University Nell Dale Mike Mulder University of Texas at Austin Servio Logic Corporation

Karen Duncan Richard Pogue Health Information Systems Medical College of Georaia

Francis Edwards Joseph Raben Towson State University Queens College/CUNY

Gerald L. Engel Nancy Roberts Christopher Newport College Lesley College

Mary Dee Harris Fosberg Alan L. Roecks Loyola University San Antonio ESC

John Hamblen Jean Rogers University of Missouri-Rolla University of Oregon

Diana Harris William Ryan University of Iowa Swarthmore College

Harry Hedges Ted Sjoerdsma Michigan Sate University University of Iowa

Paul Heller Dennis Speck EDUCOM/EDUNET University of Houston

Lawrence Jehn E.M. Stanan University of Dayton University of Missouri-Columbia Sister Mary Kenneth Keller David Stonehill Clarke College University of Rochester

Jesse C. Lewis Joe Turner Jackson State University Clemson University

iii NECC Conference Committee

General Chairperson Doris K. Lidtke Towson State University

Vice-Chairperson Robert Caret Towson State University

Program Committee Joseph Turner Clemson University

Contributed Papers Joseph Turner Clemson University Society Sessions and Panels Jean Rogers University of Oregon

Project Presentations William Ryan Swarthmore College

Special Sessions William Dorn University of Denver

Films Lillian Cassel Goldey Beacom College

Birds-of-a-Feather David Stonehill University of Rochester

Session Chairpersons Patricia Powers Goucher College

Workshops Ralph Lee University of Missouri-Rolla

Proceedings Della Bonnette University of Southwestern Louisiana

Publicity James Adams Association for Computing Machinery

Exhibits Gerald Leach-Lewis IEEE/Computer Society

Evaluation of NECC/82 Alan Roecks Education Service Center-San Antonio

Mailings Carol Edwards Towson State University

Local Arrangements Francis Edwards Towson State University

Co-chairpersons Michael Haney Towson State University Charles Parrish Towson State University

Information Desk Clarence Miller Maryland State Department of Education

Continuing Education Uoits Robert Wall Towson State University

Social Events Joyce Currie Little Towson State University Dick Austing University of Maryland

Restaurants Ann Wagner Towson State University

Student Assistants TSU ACM Student Chapter and others Press Relations Gerald Riggleman Towson State University

iv EXHIBITORS

ADDISON-WESLEY PUBL CO, READING, MASSACHUSETTS ASSOCIATION FOR COMPUTING MACHINERY, NEW YORK ANAHEIM PUBLISHING GO, FULLERTON, CALIFORNIA APPLE COMPUTER, CUPERTINO, CALIFORNIA Ase:DS, WLSRINGTON, D.C. ATARINC, SUNNYVALE, CALIFORNIA POBBS-hprinra EDUC PUBL CO, INDIANAPOLIS, INDIANA. BOYD & FgLSER PUBLISHING CO, SAN FRANCISCO, CA C & C SOFTWARE, WICHITA, KANSAS CHARLES E.MERRILL PUBLISHING, COLUMBUS, OHIO CLASSROOM COMPUTER NEWS COMMODORE BUSINESS MACHINES, WAYNE, PENNSYLVANIA COMPUTER SCIENCE PRESS, ROCKVILLE, MARYLAND COMPRESS-SCIENCE BOOKS INT'L, WENTWORTH, N.H. CONDUIT(UNIV OF IOWA), IOWA CITY, IOWA THE CONTINENTAL PRESS, ELIZABETHTOWN, PA CORONADO PUBLISHERS, NEW YORK CITY DEVELOPMENTAL LEARNING MATERIALS, ALLEN, TEXAS DIGITAL EQUIPMENT CORPORATION, MAYNARD, MASS DILITHIUM PRESS, BEAVERTON, OREGON EDUCATIONAL TESTING SERVICE, PRINCETON, N.J. EDUCOM/EDUNET, PRINCETON, NEW JERSEY EDWARD ARNOLD, BALTIMORE, MARYLAND ENTELEK, PORTSMOUTH, NEW HAMPSHIRE EPIE -EDUCATIONAL PRODUCTS, NEW YORK CITY, N.Y. FOLLETT LIBRARY BOOK CO, CRYSTAL LAKE, ILLINOIS GAMCO INDUSTRIES,tNC, BIG SPRING, TEXAS GLOUCESTER COMPUTER,INC, GLOUCESTER, MASS GOUCHER COLLEGE, TOWSON, MARYLAND GREGG/McGRAW-HILL, NEW YORK CITY, N.Y. HARPER & ROW PUBLISHERS,INC, NEW YORK CITY, N.Y. HARTLEY COURSEWARE,INC, DIMONDALE, MICHIGAN HOLT,RINEHART & WINSTON, NEW YORK CITY, N.Y. HOUGHTON MIFFLIN CO, BOSTON, MASSACHUSETTS IBM CORPORATION, BETHESDA, MARYLAND INFORMATION SYNERGY,TNC, PRINCETON, NEW JERSEY J.L.HAMMETT COMPANY, BRAINTREE, MASSACHUSETTS JOURNAL OF COMPUTERS IN MATHS & SCIENCE, K-12 MICROMEDIA, WOODCLIFF LAKE, N.J. LAWRENCEVILLE PRESS, LAWRENCEVILLE, N.J. LITTLE, BROWN & CO, BOSTON, MASSACHUSETTS BROOKS/COLE PUBLISHING CO, MONTEREY, CALIFORNIA McGRAW -HILL PUBLISHING CO, NEW YORK CITY MILLIKEN PUBLISHING CO, ST.LOUIS, MISSOURI MONROE SYSTEMS FOR BUSINESS, MORRIS PLAINS, N.J. RADIO SHACK, FORT WORTH, TEXAS RANDOM HOUSE, NEW YORK CITY RESTON PUBLISHING COMPANY, RESTON, VIRGINIA SCOTT INSTRUMENTS, DENTON. TEXAS SCHOLASTIC,INC, NEW YORK CITY, N.Y. SPRINGER-VERLAG,N.Y. NEW YORK CITY STERLING SWIFT PUBLISHING, AUSTIN, TEXAS SUNBURST COMMUNICATIONS, PLEASANTVILLE, N.Y. SYNTAURI CORPORATION, PALO ALTO, CALIFORNIA SYSTEMS DESIGN ASSOCIATES, CHARLESTON, W.VA TECHNICO,INC, BALTIMORE, MARYLAND TERRAPIN,INC CAMBRIDGE, MASSACHUSETTS WADSWORTH PUBLISHING CO, BELMONT, CALIFORNIA WADSWORTH PUBLISHING CO (INT'L DIVISION) WEST PUBLISHING CO, SANTA CLARA, CALIFORNIA WEBSTER/McGRAW -HILL, NEW YORK CITY, N.Y. JOHN WILEY & SONS, INC, NEW YORK CITY, N.Y. QUEUE,INC, FAIRFIELD, CONNECTICUT ZERO PAGE,INC, COLORADO SPRINGS, COLORADO

V COOPERATIN' SOCIETIES

The Conference is hosted by Towson State University in cooperation with:

American Association for Medical Systems and Informatics

Association for Computing Machinery Special Interest Groups on: Computers and Society (SIGCAS) Computer Science Education (SIGCSE) Computer Uses in Education (SIGCUE) University and College Computing Services (SIGUCCS)

Association for Education Data Systems (AEDS) American Educational Research Association/Special Interest Group on: Computer Applications in Instruction (AERA/SIGCAI)

ARIPS Education Committee Association for Computers and the Humanities (ACH) Association for Small Computer Users in Education (ASCUE) American Society for Engineering Education/Computers in Education Division (ASEE/CoED) Conference on Computers in Undergraduate Curricula (CCUC) Educational Computing in Minority Institutions (ECMI)

Health Education Network (HEN) International Council for Computers in Education (ICCE)

EDUNET/EDUCOM IEEE Computer Society Society for Computer Simulation

vi FOREWORD

This volume of proceedings of the Fifth National Educational Computing Conference (NECC/83), accurately reflects recent research and current trends in the field of computers and education. It embodies the critical thinking of a vast number of experts on topics that are both crucial to the whole society and especially relevant at the present time. The conference shows an excellent balance of reviewed papers, tutorials, panels, project presentations, andother sessions covering the broad spectrumof computers in education. The ideas expressed in these proceedings and in the sessions are a manifestation of the vitality of this field and provide attendees an opportunity to expand their expertise and increase their appreciation of computers in education. We believe that this conference will be beneficial to all participants and that these proceedings will serve as a valuable reference in the future. The conference and these proceedings are the culmination of a greatdeal of effort by many individuals. Particular thanks are due to

- the National Educational Computing Conference Steering Committee for guidance and support, especially those who advised and encouraged the conference committee; - all authors who submitted papers for review; - the referees for their considerp.bleefforts in reviewing the papers and for making the frequently difficult iie.:isions of whether to accept or reject papers; - the organizer of panel and tutorial sessions; - A. J. (Joe) Turner (Clemson University), who so ably chaired the Program Committee and had the awesome task of coordinating the review of papers; - Jean Rogers (Universityof Oregon), who with diligence and skill coordinated society sessions, tutorials and suggested sessions; - William Ryan (Swarthmore College), who organized the project presentation sessions; - WilliamDorn (University of Denver), who withdiscriminating sense (flair) organized the invited sessions; - James Adams (Association for Computing Machinery), who with energy and imagination handled the publicity for the conference; Gerald Leach-Lewis (IEEE Computer Society), who worked creatively to expand the quantity and quality of the exhibits; - Alan L. Roecks (San Antonio, ESC), whose superb evaluation of NECC/82 gave us excellent ideas for this years conference; Ralph Lee (Universityof Missouri, Rolla), who organized a splendid array of pre-conference workshops; Francis Edwards (Towson State University), who worked effectivelyon the broad range of local arrangement tasks; - David Stonehill (University of Rochester), who organized the Birds of a Feather sessions; - Robert Caret (Towson State University), who was always willing to assist and support the activities of the conference; - Carol Edwards (Towson State University), who with good humor coordinated the processing of nearly 100,000 pieces of mail; Iva Thommen (Towson State University), who handled all secretarial tasks cheerfully and expeditiously; - Donna Feldmann (Towson State University), who as the student assistant for NECC intuitively saw what needed to be done and efficiently did it; - All attendees who made the efforts worthwhile; - Della Bonnette (University of Southwestern Louisiana), who made this volume of proceedings possible, through her skills as an editor, her patience in dealing with the authors, and her ability to accomplish it all on schedule.

Doris K. Lidtke General Chairperson, NECC/83 Towson State University Baltimore, 'Maryland 21204

vi i TABLE OF CONTENTS

INVITED SESSION INVITED LESSION 1 The Role of Language in Teaching Programming Stephen Garland, Chair

SPECIAL SESSION SPECIAL SESSION 2 Approaches to Requiring Microcomputers of Undergraduate Students Jane Caviness, Chair 3 Accreditation in the Computing Sciences John Dalphin, Chair

ADMINISTRATIVE APPLICATIONS PAPER SESSION 4 Networking for Microcomputer Management Kenneth Forman, Carl Steinhoff Instruction in Resource 7 Spread Sheet SimulationModeling (SSSM) for Training and Allocation Ronald Lindahl, Brent Wholeben 12 Development and Validation of Computerized Adaptive Screening Test(CAST) for use in Army Recruiting Herbert Baker, Bernard Rafacz, William Sands

COMPOSITION AND LITERATURE PAPER SESSION 18 Word Processing in the Classroom Karen Piper 22 The Computer in the Writing Class: Problems and Potential C. Daiute, P. O'Brien, A. Shields, S. Liff, P Wright, S. Mazur, W. Jawitz 27 A Hybrid Humanities Application Course Rudy Spraycar

TUTORIAL TUTORIAL 31 The DISC Project Shelley Rose, Carol Klenow COMPUTING FOR THE LEARNING DISABLED OR HANDICAPPED PROJECT SESSION 32 Using LOGO with` Learning Disabled Students Rita Horan 32 Project CAISH Second Year Update Warren Brown 33 Project S.O.S. Mary Russo, Nancy Jones Instruction on the 33 Relative Effect of Microcomputer Instruction an:; Teacher Directed Performance of Hearing Impaired and Normal Hearing Students Sharon Smaldino, Patrick Schloss

COMPUTER SERVICES PAPER SESSION 35 A Guide for the Purchase of Computer Systems for a Two-YearCampus Laurena Burk 42 Extensive Computer Grading of ID-Individualized Homework Problems M. J. Maron 48 Automatic Syllabus Generator (ASG) Asad Khailany, Marc Schubiner, A.M. VanderMolen

SPECIAL SESSION SPECIAL SESSION 54 How Schools Use Microcomputers: Findings from the Johns Hopkins University National Survey of Computer-Using Teachers Clarence Miller, Chair 55 CAI in Foreign Language Instruction Carl Adamson, Chair

viii 9 COMPUTER SCIENCE CURRICULA PAPER SESSION 56 What Computer Curriculum is Right for the Small College William Mitchell 64 A New Source of Computer Science Teachers: Faculty Members from Other nepartments Keith Harrow 68 Hobby Robots as Teaching/Learning Tools Michael Moshell, Charles Hughes, Carl Gregory, Lee Wittenberg COMPUTERS IN EDUCATION PAPER SESSION 75 Ending the Isolation: Deaf-Blind and Microcomputers Dan Zuckerm'n 80 Plato Stayweli: A Microcomputer-Based Program of Health BehaviorChanges that Improves With Use Murray Naditch 85 The Neuroscience Software Project Terry Mikiten, Ronald Pyka COMPUTING IN THE NON-CURRICULAR SUPPORT ROLE PROJECT SESSION 90 A Microcomputer Based Vocational Placement add Follow-up System Spicer Bell, Alonzo Peters 90 Individualized Grade Reports: Motivational Aid and Teaching Tool Linda Royster 90 Using a Microcomputer for a Test Question Storage Bank Robert Jackson 91 How Easy to Use Can a Grade Management Program Be? Richard Cornelius 91 An Analysis of Academic Grades at the US Naval Academy Randall Spoeri, Malcolm Fordham PRE-COLLEGE COMPUTER SCIENCE PAPER SESSION 92 Experimenting with a Computer Literacy Program for Elementary School Gifted and Talented Students W. Starnes, J. Muntner 99 Introductory Computer Programming for All College Bound High School Students Ken Jones, Dennis Simms 103 A Programming Environment for Preliterate Children Charles Hughes, Michael Moshell SPECIAL SESSION SPECIAL SESSION 107 Teacher Training in Computer Education William Wagner, Chair 108 Instituting Computer Programs within a School District John Cheyer, Chair 109 Voice Input/Output: New Directions in Instructional Technologies Carin Horn, Chair 110 Educational Use of Microcomputers by Special Needs Students Joan Davies, Chair 111 Needs and Opportunities for Educational Software in Grades K-12 Edward Esty, Chair COMPUTER SCIENCE - SOFTWARE PAPER SESSION 112 Program Maintenance ... The Forgotten Topic Frank Connelly 115 An Environment to Develop and Validate Program Complexity Measures Enrique Oviedo, Anthony Ralston 122 Teaching a Software Engineering Class Using an IBM Personal Computer(tm) Ronald Frank PRE-COLLEGE COMPUTER SCIENCE PAPER SESSION 126 Crisis in Programming or History Does Repeat Itself Jacques LaFrance 132 An Evaluation of a LOGO Training Program M. Elizabeth Badger 138 Educational Computing Post Haste: A Case Study Deborah Blank

ix LOGO PROJECT SESSION 141 LOGO A Three Year Sequence, Grades 4-5-; Carolyn Markuson, Joyce Tobias 141 Development of a Program Designed to Use LOGO and a Floor Turtle in a Nursery School Environment Martin Saltz, James Gottlieb, Bobbie Gibson, Roy Moxley 141 LOGO Instructional Development Project S.Tipps, H.Evans, G.Bull, T.Schwartz, M.King, S.Taylor, S. Walker, P.Davidson 142 The Programming Styles of Fifth Graders in LOGO Leah Rampy, Rochelle Swensson 142 Modifying Papert's Vision: LOGO Lessons Barbara Hilberg ALTERNATIVE APPROACHES TO PROVIDING COMPUTING FACILITIES PROJECT SESSION 143 CompuShare: A School-Community Project Mary Sennett 143 The Central Illinois Computing Consortium Richard Murdach 143 A Relocatable Computer Laboratory Pat Kelly 144 CALL: A Multipurpose Educational Computer Facility Richard Evans 144 Cost Effective Implementation of a Microcomputer Program in Elementary School Mary DeBoer

INVITED SESSION INVITED SESSION 146 Distance Teaching of Software Engineering Darrel Ince, W. S. Matheson

SPECIAL SESSION SPECIAL SESSION 147 District Planning For Computer Use In K-11 Glenn Fisher, Chair 148 InformationTechnology and Its Impact on the United States - Overview and Implications Linda Garcia, Chair

COMPUTER USES IN EDUCATION PAPER SESSION 149 The Electronic Blackboard using a Microcomputer and Large-ScreenTelevision as a Lecture Aid James Clark 152 Results and Lessons From a Survey of Readers' Controlof Rate of Text Presentation on Computer Screens Werner Feibel 157 An Experimental Comparison of. Discovery and Didactic Computerized Instructional Strategies in the Learning of Computer Programming Brian McLaughlin

SCIENCE PAPER SESSION 163 Checking Lab Calculations William Pelham 167 TeachingUndergraduates toTheorize Through the Useof a Computer Simulation of Kidney Function David Wilcox 174 Microcomputer-Based Data Acquisition for Neurobiology Richard Olivo COMPUTER LITERACY PROJECT SESSION 180 Algebra, Basic, and Computers: The ABC's for Non-Science Majors Margaret Christensen 180 Computer Literacy in the Two-Year College Curriculum Carla Thompson, Joyce Friske 180 The Vassar College Computer Literacy Program William Pritchard, Donald Spicer 181 A Microcomputer Literacy Program Ronald Bearwald 181 Machine Language in Computer Literacy: Strategy and Supporting Software David Lewis

x 1 COMPUTER EDUCATION FOR ELEMENTARY SCHOOL TEACHERS PROJECT SESSION 183 Computer Literacy for Elementary and Middle SchoolTeachers Joyce Currie Little, Robert Wall 183 Microcomputer Simulation: An Aid in Training Elementary School Teachers Harold Strang, Ann Loper 184 Toward Curriculum Development: A Case Study in ComputerIn-Service Training Alice Ann Winner Program for 184 Incorporating the Microcomputer into the Department of Mathematics Prospective Elementary School Teachers Muriel Wright, Helen Coulson

COMPUTERS IN EDUCATION PROJECT SESSION 186 Real-Time Microcomputer Programs for Teaching Statistics C. Michael Levy 186 High Schdol Science Microcomputer Project John Pancella, John Entwistle, Carol Muscara 187 The Function Game: Using Microcomputers to Improve. Grading Skills Edward Zeidman 188 Computer Chronicon Project Melvin Wolf

INVITED SESSION INVITED SESSION 189 Where We Are Going in the Use of Computers in PublicEducation Sylvia Charp

SPECIAL SESSION SPECIAL SESSION 190 Computers in the Undergraduate Mathematics Curriculum Sheldon P. Gordon, Chair 191 Simulation: A Teaching Strategy X.-College Beverly Hunter, Chair Education - Why? 192 Considering the Lack of Instructional Computing in Higher Lincoln Fletcher, Chair

TUTORIAL TUTORIAL 193 The Funding Game: Playing to Win John T. Thompson PAPER SESSION COMMERCE 194 Designing a Programming Course for MBA Students David Cossey, David Rossien 200 A Curriculum for a Master's Program inComputerized Materials Management Daniel Shimsak, Dean Saluti 204 Information Literacy Course: A Recommended Approach Eileen Trauth COMPUTER SCIENCE - TEACHING PROGRAMMING PAPER SESSION 208 A System for the Automatic Grading of ProgrammingStyle Patricia Van Verth, Anthony Ralston 214 Teach Top-Down Programming While You Teach BASIC Michael Streibel 220 Using Computer Simulated Models to Teach Programming Languages Bogdan Czejdo

COMPUTERS IN SCIENCE EDUCATION PROJECT SESSION 224 The Use of an Apple/Corvus Networking Systemin an Elementary Physics Course Raymond Bigliani 224 Program Development by a Biology User's Group forMicrocomputer-Assisted Instruction L. Dove, S. Bryant, H. Edwards, K. Kendell, P. Nielsen,G. White 225 A Scientific Instrument Trainer Robert Henkins 225 Concentrated Physics Concepts: A Comprehensive Package of Tutorial Problem Solving David Alexander.

INVITED SESSION INVITED SESSION 226 Courseware Development from a Publisher's Perspective M. D. Roblyer, Chair

xi SPECIAL SESSION SPECIAL SESSION 227 Trends in Interactive Data An?lysis Jon Christopherson 229 Science Education and the Growth of the U. S. Computer Industry Dorothy Derringer, Chair 230 Computing Curricula Prepared by the Professional Societies Joyce Currie Little, Chair

COMPUTER SCIENCE - TEACHING PROGRAMMING PAPER SESSION 231 Augmenting Self-Study Materials with Microcomputer-Based Lessons Ernest Giangrande, William Sregar 239 Bridging from Non Programmers to Programming Jeffrey Sonar, Elliot Soloway 244 Predicting Student Successs in an Introductory Programming Course Terry Hostetler

CAI PROJECT SESSION 249 Computer-Assisted Sentence Combining Michael Southwell, Carolyn Kirkpatrick, Mary Epes 249 Writing Computer-Assisted Instructional Programs to Support a Textbook J. Kenneth Sieben 250 Project Better Chance; A Comprehensive Approach to Basic Skills Improvement Ellen Leahy 250 Appropriate Technology for Computer Education R. K. Wiersba

COMPUTERS IN EDUCATION AT AN EARLY EDUCATION LEVEL PROJECT SESSION 252 The Magic Crayon Carol L. Clark 252 Effectiveness of Computer Usage on Achievement of Specific ReadinessSkills of Preschoolers Elizabeth Legenhausen 253 The Oak Street Interns: An Experiment Stewart Denenberg 253 Wily Computer Education in the Elementary School? A Model for Maximum Use Marilyn Pollock COMPUTER EDUCATION FOR SECONDARY SCHOOL TEACHERS PROJECT SESSION 254 Infusion of Microcomputer Training into the Existing School of Education Undergraduate and Graduate Curriculum Susan Zgliczynski 255 Certification of High School Computer Science Teachers Harriet Taylor 255 Introduction of Computers and Educational Computing - A CAI Approach Dale Johnson, Carla Thompson 255 planning and Training for Effective Use of Computers Sandra Crowther, Linda Hyler, Michel Eltschinger

MATHEMATICAL NEEDS OF COMPUTER SCIENTISTS INVITED SESSION 256 An Overview of the Mathematical Needs of Computer Scientists Anthony Ralston 258 Mathematics in Computer Science and the Applications Programmer A. T. Berztiss 261 MatSematics Service Courses for the Computer Science Student Martha Siegel 263 Stirrings in the Mathematics Curriculum: Changes Mathematicians are Thinking of Making Stephen Maurer

SPECIAL SESSION SPECIAL SESSION 266 Using a Large Screen Computer' System to Improve Teaching David Lundstrom 267 Educational Software Copyright Issues Ronald Anderson, Chair 268 Teaching Structured Programming in the Secondary School Jean Rogers, Chair 270 Nationwide Computer Literacy Project Daniel Updegrove, Steveb Gilbert TUTORIAL TUTORIAL 272 Using the Microcomputer Creatively with Young Students Marilyn Church, June Wright

COMPUTER-BASED EDUCATION PAPER SESSION 273 Huntington III:. Microcomputer Courseware Development Project Thomas Liao 279 A Universal Compute:: Aided Instruction System Henry Dietz, Ronald Juels 283 A Study of Student-Computer Interactivity David Trowbridge, Robin Durnin

TEACHER TRAINING PAPER SESSION 290 The Implementation of Technology and the Concerns-Based Adoption Model Cheryl Anderson 294 Elementary Teacher Education: Including LOGO in Teaching Informal Geometry M. Moore, W. Burger 298 A Computer Literacy Curriculum for Preservice Teacher Education Candidates Brent Wholeben

PRE-COLLEGE INSTRUCTIONAL USE OF COMPUTERS PAPER SESSION 302 Dynamics of Learning and Mis-Learning in a Simulated Micro-World Andrea Petitto, James Levin 308 Observation and Inference - A Computer Based Learning Module Alfred Bork, David Trowbridge, Arnold Arons 311 Does Use of Microcomputers in Junior High School Increase Problem Solving Skills? Barbara Kurshan, Joyce Williams, Nancy Healy

INVITED SESSION INVITED SESSION 316 Divergent Answers to the Question, "Where Should Computer EducationDollars Be Spent?" Arthur Luehrmann, Eric Burtis, Beverly Hunter

SPECIAL SESSION SPECIAL SESSION 317 An Evolving Model for Providing Computer Education for Gifted Children Mary Crist, Chair 318 Training University Faculty in the Use of Computer Graphics Richard McGinnis 320 Recommendations for Programs in Computing at Small Colleges John Beidler, Chair

COMPUTERS IN EDUCATION PAPER SESSION 321 Computers and Quantitative Methods; Healthy for the Humanities? Rudy Spraycar 326 A Personal Computer for Every College Student David Bray 330 Computer Assisted Simulation in Politics of Reapportionment/Redistricting (CASPOP) Jerry Bolick, James 0. Icenhour

MATHEMATICS AND STATISTICS PAPER SESSION 336 Integrating Computing Packages and Statistics Instruction William Schafer, C. Mitchell Dayton 342 A Computer Based Tutorial on Mathematical Induction J. Mack Adams, Marvin Landis 345 Implicit Functions and Computer Graphics Sheldon Gordon COMPUTER SCIENCE - MISCELLANEOUS PAPER SESSION 350 Interrupt Drive I/O Projects in an ACM '78 CS4 Course Greg Starling 355 Assembly Language on the APPLE: A Thorough Introduction W. D. Maurer 360 Student-Down System Design Robert Geist COURSEWARE DEVELOPMEMNT AND EVALUATION PROJECT SESSION 364 Computer Literacy and the Liberal Arts L. Carl Leinbach 364 Courseware Evaluation Techniques Barbara C. Garris 365 The California Courseware Clearinghouse Ann Lathrop 365 Let's Write Usable Courseware: The City College Algebra Project Jon C. Miller SPECIAL SESSION SPECIAL SESSION 366 Request for Equipment Proposals Joseph Wolfsheimer 367 Courseware on Social Issues of Computers Ronald Anderson, Chair 368 Word Processors in the Composition Classroom Mary Dee Harris Fosberg, David Ross, Chairs 369 Interactive Computer Graphics and Computer Animated Films in Education Maria Mezzina, Chair 370 Teaching Ada Via Computer George Poonen, Chair 371 Electronic Main and Computer Conferencing Paul Heller, Chair COMPUTERS IN EDUCATION PAPER SESSION 372 Sex Difference in Microcomputer Literacy Marlaine Lockheed, Antonia 'lielsen, Meridith Stone 377 Computers: Less Apprehension, more Enthusiasm Janet Parker, Constance Widu...r 381 The Microcomputer as a Tool in Educational Research: A Case in Point Scott Brown, Daniel Kaye

PRE-COLLEGE COMPUTER SERVICES PAPER SESSION 385 Strategic Concerns in Establishing and Elementary School Microcomputer Instructional System Ronald Bearwald, Theodore Bargmann 391 Evaluation of Microcomputer Software:How Valid are the Criteria Procedures? Robert Caldwell 394 Micro-Networking - Some Practical Applications David Rieger SPECIAL SESSION SPECIAL SESSION 403 Computers in the Elementary and Secondary Mathematics Curriculum Sheldon P. Gordon, Chair

xiv THE ROLE OF LMGUAGE IN TEACHING PROGRAMMING

Stephen J. Garland Dept. of Mathematics and Computer Science Dartmouth College

ABSTRACT When teaching students how to write, we enable us to saywhat we want naturally must teach them howto write in a specific and easily, so that we can write language such as English or French. When programs to fit problems, not to fit the teaching them how to program, we must teach language; them how to program in a specific language such as Basic or Pascal. In both cases, a help us organize and convey our language is the vehicle, not the object, of thoughts, so that we can understand and instruction. be understood; Teaching a language involves instruction in vocabulary, spelling, grammar, and be used in a uniform manner by many punctuation. Teaching writing or programmers, so that we and they can programming, on the other hand, also share our knowledge. involves instruction in logic, organization, expression, and style. No programming language is perfect by these The reason language becomes an issue in criteria. Basic is easy to learn, but most teaching programming is simply that we have of its common dialects cause programmers to the a choice. Students have learned their obscure, rather than illuminate, native tongue much before they learn to structure of their programs. Pascal has write, but generally theymust learn a fewer divergent dialects, and it allows us programminglanguage when they learn to to express many constructsquite nicely; program. yet it can make other constructs extremely A good programming language should awkward. enhance our ability to teach programming, Thebest teaching strategy is to turn not distract our attention from that task. this lack of perfection into an asset. It should: Teaching the limitations of a language along with its virtues illustrates be easy to learn, so that we can devote dramatically that programming transcends time to teaching programming and not language. just to teaching the language;

1 APPROACHES. TO REQUIRING MICROCOMPUTERS OF UNDERGRADUATES

Chaired Dy: Jane Caviness University of Delaware

ABSTRACT: Approaches to Requiring Microcomputers of Undergraduates

The use of microcomputers is growing rapidly, while the age of the users and the cost of the microcomputers have been decreasing.Secondary schools are discovering that many of their students have acquired microcomputers and desire some general computing instruction. Colleges and universities are discovering that many entering students already have computing experience, most often with microcomputers, and they wish to continue using microcomputers. This presents a challenge to those involved in Computing Services, since they are accustomed to providing services through the use of timesharing on medium to large scale machines. How are they to deal with the change in the type of demand for computing services? Answers to this challenge cover the spectrum from ingoring the problem totally, to turning it around and requiring undergraduates to have their own microcomputers. The panel members are all from institutions that have taken, the latter approach. They will discuss many aspects of such an action: the decision to do so, the planning involved, the choice of hardware, the costs involved, the expected benefits, the difficulties of implementation, student reactions, and perceptions of first experiences. Discussion and questions from the floor are encouraged.

M. Peter Jurkat Stevens Institute of Technology

Robb Russell Drexel University

Wilson Dillaway Rennselaer Polytechnic Institute

Doug Van Houweling CarnegieMellon University

David Bray Clarkson College Accreditation in she Computing Sciences

John F. Dalphin, Moderator Purdue University

ABSTRACT A joint task force of the ACM and IEEE administer them, these tend to be directed Computer Society is meeting regularly to to specialized programs and the field is so discuss issues relating to accreditation or broad that a widerview must be taken. It approval in the computing sciences. In is estimated that as manyas 500 programs addition to considering various mechanisms not presently served by existing mechanisms to implement the important qualitative and agencies would benefit from such review and certification, the Joint guidance. Committee is developing a preliminary set This panel will discuss some of the of Computer Science Program requirements. issues relating to implementation of Increasing requests are being made to accreditation or approval as well as the professional societies to provide quantitative criteriafor computer science guidance in computer science programs. programs that provide competency in the While certain guidance and evaluation profession. Audience participation and mechanisms exist, and agencies to discussion will be encouraged.

PANELISTS: Michael C. Mulder Servio Logic Corporation Tom Cain University of Pittsburgh

George Davida University of Wisconsin - Milwaukee Gerald L. Engel Christopher Newport College Terry J. Frederick University of Central Florida

Norman E. Gibbs Arizona State University Harvey Garner University of Pennsylvania

SPONSORS: IEEE and ACM

3 NETWORKING FOR MICROOOMPUIERlama :maw

Kenneth Forman Carl Steinhoff

Community School District 27 New York University

A network of several microcomputers peripherals(printerhaXlen)1, connected to a common hard disk storage system provides several administrative functions for Current research defines three basic types effective management. of nerks: Star, Daisy Chain and Dlp Line is) ' , A microcomputer network constructed JliStar configuration, consists of a central Community School District 27 is one of the intelligent microcomputer, termed the host, with 32 public school districts within the Cityof other devices connected in a radial or starlike New York with approximately 27,000 pupilsin pattern. All devices are directly connected to grades K-9 and 1,400 employees. Early in 1980, the host. If one device becomes inoperative, we began to investigate the feasibilityof other devices still function. A Daisy Chain using computers for evaluative and management configurated network consists of a single host with all other devices wired in series to the purposes. This investigation was a collaborative effort with our evaluation host. If one device becomes inoperative, all consultant, New York University, under the devices past the inoperative device cease to leadership of Dr. Carl Steinhoff.We function. The third network type, a Drop Line with determined that microcanputers would most or Bus configuration, consists of a host effectively manage all the various applications a single cable. All devices are connected in we desired to implement for fiscal,information parallel to the main cable via junction boxes leave and evaluative reporting. so that failure of any one device will the remainder of the system operative. Several microcomputer systems were in- Our investigation of types of networks vestigated including:Apple, Atari, Bell and available for use with microcomputers led us3to Howell, Commodore/Pet, Radio Shack, etc. Upon reviewing the literature, we came upon the select the "Omninet" of Corvus Systems, Inc. Cmninet is constructed essentially in a Drop successful experiences of NBC (Minnesota Educational Computing Consortium) with Apple Line configuration, with each intelligent device connected to a hard disk storage system microcomputers.The Apple microcomputer offered numerous applications in the areas of business, via an easily installed piece of hardware The advantage of this financial reporting, personnel and information termed a "transporter". management, as well as its ability to interface type of network is obvious, if a problem arises in one intelligent device, all other devices with larger systems. Therefore, we decided not remain functioning. All intelligent devices to "reinvent the wheel", but to improve on the In applicability of Apple microcomputer systems. Share the resources of hard disk system. addition, a primary intelligent device or host Our plan involved creating a microcomputer through a network for administrative and fiscal reporting. can control access of other users user defined security system, giving users(up to 64) different levels of access toinformation; With the support of our Community School Board and District Superintendent, read only, read and write and manager level Marvin R. Aaron, we were ready to implement our access for security purposes. microcomputer network design (see Appendix). Therefore, our initial microcomputer network Networking refers to connecting several design consisted of the following equipment: microcomputers together through a common 5 Apple II Plus (48k) Microcomputers with transmission line and central source so as to Language Cards, 5 Disk Drives with Controllers, allow the sharing of information and peripheral 5 Zenith Data Monitors, 1 Qume 5 Printer, 1 Corvus 10 Megabyte Hard Disk System, 1 Corvus devices (mass storage, printer, modem). Disk Server, 8 Corvus Transporters, Panasonic Devices connected to a network have been Video Cassette Recorder (NV 8200) and D.C. Hayes termed "nodes" by network users Currently, Micranodem (see Appendix for approximate costs). nodes must be intelligent.They must not be Our financial commitment toward developing individual microcomputers orb intelligent a microcomputer management network for district hardware attachment, can also function as a use was. further supported through creating a remote intelligent terminal with the Central district position of "Computer Specialist", that Board of Education's IBM mainframe computer. is, a person to provide support in implementing this network design. Recently, we have begun to experiment with an optical scanning device, the Scantron 2700, To allow for rapid implementation of our for mass entry of information into our data base. management network, we chose to use commercially Eventually this device will minimize data entry developed software packages developed for hard time permitting immediate use of a data base at disk systrms rather than have our Computer each participating site as well as facilitating Specialist develop customized management soft- creation of a master district data base. ware (which would have delayed implementation over several months). These packages include Additional software applications for our DB Master (a data base management package), network will be forthcoming; for example, we Word Handler (a word processing program) and are investigating general accounting software Visicalc (a numerical data manipulation program). (accounts payable, receivable, general ledger) DB Master was selected for our data base which will be interactive with other programs. management for several reasons, same of which As an additional support to participating sites, include: automatic data compaction upon we are investigating student attendance and storage, sophisticated report generation and scheduling programs that would be interactive password protection system. with our student data base.

Let's take another step back and discuss Lastly, one must consider confidentiality of methods of data base management.Data base information, security precautions and levels of management can be viewed as a pyramid structure. access to information within a network.We have In a tap/bottom structure, all information is structured Omninet with four levels of security collected by a central authority for creation protection to authorize only approved users of a data base. Subsequently, information is access to the network. First, Omninet only reported to collection sites for verification. recognizes authorized users by an individual In a bottom/tap structured data base, each "name" assigned to each user; illegal users are site manages its own data base.We used the denied access to the network by Omninet. Once floppy disk version of DB Master, which is, of the network recognizes the user's "name", the course, compatible with the district hard disk user must then enter an identification code for version used for data base management. In a further access to the network. Then, authorized bottom /top configuration, each site shares its users must place the DB Master Management Disk data base with the district producing a more into the disk drive to access the data base accurate data base. Each site has a vested management program stored within the hard disk interest in management of its own information. system. Finally, the user must enter another District supports each site and a cooperative password to gain access to data files. Once working partnership has developed. within the network, Clarinet is structured to permit differentiated levels of access, that is, Each site manages its data base using the read, read and write, and manager level access following equipment: Apple II Plus (48K) to information. Microcomputer with dual disk drives, Zenith Data Monitor, Epson MX 80 Printer and DB Master If networking will best suit your management (floppy version). Recently, we have provided needs, then consider the following questions : selected sites with DC Hayes Micramodems for eventual telephone hookups. Furthermore, for 1. What is the greatest distance from one end more efficient use of information management, of the network to the other? sites are also adopting word processing using Word Handler (floppy version) which is inter- 2. What is the maximum number of microcomputers active with our data base management program. to be networked?

Our microcomputer management network functions 3. How much mass storage is required? to support the instructional process through various management applications, which include: 4. Which network offers capability for expanding or upgrading? word processing

. information storage and retrieval 5. Are there current users that I can speak inventory with? mailing lists/labels vendor reports 6. What kind of service and support are

. personnel records available? student records ad hoc reporting fran larger datafiles 7. Is the manufacturer reliable?

Individual microcomputers within the network 8. Is the product supported by on-going can function independently, and with a small development?

5 9. What is the cost of installation?

10. Can microcomputers of different manufacturers and/or peripherals be connected within the network?

REFERENCES

1. Minnesota Educational Computing Consortium (AECC), "Spotlight on Local Networking with the Apple II Ccmputar", September 1982.

2. Charp, Sylvia, "Trends - Time Sharing, Microcomputers - Networking", T.H.E. Journal, November, 1981.

3. Corvus Systems, "Winchester Disk Systems for the Apple II and Apple II Personal.COmputers", Corvus Systems, San Jose, California, 1982.

Connell, Cassie, "Networking" What are the Alternative Systems", T.H.E. Journal, September 1982.

APPENDDC Crnninet Hardware Costs 441C12001)44P1JI TIP Network 5 Apple II Plus (48K) $5730. !District 27 Microcomputers 1146 X 5

5 Apple Disk Drives/ 2415. Controllers 483 X 5

5 Apple Language Cards 730. 146 X 5

5 Zenith Data Monitors 675. 135 X 5

1 cone 5 Printer (with 2245. cable/interface)

1 Corvus 10 Megabyte Hard 3495. Disk System

1 Corvus Disk Server 990.

8 Transporters 3790. (1895/4 units) OISTSICT LLSNI SCHOOLS (1895 X 2)

1 Panasonic V.C.R. 1000. (NV - 8200)

Wiring/Installation 1000.

TOTAL $22070. SPREAD SHEET SIMULATION MODELING (SSSM) FOR TRAINING AND INSTRUCTION IN RESOURCE ALLOCATION

Ronald A. Lindahl, Assistant Professor Brent E. Wholeben, Associate Professor

Department of Educational Administration and Supervision The University of Texas at El Paso (El Paso, Texas 79968)

Withtheemergenceof microcomputer technology modifications. Where entering a new value onto a over the past decade, managers in all fields have traditional "paper and pencil" worksheet might foundpowerful newtools at their disposal to require the recalculation, erasure, and re-entry assist them in decision-making. One of the most of numerous interrelated figures, the electronic utilitarian and universally accepted of these new spread sheet requires only a re-entry of the ini- resources hasbeen the electronic spread sheet. tial figures; all subsequentimpactswould be However, only recently have the programs respon- reflected automatically. sible for the initial preparation and continued development of such managers have only recently This apparently simple technological advance begun to appreciate and explorethe valueof has dramatically altered the manager's perspec- incorporating spread sheet simulationmodeling tives for simulation modeling and sensitivity (SSSM) intotheir instructionalprograms. The testing. No longer faced with, hours of tedious very same features of flexibility, adaptability, recalculations and revisions for eachdesired and facility of operation which mark theuse- change,and the inherent potential for clerical fulness of electronic spread sheets for managers error, the manager now hasaccess to a spread are equally relevant to their value as an instruc- sheet which facilitates and encourages manipula- tional device. tion and creative experimentation, without sacrificing its integrity as a tool for data registration and analysis.

Spread Sheet Simulation Modeling SSSM as an Instructional Resource Electronic spread sheets havetremendously expanded the manager's capabilities beyond tradi- However significant atool SSSM may be for tionalpaper, pencil, and calculator worksheet today's managers, it proves to be of even greater methodology. Prior to the advent of microcomputer worth as an instructional aid in the preparation technology and the development of appropriate and development of these managers. software, spread sheetswerecommon managerial tools for preparing and managing budgets, making Oneof thekey featuresof the electronic financial projections, forecasting sales and pro- spread sheet is its relative simplicity, a charac- fits,andcontrollinginventories, etc. These teristic also of prime importance in considering spread sheets were basically a grid of columns and SSSM for instructional purposes. It is not rows, organized in such a manner as to allow the surprising thata significant proportion of the manager to view both horizontal and vertical existing managerial labor force has not yet had interrelationships between cell components of the the opportunity tobecome "computerliterate." grid. However,it should also be noted that a significant proportion of the professors,instructors, More sophisticated applications called for the trainers, and consultants engaged in the prepara- master grid to be divided into various semi- tion and developmentof managers is similarly independent sub-matrices. Under such an arrange- unacquainted withthisnew technology. Conse- ment, the manager could perform all calculations quently, for SSSM to be incorporated successfully in onesub-matrix, record a table of figures into instructional programs, both instructors and (e.g., a tax table, amortization table, or depre- students must be able to grasp itsbasic tech- ciation index) in another sub-matrix and maintain nological aspectswitha minimalinvestment of a third sub-matrix as the primary exhibit of tran- time and effort. While most major producers of sactions. software offer their own version of the electronic spread sheet, many share common features; and most Consequently, today's manager has the ability arestructuredand documented wellenough that to be able to make virtually limitless changes in novices can readily master the rudiments of their the worksheet in order to witness instantaneously operation. Obviously, some of the more sophisti- the sensitivity ofrelated components tothese cated features may requirein-depth studyand

7 hands-on experimentation; however, their use Basic SSSM: Cohort Survival Analysis would be optional and need not be a deterrent to the introduction of SSSM as an instructional As discussed earlier, the spread sheet may be approach. configured as a single, master matrix or, for more complex situations, as a compositeof several Another featureof electronic spread sheets interrelated sub-matrices. For less complex which has facilitated their acceptance among mana- models,the single matrix format facilitates an gers is that of versatility. Managers have exhi- understanding of the whole, while stillallowing bited considerable creativity in developing new the student to perceive and follow the interac- ways to use electronic spread sheetsin solving tions. problems and making decisions. Rather than learning the intricacies of several different, One example of the successful utilization of a dedicated softwareprograms toaddress specific single matrix in the instruction of resource allo- applications, managers have found the spread sheet cation isin the teaching of cohort survival ana- to be versatile and adaptable to a wide variety of lysis. Cohort survival analysis is a forecasting situations. The savings, in termsof time and techniquein which thesurvival, attrition,and energy required to master the rudiments of the new membership rates of certain specific popula- program and make it operational, have been signi- tion groups are examined oversuccessivetime ficant. periods. By extrapolating this data, and assuming a continuity of past trends, projections can be Obviously, this feature is also of prime made regarding the future population (membership) importance for instructional applications.Having of the group. mastered the rudiments of the electronic spread sheet,the instructor can successfully employ it The cohort analysis technique has been effec- for a varietyof simulations without having to tively employed in the forecasting of school invest further time on the operational aspects. enrollments in those districts which have achieved At the same time, students who have familiarized some consistency or stability of population growth themselves with the basics of the spread sheet can or decline. The matrix consists of both histori- better concentrate on the content aspects of all cal and extrapolated data, with the columns being subsequent simulations. In viewof the time defined as the years under consideration and the constraints and need for instructional efficiency rows being configured as the grade levels being which characterize most managerial science examined. The individual cells contain the actual programs, this is a highly significant feature. class membership figures for past years and formulae topredicta weighted average for future years. The computer then automatically calculates The Value of Simulation Modeling the projected enrollment for future years, either bylinear extrapolation ofpasttrends or by In his book, Computer Modeling and Simulation, systematically weighting the most recent trends Francis Martin defines simulation modeling as "a more heavily than earlier performance of the logical-mathematical representation of a concept, system. Figures 1 through 3 illustrate the three system, or operation programmed for solution on a major phases of this SSSM process: (1) entering high-speed computer." In ComputerSimulation background data; (2) derivation of survival Applications, Julian Reitman describes simulation ratios; (3) and projection of future enrollments. as"a practical,application oriented procedure" in which "one must construct an abstraction of the At the same time, itis extremely simple for problem, transfer the problem to a foreign device, students to construct a matrix which would allow the, computer, and thenobtainindications per- them to change this formula, enter in their own taining only to the representation of the system." survival ratio predictions, andtest the sen- Both authors then dedicate considerable discussion sitivity of the projected enrollment to such to the problemof identifying those conditions variations. Figure4 illustrates a simplified under which simulation modeling can effectively be version of such a learner-manipulated matrix. employed. Although the above example serves to Simulation modeling enables the manager, pro- illustrate the utility of the SSSM inits most fessor, or student to: (1) examine a complex simple form,it by no means exhausts the capabi- situation; (2) state the problem in a simplified, lity of this toolin helping students understand understandable form; (3) define therules and cohort survival analysis. For example, itis a relationships which govern the situation; (4) and relatively simple matter for the professor to set experiment withoperating and manipulating the up the matrix as an interrelated series of subsystem. As Reitmanpointsout, this affords matrices,which thestudentcan study indepen- several significant benefits, including: "aid in dently to better understand the various components problem definition; help in relating numerous fac- of the cohort survival analysis process. One such tors with their influences on the design; insight sub-matrix which hasnot yet been demonstrated, into the sensitivity of the design to wide ranges but which would be of great value in training stu- ofparameters; support for selecting thefinal dents to utilize this resource allocation tool, design from among the alternatives uncovered; and would be that of multi-year forecasts. Figure 5 guidance in predicting system performance." presents such an extended projection, based on the assumption that the trends upon which the single-

8 Figure 1. Background Data

;111111111Mulaff"----:1- Figure 4. learner-Manipulated Survival Ratios

;

Figure 2. Survival Ratios

Figure 5. Multi-year Enrollment Projections

Figure 3. Enrollment Projections

9 year projection were based wouldcontinue in a greater than 64K, the State minimum salary matrix linear pattern throughout the total period being can be entered as a separate sub-matrix. This, in considered. turn, canthen bereferenced directly from the sub-matrix created to reflect the staffing pattern selectedbythe district. Anothersub-matrix Extended SSSM: State Allocation Formula would be designed to record the participation of students in special programs; thecalculations Inlight of the flexibility and capacity of employed in this matrix would then feed into the such multiple sub-matrix models, there are a principal matrix to indicate funding for classroom myriad of instructional uses which can be served units, additional weighted personnel units, etc.. by SSSM, especially in the area of resource allocation. For example, students at The University Such a SSSM structure allows the students to of Texas at El Paso have responded very positively manipulate any ofthe variables of the formula to the use of this tool in investigating the pre- alone or concomitantly and to test the sensitivity dicted effects of changes in State legislation on of the overall allocation to such modifications. the allocations of operating funds to local school For examples, students cancompare the several districts. proposals which will be considered for adoption at the nextLegislative session anddeterminethe At In thisapplication, the electronic spread effects of each on local district allocations. sheet is configured as an inter-related composite thesametime, they canexamine the financial of sub-matrices. The principle sub-matrix presents implicationsofvarying the district's staffing a summary of the major componentsofthe State pattern, test the sensitivity of the allocations allocation formula, including the: to various enrollment projections, or make inter- district comparisons, all with great ease. - flatgrantstipend for each unit ofaverage daily attendance; Figures 6 through 8 illustrate typical screen formats from a simplified version of this applica- - State contribution topersonnel costs, based on tion. the maximum allowed number of weighted personnel units andthe State minimum foundation salary scale; SSSM: Instructional Transition to Application most benefits of - specialprogramsupplemental funds,forsuch One of the significant areas as special education and vocational employing SSSM for instructional purposes is the knowledge, education; high degree of transferability of allowing the learner to make the transition from - transportation allotments, againdifferentiating the academic environment to managerial decision- between regular and special program costs; making reality.

- State compensatoryfunds, providing program The relatively low cost of microcomputer hard- enrichment for children of low-income families; ware, coupled with the almost universal availability of electronic spread sheets forall major configurations of this hardware, suggests that it - cost-adjustment allocationsfor small and/or sparsely populated districts; and, would not be unreasonable to expect that a high proportion of today's managers would have access normal work - State minimumfoundation program funds for to this technology within their district operation and maintenance expenses. environments. Previous ugemlrations" of learners were faced with tremendous di'ficulties in By limiting this principal matrix to a flow- attempting to transfer their exP iences on the through listing of the total allocations available mainframe equipment utilized in the instructional con- fromeach of these components, thestudent is environment to often radically different better able to view the formula from the systems figurations of equipment within their professional of perspective. Concomitantly, a separate sub-matrix work environments. With theaccessability would be extablished for each component, so that microcomputers in virtually all professional set- the student might clearly understand how the allo- tings, and with the high degree of transferability cation is generated. This sub-matrix system also of knowledge between the various electronic spread facilitates the sensitivity testing of minor sheet/microcomputer hardware configurations, changes in any given component which might be pro- today's students' time investment in learning the posed by the Legislature. For example, one sub- basics of SSSM may well be translated into matrixwould consist of dataon averagedaily increased productivity upon entry into or return attendance at the various grade levels identified to the work situation. in the State formula. As such, the manager can Orsonally configure Calculations would be shown in this matrix to electronic spread sheets to address specific real illustrate how these figuresare converted into decisions. At this point, the learning tool is weighted personnel units. On those microcomputer transformed intoa decision-making aid, allowing systems which would afford sufficient memory to the manager to combine professional expertise with oftheSSSM to accommodate a matrix of sufficient size, e.g. the dataanalysiscapabilities

10 simulate the situation in question under a variety of assumptions andhypotheses. The cycle, in turn, becomes complete when the learning and discoveries gained through these practical applications of SSSM are brought back into the instructional setting to enhance the education and re-cycling of other managers.

Summary

In conclusion, educators responsible for the training and instruction of topics related to the area of resource allocation will find Spread Sheet Simulation Modeling to bea versatile, effective instructional technology. In addition to its ver- satilityandapplicability toa wide range of managerial decisions, SSSM offers the instructor the desirable characteristics of being relatively simple to learn, even for the non-computerliterate, and of promoting interaction between the learner and the technology.

At the same time, students who become familiar with this technology in the classroom setting will actually reap a double benefit; not only will they have the opportunity to better understand the content area being presented, buttheycanalso become conversant with a powerful decision-making tool which can be of invaluable assistance to them throughout their managerial careers.

Bibliography and Related Readings

1. Francis F. Martin, Computer Modelin and Simulation. (New York: JohnWi ey& Sons, Inc., 1968).

2. Julian Reitman, Computer Simulation Applica- tions. (New York: John Wiley & Sons, Inc., 1971).

3. Brent E. Wholeben, "Operational Network Displays" in Communication Strategies in Evaluation (Nick L. Smith, Ed., Beverly Hills, California: SAGE Publications, 1982).

4. Brent E. Wholeben, The Design, Implementation and Evaluation of Mathematical Modeling ProceduresforDecisioning Among Educational Alternatives. (Lanham, Mary and: University Press of America, Inc., 1980).

11 DEVELOPMENT AND VALIDATION OF A CU211TERIZED ADAPTIVE SCREENING TEST (CAST) FOR USE IN ARMY RECRUITING

Herbert George Baker, PhD, Bernard A. Rafacz, and William A. Sands

Navy Personnel Research and Development Center San Diego, CA 92152 619-225-2408 Abstract competition for available personnel among colleges and the several armed forces is probable, a competition that will NAVPERSRANDCEN designed and developed increase the difficulty of recruiting. a Computerized Adaptive Screening Test best available candidates for (CAST) that predicts an applicant's score The on a selection test. CAST is capable of enlistment must be located, enlisted, and operating on a stand-alone microcomputer optimally assigned; neither fiscal nor in Army recruiting stations, and personnel resources can be wasted; and correlates at .866 with the criterion. those mundane tasks of the recruiter, Further data collection and validation which detract from the primary mission of efforts are currently in progress. This locating propects and "selling the development suggests that a shorter, service," must be reduced. adaptive test is feasible for use by organizations employing an aptitude test The screening of applicants for in the selection processes. enlistment, which takes place near the end of the recruiting process, includes as a major element the Armed Services Introduction Vocational Aptitude Battery (ASVAB), a 10-part test given to all armed services The Navy Personnel Research and applicants. To be eligible for Development Center (NAVPERSRANDCEN), enlistment, an applicant must achieve a through an agreement with the U. S. Army minimum qualifying score on the Armed Research Institute for the Behavioral and Forces Qualifying Test (AFQT), which is a Social Sciences, is conducting research linear composite of the scores obtained and development in support of the Joint on four ASVAB subtests. The ASVAB is Optical Information Network (JOIN) System given either through the Department of being implemented nationwide at all Defense High School Testing Program, or levels by the U.S. Army Recruiting to service applicants at a Military Command (USAREC). Army enlisted Entrance Processing Station (MEPS) or applicants will directly interact with Mobile Examining Test (MET) site. this system in the course of recruiting and accessioning. The objective of this Applicants who have no previous effort was to design and develop the qualifying AFQT scores must be sent to ComputerizedAdaptiveScreening Test the MEPS/MET for ASVAB testing. For (CAST) that could:(1) operate on a those who must be transported from a stand-alone microcomputer system in recruiting station to a MEPS/MET site, recruiting stations as part of the JOIN costs are entailed for transportation, system, (2) reduce recruiters' and in many cases for meals and lodging. administrative burden, and (3) predict Costs are also incurred for personnel Army applicants' scores on the ASVAB more time at' the recruiting station. If efficiently than the paper-and-pencil applicants are sent to the MEPS/MET site Enlistment Screening Test (EST) used at and subsequently fail the ASVAB, there is present. a significant waste of money. Conversely, if applicants who would have Armed services recruiting faces passed are erroneously deniedASVAB serious challenges in the future, due t9 testing, their talents are lost to the a shrinking pool of military eligibles.i se:v recruiting quotas are not met, The All-Volunteer Force (AVE) concept has and costs- accrue to the led to vastly increased expenditures in applic!,1' recruiting. A fierce and costly Currently, all armed services use the Enlistment Screening Test (EST) at *The views expressed herein are those of recruiting stations to predict applicant the authors and do not necessarily AFQT scores. This test is composed of 3 reflect those of the U. S. Navy. subtests of 15 items each, with a total

12 burdensome in its of 45 minutes. It was method less time limit administration, to reduce the recruiter's developed by the Air Force Human Resources Labopatory in 1976,' and task load. revised in 1981. Computerized adaptive testing combines recent developments in latent Pure measurement error does not seem trait theory with the ever-increasing to be a problem; it appears that the power and efficiency of computers. The newer forms of the test (81a and 81b) result isa return to individualized measure quite well throughout the score without the loss of and testing range where most selection administration efficiencies gained classification decisions are made, aid it through group testing, combining advances correlates at .83 with the AFQT. The in computer technology with those in EST, however, does share in the problems psychometrics. of adverse psychological effects. The EST is a conventionally administered Only a limited subset of test items paper-and-pencil test, which has been is needed to establish an aptitude shown to increase guessing, ;rwstration, estimate. In adaptive testing, each and boredom in subjects. In its subject receives only those questions customary delivery mode in the recruiting station, the EST actually combines the appropriate to his or her ability level. individually The result is an individualized test, worst features of an "tailored" or "adapted," and actually administered test (i.e., it is heavily constructed for each examinee. While dependent upon variables associated with there are several "branching" strategies the examiner or the examiner-examinee available, in tailored testing the relationship) and the group test (i.e., selection of each question is based on item arrangement, item setA answer sheet the subject's response to the previous effects, and imprecision). ° It is scored question. Typically, a correct response on a pass/fail basis. The EST, being a is followed by a more difficult item; an timed test, also exerts differential incorrect response is followed by an item pressure on the examinee (i.e., results of less difficulty. Therefore, the are based partly upon the reaction of the difficulty level of an adaptive test is individual to time constraints.? dynamically tailored to the ability of And, because each subject on each subject. Major concern focuses receives a tailored subset of items, the administrative error and clerical burden. chances of test compromise through The EST requires approximately 45 minutes copying, memorization, or coaching are to administer, as well as time to score and interpret results (by a recruiter considerably lessened. already investing many hours in the With adaptive testing, shorter tests potential enlistee). To this must be may be used wi0out loss of reliability added the time required to manage the or validity.9,7 Adaptive tests are test supplies. At pre.ent, besides untimed, reducing pressure on the storing, filing, retrieving, and ordering subject, without hindrance to the replacements, the recruiter is required proctor.1° Adaptive testing reduces to take frequent inventory, make numerous guessing,9 real checks and corrections for unauthorized boredom and frustration,7 or perceived proctor-subjectbias,11 acq markings in the test booklets, and racial bias. safeguard used answer sheets. The ESP is culturally specific Adaptive testing is more motivating, thus highly labor intensive, consuming from the time of a senior noncommissioned thereby eliciting "best results" officer in quasi-clerical tasks. subjects.° and more 4qcurately reflecting subject competence. Only two forms of the EST are in use, testing has its own a situation that offers the failing Computerized subject hope of eventually passing the merits. Interactive dialogue provides results without manual test by repeated testing and item immediate scoring. Immediate knowledge of results memorization. Furthermore, the recurrent a, motivator to problem of "malicious error" appears, has been shown to b better performance;13,i4,10 Computerized involving recruiter malpractice. item Recruiter influence through pretest testing lessens test bias through selection and increases test fairnessby coaching is a simple way to manipulate and the Initial and replacement costs the nature of the test itself results. modality.15 The and short materials life associated with test's administration paper-and-pencil tests; poor impression computer makes it possible to eliminate printed test materials and created by dilapidated materials; and security, an security, custody, and control associated logistical, problems, difficulties. In short, there is a administrative critical need for both a more efficient facilitating item replacement, whole test predictive instrument and a screening construction, and the capturing of data

13 d for validation purposes. Test Administration

Stated simply, the computer can In each CAST subtest, a provisional administer a test item, accept a ability estimate is made, a test item response, score that response, and keep a to that ability level is record of the subject's response presented, and then the ability estimate history. After each question, the is updated based on the response to the computer can use the response information test item. The computer program for this to update the ability estimate and then iterative ability estimation process use the new estimate of ability to select starts by associating each examinee with the next item. With each successive an ability level of the information response, the ability estimate gains matrix. Selection of a test item starts reliability. The process can continue from the top (highest information value until some stopping rule is satisfied within a level) and searches for the (e.g., a fixed number of questions first item not yet presented. The item administered, or a prespecified level of that results from this search is then reliability). Computerized adaptive presented. Based on the examinee's testing can effectively shorten testing correct or incorrect response, the time without effectiVeness loss as well ability estimate is updated. This as eliminate scoring an0 recording errors associates the examinee with a new due to clerical error.I1 ability level of the information matrix, (potentially) from which a new test item CAST Design and Development is selected for presentation. This process continues until the limit Designing and developing CAST established by the stopping rule is required parallel work in psychometrics reached. and computer programming. All design work was accomplished on an Applied To facilitate user acceptance, the Computer Systems (ACS) microcomputer with time between item response and a Perkin-Elmer Data Systems 1200 video presentation of the next item was display terminal (VDT). The effort established as less than or equal to proceeded in the following steps: three seconds. Software documentation for presentation of CAST on the ACS Test Construction microcomputer system was completed.. Subsequently, the program was converted CAST was envisioned as incorporating to operate on the JOIN developmental three subtests that would correspond to system, an Apple II-Plus microcomputer, three of the four ASVAB subtests used to with two disk drives, and a VDT. calculate the AFQT composite score: Word Knowledge (WK), Arithmetic Reasoning Test Validation (AR), and Paragraph Comprehension (PC). These subtests were determined to be the Concurrent NAVPERSRANDCEN research to best predictors of AFQT score. Item assess the relationship between banks for each subtest were assembled paper-and-pencil ASVAB tests and an under a contract with the University of experimental battery of three Minnesota. These item banks included 78 computerized adaptive subtests (WK, AR, WK items, 247 AR items, and 25 PC items, and PC) provided the opportunity to together with the estimates of three conduct a pilot test of CAST. The three parameters (discrimination, difficulty, CAST subtests were administered during a and guessing) for each item. A Bayesian 90-day period late in 1981 to 356 male ability enimation procedure described by Marine Corps recruits at the Marine Corps Jensema'° was chosen for scoring and Recruit Depot, San Diego. Each recruit determining the selection and had taken the ASVAB before enlistment and presentation sequence of test items. had been retested on a parallel form of ASVAB during recruit processing. After Computer Software eliminating subjects with missing scores on any test and those who had been CAST computer programs were written administered obsolete forms of ASVAB, the to provide interactive, user-friendly remaining sample was 270.17 software that presumed no previous computer experience on the part of either In the CAST pilot test, each subtest recruiter or applicant. In addition, VDT was administered with a fixed length: 15 screen text displays were written to items each for WK and AR, and 10 items conform to readability (reading grade for PC. All, examinees began with the level) standards. same item, which was of medium difficulty. While the paper-and pencil ASVAB is a timed test, the adaptive tests were conducted without time limit. Preliminary introduction to the testing

14 Work in Progress situation was delivered orally by the proctor, while all other instructions, Data analyses are now underway to including use of the terminal and evaluate the validity of CAST with its procedures for answer entry and answer intended population. The CAST was changes were delivered on the terminal administered to 364 Army applicants at screen. Practice preceded each subtest; the Los Angeles MEPS between 29 November the successful response to these items 1982 and 7 January 1983, using Apple II was a condition of beginning the subtest. Plus microcomputers operating under CPM, The test was administered by computer, on with two disk drives and a VDT. Results four terminals in a specially designated are pending completion of analyses. testing room. The terminals were on-line with the Hewlett-Packard 21 MX computer Discussion at the University of Minnesota, through a data communications line. A complete microcomputer-based CAST demonstration system is now in operation, There were several differences with complete documentation for all between CAST as it was designed to be software. The test item bank for AR has given and the subtest administration been reduced in size to 225 items. The during the pilot test described herein. WK and PC item banks remain at 78 and 25 A true backspace key was not available on items respectively. User-friendly, the terminal, requiring the recruit interactive software provides full screen subjects to use the "Rubout" key" to make display, clearing the screen after each corrections. In the PC subtest, the display. Respoose time is three seconds stimulus paragraphs did not remain on the or less. An easily used backspace key, screen while the response alternatives feedback to the user after each answer, were displayed. The pilot test was and an error-trapping capability that administered on terminals communicating ensures recruiter assistance after with a host computer, while CAST is repeated procedural errors by the designed for use on a stand-alone applicant have been added to the system. microcomputer with attached disk drives. Finally, because of restriction in range, Remaining research and development in the Marine Corps recruit test subjects adaptive testing involves evaluating the are not representative of an unselected interactive screen dialogues for applicant population. readability and user friendliness with actual applicants or recruits, and field After test scores were collected, the testing the CAST system in a recruiting relationship between CAST subtests AR, station. Capitalizing on the results of WK, and PC and their paper-and-pencil the pilot testing of the subtests, it counterparts was evaluated through appears entirely feasible to eliminate correlational analysis. The ability of the PC subtest from CAST, without a the CAST subtests to predict AFQT significant decrement in the correlation composite (AR, WK, PC, + 1/2 Numerical between the CAST linear composite and Operations) scores was assessed using AFQT (see Table 1).17 Examination of the multiple regression analysis. data suggests that the length for the AR and WK subtests could be set at 7 and 15 Results items respectively, without appreciable .mar loss in predictive validity. Were both Each CAST subtest correlated as well procedures to be implemented, the with its ASVAB counterpart as did the correlation between AFQT and CAST would parallel form ASVAB retest score. drop only to .865 (from .866), still Multiple regression analysis to evaluate comparing very favorably to that of the the relationship of the CAST subtests current EST (.83). The result would be a with AFQT score indicated a multiple CAST of only two subtests, requiring an correlation of .866 average of 16 minutes for complete administration, scoring, and Notwithstanding the differences interpretation, as opposed to 45 minutes mentioned previously between CAST for the EST administration alone, a administration as originally designed and savings of approximately 65%. as carried out in the pilot test, the results of the pilot test are Conclusions significant. It was clearly demonstrated that military recruits, and by CAST can be regarded as successfully implication, military applicants, could developed, requ,iring only minor be tested by computer terminal with refinements in both psychometrics and minimal intervention by a proctor. The programming. It is superior to the EST CASTsubtests measured the same abilities in terms of administration and management as the corresponding ASVAB subtests, but and about equal in predictive power. with abov4 half the number of CAST eliminates the need for traditional questions. test materials, thereby saving storage

15 battery in a selection environment may space, replacement costs, and recruiter profit from development of a shorter, time formerly used for administering the adaptive test to predict success on the tests and controlling and maintaining battery. There is nothing inherently test materials. With CAST, test loss, military about the ASVAB or the CAST. theft, and compromise would be all but This research may therefore be eliminated. Security would be maintained generalized in large measvre t.o analogous or by a built-in user password situations. identification. Rather than serving as a 11.11=1.11111.References test proctor-scorer, the recruiter simply would manage a computer-subject dialogue, 1. Congressional Budget Office. Costs of with a self-scoring test for which manninatheactivadty_military (Staff results are immediately available and may working paper). Was6ington, DC: Author, be automatically stored for later use. May 1980.

Cost effectiveness would result from 2. Office of Naval Research, Psychological eliminating traditional test materials Sciences Division. 1979 programs and from reducing recruiter time spent in (450-11). Arlington, VA: Author, testing. In short, CAST will decrease November 1979. negative psychological effects, decrease administrative error, and increase test 3. Joint Chiefs of Staff. United States security. CAST is important for its military. posture for FY 83. WaShington, present economizing service, but DC: Author, 1982. important, too, for its enabling functions. Implementina CAST will allow 4. Jensen, H. E.,& Valentine, L. D. the Army to be highly responsive to Development of the Enlistment Screening. (Tech. Rep. advances in psychometrics and managerial Tes-,.....(EST)- forms.-.5 and 6 science, as well as enable it to 76-42). Brooks Air Force Base, TX: Air implement further applications rapidly Force Human Research Laboratory, May when they are needed. Computerized 1976. ability testing systems are predicted to find their optimal use in organizations 5. Mathews, J. J., & Ree, M. J. serving populations of wide-ranging Enlistment screenina_test_forms_81a_and ability,18 and with CASP, the Army will 81b: Development and calibration be in the forefront of Computerized (AFHRL-TR-81-54). Brooks Air Force Base, Adaptive Testing (CAT) implementation. TX: Manpower and Personnel Division, Air Force Human Resources Laboratory, March In today's recruiting climate, where 1982. increased screening capabilities assume ever greater importance, the Army will 6. Vale, C. D., & Weiss, D. J. A study of have the technological base upon which to computer-administered stradae.t.ive abilit mount other screening instruments-for testing (Res. Rep. 75-4). Minneapolis: both selection and classification. These University of Minnesota, October 1975. might includeipMytors of tenure and Strate.gies of adaptive effectiveness;1. 9I I assessment of 7. Weiss, D.J. expectations, intentions, job abilitymeasurement (Res. lkep. perceptions, and attitudes;22 and Minneapolis: University of Minnesota, screening for specific placement. Dept. of Psychology, December 1974. Screening could be significantly improved by expanding the array of measures to 8. Weiss, D. J.,& Betz, N.E. Ability include special abilities and, even measurement: Conventional or adaptive? biodata,23 since the administration and (Res. Rep. 7S -1). Minneapolig: motivational problems associatedwith University of Minnesota, February 1973. lengthy testing and examinee fatigue would be reduced by automation. 9. Betz, N. E.,& Weiss, D. J. Empirical and simulationstudies of fIgxfie7gf Tangentially, future test development abiliEy. testing. Minneapolis: costs and intrusion on operating systems University of Minnesota, July 1975. will be reduced because experimental test items can be introduced within CAST, in a 10. Weiss, D. J. Final report: manner that is transparent to the field Computerized ability.....testina, 1972-1975. user. This will facilitate the Minneapolis: University of Minnesota, development and evaluation of new items. April 1976. Test administration will be standardized, fairer to all applicants, and far more 11. Gorman, S. Computerized adaptive In efficient in scoring and recording testing with a military population. methods. D. J. Weiss (Ed.),Proceedings of the 1977 Computerized Adaptive Testing_ Importantly, CAST demonstrates that Conference. Minneapolis: University of any organization using an aptitude test Minnesota, September 1977.

16 22. Horner, S. u., Mobley, W. H., & 12. Pine, S. M. Reduction of test bias by Meglino, B. M. An____expeOmental adaptive testing. In D. J. Weiss, (Ed.), evaluation of effects of areafistic to Proceedings of' the 1977 Computerized preview on Marine recruit affect, Adaptive Testing Conference. intentions, and behavior. Columbia: Minneapolis: University of Minnesota, University of South Carolina, September September 1977. 1979. D. J. 13. Betz, N. E., & Weiss, 23. Swanson, L., & Rimland, B. --A Psychological effects of immediate preliminary evaluation of brief Navy knowledge of results and adaptiveability enlistment classification tests (Tech. testin.g._. Minneapolis: University of Bul. TrSi. San Diego: Navy Personnel Minnesota, 1976. Training Research Laboratory, January 1970. 14. Prestwood, J. S. Effects of knowledge of results and varying proportion correct on ability test performance and psychological variables. In D. J. Weiss, Proceedings of the 1977 (Ed.), ..... Computerized Adaptive Testing Conference. Minneapolis University of Minnesota, September 1977. 15. Pine, S. M., Church, A. T., Gialluca, K. A., & Weiss, D. J. Effects of comeuterized adaetive testinaonViick and white students. Minneapolis: University of Minnesota, March 1979.

16. Jensema, C. J. Bayesian tailored testing and the influence of item bank characteristics. In Applied Psychological Measurement. Minneapolis,

17. Moreno, K., Wetzel, C. D., McBride, J. R.,& Weiss, D. J. Relationship of the Armed Services voal-i-onaf Aptitude Batter to three computerized adaptive subtests (Tech. Rep.). San Diego: Navy Personnel Research and Development Center. (In preparation). 18. DeWitt, L. J.,& Weiss, D. J. A computer software system for adaptiveability measurement. Minneapolis: University of Minnesota, January 1974. 19. Sands, W. A. Development of a revised odds_ _for effectiveness (OFE) table for screening male applicants for Nag enlistment (Tech. Note 7-6=5). San Diego: NavyPersonnel Research and Development Center, April ]976.

20. Sands, W. A. Screening male ajplicants for Nau enlistment (Tech. Rep. 77-34). San Diego: Navy Personnel Research and Development Center, June 1977.(AD-A040 534)

21. Sands, W. A. Enlisted personnel selection for the U.S. Navy. A Journal of _Applied Research, Spring, 1978, 31(1),

17 WORD PROCESSING IN THE CLASSROOM: USING MICROCOMPUTER - DELIVERED SENTENCE COMBINING EXERCISES WITH ELEMENTARY STUDENTS

by Karen Piper

Texas Tech University Texas Instrwents Incorporated Lubbock, Texas

Abstract a study using sentence combining treatments This article investigates the feasibility presented on the microcomputer to fifth gra- of using the microcomputer to deliver sentence ders in Abernathy Elementary School. Of par- combining instruction to upper elementary stu- ticular interest to the researchers was the dents. Background research on traditional sen- effect of using the microcomputer as a wri -. tence combining instruction is reported, and ting tool on the areas of writing motivation the use of the microcomputer in writing in- and syntactic maturity. Growth in syntactic struction is discussed. The author describes maturity, a basis for writing ability and rea- recent research in a small elementary school, ding comprehension, might also foster gains in which yielded positive results for microcom- these important areas. To date, results from puter-delivered instruction in sentence com- this study exceed expectations, especially in bining and expansion activities. Positive re- the areas of motivation to write and revise sults were evident in the areas of writing, and writing improvement. revisions, attitude toward composition, attitude toward sentence combining instruction using Sentence combining is usually presented the microcomputer, and student ability to ef- in a large group, followed by oral work in fectively use word processing programs. small groups, and finally, with the student writing out his combinations. Using the microcomputer in this scheme has many benefits. Ef- fectively used, the microcomputer should enhance teacher abilities to provide high qual- Introduction ity instruction in sentence combining. Areas As the microcomputer becomes more common particularly affected by this treatment.include in classrooms, language arts teachers will be writing motivation and syntactic maturity as less enchanted with it solely as a drill and evidenced by growth in writing ability snd practice delivery system. They will desire reading comprehension. The use of microcompu- other uses for it that will increase its ap- ters as a delivery system in this task should: plicability to the total curricular program. 1) allow for individualization of instruction One direction for the expansion of microcom- with student self pacing, 2) enhance student puter use is in the area of structured wri- motivation and interest, 3) enhance student ting instruction such as sentence combining. awareness of the manipulative quality of len, guage, 4) provide students with immediate Recent improvements in the quality and feedback through print-outs end spontaneous user-friendliness of word processing programs interaction with video display, 5) motivate have already stirred an interest in the micro- students to write, and 6) provide students computer as a tool for writing instruction. with a record of student processing abilities. Yet the studies that examine the role of the microcomputer in writing most often deal with The major goal of sentence combining in- open, creative writing, not structured instruction is an increase in syntactic matur- struction in writing technique. Sentence com- ity. Based on research in traditional sentence bining instruction which uses the monitor as combining, this desired increase in syntactic paper and the keyboard as pen can aid students maturity might result in an increase in read- in several areas, some of which include lin- ing comprehension. Word processing systems guistic development and structure, comprehen- should be useful in teaching sentence combi- sion of complex structures, and motivation to ning because of the ease of text manipulation write and revise. and print-out capabilities. However, this has not been previously investigated in regards to Research in the Classroom the feasibility of such instruction, the ease In order to investigate some of the effects of implementation, the motivational aspect, of microcomouter-delivered instruction in sen- and the' overall effect of such instruction on tence combining, the author and Dr. Michael student abilities and attitudes. The major Angelotti, of Texas Tech University, designed goals of this research included determining

18 33 the feasibility and examining the total effect sion. Stotsky (1975; 1982) reviewed studies of this instruction on writing growth, reading that attempted to show that enhanced syntactic comprehension, and attitudes toward writing. skills through writing activities improved In order to provide necessary background for reading comprehension. She confirmed that an the use of sentence combining activities de- increase in syntactic skills does effect read- livered by the microcomputer, the results of ing comprehension (1977). The overall effects traditional sentence combining techniques will of sentence combining practices on reading are be reported. This information will be followed still inconclusive, and several researchers by a brief description of the study conducted urge further investigation (White,1980; by the researchers during the spring of 1983 Stotsky,1982). and a discussion of some of the implications of this exciting technique. Sentence Combining with Microcomputers Little research has been found that speci- Traditional Sentence Combining fically addresses the use of microcomputers to Research in sentence combining supports it deliver instruction in sentence combining for as an effective way to improve student writing the purpose of enhancing the syntactic maturi- and although the evidence is inconclusive, ty and reading comprehension of school aged sentence combining also seems to enhance read- populations. Most of the research focuses on ing comprehension. Sentence combining theory either college populations, or is restricted features the use of embedding techniques to to creative writing. Results of microcomputer form more complex sentences. These techniques use in creative writing, however, provide a encourage students to embed modifiers, make necessary foundation for the extension of deletions, and perform various transformations writing instruction to sentence combining ac- (0'Hare,1973; Strong,1976; Combs,1977). Sen- tivities .For example, Schwartz (1982) stated tence combining activities force students to that the use of a computer-based text editor use their linguistic knowledge to manipulate encouraged greater manipulation of written ma- language. An additional benefit of sentence terial. Bradley (1982) reported positive re- combining is that students are asked to embed sults when students used word processing sys- kernel sentences, which forces them to keep tems for creative writing. In addition, she longer discourse in their heads. This activi- compared student ability to use various word ty promotes chunking, a memory technique which processing systems. Seymour Papert in can lead to improved organization and enhanced Mindstorms (1980), reported that children at cognitive maturity (O'Hare,1973; Sternglass, his MIT Computer Lab often went from "total 1980). rejection of writing to an intense involvement accompanied by rapid improvement of quality Due to the interrelationships among the within a few weeks of beginning to write with language arts, sentence combining may improve a computer." Schwartz (1982) spoke of the pos- reading comprehension as well as writing abil- itive effect of the microcomputer on student lity (Combs,1977). Encoding written language revisions. Educators and researchers are be- is not simply the process of writing down oral ginning to examine the feasibility of using language. Written language is more grammati- the microcomputer in writing instruction and cally complex, reflecting more embeddings. In to suggest it as a possible delivery system addition, written language requires specific for sentence combining instruction (Bradley, rules (i.e., punctuation and spelling) for 1982; Cronnell, 1981). correct implementation. Decoding written language (reading comprehension) is equally com- The manipulative characteristics of compo- plex, involving student experiential, intel- sing with word processors, coupled with the lectual, and linguistic resources in relation motivational value of such a tool has great to content, vocabulary, and sentence struc- potential to tempt students to write more, re- ture. A common denominator in these language vise more often, combine and embed language processes is syntactic maturity. Researchers elements, and to increase in syntactical ma- were initially interested in establishing a turity as evidenced by writing ability. It relationship between syntactic maturity and seems plausible that as students are trained writing ability (Hunt,1965). Once established, to deal with more complex written syntactical they directed their attention to the relation- units, an increase in reading comprehension ship between syntactic maturity and.more ef- might occur. The microcomputer seems to be a ficient reading comprehension. Some research- reasonable delivery system for instruction of ers felt that sentence combining might also this sort, and can help teachers more ade- foster reading syntactic maturity as evidenced quately incorporate instruction in sentence by comprehension (Stotsky,1975; Ney,1980; combining into their classrooms. One way sen- White,1980). tence combining instruction delivered by the microcomputer can be implemented is the method Research by Evanecko, 011ila, and Armstrong used in this study. Use of this technique or a (1975) indicated that two language competen- related procedure can help the teacher expand cies, fluency and control of syntactic com- the microcomputer from simply an instructional plexity, underlie reading comprehension profi- tool to an instructional medium which holds ciency. They suggested that building these two great promise in the area of structured writ- competencies would improve reading comprehen- ing technique.

3 _Li made his own revisions on the microcomputer. The Research Technique Final print-outs were obtained and shared with The purpose of this research was to sift the group. Sharing written stories was a spe- through several notions concerning the feasi- . cial time, and the pride of student achieve- bility of sentence combining exercises and sentence expansion activities delivered by the ment was evident. microcomputer. Of specific interest were growth in syntactic maturity in the writing and reading comprehension realms, the motiva- Conclusions Due to the nature of small group or case tional quality of sentence combining instruc- study research, few statistical measures were tion delivered by microcomputer, and student employed, and the reader is cautioned against attitudes toward the technique. generalizing the results of this study. How- ever, general conclusions can be drawn rela- Due to funding constraints, microcomputer tive to this study and graphs of individual availability, and a desire to investigate the student achievement will be presented at NECC. feasibility of large-scale implementation of this technique, a series of case studies (N of While further research in this area is 1 design) was chosen. Several case studies of needed, it appears that this method of wri- students representing different ability lev- ting instruction can be used successfully els were needed to help determine the overall with upper elementary students. Although six effect of these sentence combining exercises. weeks of sporadic treatment may not be long Fifth grade students comprised the population. enough to draw final conclusions, initial re- Student scores on the Comprehensive Test of sults look positive. Fifth grade students in Basic Skills (CTBS), language and reading sub- this study were able to quickly learn both the tests, and ethnicity were used in drawing a word processing system chosen and the sentence random sample of matched pairs (names were combining techniques. Students in the group masked). The control and experimental pairs that used microcomputers tended to show a were chosen to reflect four students who were gradual increase in semantic maturity as meas- above grade level in their scores and four ured by T-units, or units of the "shortest students who were below grade level in their grammatically allowable sentences into which scores. Student pairs were then randomly as- written language can be segmented (Hunt, signed to either the control or experimental 1965)." In addition, students in the experi- group, so that each group was composed of two mental group were observed using structures above average and two below average students. not sec.,n in the writing of the control group (i.e., appositives, many complex sentences, Pretesting and posttesting consisted of a and items a series). Each of these struc- timed writing sample, a cloze-type reading tures had been presented in sentence combining comprehension test, and a semantic differen- exercises. A final analysis of the results of tial attitudinal measure designed to reflect reading comprehension growth is pending and attitudes toward: school, writing assignments, will be presented at NECC. writing revisions, reading, and the use of microcomputers for instruction. Perhaps the most notable change for the experimental group and the greatest differ- The Treatment ence between the groups was in theattitudinal Sentence combining lessons and exercises realm. Attitudes toward writing and writing were presented in forty-five ninute segments revision using the microcomputer were very twice weekly by the researcher. As a group, positive. Teachers visiting the computer cen- students discussed each new technique with the ter were often surprised by the intent with researcher and later with a student rertner. which the students tackled the sentence com- Each student was asked to do the lesson on an bining activities. Even the two students pre- Apple II Plus (48K) microcomputer; using the viously lableled as "below average" showed Apple Writer word processing pr:-.!gram. Strong's great enthusiasm for the project and were mo- method of open-style cluster rresentation was tivated not only to complete the sentence com- used. His text, Sentence Combining and Para- bining activities, but also to revise and ex- graph Building (1981) served as a methodologi- pand what they had previously written. Another cal model for the exercises. After the indication of the positive responses to the students had completed each combination and sessions was that students revised all aspects typed in "writeouts" (combinations), they were of their work, including content, punctuation, asked to expand the group of writeouts into a grammar, and even spelling. Student interest story. was high, and they were extremely reluctant to leave the sessions.This group viewed with The second lesson of each week served as a disdain the sessions at which both groups review of the =sentence combining technique (control and experimental) were asked to give presented earlier and a chance for revision of written samples using pen and paper. In other student stories. Each student received a words, they vigorously preferred using the mi- print-out o his combinations and resulting crocomputer as a writing tool over pen and pa- story. Student editing teams (a low ability per. An investigation of attitudinalratings student paired with a high ability student) toward school, writing instruction, writing read and edited each other's work. Each author revision, using the microcomputer, ease of REFERENCES use of the word processing system, and sentence combining' exercises was very encourag- BRADLEY, Virginia. Improving students' writing ing. with microcomputers. Language Arts, 1982, (October),59,7. It appears that not only is sentence com- COMBS, Warren. Some further effects and impli- bining delivered by microcomputer effective, cations of sentence-combining exercises for but it is also a popular form of instruction. the secondary language arts curriculum. Un- This added incentive encouraged students to published doctoral dissertation,University write, revise, and read what they and others of Minnesota, 1975. had written, making them particularly open to COMBS, Warren. Sentence-combiningpractice instruction. The researchers feel that the of conclusions of this study and the related im- aids reading comprehension. Journal Reading, 1977, 21,1,18-24. plications of these findings can be very im- pactful on language arts instruction and fu- CRONNELL, Bruce, ,et.al. Using microcomputers ture research. for composition instruction. Paper presented at the Conference on College Composi- Implications tion and Communication (Dallas, Tx., March, Although final results are pending, this 1981). research indicates several positive effects EVANECKO, Peter, et.al. An investigationof of the use of microcomputers for sentence com- the relationships between children's per- bining instruction. The motivational aspect in formance in written language andtheir rea- and of itself might prove highly useful in ding ability. Research in the Teachingof countering the negative feelings many students English (1975),8, 3, 315-326. hold toward writing ,and revisions. Students who wrote with the microcomputer tended to HUNT, Kellogg. Grammatical Structures Written of view the instruction and the writing positive- at Three Grade-levels. National Council ly. Their stories were generally more complex Teachers of English, Urbana,Illinois, 1965. than their matched pair in the control group. NEY, James W. A short historyof sentence-com- They enjoyed the sentence combining activi- bining: Its limitations and use. English ties, and by the end of the second treatment Education, 1980, 11, 3, 169-177. phase they were able to use fairly advanced combinations without prompts from the re- O'HARE, Frank. Sentence-combining: Improving searcher. Other positive side effects were ev- Students' Writing Without Formal Grammar ident. Absenteeism was non-existent throughout Instruction. Research Report No. 15, the study. Students became more cognizant of National Council of Teachers of English, how to dissect and read complex structures be- Urbana, Illinois, 1973. cause they knew how to 'build' them. Profi- PAPERT, Seymour. Mindstorms. Basic Books, ciency with the word processing system in- Inc.: New York, New York, 1980. creased quickly and the students became much SCHWARTZ, Mimi. Computers and the teaching of more efficient in their use of microcomputer. writing. Educational Technology, 1982, time. At this point they were able to concentrate more intently on their writing. (November), pp. 27-29. STERNGLASS, Marilyn. Sentence-combining and While enthusiastic about these results and the reading of sentences. College Composi- the implications they hold for language arts tion and Communication, 1980, 31, 3, pp. education, the researchers recognize that they 325-328. have simply scratched the surface. The possi- STOTSKY, Sandra. The role of writing in devel- bilities for the expanded use of microcompu- opmental reading. Journal of Reading, 1982, ter instruction in reading and writing are exciting and ever increasing. Word processing 25,4,pp. 330-340. in the classroom is not a panacea, but it does STOTSKY, S.L. Sentence-combining as a curricu- offer opportunities for growth in the language lar activity: Its effect on written lan- arts processes. The expansion of the pilot guage development and reading comprehen- study described here that is planned for the sion. Research in the Training of English, fall will help lay the groundwork in this in- 1975, 9, pp. 30-71. vestigation, and far-reaching results are anticipated. However, these studies can only STRONG, William. Close-up: Sentence-combining. serve to provide a knowledge base and spur in- Back to basics and beyond. English Journal, terest in this area. Further research needs to 1976, 65, 2, pp. 56-64. be done so that we may fully investigate the use of word processing microcomputers as tools WHITE, Regine, et.al. Reading, writing, and for structured writing and reading instruc- sentence-combining: The track record. tion. Teachers and researchers must be tempted Reading Improvement, 1980, 17, 3, pp. 226- to explore the applications of such an in- 232. structional program so that students might more fully realize the vast educational possibilities offered by the word processing microcomputer in the language arts classroom.

3 THE COMPUTER IN THE WRITING CLASS: PROBLEMS AND POTENTIALS

C. Daiute, P. O'Brien, A. Shield, S. Liff, P. Wright, S. Mazur, W. Jawitz

Teachers College, Columbia University

Writing teachers have recently become would learn the technology we need enthusiastic about the use of word processing to know, how to fly airplanes, and This for improving their students' writing. all that stuff. I think I would rule paper offers guidelines for introducing the world better than president Regan computers in the writing class. We note pre- rules the United States. I hate to liminary results of our research on the effects say it but its what I feel. of computers for developing writing skills, but our focus is on the important practical consi- Janie to improve drafts of her texts derations that make the computer a useful when she !rot: the compu- Unlike most writing tool rather than a problem. eleven yer c who have , has Axe tice revising their writing, Janie - quite few significant as well as superficial changes in the piece above. Janie did not simply read her text. She saw it more objectively than when she had first written it and read it over thinking "it was fine." She deleted words, moved sections, changed words, and added details, such as the President's name. She formed a new, more focused piece.

Revising behaviors like Janie's are not typical of most children. Using computers can stimulate children's revising activities, which many writing teachers feel are extremely important. Janie Nevertheless, using computers is not as easy as Draft using other writing instruments in the classroom.

This may sound pretty weird and sad The authors of this paper are all working on but sometimes I dream that the whole the Computers and Writing Project at Teachers world would blow up. Another person College, Columbia University, but we have varied and I would be the only survivcrs backgrounds and varied roles. We are researchers, and we together would start the teachers, and students who have used computers for world all over again. The other our own writing. Our research has taken place in person and me would rule the world. three school settings in Manhattan: at J.H.S. 118, We would learn all the technology the United Nations International School,. and we need to know, how to fly and Teachers College in the Department of Communica- airplanes, and all that stuff. We tion, Computing, and Technology in Education. would also learn to care for eachother. I think I would rule the Our overall goal of the research is to find world better than the president ways to make writing easier for children and adults rules the United States of America, alike. Our specific goals differ. The researchers I hate to say it but its what I on the project are focusing on the ways in which think. computers relieve physical and cognitive burdens that affect young writers as they compose and Revision revise. The teachers, however, are more interested in methods of implementing computers in their This may sound weird but I dream writing classes to free student writers to express that the whole world would blow up. their ideas more easily than they can with tradi- Another person and I would be the tional writing tools. only survivors. We would start the world all over again. We would rule The purpose of this paper is to suggest why the world together. We would also computers may be valuable tools to use in the care for eachother a great deal. We

3 writing class and to offer specific suggestions on Getting Started implementing them as writing tools. Using computers in the writing class takes money, time, and Teachers have expected that the dynamism and patience, so we would like to offer guidelines. precision of writing on the computer would encour- In addition to presenting the researchers' and age their students to reconsider the purpose and teachers' reasons for using computers, we dis- style of their papers when they revise because cuss the logistical problems involved; we offer they would not have to recopy, as well as to pay solutions to these problems by presenting our closer attention to details of spelling and gram- methods for overcoming them. Finally, we discuss matical mechanics. The following quote by Peggy preliminary results of the effects computers have O'Brien of the United Nations International School on the writing of school children. expresses some of teachers' expectations.

Research Goals As soon as I felt my own response to using the computer as a writing tool Our research goals have been to explore the ...I was on the look out for a way to cognitive causes of writing difficulties, espe- mo computers in my English classes. cially revising. The computer seemed the ideal When I discovered that you could do writing instrument to use because it offers the word processing with micros, I made possibility to reduce certain physical and cogni- every effort to get hold of one. The tive burdens that make revising difficult. big "pay off" I expected was the re- Computer word processing programs allow writers inforcement of all English skills. I to type text into the computer, and make small or theorized that there woald be no large changes by giving commands rather than by stopping the kids who already liked recopying. The program automatically incorpor- writing and kids whose slowness or ates each change into an updated, neatly-formatted lack of coordination kept them away version of the text. Thus, the dread of recopying from expressing the fantastic ideas does not discourage children from revising. they had, or who hated the messy page they always ended up with, would par- Word processing programs are also useful ticularly benefit from using compu- tools for stimulating revising because they -an ters. I also suspected that intro- be augmented with automatically-presented comments ducing the students to the range of that can guide writers in self-questioning stra- options available from rcformatting tegies, modeled on those that students have with would dramatically affect their teachers and peers. Conferencing -- discussing written work. texts with student writers -- has been found an Ms. O'Brien initiated the process on her own: effective method of instruction (Calkins, in Right from the beginning, I received press). We thought some form of conferencing on strong encouragement, support and the computer would be helpful because the writer involvement in future planning from would initiate it and would thus be in charge of the Dean of Studies at the school. the self-monitoring and rewriting processes. We I knew we had machines in the school were, in short, using the computer to show which were underutilized. By a cer- children how writers talk to themselves to improve tain amount of conspiring and begging, their texts. The program we developed to present I managed to get the TRS-80 for use in young writers with prompts to stimulate revising my classes and persuaded the school to is called Catch. buy Scripsit and a trolly to make the thing mobile. That got us started. At any time during or after writers compose We used the machine without a printer with the word processing program, they can give a for several months, putting pressure command to see a list of analyses and prompts on all the people we could to help us Catch gives to guide writers as to get a printer. In the meantime, I they revise. If a writer, for example, selects was encouraged by the Dean of Studies the "empty word" option, the program identifies to write a proposal for an Apple II unnecessary words such as "sort of" and "well." plus with disk drives, printers, and As these words are highlighted on the monitor, a monitors. I proposed teaching a prompt appears at the bottom of the screen. The course in language enrichment (making prompt that appears with "empty words" is "The a grademagazine) for the new academic highlighted words may not be necessary. Can you year which entailed the use of all make changes?" The writer can immediately make the technology available in the school, changes if he or she wishes. Writers using the and so managed to get my hands on the program are aware that it may identify words that Pets temporarily. Funding has been are empty in some contexts such as "well" in "Well, done from the school's overall budget. I will begin with my childhood," but not in others, such as "My first memories are of throwing pennies On the other hand, school administrators often into a wishing well." present the impetus for encouraging teachers to be part of the technological age.Arthur Shield of J.H.S. 118 describes the process:

23 The principal of my school asked me The teacher also has to have con- to teach a word processing course. tingency plans for times when one of The principal informed me that the machines does not function. In through the Division of Curriculum order to expedite the process of ser- and Instruction of the New York City vicing the microcomputers, I had to Board of Education, our school would learn whom to contact within the acquire five micro-computer systems, company when problems arise. to be specifically utilized for the improvement of writing skills. These Finally, knowledge of the hardware systems would be purchased with New is absolutely necessary. Often York State Umbrella Funds.The times, I had to determine which Principal was concerned about the piece of hardware was malfunctioning poor writing ability of students and needed servicing. When dealing and the traditional writing course with computers, the guiding precept had not been successful. Initially, is to become as knowledgeable as I was against teaching such a course possible, which in turn tends to for two reasons. First, I recog- avoid the waste of time. Students nized that for me change is diffi- have been learning to become inde- cult and adjusting to a different pendent when using the computers. set of conditions would be quite The goal is to have the students arduous. Secondly,I felt intimi- depend less upon the teacher when dated by computers. I had no pre- dealing with the practical matters vious experience with computers and relating to the computer. further realized that within a bureaucracy there would be little I strongly recommend that a teacher opportunity for me to adequately who undertakes computer instruction learn the intricacies of the hard- be well-prepared and well-versed re- ware, let alone the software for garding the programs that are selec- word processing. On the other ted for the word processing course. hand, I was ready and even anxious Access to a computer on a regular to undertake new teaching objectives. basis is fundamental in order to master the available software. Logistical Problems and Solutions Registering for college courses dealing specifically with computers, Managing Computers and the Writing Class particularly those related to the field, is vigorously recommended. Managing the computers, human and intellectual energies, and curriculum in the classroom is a Training Teachers major job. Arthur Shield has five Apple II plus machines with disk drives and one printer. His Teachers must be comfortable with all the We comments highlight the major problems and suggest tools their students use in the classroom. ways around them. have found that the teachers who use word processing programs for their own writing before they One must be extremely organized in introduce them in the classroom travel the smoo- order to be successful with the pro- thest road. First, adults and children have gram and having only five computers different styles of approaching and mastering and approximately twenty-five stu- computing tools, so they should probably not have dents in the classroom has created their first experiences at the same time (Daiute, various problems. There is never 1983). Second, teachers who use word processors enough preparation time. I have to themselves understand the benefits and drawbacks prepare lessons so that those stu- and often dream up'innovative applications to Finally, dents who are not on the computers their writing curriculum as well. are actively involved in a motivating teachers who have used programs extensively, know activity. Since the majority of the the quirks, bugs, and ways around them - which in the students had had absolutely no expo- occur in many programs at this early stage sure to computers,I had to teach the state of the art. Teachers who intend to use basics. A typical class period in- computers in their writing classes should thus volves my running back and forth spend several months training and practicing. within the classroom to deal with questions, bugs, etc., that continu- Training Students ally arise. I have had to interrupt several lessons at crucial points in In this project, research assistants order to answer'a question of the helped to prepare materials and to train students working on computers. You the students in using the editor.Research also have to make sure that you have assistants helped in copying the required number adequate supplies, such as printer of diskettes, preparing and copying necessary paper and diskettes, essential ingre- writing material, and helping the teacher to orga- dients for any word processing program. nize a large number of students onto the limited

If.

3,r) number of machines. For example, it was important However, if the instructors were not available, that the assistants work in conjunction with the the students would refer to their worksheets and teacher to ensure that two students in the same find their own solutions. class were not using the same diskette. Eventually, students called upon other mem- The research assistants also helped train the bers of the group to collaborate and work toward students to use the word processing program. They their own solutions. More group interaction was were available to answer any questions that the observed among students in the higher grade levels, students might have, and to deal with any system again perhaps due to the smaller group size. It crashes that might otherwise have frustrated a .lew is important to provide each student sufficient learner. time to work on the computer.

Of course, classroom teachers do not have Using the Computer college or graduate students available to assist for Meaningful Writing Activities them. This need not present a problem, for the students themselves act as assistants. We noted Children need meaningful topics and composing from our training session that students who were strategies so they can benefit from the physical familiar with a word processing program, or who and cognitive aids offered by the computer writing had had previous experience on computers, were tools. It is always important to present children very anxious and willing to help their fellow with good writing topics to enhance the develop- classmates who were having problems. A typing ment of writing skills. It is especially true tutor program was developed to familiarize stu- when they are learning to write on the computer. dents with the keyboard. Students were encouraged It is particularly important to present a topic to develop their typing skills since this would that is intrinsically motivating because writing facilitate their writing on the computer. Stu- on the computer is difficult at first. There are dents were exposed to the typing program for many new things to keep in mind, and if the child several weeks before being introduced to the must dwell on an uncomfortable topic or style, the word processing program, and although students writing task easily becomes overwhelming. did not learn to touch type, most students felt comfortable with the keyboard when the word pro- The child who is interested in the topic is cessing program was introduced. Students who much more likely to stick at it longer, produce were not comfortable with typing were encouraged more, and master the technique. Good topics seem to continue using the typing tutor program. to be those that have the child write in an informal style about something he or she feels strongly A text containing errors in spelling, word about. Opinion essays such as "Why I like/dislike repetition, word omissions, and organization was New York City" are good. This kind of topic seems copied onto the students' diskettes. A step-by- to provide enough structure, still allowing each step worksheet was created to lead the students child choices about content and style. through several stages from simple machine operation to word processing commands. The worksheet Other writing tasks that work well are introduced commands which ranged from the deletion autobiographies or stories about the child's group of one character to the more complex movement of a of friends. Most children will write comfortably block of text. on those two, but for some children they are too personal. When a child shows signs of such dis- Effectiveness comfort, allowing him to invent a fictional character or characters usually solves the problem. Students worked in groups of 3-5 at each computer for a 35 minute session. This apprach Effects of Computers on Writing: Preliminary Results 1109 MEHRtinth thh rreedom to experiment within a structured framework, We used one practice assignment to introduce the students to Our preliminary findings about the effects the word processing commands. Our training of computers on the children's writing skills are methods appeared to be more effective for the based on classroom observations and analyses of higher grade levels (i.e. 8th and 9th) than for the children's texts. An observer asks a student the lower grade (i.e., 7th). This seemed to for permission to watch while the student works on occur because the lower grades had larger classes, the computer. Occassionally, a student declines resulting in less computer time for each student. the request to be observed, but most are clearly There was also a noticeable difference in computer happy to display their work. The observer has a -student interaction. The lower grade students list of observation points from the student's often executed commands mechanically without overall enthusiasm, to his or her typing skills, regard to function and they rarely watched the and from the types of revisions made, to the over- screen when correcting errors. all quality of the text.

Groups were selected at random and asked to Our preliminary observations confirm three work together on the lesson. At first, students crucial elements in bringing children and computers were dependent upon the teacher or assistants for together. First, and in a sense most importantly, a quick solution to any difficulties that arose, the students have an overall positive attitude such as forgetting what key to press for a deletion. about the computer. They like the idea of having

25 computers in their classroom, and they are excited Effects of the computer depend on the about using them. For whatever reason, their writer's abilities. Even though the most skilled excitement serves to diminish the more frustrating young writers approach the computer with excite- aspects of working on the computer. Namely, the ment, they find the mechanics of word processing sheer mechanics of using a computer presents somewhat frustrating. They sometimes feel that some problems for children. For example, we writing with pen and paper is easier and more found that the students who type better (regard- reliable. Usually, enough of the advantages of less of method), seem to be more generally enthu- word processing are apparent to help them pass siastic about writing on the computer. Also, this phase. Once they master the skills of actually starting the writing process can be interacting with the computer, the tool helps tricky because of preliminary steps required them write and revise efficiently. to create and maintain a file, such as finding, writing, saving, and transferring text from the Less skilled writers tend to come to the diskette. computer with the view that writing is a laborious process. Their attitude towards writing is It is also clear that at least in the early generally negative and the computer does much stages of writing on the computer, students favor to improve it. They also approach the computer the less abstract word processing commands over with excitement and their intitial excitement the more abstract and powerful commands. For never seems to wane. instance, though they are aware of commands that move the cursor by a line, sentence, or even a In contrast to the majority of students, many paragraph at a time, the large majority of stu- teachers approach computers with fear. They feel dents move the cursor a character at a time for that many of their students already know more sixty or seventy strokes to get to a wordthey want. about computers than they do. They learn to use to correct in the previous sentence orparagraph. the machines and decide how they will be most However, it is important to remember during this beneficial to their students' education. But, first and often frustrating stage, that as stu- when they start with the computers in the class- dents become more familiar with the keyboard, room, they often find the extra logistic chores hardware, and editing commands, they more fre- overwhelming. Nevertheless, once the major class- quently experiment, and become comfortable with room management problem is solved, they become the higher editing commands. excited about the fact that their students find writing to be fun. Our preliminary experimental results suggest that the word processing program and Catch generally stimulated children to revise.

Most of the children did more revising with pen after they had used Catch for aboutfour weeks. Janie, however, was one of the children References who did more revising with the computer, but did not do as much when she used pen after havingused the computer and Catch as a guide. Calkins, L M Ltaaana ExaktLChild: QM till Teaching and Learning of Writing. Forthcoming. We have several interpretations of these Weak writers like Janie may find the new Daiute, C. Computers and writing. Reading,Mass.: results. Addison-Wesley Publishing Co" in press. tool makes writing more comfortable. Often weak writers simply have not read or written very much, so they have not developed a styleand do not have text models to imitate. Children with strong writing skills, on the other hand, who have written extensivelyhave developed a sense about writing. Such writers have also become comfortable with a traditional writing instrument, usually the pencil or pen. The computer may offer them less relief as a writing instrument until they are as comfortable using it as they are with pencil or pen -- the instrument they had been using all along. Thus, initially the computer may help them less than it does children who had not become comfortable with other writing instruments, and it takes longer for experienced young writers to benefit greatly from its capacities. Similarly, good writers bring composing and revising strategies to their first use of the computer, so prompts like those in Catch may at first have less impact.

26 A HYBRID HUMANITIES APPLICATIONS COURSE: "THE COMPUTER IN LITERATURE" AND "LITERATURE IN THE COMPUTER"

by Rudy S. Spraycar

Data Processing Department United States Fidelity and Guaranty Company Baltimore, Maryland 21203

Abstract undergraduate instruction. The segment of this course entitled "Literature in the Computer" This paper reviews the development of an inter- surveyed current applications of the computer in disciplinary course whose aims were threefold: 1) literary analysis and the preparation of literary to use science fiction and non-fiction as vehicles research tools; the course culminated in student for discussion of the computer's impact upon society team projects in this area. and societal attitudes toward the computer; 2) after reviewing programming, to teach computer-aided tech- Background niques of literary analysis; and 3) to require the class to undertake team projects in analyzing lan- The gradual acceptance over the past decade or guage or literature with the computer; most teams two of science fiction as legitimate English included students both a) familiar with the computer curriculum content needs no rehearsal here. I would but innocent of literary analysis and b) familiar only note that the originality of this course with literature but naive about the computer. consisted largely in its hybrid approach, stressing 1) the mutual interaction of computers and literature, and 2) successful communication between devotees of both literary and computing cults Introduction within a single classroom.

A curriculum task force report prepared on the Slower initially, but recently booming, has Louisiana State University campus a few years ago been the recognition on the part of humanities noted the growing need for disciplines besides scholars at large of the value of instructing engineering and computer science to offer courses students in computer methods for literary research. about the computer. In particular, the social 12-15 To be sure, such pioneers as Robert L. Oakman impact of the computer needs to be addressed from a recognized as early as 1968 the need to teach such variety of disciplinary perspectives. Having taught research methods, as distinguished from the use of an English department course in science fiction at the computer in computer-assisted instruction (CAI), L.S.U. for some years, I had noticed how effective which has encountered less pedagogical resistance, science fiction is in stimulating class discussion owing to the celebrated crisis in the teaching of about the man-machine interface in general, and about man's relations with the robot and the English composition. computer in particular. Thus it seemed to me that The recent publication of textbooks in an English department course in computing ought to 16 include a unit of readings in science fiction and computer-aided literary research by Oakman and selected non-fiction works that examine 1) the 10has stimulated interest in the area. computer's impact on society and 2) the panoply of Susan Hockey Nevertheless, the recent third edition of Robert D. current attitudes toward computers. This portion of the experimental course that I deN,sloped and taught Altick's1very fine and otherwise comprehensive at L.S.U. in the fall of 1981 was entitled "The textbook on literary research methods devotes but Computer in Literature." five pages exclusively to computer methods, despite the dust jacket's promise of the book's expanded On the other hand, there has been great treatment of such methods. If resistance to interest in the application of the computer to the interdisciplinary research methods remains among development of traditional tools for literary humanists, perhaps a hybrid course like the one research, such as concordances, indices, and described here may prove a reasonable effort to bibliographies. At first, however, only a few bridge the gap of understanding. pioneers recognized the machine's potential to extend radically the application of quantitative and Course Structure statistical research methods to problems of authorship and attribution and to fundamental The readings included two novels, several research into the character of literary style. In short stories, an anthology of computer-related many respects, this field is now approaching fiction and non-fiction, and two textbooks on maturity and is a suitable subject for graduate and programming and literary computing (see Course

27 4 Materials, below). The research projects brought together two kinds of students: those without previous The Coputer in Literature experience with the computer helped to define the problem, design an algorithm, and participate in Non-fiction and science fiction materialswere data entry when necessary; those who had suitable used to define the man-machine relationship. The prior experience developed the program, generally computer's inpact on society was treated both in receiving more credit for their participation in terms of changes in the life of the individual and the projects. The team's joint development of the broad social and public policy issues. Of course, approach served two purposes: 1) it provided a these two levels of the problem cannot, in the end, sound basis for demystifying the computer in order be separated: such issues as privacy, whether jobs to critique social myths about it; and 2) it gave are created or lost (with some reference to the both the knowledgeable and uninitiated students an parallel issue of robotics), and military, social, opportunity to be a part of a "user/analyst" and industrial impacts are often of public relationship. The latter experience may have been consequence precisely because of their effects upon the most valuable portion of the course for the individuals. The non-fiction readings stressed the computer science majors who enrolled. present potentials and possible threats posed by information banks, networking, and such; the science Course Materials fiction readings tended to force speculation about the future by extrapolating one or another trend to Several science fiction anthologies devoted to its logical extreme. fiction about the computer were out of print. The 21 Ironically, the readings in science fiction, Van Tassel anthology offers a wide range of often obsessed with technology, provided some of fiction and non-fiction, but it is largely out of the best examples of the paranoid view of the date. Students and instructor alike, however, computer. As is the case with robots, computers supplemented this anthology with up-to-date articles are more threatening the more they are perceived as culled from current popular magazines and newspapers; human-like, for through the ages what is almost, many of these materials were-presented in the course but not quite, human has often been regarded as of the oral reports. monstrous. Out of the critique of myths about computers emerged an understanding of both the Particularly useful were two science fiction limitations and the true potentials of the computer, stories about the survival of man and the computer after a nuclear holocaust:Roger Zelazny's "For a leading into a brief account of how the machine 22 works. Breath I Tarry, a charmingly humorous, optimistic view, and Harlan Ellison's "I Have No Mouth, and I Literature in the Computer Must Scream, "7 a fiendishly pessimistic vision. Literary data processing was thus introduced as an object lesson in how the computer can be used. A good introduction to programming and the The course then took up in some detail the problems concept of the algorithm was provided by Richard in literary research that have proven most amenable Conway's Programming for Poets:A Gentle 3 to computer assistance. Computer-aided instruction Introduction Using PL/I; the book is also (CAI) was treated briefly, using as an example the available in versions using Basic, Fortran, and course authoring and instructional system that I 4-6 designed, wrote, and tested at L.S.U., CALAIS Pascal. I chose PL/I because it is a relatively (Computer-Assisted Language Arts Instructional widely implemented language, especially at academic 19-20 IBM installations, and is amenable to string System). Finally, we glanced briefly at the processing. (However, see discussion below of frontiers of both computer-assisted statistical revision of the course.) analysis of literary style and artificial 16 10 intelligence. Throughout the first two units of Both Oakman , s and Hockey's texts on the course, each student presented an oral report literary computing are excellent; because my on some aspect of either social attitudes toward institution had an IBM system, and Oakman's book is the computer or computer-aided literary research. somewhat IBM-specific, I chose it. The students then formed research teams to define problems in literary analysis and wrote programs to Finally, many science fiction novels would solve them. serve well for the final synthesis of the course. 9 Synthesis I chose Heinlein's The Moon Is a Harsh Mistress 2 and Brunner's The Shockwave Rider. The former at The study and discussion of two science fiction once humanizes the computer and characterizes it as novels about the computer (while the teams worked on a tool that has become a friend. The latter their projects) provided an opportunity to dramatizes the potential hazard::: of widespread synthesize the main themes of the course and to sort networking, then falls back upon the conventional out myth and fantasy from the issues surrounding the moral position that the final responsibility for relation between computer and society. In man's tools and his social organizations, especially particular, the students attemptee to decide whether bureaucracies, always rests with the individual; the computer is a monster, a friend, or simply a here a brave hero saves mankind from a bureaucratic tool. behemoth objectified as a computer, network. Other

28 42 science fiction novels could provide grist for Most popular (indicated by the lower numbers) biases other than my own, of course. were the social issues topic, the science fiction short stories (not the novels), and the special Syllabus projects, in that order. All the other items were ranked favorably except the Conway book, about I append a course outline; for Conway's PL/I, which the students were, on the average, neutral. 8 one might substitute Snobo14. This neutrality was reflected in the students' responses when they were asked whether they would Student Projects have preferred a standard programming text: yes, 50 percent; no, 43 percent; not sure, 7 percent. One group compared samples from Robert 49 percent favored substituting for PL/I a string- Heinlein's novels, written between 1939 and 1980, in manipulation language, Snobol4, while 20 percent terms of twenty-two variables. Significant changes disagreed, and 33 percent were uncertain. were observed: sentence length increased over the years, as did the proportions of the words in the Finally, 93 percent of the respondents favored text that were enclosed within quotation marks. - the addition of a one-credit laboratory period for The students interpreted the latter feature as a instruction in programming and a start in hands-on sign of Heinlein's improved ability or desire to experience for those who needed it. characterize through what a character said rather than through authorial comment. Revision of the Course Another team drew five samples of journalistic Unfortunately, the only course number available prose written in 1961 from the Brown University in my department for such an experimental course was 11 Standard Corpus of American English; each word in at the sophomore level; of the sixteen students who the Corpus has already been labelled or tagged completed the course, eleven were upperclassmen. I grammatically. The group confirmed the professional recommended that the course be formally instituted rule of thumb that average word length in newspapers at the senior level, with enrollment open to is about five letters (their samples ranged from 4.7 graduate students. While most of the projects were to 5.1 characters per word). The group failed to completed satisfactorily, the aims of the special find significant patterns in distribution of parts team projects outlined above would be better of speech either between samples or within an achieved if all the students brought to the course individual news story as a function of "pyramidal" more special competence in either programming or journalistic style. potential applications, e.g., literature, music, etc. Students without prerequisite courses in A third group interested in computer-composed computer science should be required to enroll in the poetry rejected a purely random approach to word one-credit laboratory that the students favored so selection in favor of a program with conversational strongly. The assignment of science fiction readings natural - language. prompts for parameters of verse as an occasion for discussion of social issues should form, theme, end-rhymes, and so forth. Finally, one be retained, as should the oral reports and, above student wrote a concordance-generating program, and all, the special team projects. another worked at debugging and refining the EYEBALL literary analysis software package, written in Appendix 17-18 Fortran, for local use. Course Outline: Literature and the Computer Student Response Week Readings The students were asked on a questionnaire to I. The Computer in Literature rank each component of the course on a scale of one (must keep), two, three (neutral), four, andfive 1 Introduction (drop it). The results were as follows: 2 Van Tassell, sections 1-4 3 Van Tassell, sections 6-9 Table 1, Student Questionnaire. 4 Science Fiction Short Stories Averaged Rankings of Course Materials. II. Literature in the Computer Social Issues 1.57 5 Oakman, chapters 1-2 Van Tassel 2.42 6 Oakman, chapter 3; Conway, part I, chapters 1-2 7 Conway, part I, chapters 3-6 Science Fiction Short Stories 1.67 8 Conway, part I, chapters 7 & 9; mid-term exam. Heinlein Novel 2.33 9 Van Tassel, section 5; Conway, part II, chapter 1 10 Conway, part II, chapters 2 :2 4; Oakman, chapt. 4 Brunner Novel 2.90 11 Conway, part II, chapter 5; Oakman, chapters 5-6 Computer Methods. 2.33 12 Oakman, chapters 7-8

Oakman 2.80 III. Synthesis Conway 3.00 13 Heinlein, The Moon Is a Harsh Mistress Oral Reports 2.73 14 Brunner, The Shockwave Rider 15 Results of Special Projects; Conclusion Special Projects, 1.73

29 References 17. Ross, D. and Rasche, R. H."EYEBALL: A Computer Program for Description of Style," Computers

1. Altick, R. D. The Art of Literary Research. and the Humanities, 6 (1972), 213-21. Third ed., rev. J. J. Fenstermaker. New York: Norton, 1981. 18. Ross, D. and Rasche, R. H. User's Instructions for EYEBALL. Rev. ed. Minneapolis:University 2. Brunner, J. The Shockwave Rider. New York: of Minnesota English Department, 1976. Ballantine, 1976. 19. Spraycar, R. "CALAIS/Teach: The Lesson-Writing 3. Conway, R. Programming for Poets: A Gentle Component of a Computer-Assisted Language Arts Introduction Using PL/I. Cambridge, MA: Instruction System," 1982 ASEE Annual Conference

Winthrop, 1978. . Proceedings, 247-54.

4. Conway, R. and Archer, J. Programming for Poets: 20. Spraycar, R. "From CALAIS to DOVER: Using A Gentle Introduction Using FORTRAN, with WATFIV. Computer-Assisted Programmed Instruction to Cambridge, MA: Winthrop, 1978. Gather Data on Composition Deficiencies Automatically," Proceedings of the ASIS Annual 5. Conway, R. and Archer, J.Programming for Poets: Meeting, 19 (1982) (in press). A Gentle Introduction Using BASIC. Cambridge, MA: Winthrop, 1979. 21. Van Tassel, D. The Compleat Computer.... Chicago and Palo Alto: Science Research 6. Conway, R., Archer, J., and Conway, R. Associates, 1976. Programming_ for Poets: A Gentle Introduction Using PASCAL. Cambridge, MA: Winthrop, 1980. 22. Zelazny, R. "For a Breath I Tarry," in Spinrad, N., ed., Modern Science Fiction, Garden City, 7. Ellison, H. "I Have No Mouth, and I Must NY: Anchor Doubleday, 1974. Scream," in Silverberg, R., ed., The Mirror of Infinity, New York: Harper and Row, 1973.

8. Griswold, R. E., Poage, J. F., and Polonski, I. P. The SNOBOL4 Programming Language. Second ed. Englewood Cliffs, NJ: Prentice-Hall, 1971.

9. Heinlein, R. The Moon Is a Harsh Mistress. New York: Berkley, 1968.

10. Hockey, S. A Guide to Computer Applications in the Humanities. Baltimore: Johns Hopkins, 1980.

11. KuEera, H. and Francis, W. N.Computational Analysis of Present-Day American English. Providence, RI: Brown University Press, 1967.

12. Oakman, R. L. "Computer Methods for Humanities Research: An Interdisciplinary Approach at South Carolina,"Proceedings of the IBM Symposium on Introducing the Computer into the Humanities, June 30-July 2, 1969, Poughkeepsie, NY:. IBM, 1969.

13. Oakman, R. L. "Computers for the Humanities," Humanities in the South: Newsletter of the Southern Humanities Conference, 35 (Spring 1972), 4 & 10.

14. Oakman, R. L. "Computer Education for the Humanities:Multiple Possibilities at the University of South Carolina," Proceedings of a Fourth Conference on Computers in the Undergraduate Curricula, June 18-20, 1973, Claremont, CA: The Claremont Colleges, 1973.

15. Oakman, R. L. "A Videotape Course for Computer Education in the Humanities," Computers and the Humanities, 9 (1975), 123-26.

16. Oakman, . L. Computer Methods for Literary Research. Columbia: University of South Carolina Press, 1980.

30 4J The DISC Project

Shelley Yorke Rose Carol Klenow Instructional Computing Oakland Schools Pontiac, MI 48054

ABSTRACT The DISC (Documentation and Integration of superintendents and curriculum directors Software into the Classroom) Project, selected potential DISC participants using administered by the IICD (Interactive and project criteria (computer literacy, Instructional Computing Department) of ability to load and run a program, access Oakland Schools in Pontiac, Michigan has to a classroommicrocomputer, etc.) who, produced the DISCCompendium, a collection after an orientation session, attended four of 91 software evaluations and DISC workshops. DISC participants were documentation. The 565 page Compendium trained in theuse of MicroSift evaluation contains materials for the PET, APPLE, form and a format for the development of TRS-80, and Atari. Through DISC, 90 classroom viable support materials which Oakland County educators were trained in was loosely based on MECC documentation. evaluation techniques using an adapted Their efforts inevaluating, pilot testing MicroSift form and MECC documentation and creating support materials for selected format. Combining these tools has provided commercial programs were rewarded with teachers using the programs contained in release time and the programs they the Compendium with a means to better evaluated. implement the commercial programs towhich To ensure the continuing contribution of they have access. the DISC Project tothe area of commercial In addition to the Compendium, the DISC software evaluation and support materials Project has produced a DISC Training Manual development, the IICD is training and Project Reportwhich details each of additional Oakland County educators using the four workshops and includes presenters' the DISC Model. Resulting evaluations are outlines, hand-outs and transparency being gathered and published in the IICD's masters. Suggested session outlines and newsletter "Remote Notes". Software activities are include in the Training publishers are being contacted with DISC Manual. results in the hopes that they will begin The Disc Project was developed by the incorporating teacher needs into their IICD during the 1981-1982 school year under documentation design. a Title IV-C grant for $15,000. District

31 4 6 COMPUTING FOR THE LEARNING DISABLED OR HANDICAPPED

Rita Horan Warren R. Brown Mary Russo Dr. Nancy Jones Sharon Smaldino Partick Schloss

ABSTRACT: Using LOGO with Learning Disabled This study proposes to use LOGO as a tool to Students develop problem solving skills for Learning Disabled children. Tasks initiated either by Rita Horan, 10 Bayard La Apt 5, Princeton, NJ 08540 student or teacher will be defined and each student will finish the task within a certain amount of The Learning Disabled child has been typically time. Problem solving strategies based upon those described in the literature as having poor problem used in cognitive modification will be used as well solving skills.Other descriptors of Learning as LOGO. Disabled children include such terms as: This study will compare three groups. The inpulsive, deficient in selective attention, an subjects will be Learning Disabled students inactive learner, and users of deficient strategies ramdomly assigned to the three groups.One group in solving problems. will receive LOGO instruction with a strategy based Much of the research on problem solving has on cognitive modification techniques for problem focused or whether or not such children possess the solving. One group will receive LOGO instruction underlying skills or abilities to solve problems. without problem solving strategies. The thrid Studies have shown that learning disabled children group will receive neither LOGO nor problem solving have the ability to acquire and store information strategies. Each group will be give a pretest but lack spontaneous access to these processes. prior to instruction involving a problem solving And furthermore they (Learning Disabled) seem to task and a posttest problem solving task upon have an inability to generalize a previously completion of instruction.A comparison will be learned problem solving strategy to a new problem. made of the three groups based on pretest and Scientists and researchers in the data posttest scores to determine whether there is a processing discipline have been studying the difference between the groups' abilities to solve underlying processes of problem solving and refer problems. to the process as information processing. The independent variable will be the method of Educators and researchers in the education instruction or noninstruction.The dependent disciplines address problem solving as a cognitive variable will be the posttest scores. function which can be modified and trained to. This study hopes to answer the following Cognitive modification research has provided questions: Will Learning Disabled students who educators with procedures that have shown some receive LOGO instruction with problem solving durable and generalizable effects for Learning strategies differ from those who receive only LOGO Disabled children. Specifically they have shown instruction? Will students who receive LOGO success with producing behavioral changes in instruction differ from those who do not receive hyperactive and impulsive children and children who LOGO instruction? have difficulties attending to school tasks. During the 1970's a group of researchers at MIT developed a language called LOGO or Turtle ABSTRACT: Project CAISH Second Year Update Geometry. This language was specifically desgined to be something children could make sense of, to be Warren R. Brown, School Board of Sarasota County, something that would resonate with their sense of 3450 Gogio Road, Sarasota,FL 33580 what is important: in order to learn something, first make sense of it.Turtle Geometry is Project CAISH (Computer Assisted Instruction learnable because it is syntonic. And, it is an and Support for the Handicapped) is currently aid to learning other things because it encourages focusing on the needs of the mentally and the conscious, deliberate use of problem solving emotionally handicapped population, ages 3-21. and mathematical strategies.The language is Continuing support is being provided for graphic oriented and users are encouraged to choose orthopedically handicapped students. The project a project (problem) and through commonly used uses systematic methods to provide microcomputer English commands (back, forward, left, right) systems,. software and courseware, and specialized create the steps which will solve the problem or hardware interfacing to address educational complete the project.When the child hits a snag objectives of students that are specified by the he/she is encouraged to use their own knowledge of teachers and specialists involved. the commands and previously learned information to New interface devices will be demonstrated to fix the snag. simulate student interaction with the

32 4 microcomputers. New and modified software will be ABSTRACT: Relative Efect of Microcomputer demonstrated, to include single switch control of Instruction and Teacher Directed Instruction on the LOGO graphics, drill and practice of Blissymbols, Performance of Hearing Impaired and Normal Hearing and other MECC software modifications. Students A description of a 15 station microcomputer laboratory for mentally and emotionally handicapped Sharon E. Smaldino, Patrick J. Schloss, TriCounty students will be given. Security, accessibility, Special Education District, Dewey Building, 1000 N. adaptive device needs, and software recommendations Main, Anna, IL 62906 will be discussed. Anecdotal summaries of student uses of the microcomputer systems will also be Scope and Purpose: With the increased presented. interest in the use of microcomputers in the One of Project CAISH's goals is dissemination schools and with their value as an instructional of information. In this regard, both hardware and tool still in infancy, the intent of this study was software experience can be useful to others in the to raise the issue of time of instruction in its educational community. A 46page Interim Report relation to performance of the student. A major will be available for distribution. assumption supporting the use of microcomputers is that teachers' instructional time is limited and the use of the computer may serve to augment and ABSTRACT: Project S.O.S. expand that instructional time. To date, this Mary Russo, Project Coordinator, Dr. Nancy Jones, assumption has not been empirically validated. Media Specialist, The School Committee of Boston, Comparing teacher direct instructional time with 75 New Dudley Street, Boston, MA 02119 computer instructional time with the performance of the students as the dependent variable should Project S.O.S. (Support for Occupational provide educators with guidance for judicial use of Students) at the Humphrey Occupational Resource microcomputers. Center uses computerassisted instruction to help The unique learning and behavioral secondary level handicapped minority students characteristics of hearing impaired students makes develop the competencies in math, reading, language them particularly germane to the study of relative a'ts and problem solving needed for their instructional efficiency. Traditionally, the occupational training programs. Using teachers of the deaf have experienced the need to minicomputers donated by Digital Equipment be highly redundant in their instruction of new Corporation and Apple microcomputers, students material. The verbal nature of instruction makes learn basic skills while developing the computer teaching time lengthy and student performance a awareness needed in the increasingly computerized poor indication of comprehension. This study was world of work. With the increasing use of to examine the efficiency of time of instruction by microcomputers in business and industry for basic the teacher or the microcomputer. The performance tasks, such as storage of typing in a word of the deaf students compared with the time of processing system or training programs for brake instruction will support the instructional repair in the automotive industry, efficiency hypothesis. computerassisted instuction is a bridge between Method: The subjects of the study consisted basic skills development and occupational training. of students in the school who had minimal exposure The content of thirty instructional software to computer assisted instruction prior to the programs was aligned with a set of 104 basic skills study. The subjects of the investigation ranged in competencies identified as necessary for the age between 13 and 19. There were two categories occupational training programs at the Center. of students, those with normal hearing and those Students take a computerized survey test and are with hearing loss averages in their better ear scheduled for two 40 minute sessions per week based ranging from 60dB to 105dB. Six of the subjects in on their performanbce and the number of sessions the hearing impaired group were male and six were required for task completion. The goal of the female. The normal hearing group had no identified Project is the placement of 240 handicapped loss of hearing. Four of the subjects were female minority students into skills training jobs and eight were male. involving computer utilization. The procedure was conducted over a fiveday Students are selected for the Project based on period with daily instructional segments of ten their scores on the Metropolitan Achievement Test: minutes in length. The computer program utilized Reading and Mathematics, the Massachusetts Test of for this instructional format was the unit "Hurkle" Basic Skills, referrals by instructors, and from the ElementaryVolume 1diskette developed by selfreferrals. The course of instruction, the Minnesota Education Computing Consortium. An designed to integratethe achievement of basic Apple IIplus microcomputer with a green skills with occupational training, computer phosphorous screen was the only equipment employed. awareness and employability skills, is The teacher direct instructional material individualized through the use of student coordinated with the material in the microcomputer contracts. Further Project development includes program. the development of an item pool for computerized Each student was seen by the same teacher for assessment specifically related to occupational a tenminute instructional period. The groups of training programs. students were randomly split into those who would receive instruction directly from the teacher and those who only received instruction on the computer. A total of four groups were identified: the hearing impaired receiving teacher directed

33 instruction (HITD); hearing inpaired receiving microcomputer instruction (HI H); normal hearing receiving teacher directed instruction (N-1D); and normal hearing receiving microcomputer instruction (NH). A script was used by the teacher to keep the instructional presentation for all groups the same throughout all conditions. Each student was given an opportunity to learn how to use the microcomputer by playing a hangman game, written by Wendell Bitter. Familiar spelling words were used. The students receiving microcomputer instruction were instructed that they would be learning new math material and to attend closely to the computer's instruction. Those students receiving the teacher directed instruction were told that to help them better understand material to be utilized on the computer, they would benefit from some instruction from the teachers. Each instructional period was timed. Both the number of guesses and the total number of trials in each instructional segment were tabulated.A 2x2 ANOVA technique contrasting the rate of guesses and correct responses for the four identified groups will be conducted.

34 A GUIDE FOR THE PURCHASE OF A COMPUTERSYSTEM FOR A TWO-YEAR CAMPUS

by Laurena A. Burk

Department of Systems Analysis Miami University-Hamilton, Hamilton, Ohio45011

an integral part of virtually every degree ABSTRACT program, is no easy task. The following paper therefore suggests a strategy that The increasing number of students will enable a campus to make an informed in computer-intensive programs, decision about purchasing computing together with a need for various equipment for its specific needs. levels of computing power in virtually all disciplines, has forced schools to GETTING READY allocate many thousands of dollars for the purchase of computing equipment. A key person or group of people This paper suggests a strategy for should be designated responsible for all selecting computing equipment or for specific parts of this decision appropriate to the campus' needs based making process. The total group, or on a data processing method known as project team, should be computer literate IPO. and have a vested interest in the utility of the computer system to be purchased. Otherwise, there is little hope that the campus will acquire the most practical and INTRODUCTION cost effective system. As the only Despite a general shortage of funds full-time instructor in the computer for education and declining enrollments in technology program, I uas the project many college programs, there has recently manager and team at MUM by default. I been an increase in students in technical managed the small but extremely busy Associate Degree programs.Most of these computing facility, and my students were programs require moderate to heavy use of the heaviest users of the facility. Thus computer equipment. At Miami I became the campus computing resource University-Hamilton, for example, the person. Computer Technology program alone has grown 30% in the last year. Students must Since not all team members will have learn to program in four computer a technical background, this discussion is languages: Fortran, Assembly, Cobol, and aimed at the intelligent but uninformed PL/1. The increasing number of students reader. While systems analysts have in computer-intensive programs, together written many highly technical papers on with a need for various levels of computer procurement, these may be computing power in virtually all technically sophisticated for some members disciplines, has forced schools to of the project team. A simple allocate many thousands of dollars for the process-oriented approach such as IPO and purchase of computing equipment. a few chapters of a good data processing text may be sufficient to focus the Since two-year schools are frequently involved educator's background.: These diverse in their degree and course texts are easily obtained throUgh most offerings, any computing system that is to libraries. A sample of appropriate texts serve the entire campus must have is listed in the bibliography.. extensive capabilities. At MUH, the computing needs are expanding as those THE OVERALL PROCESS' teachina both technical and non-technical courses discover the computer as a helpful The data processing method known as educational tool. As faculty recognize Input-Process-Output(IPO) is a useful the value of the computer to their technique for developing an orderly curricula, programming languages as well decision making strategy for a computer as word processing, graphics, and system purchase. Since an If0 chart specialized software packages become a provides a clear picture of what a program must. Yet, finding and purchasing is to do, application programmers affordable equipment that will serve frequently find it helpful to prepare IPO faculty and students adequately, becoming charts when planning programming tasks.

35 5 Working out the logical steps in a program COLLECTING FACTS prior to actually writing and testing it on the computer saves time and prevents problems. If programming changes are I recommend beginning to gather facts required, the IPO charts provide the as soon as there is evidence new equ!pment documentation necessary to describe the will be needed in the near future. The tasks performed by the program logically, following tasks should be completed before enabling the programmer to update the contacting potential vendors: program quickly and easily. An IPO chart divides the work into three basic 1. Compile a WISH-LIST by seeking funcions: INPUT, PROCESS, and OUTPUT.The input from all computer users, INPUT section of the chart shows all of current and potential. A simple the data needed to run the program.The questionnaire can provide uniform PROCESS section describes the manner in facts. Put no cost restrictions which all of the inputs will be on wishes. manipulated. The OUTPUT section shows what the result of the process will be. 2. Separate the list into two parts: This approach can be used in similar a MUST-HAVE list and a fashion to select a computer system. NICE-TO-HAVE list.

First, begin gathering the facts 3. Convert the WISH-LIST into a list governing the selection of a computer of functions the computer system system and organize these into a useful should be able to perform. format for study and comparison. The Separate hardware and software result will be a gradual distillation of functions. information necessary for a practical decision. This collecting of facts is a 4. Prioritize these functions. continuous iterative process. The initial input will not provide enough information to make a final decision but will point A sample of questions is given in out what additional facts must be Table 1. These questions do not need to gathered, studied, and organized before a be extensive or statistically processed final decision can be made. The form of for a small school. Faculty are less an IPO chart is shown in Figure 1. likely to complete a lengthy questionnaire than a brief one. Additional faculty input can be sought as needed. If specialized needs are indicated, the faculty most closely associated with those needs should begin to research them. For example, if a faculty member foresees a IPO CHART need for computer-aided-design (CAD), he or she should begin to investigate PROGRAM: PREPARED BY: available options in CAD computing and DATE: should ultimately specify the CAD functions. INPUTS: OUTPUTS: Dividing the WISH-LIST into MUST-HAVE and NILE-TO-HAVE subcategories is necessary to avoid buying frills with no PROCESS DESCRIPTION: substance and to discourage buying glamorous but nonessential items. A digitizer tablet may appear to be essential to CAD functions, for example, but it is probably not essential for students and faculty who create drawings from scratch at a terminal. In fact, the digitizer would probably be used only to demonstrate how to enter an exising drawing into the computer via the An example of an abbreviated IPO chart digitizer. This does not mean necessarily often used by programmers when planning a that a digitizer is a frill but rather new program. It is changed as the input that it should have a lower priority than data, the output data, or the process a plotter in a CAD application. changes. The MUST-HAVE and NICE-TO-HAVE FIGURE 1 categories should be subdivided to show hardware and software needs. Hardware is

36 SAMPLE_OF QUESTIONS MUST-HAVE LIST HARDWARE FUNCTIONS 1. Do you. currently use the computer? 1. support 20 users 2. upgradeable to 60 on-line users a. sc. Ana.z are your r-rojec=mA meeds over the next tn-me -marm? 3. prograrmable ports 4. easy daily backup b. IfmD1what are your -ro,sct!d interests over the 5. output to large screentelevision next. three years? 6. dial-in capability c. How many students in your classes do you estimate would be involved in using the 7. graphics terminals computer? 8. adequate disk storage 2. If you are not currently a user, do you plan to use a computer in 9. 2 color plotter the next three years? 10. communicate with other computers a. If so, what type of computing for batch processing would you like to have access to? SOFTWAREFUNCTIONS b. How many students in your classes do you anticipate 1. User-friendly operating system would use the computer? 2. Compilers needed: TABLE 1 Fortran a. Cobol the electronic and mechanical equipment b. PL/I needed to support the specified functions. Software is the set of programs which 3. Ability to create graphs control the hardware and enable it to perform specified tasks. Separation of 4. Ability to program in interactive these hardware and software functions Basic makes it possible to choose the best type of hardware configuration for the campus 5. Financial software package needs. Once the general hardware configuration has been chosen, vendors and appropriate software can be sought. A TABLE 2 sample list of hardware and software functions, in order of priority, is shown in Tables 2 and 3.

curriculum, however, more computing power AVAILABLE HARDWARE OPTIONS is usually needed. Thus, the hardware option chosen for MUH was the multi-user minicomputer system.Only minicomputer Two major hardware options are vendors were sought. If the major available to schools purchasing computing hardware option is not easily decided by equipment. The first is to develop a the project team, a checklist can be used, microcomputer lab with one or more brands similar to that found in Davis.* of microcomputers. The second is to purchase a multi-user system that supports all the major functions and to plan on using microcomputers as remote workstations or for special purposes. In William S. Davis, Systems a school where computing supports Analysis and Design A Structured non-technical programs, a microcomputer Approach (Reading, Mass.: Addison lab can be a viable hardware option. If a Wesley, 1983),p 277. computer-intensive major is part of the

37 CHOOSING VENDORS 5. Seek recommendations from other institutions. The first four tasks of the decision making strategy are actually preparation 6. Contact existing users' groups for approaching potential vendors. Since familiar with the equipment. fact gathering leads to meaningful organization and study of those facts, the organizing process produces functional Once the vendors have been chosen, specifications useful in choosing a vendor ask them to provide two general and in gathering important technical and configurations. One should inlude all of cost information from that vendor. the MUST-HAVES, the other should add the Comparing computer systems is easier and NICE-TO-HAVE hardware and software. Both more accurate when specific and detailed configurations should include itemized information is at hand. The two most pricing and monthly maintenance. critical steps in the decision making Supplying vendors with a form makes the process are selecting potential vendors responses uniform and easier to evaluate and choosing the right one. later. A sample list of hardware and software items derived from the functional Determining the practicality of specifications is given in TABLE 4. functions can be an overwhelming task. Priorities change when specific hardware, software, and price are considered. For example, a CAD-oriented system may be considered a MUST-HAVE. If projected CAD NICE-TO-HAVE LIST usage is significantly lower than demand for an application software-oriented HARDARE FUNCTIONS system, however, the CAD priority may be infeasible as a MUST-HAVE. CAD equipment 1. Color graphics terminals was projected to affect only one faculty member and twenty students at MUH, while 2. 8 color plotter twenty faculty, three degree granting programs and more than two hundred 3. Digitizer students per semester demanded software development capability. The CAD system 4. Additional disk storage had to be moved to a NICE-TO-HAVE priority because of high expense and lesser impact 5. Communicate with other computers within the school. Yet this shift in for interactive processing priority was not agreed upon until after the vendors contacted had provided 6. Ability to transfer files from at specific information . Potential vendors least one brand of microcomputer. can provide essential facts that will enhance the final decision.Failure to identify appropriate vendors and to gather SOFTWARE FUNCTIONS enough of these facts will hinder informed decision making. 1. CAD software At first glance, the number of 2. Database software vendors possibly meeting the campus' needs may be far too great to permit reasonable 3. Debugging software comparison. Although a total of no more than five vendors should be sought, it is 4. Electronic mail preferable to limit your vendor choice to three. There are six vendor 5. CAI packages characteristics to look for before calling any sales representatives: TABLE 3 1. Limit your choices to vendors with a sales and support office Additional data on each vendor can be within 100 miles. obtained from users of that vendor's equipment. Request lists of sites with 2. Check vendor support record and systems similar to the preferred reputation before contacting. configurations and visit these at the busiest possible times. Ask the operators 3. Verify each vendor's financial and system managers about performance, status. vendor support, ease of use, and response time. Note any problems mentioned and any 4. Choose vendors which are outstanding features, either positive or technologically up-to-date. negative.

38 Non-biased technical information can HARDWARE ITEM PURCHASE MONTHLY be obtained from Datapro reports. Datapro PRICE MAINTAINENCE publishes objective technical summaries of 1. Central all classes of computer systems and polls processor users of the equipment for their opinions. 41 These user summaries are 2. Communication to particularly noteworthy in ease of IBM 370 operation, maintenance service, and documentation. Documentation is the 3. Tape drive (45 vendor provided set of manuals which give ips) the user the directions needed to use the system components appropriate to the task. 4. Disk drive (200 Quality documentation is extremely megabyte important in an academic environment minimum) because of the variety of levels of faculty and student experience. Without it 5. Terminal (with the system will be greatly under-utilized. printer port and advance video) MEASURES OF EFFECTIVENESS: THE FINAL DECISION 6. graphics terminal After the input from users and vendors has been collected, the project 7. line printer team will has several alternative computer- (300 1pm) systems to choose from. Occasionally, feature of a computer system may make it 8. battery backup so far superior for a particular campus' purposes that it is the apparent choice. 9. modem (1200 Ordinarily, however, the key people baud) responsible for organizing and studying will find additional prioritizing 10. cable for 20 decisions necessary before they can decide terminals on a vendor. Choosing between vendors, when two or three appear to be able to SOFTWARE ITEM supply appropriate hardware and software within budget constraints, requires 1. Basic further careful measurement of system capabilities. The next step in the 2. Cobol decision making process, therefore, is to analyze these alternatives, using 3. PL/1 technical, economic and operational criteria. These criteria may be 4. Software for summarized as follows: communication to IBM 370 1. Technical criteria 5. Financial a. Functionality software package b. Level of technology of the TABLE 4 hardware and software Preliminary form for vendors to complete _ prior to bidding' Richness of available software

Upgrade costs should be investigated, d.. Security as well as purchase and monthly maintenance costs, since upgrading will 2. Economic criteria inevitably be necessary at some future time. Ask prospective vendors to include a. Purchase or lease cost projected upgrade costs on their information forms. Knowing the cost per b. Maintenance contract costs additional workstation, including cables, ports, and memory, is helpful in c. Upgrade costs estimating future costs and determining if the proposed configuration is adequate. If upgrade costs are too high, the proposed system is probably too small. Datapro Reports. Delran, New Jersey: NEVER consider a system that cannot be Datapro Research Corporation. upgraded.

39 3. Operational criteria system becomes unneccessarily obsolete without them. Documentation too falls a. Vendcr support under technical support. Verify that all manuals are readable. b. Security Although the type of computer system c. Implementation schedule security needed in an academic environment should be readily available on most multi-user systems, the question of The first technical measure is the adequate security merits careful scrutiny. vendor's response to the question, "How Passwords and the ability to prevent well does the system perform the needed students from accessing each others functions?". For example, two systems may programs are absolute minimums. Bemire supply a Fortran compiler, but one is the proposed system offers appropriate Fortran '66 and the other is Fortran '80. security. Examining security procedures Thus the second system performs that within another school Anvironment may give function better. additional insight into methods your institution can adopt. The level of technology of equipment is another important consideration. The The third operational criteria is the quality of education an institution implementation schedule. The neccesary provides is reflected in how up to date equipment should be installed in its equipment is. Better computing sufficient time to give faculty members equipment attracts more and better adequate preparation for classroom use. students and their skills are more Those faculty already familiar with marketable upon graduation if learned on computers will need one semester to state of the art equipment. In addition, adapt to a new computing environment. newer computing equipment is more reliable Introducing unfamiliar faculty to the and cheaper to maintain. New equipment computer as an educational tool will becomes obsolete quickly enough; saving a follow. If curriculum objectives are few dollars on older technology is penny, affected by the availability of a computer wise and pound foolish. system, then the implementation schedule should be a criteria in evaluating the The richness of software available on vendor proposals. the market should be considered at the time of purchase. As more faculty and Once it is apparent that two or three staff learn to use the computing resources of the vendors meet the campus' needs and at hand, and as the variety of measure up fairly well, request formal applications software increases, bids. Although this may delay delivery by additional software purchases will become a few weeks, it will most assuredly reduce cost effective investments. Our new the price or, bring a few of the computer system has not yet been unaffordables within reach. When delivered; yet those who expressed special requesting these bids be sure to demand a needs are already searching for additional complete installed system price. This software. includes delivery, installation, and bringing the system on-line .At this point, price can be the final, dadding Cost is the first economic measure. factor. Let the vendors know that' it is The number and quality4of functions not the only criteria for final selection. performed within budget limitations must The results of the bidding process can be be considered, as well as additional costs surprising. At MUM, the three vendors such as maintenance and upgrade. Further that were asked to bid were IBM, cost-related consideration is the Hewlett-Packard, and Digital Equipment availability of compatible equipment. Corporation (DEC). Only Hewlett-Packard Frequently, compatible equipment such as and DEC submitted bids. Both vendors terminals or printers, can be obtained measured up well, DEC's richness of more cheaply through a distributor. This software and newer technology winning the can be factored into cost of upgrade. purchase contract. Hewlett-Packard is currently offering free software to Smooth daily operation of a computer educational institutions, whiola makes them system depends heavily on vendor support. extremely competitive. The vendor's support record, both technical and maintenance, is a good The decision making process does not indication of how much the system will be end when the equipment is purchased and used. If technical problems can be ironed installed. Instead, as long as computing out easily, minimizing machine repair is seen as an integral part of the time, more people will use the machine. educational process, the effectiveness of Software updates are a must and should be the system must be maintained through included in software maintenance since a

40 continuous reassessment. This means project teams keep abreast of changes in campus' needs and new developments in the industry which can better meet those needs. Here again, the IPO process facilitates efficient decision making, building on existing knowlege to achieve consistent effectiveness.

BIBLIOGRAPHY

Davis, William S. Systems Analysis and Design. Reading, Mass.: Addison Wer:::1?.y Publishing Co., 1983. pp. 274-279

Davis, William S. Operating Systems A Systematic View. Reading, Mass.: Addison Wesley Publishing Co., 1983. Fitzgerald, Jerry; Fitzgerald, Ardra; Stallings, Warren. Fundamentals of Systems Analysis. New York: John Wiley and Sons, 1951. chapter 7 Weinberg, Victor. Structured Analysis. Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1980.

41 EXTENSIVE COMPUTER GRADING OF ID-INDIVIDUALIZED HOMEWORK PROBLEMS

M. J. Maron

Department of Applied Mathematics and Computer Science University of Louisville, KY 40292* at a desk, not at the computer terminal. It is only after solutions are Obtained that the student uses the grading program. A grading session con- Abstract sists of identification, submitting answers to some or all problems of a ringle assignment, This paper describes an efficient methodology grading, and getting a summary (Figure 2). It in which a computer is used to grade ID-individual- takes about 3 to 6 minutes to submit and grade ized homework problems on demand while keeping 16 to 18 answers. The program does the grading The methodology was records for these homeworks. by accessing the file of student IDs and checking developed at the University of Louisville and used Whether the submitted answers agree with the there with undergraduate courses in statics and "correct" ones calculated by an appropriate sub- Results so far indicate that numerical analysis. routine of the program to within a prescribed it can be an effective pedagogical tool for any accuracy. course in which homeworks have numerical answers. The student has a fixed time period (e.g. 10 Rationale days) to get the best possible mark on a given assi-gnment. During that time he or she may con- The following feature distinguishes this sult with the instructor and/or classmates to methodology from most existing schemes for cpm- correct and resubmit those answers that were puterized grading of individualized problems': wrong. Such a time limit effectively eliminates The student does not have to go to the computer the problem of procrastinatign that generally to get individualized problems. Instead, home- plagues "self-paced" schemes 4, while still giving work assignments are given in the "traditional" the student a chance to achieve proficiency on way, that is, on a single duplicated sheet each segment of the course. which is distributed to all students in the class (Figure 1). The individualization is The methodology described above ensures the imposed by the following simple device: following: Problems are given in terms of A, B, C, D, E, and F, where ABCDEF denotes a 6-digit student Each student does his or her own work; but stu- ID. For example, a problem may require the dents may consult with each other about how to student to evaluate get correct answers. (This type of collusion is considered as desirable.) Simply copying some- (3.AB)(5.CD + 12.F - E) one else's work (considered undesirable) does the student no good whatsoever. A student whose ID is 123597 will calculate The grading is uniformly fair and immediate. (3.12)(5.35 + 12.7 - 9) = 28.236 Rather than wait for a human grader, the student immediately knows which answers are wrong and Any other student, having a different ID, can begin to correct them at once. Answers can should oet a different answer. In this way, be re-submitted repeatedly if necessary until all students solve the same problem but no either the student is satisfied with the grade on two students should get the same answer. that assignment or the time limit for that assignment has expired. The 6-digit IDs are stored when the student files are created. Once this is done, there is The instructor can assign as much routine drill no further need to store or generate individualized as-deemed necessary without having to face the randomized numbers for each student for each tedium of grading it and recoreing the results. homework. Instead, human grading energy cal be expended where it is needed, namely on projects and assign- The problems are solved in the "traditional" ments which require judgment, analysis;\comparison, way, i.e., with pencil, paper, calculator, etc., etc.

1 *This material is based upon work supported by the There is no major crisis when the computer goes National Science Foundation under LOCI Grant No. down because the work is done away from the termi- 79-00864. TOT If necessary, a simple adjustment of the

42 1 1 DATE ASSIGNED ; January 24,1983 + 14 COURSE #1207 1 ASSIGNMENT #1 2 ---1!

CLASSLIST 01 ID= 1A1 BICIDIEIF1

1 1 1 1 1 1 1

REMINDERS: 1. Substitute digits of your ID for ABCDEF.

2. Unless otherwise indicated, use your calculator accuracy in all intermediate steps, storing intermediate values if possible.

3. SVJmit answers (Al, A2,...) to at least 1 5 1 significantdigits.

AnswersPoints

Q1. Let f(x) = z3 (50.E)x s x3 (50. )z. Al = 6

If the Newton Raphson (NR) method is used A2 = 6

with zo = 4.2CD = 4.2 then A3 = 6

A4 = 6 xi = Al:_, x2 a 1 A2 1.x3 =i,A3 1

whereas if X0 2 3.FB = 3. . then A5 = 6

6 xi = A4 1 . x2 = :A5 1, z3 = 1A6 1 A6 =

Q2. Let f(x) = z3 (50.E)x = z3 (50. )x.

If the Secant (SEC) Method is used with A7 = 6

z_i = 4.3DE = 4.3 and x0 = 4.2CD = 4.2 , A8 = 6

Ag = 6 xi a 1 AT 1,x2 s 1 A8 1.13 = 1A9 1

whereas if it is used with z_t 2.9E 2 2.9 A10= 6

and x0 = 3.FB = 3. , All= 6

rt w 1A10 1. x2 = 1All 1. z3 = 1Ail !. A122 6

Q3. Let xo = 3.D = 3. , z1 = /.0 = /. , and

z2 = 0.4F = 0.4_.

If the convergence is exactly linear, than A13= 6

x3 = 1 A14 1 A14= 8 iixklAxic =C1 a 1 A13 1and If the convergence is exactly quadratic, then A15= 6

Al6= 8 Axic/(&zk_1)2 = C2 : A15 1 and x3 = 11C16-1

[ Figure 1: Sample Computer Graded Assignment]

43 HI THERE. PLEASE ENTER SECTION 4, HOMEWORK *: 1. 1

ENTER CLASSLIST NUMBER, ID: 1, 123456

ENTER LAST NAME: MARON

DO YOU WANT TO 1) SUBMIT ANSWERS OR 2) GET SUMMARY? 1 OR 2? 1

REMINDER: HIT KEY TO SKIP A QUESTION CTRL/U TO REENTER AN ANSWER CTRL/Z TO DISCONTINUE 'Am' PROMPTS.

Al- OK A2= JK A3= OK A4= OK A5= OK COMAcarnag A6+ OK ger( ONDe'ilt-midts A7= OK sJerze r 4 A8= OK A9= 14.9 renuommt.

AlOm OK All .4125

Al2= OK Al3= OK Al4= OK Al5= 5.24448

Al6= 8.26305

Al7= 37.2045

PLEASE WAIT WHILE GRADING:

A9= 14.9000 All=0.412500 Al5= 5.24448 A16= 8.26305 A17= 37.2045

CORRECT ANSWERS:

A9 All Al5 Al6 All

SUMMARY FOR HOMEWORK 41

114 1 2 3 4 5 6 7 8 9 POINTS: 5 5 5 5 5 5 7 7 7

A4 10 11 1T 13 14 15 16 17 POINTS: 7 6 6 6 6 6 6 6 TnTAL POINTS:100 OUTOF A POSSIBLE100 WELL DONE!!!

ENTER ANY COMMENTS OR COMPLAINTS (THEN HIT). IF NONE, JUST HIT

? THANKS.I NEEDED THAT

Y'ALL'COME BACK SOON.

Figure 2:Illustrative session with grading p,rogralyI

44 program files can extend the grading period for able programming expertise can write a subroutine selected homeworks. for a new problem in 20-40 minutes by merely mimicking Figure 3. This is more challenging, Applicability more interesting, and certainly no less of a learning experience than grading the same prob- Homework problems and the associated subrou- lem(s) repeatedly for all students in one or more tines for grading them have been written for a classes. freshman level course in statics and for a junior level course in numerical analysis. However, Once the grading subroutines are written, a teacher the grading program can be used with any course can choose to re-use assignments of previous for which there are routine problems which have semesters without fear of students copying the numerical answers. This attribute characterizes homework of the previous semester. most -"core courses" in an engineering or natural science curriculum, and most quantitative courses Results in a social science or business curriculum. The grading program was used intensively with Desirable Operational Features three sections of ESC 205, a freshman-level statics course, and two sections of AMCS 207, a junior- Aside from the general benefits described in level numerical analysis course in the Spring 1983 the Rationale, the use of the computer grader in semester. All sections had an initial enrollment several courses simultaneously offers the fol- of about 35. Each week, two short computer-graded lowing operational benefits: homeworks (typically 1-8 answers) were assigned in ESC 205 and one longer computer graded assignment Student files for each course are easy to set up. (typically 12-18 answers) was assigned in AMCS 207. The information required for a student record The grading period for these assignments varied (viz. name and an ID) is precisely the rormation between ?'and 14 days. Although the final results provided by the registrar on alphabetical class- are not, yet available, some results are evident. lists. So this classlist information can be dumped onto a temporary file which can then be read into Pedagogical Effectiveness the appropriate fields of the grading program files by a simple utility. Only minor modifica- After eight weeks, 89 students in the three tions (late registrants, etc.) need be keyed in sections of ESC 205 and 68 students in the two manually. sections of AMCS 207 were still using the computer grader. The grades they received on the first The storage requirements are minimal. A single 15 homeworks in ESC 205 and the first 6 homeworks grading program can serve several courses. Since in AMCS 207 are tallied in Table 1. there are no randomized numbers for the computer to generate oo store for each homework, the program Grade Number Number and files ay-42, quite small. Object code for the Range in ESC 205 in AMCS 207 program requires 65 blocks (about 40K) of disk storage on a DEC 1090; the grading subroutines =100% 908 (68%) 351 (86%)

are generally short and should add no more than 90-99% 16 ( 1%) 13 (3%)

10-25 blocks per 3 semester hour course (Figure 3). 80-89% 37 ( 3%) 15 (4%)

With the use of bit packing, all necessary infor- 70-79% 47 ( 4%) 10 ( 2%)

mation for one student's performance on one assign- 60-69% 41 ( 3%) 7 ( 2%) ment is stored in a single word (which can be as up to 59% 286 (21%) 12 ( 3%) small as 32 bits). As a result, the file for a class of 35 students who are given 25 computer- Table 1 graded assignments requires only 20 blocks (about 12K) of disk storage on a DEC-1090. Evidently, most students were motivated to get the highest possible grade on each assignment. This Teachers of different sections of a particular is further supported by the fact that students course can assign homework independently. This used an average of about 4 to 8 grading sessions flexibility was written into the grading program to get these grades, and some students used more to ensure that instructors using it do not feel than 15 sessions for some homeworks! that someone else's "gimmick" is being forced upon them. Rather, each instructor can teach the This high degree of motivation resulted in course in the same order and with the same empha- many students coming to my office when they (and sis that he or she would use without the computer perhaps their friends as well) ran out of,ideas grader. This flexibility also makes it easy for -Bout how to do problems I an instructor to modify the syllabus from semes' defrom the obvious benefits` to semester if desired. ',,ihese encounters pointed out se,.,eal p+1!Aycrn- ce' "r °ed problems and some shortcomings in the Some of the time which student assistants current- graiiing subroutines. ly spend as graders can be spent programming grading subroutines. These subroutines can be Results of student questionnaires during an written without any knowledge of the design of the earlier trial with only AMCS 207 indicated that grading program. Indeed, a student with reason- (a) the students enjoyed and believed they

45 SUBROUTINE HW2 COMMON /HWK/ A, B, C, D, E, F, IPROB, ANS(9) C GO TO Cl, 1, 2), IPROB C C **** 01 AND 02 **** 1 Cl = 50. + E/10. X0 = 4.2 +.01*C +.001*D X00 = 4.3 +.01*D +.001*E IF (IPROB .E0. 1) X00 = X0 CALL ROOT(IPROBy X00, X0, 3, Cl, ANS, 0) X0 = 3. + .1*F + .01*B X00 = 2.9 + .01*E IF (IPROB .E0. 1) X00 = X0 CALL ROOT(IPROBy X00, X0, 3, Cl yANS, 3) RETURN C C **** 03 **** 2 X0 = 3. + D/10. X1 = 1. + C/10. X2 = .4 + F/100. DXO = X1 - X0 DX1 = X2 - X1 ANS(1) = DX1 /DX( ANS(2) = X2 + ANS(1)*DX1 ANS(3) = ANS(1) /DXO ANS(4) = X2 + ANS(3)*DX1**2 RETURN END C SUBROUTINE ROOT(Mg XPREV, X, ITER, Cl, ANS, IBUMP) DIMENSION ANS(9) FUNC(X) = (X*X - C1)*X DERIV(X) = 3.*X*X - Cl

FPREV = FUNC(XPREV) DO 50 I=1,ITER F = FUNC(X) IF (M .E0. 1) SLOPE = DERIV(X) IF (M .EQ. 2) SLOPE = (FPREV - F)/(XPREV - X) XPREV = X X = XPREV - F/SLOPE FPREV = F ANS(I+IBUMP) = X 50 CONTINUE RETURN END

(Figure3: Subroutine to grade assignment in Figure 1

46 learned from the computer assignments, and 6,, their attitude toward computers was more positive after the course, perhaps due at least in part to the computer gradingi.

Time Involved

The terminals used were connected to a DEC- 1090 via 300 baud lines. The average connect time over all sections seems to be around 18 minutes (total) per student for one assignment, or about 3 minutes per grading session. Most likely, initial sessions took longer and later "fixup" sessions were shorter. Grading sessions can be made somewhat shorter by using higher speed lines.

Another factor tha will reduce the total grading time is the removal or correction of problems that were poorly worded, ill-conceived, or improperly graded. (All of the ESC 205 problems and about half the AMCS 207 problems wer written about a week before they were assigned.) These caused many unnecessary grading sessions.

Cost Involved

At a university that already has remote terminals available to a host computer and makes no real dollar charge for instructional computer use, the only possible major expense is for the time spent preparing and debugging the subroutines for grading problems for a particular course. Since problems can be re-used effectively, and subroutines to grade new problems can be written by student assistants, this charge should be reasonable. Even if disk storage and connect time are billed to a real dollar account, the cost should compare favorably to the cost of paying a student the minimum wage to grade papers (with not nearly the effectiveness of the computer grader).

Summary and Conclusions

This paper described an ID-individualized methodology in which homework grading, but not learning, is done interactively at a termini The results so far indicate a high degree of both pedagogical and cost effectiveness. Although more extensive study is needed, these results suggest that for courses in which there are routine problems having numerical answers, this methodology may prove to be superior to both "traditional" hand-graded schemes and more extensive CAI schemes.

Bibliography

1. "Computer Aided Assignments in Electrical Engineering Education", W. J. Smolinski, ASEE Educational Research Methods Journal, Vol. 10, No. 3, (Spring 1978).

2. "Does PSI Work Because it is Self-Paced?", M. J. Maron and L.D. Tyler, Proceedings of 1976 Frontiers in Education Conference, Tucson, AZ.

3. "The Effects of Computerized Grading on Student Attitudes Toward Computers", m.J. Maron (To Appear).

47 62 AUTOMATIC SYLLABUS GENERATOR (ASG)

Asad Khailany, Marc B. Schubiner, A.M. VanderMolen

Department of Operations Research and Information Systems Eastern Michigan University, Ypsilanti, Michigan

ABSTRACT Many instructors are facing problems in preparing syllabi for multi-section courses. Often these The situation often arises in large educational are core courses, such as introduction to computer institutions where a course is taught in several science, information systems, mathematics, and sections by different instructors. This situation lends statistics. It is difficult for the various instructors itself to several problems. Among these problems are that of consistency of material taught in each to be consistentinteaching required course material established by the administration. Fre- of these sections and in subsequent semesters. Also quently, individual instructors prepare their own there is an increased work load on the secretarial syllabus for a section, leading to many different staff in preparation of several different syllabi. The difficulty that new or.p,art time instructors have in syllabi for the same course. Furthermore, typing the determining exactly what should be taught in the syllabi is an added burden to the departmental sec- course is an additional problem. Another problem retaries. The Automatic Syllabus Generator (ASG) is a menu driven software package written to gen- is that with varying syllabi content, students may erate any type of syllabus. Such a system has a not meet their objectives in taking the course. number of economical and administrative ad- Finally, problems can arise with meeting accredita- vantages. However, it may raise some questions tion standards and student preparation for sub- with respect to academic freedom. This paper dis- sequent courses. cusses the design, implementation, advantages and The system can have its drawbacks. The most disadvantages of ASG. striking, in the case of the unified syllabus, is that of academic freedom. The question arises, how much can be dictated to the instructor as to how the class will be taught and the exact subject matter covered. This is a fine line and could pos- INTRODUCTION sibly destroy a course whose reputation was built on the personality of the instructor. Furthermore, In the last decade considerable emphasis has if the administration does not update the required been placed on the content and quality of syllabi. documents, the syllabus may not keep up with Usually, the administration requires instructors to new innovations and future technological changes. prepare their syllabi in accordance to the institu- It is possible that course material could become tions' guidelines which may contain all or some of stagnant or' ated. the items in figure 1. Further, the administration expects that instructors hand out their syllabi to A solution to some of these problems is presented the students at the beginning of the semester. In here. A computer-aided Automatic Syllabus Gen- many institutions, the syllabus is considered to be erator (ASG) was written to assist the instructor in a contract between the instructor and the students. creating a course syllabus. Additionally, to insure It is used as the base to settle any dispute between standardization of the syllabus, certain aspects are the instructor and students during or after the fixed such as institutional policies and course goals. semester period. Therefore, a syllabus is an im- Other requirements are fixed as to their inclusion portant document not only for the student but for in the syllabus,/ however, their content may vary the instructor as well. according to the individual instructor.

48 The general outline of the syllabus develops with puter and students could print their own copy. In the appropriate body (faculty, administration, etc.) both cases, considerable secretarial time and ex- drafting the objectives of the course and other re- pense can be saved. quired documentation. These results are placed in Figure1 shows the list of parameters for the Auto- files which are used by the ASG to automatically matic Syllabus Generator (ASG). Those denoted as generate a syllabus. Prior to creating a syllabus, the fixed are parameter which are consistent through- individual instructor would include their personal out all syllabi for a course. Those denoted as re- criteria and information. This includes office hours, quired are parameters which must be included in grading system, project, text book, etc. The resulting the syllabus but the content may vary from in- syllab' can then be printed en masse by the com- structor to instructor. puter. Also the syllabus can be stored on the com-

MINIMUM&

SYLLABUS OUTLINE

Name of the course Semester Name of the professor Office room and hours Telephone Meeting time and place * Goals and objectives * Text book * References * General outline of the course Detailed course outline Grading system Homework/Term project Policies: * Grading homework projects * Absenteeism of exams * Make up exams - Class contribution Class absenteeism * Late homework * Cheating * Granting Incomplete

Figure 1: Items recommended to be included in the syllabus * denotes the fixed items

49 SYSTEM DESCRIPTION THROUGH THE the GSS prints the des..*<.c1 number of copies of the DATA FLOW DIAGRAM syllabus. The source of the stable-information data flow is the stable files. The Automatic Syllabus Generator (ASG) is composed of five subsystems, see figure 2. These are: A copy of a syllabus produced by the ASG is in

1.1 The Initial Process the appendix. 1.2 Create Processes 1.3U pdate-Time-Bound-I nf ormation The following nine files make up the stable files: Update Syllabus-Stable Information 1.4 This file contains the required and Generate Syllabus TEXT file. 1.5 optional text books. The Initial Process Subsystem (IPS), see figure 3, This file contains the general and has two components: Enter and Verify subsystems. GOAL file. specific goals and objectives of the The IPS accepts the course-information data flow, which is comprised of course prefix (such as MAT, course. ORI, CCS, etc.), course number, course title and REFS file. This file contains all recommended credit hours, from the user. Before further proces- references for the course. sing of the data flow, the IPS displays it back to the DTOL file. Thisfile contains detail course user for verification. Afterconfirmation from the outline. user, the IPS passes the course-informationdata flow to the Create Process Subsystem (CPS). GRSM file. This file contains grading system and grading scale. The components of the CPS are: Input, Verify and Write subsystems. The Input subsystem ac- HMAS file. This file contains the required term cepts the time-bound-information data flowfrom projects and homework assignments the user, and calls the Verify subsystem to confirm and their due dates. the user's input. Next, the CPS calls the Write sub- GRPL file. This file contains homework and system to write the time-bound-information, which term project grading policies. consists of the semester (such as fall, winter, spring TERM file. This file contains the title and detail or summer), year,instructor's academic title, office description of the required term name, office location, telephone number, paper and projects. hours, course section ID, section number, class room number, building and meeting timein the POLS file. This file contains policies regarding SSYL file. absenteeism from exams, make up exams, cheating, granting incom- The Update-Time-Bound-Information Subsystem plete and class absenteeism. utilizes three subsystems: Read, Update and Verify subsystems. The Update-Time-Bound-Information Subsystem is used by the instructor to modify an already existing syllabus or to modify the time- IMPLEMENTATION bound-information data flow. The Automatic Syllabus Generator (ASG)is The fourth subsystem, Update-Syllabus-Stable- written in Fortran-10, Version 6, 'andis imple- Information Subsystem (USSIS), utilizes three sub- mented on a DECsystem-10 with a KI.10B processor. systems: Read, Edit, and Write subsystems.This The code follows the ANSI 1966 Fortran standards subsystem is usedtoupdatethestable- permitting ease of transfer to different processors. information files. The system is totally modularized and is comprised The last subsystem of ASG, Generate Syllabus of 55 modules. On the DECsystem-10, each block Subsystem (GSS), utilizes two subsystems: Write is 128 words or 640 characters and the processor Syllabus and Print Syllabus subsystems. This sub- has 36 bits. The ASG source code requires 120K or system retrieves time-bound-informationdata flow 186 blocks of storage on the DECsystem-10. The from the SSYL file and stable-information data flow ASG in relocatable form uses 90K or 140 blocks and from the stable files and generates the actual the execution form occupies 50K or 80 blocks. syllabus by writing it in the file SYLLABUS. Further, The compilation time is approximately 10 seconds.

50 6 1. Automatic Syllabus Generator

1,1 1 1 1.2

1.1 Initial Process

TEXT 1.2 Create Process

1.3 UpdateimeBound.Information

USER D3OAL

1.4 UpdateSyllabusStableInformation

D4 REFS 1.5 Generate Syllabus

05 DTOL 1.4

D6 GRSM

D7 HMAS

08 GRPL

D9 TERM

0 POLS

D111SYLLABUS

.6u

Figure 2: The data flow diagram of the ASC. 1 -. Course Information

A -- Course Prefix

B -- Course Number

11551$

CPS IPS C -. Course Title

D -- Credit Hours

/111...10N eMMI111., pne=, im111 TENT 2 .- Time Bound Information 2 VERIFY 4411410.1 2 VERIFY 1 ENTER I INPUT

USER USER 03 COAL A -- Semester

B -- Year IHD4IREFS

1- C -- Title OS DTOL

WRITE D -- Name

D6 GRA

01 SSYL NM. 2 E -- Office Room 1.1.1D7 1111AS F -- Building

GRPL

GSS --Telephone

D91TERM H -- Office Hours WRITE

DliPOLS 4 --- I -- Hours By Appointment J -- Section ID )111 SYLLABUS

4 K -- Section Number

VERIFY PRINT USER UPDATE USER

L -- Class Room F

M -- Building 2

N -- Meeting Time

3 .. Stable Information

4 -- Syllabus Information

Figure The detailed data flow diagram of the ASG, A BRIEF DESCRIPTION FOR THE USER user can update each element of the SSYL file individually with the new value written into the SSYL The user begins the Automatic Syllabus Gen- file. The ASG generates a new temporary copy of thecourse- erator(ASG) process by entering the syllabus if no update is required in the SSI files. information. The course-information consists of the However, the user must continue on to the next course prefix, course number, course title and option if SSI requires updating. credit hours. Next the user has the choice of: Updating the syllabus-stable-information option 1 creating a new syllabus, is chosen if the user needs to update the SSI in the 2 updating the time-bound-information (TBI), SSI files. The user can edit each file separately and 3 updating the syllabus-stable-information when updating is completed a new temporary (SSI), copy can be generated. 4 writing the syllabus on the disk and printing it on the line printer, Option 4, writing to the disk and the lineprinter, 5 print multiple copies of a syllabus. is chosen whenever the user wants to generate a temporary copy for inspection. The TBI consists of the file SSYL, the file has the following elements: semester, year, academic title Printing multiple copies option is chosen when and name, office,roon-1 number and building, tele- the user is satisfied with the syllabus created, up- phone number, office hours, hours by appoint- dated, edited and written by the Automatic Sylla- ment, section ID and number, class room number bus Generator. and building, and meeting time. The. SSI consists of the following files: TEXT, GOAL, REFS, DTOL, CONCLUSION GRSM, HMAS, GRPL, TERM,and POLS. The Sylla- bus consists of the file containing the TB!, and all The Automatic Syllabus Generator (ASG) can be the files containing the SSI. very helpful to instructors preparing class syllabi. It can be used by a new instructor to create a sylla- Creating a new syllabus option is chosen if the bus containing consistent course material with user has never generated a syllabus for the course. what is being taught in other sections and in pre- The ASG stores the TBI, input by the user, into the vious semesters. Instructors who have been teach- SSYL file. Next, the ASG generates a copy of the ing the course for several semesters who need to syllabus on the disk and the lineprinter. This sylla- update their syllabi can also benefit. Although the bus is considered a temporary copy for the users ASG has some drawbacks, such asrestricting inspection. After receiving verification from the academic freedom of the instructor, it should be user, the ASG generates a permanent copy. noted that an instructor can still use the system Option 2, updating the time-bound-information and dictate most of the parameters. This can allow option, is chosen if the user wants to update an the instructor to take advantage of the time and existing syllabus and needs to change the TBI. The economic savings produced by the ASG.

53 How Schools Use Microcomputers: Findings from the Johns Hopkins University National Survey of Computer-Using Teachers Clarence Miller, Chair Maryland State Department of Education Baltimore, MD 21201

ABSTRACT Thissymposium discusses the major findings and the implicationsfor schools of a national survey ofhow schools are currently using microcomputers in their instructional programs. The survey, a national probability sample of 2,259 public, private, and parochial elementary and secondary schools, including about 1,400 with a microcomputer, was undertakenbetween December, 1982 and February, 1983. Respondents included the principal at each school and the primary computer using teacher inthose schools with a microcomputer. The symposium consists of a presentation of some of the major statistical findings from the survey, and a discussion of the implications of these results for schools planning their future involvement with computers. The paper presents a variety of data from the survey. The principal focus is how schools in different locations, having differenttypes of student bodies, in different financial situations, and enjoying different degrees of district and administrative support have had correspondingly different histories of involvement with microcomputers, possess different types and quantities ofmicrocomputer equipment, and use this equipment in different ways. The paper also shows howschools that have had a microcomputer for over a year have changed their approach to using the computer and how these changes are related to factors in their environment. The discussion which follows the paper presentation will present the reactions of two knowledgeable people in the field of educational computing - reflecting academic, governidental, and commercial perspectives. The discussants will focus on some implications of this national survey data for schools planning new or further investments in microcomputer technology in the services of their instructionalgoals.

PRESENTER Henry Jay Becker, Project Director The Johns Hopkins University Baltimore, MD 21218 DISCUSSANTS

Arthur Melmed U.S. Department of Educaton Washington, DC 20208 Charles Blaschke Educaton Turnkey Systems/MEAN Falls Church, VA 22046 CAI in Foreign Language Instruction

Carl Adamson, Chair Wichita State University Wichita, KS 67208

Michael Bush USAF Academy USAF Academy, CO 80804

ABSTRACT The mirocomputer offers foreign language educators a unique and powerful opportunity to enhance their language learning programs. Their interactive qualities provide students with feedback, evaluation and aboveall controlled stimulus-response learning which, when coupled with their impressive audio and graphic capabilities, may well make them an even more effective laboratory too) than audio tape. The first presentation offers a brief look at some of this potential with respect specifically to the Atari 800/400/3200 computers. After a brief discussion of some factors relevant for language laboratory implementation, the unique capabilities of theAtari will be presented. Features such as a modifiable character set, more than a dozen graphic/textmodes, 256 color flexibility, graphics indirection through an easily modifiabledisplay list, simultaneous audio, processing, and di ?lay, all lend themselves well to the highly interactive methodology of foreign language acquisition. In addition, the low cost and relative programming friendliness of the microcomputer rake it particularly attractive for language learning applications. The second presentctich will discuss microcomputer based interactive videodisc in basic French instruction. The videodisc is the densest storage medium available for use on computers of any size. This new technology is being combined with the control capabilities of the microcomputer for the presentation of audiovisual material in basic language instruction. a Using a methodology developed by an agency of the French Government during the past three decades, the Department of Foreign Languages at USAFA is studying how the "old" can be combined with the "new" to create an exciting environment for foreign language instruction. The randomaccess capabilities of the intelligent videodisc system will be used to good advantage in the presentation of new material to beginning language students. It is anticipated that by using this approach, studentswill individually accomplish the initial phases of leerning newmaterial, this freeing time presently being spent in the classroom on these more rote aspects of instruction With the technology being used in the manner for which it is best suited, the classroom experience will be enriched for instructors and students alike and in ways possible only through the added human interaction that results from such use.

55 What Compcter Curriculum Is Right For The Small College?

Dr. William Mitchell

The Univerity of Evansville Evansville, Indiana

ABSTRACT The academic approach to computing There are nearly2000 colleges in the reflects the diversity and fluidity of the young to have United Stateswhich enroll 5000 or fewer field. The field is too students. The majority of these collegesare .produced many great generalizers, so it lacks of thought" liberal arts colleges with less than 2000 the unity provided by "schools There students. Most of these schools have acquired which characterizeother disciplines. computer facilities and roost offer one or more are no fundamental problems whichchallenge courses in which the computer is used. Many computer scholars (only an endless list of w,,ald like to ma%e a computer-related major or interesting or presssing ones), nowidely minor available to their students. There are acceptedparadigmswhichguide their work breadth to at least four nationaly published curricula [14,16], and consequently, little which these school might consider but noneof computer research. Like today's computer these curricula were formulated with the small applications, the academic study of computing college environment in mind. This paperwill is particularized. discuss the reasons why most small colleges cannot adopt any of the present curricula, and Little wonder, then, that so many will offer a compromise curriculum which is a definitions abound for undergraduate computing years the practical alternative. curricula. In the past five Association for Computing Machineryhas published five different and discrete Introduction computing curricula, and these havebeen augmented by the recommendations of dozens of It is increasingly evident thatcomputer other professional societies and individuals, anyof the applicationswill dominate our society during frequently t/ith rogard .f)r the close of the 20th century. Because of the previous wax. Asidc. from broadcasting the the plethora of perv-iveress of the computerphenomena, a views of each constituency, field which is barely 30 years old isalready curriclum recommendations have begun, by their fragmenting, rupturedby the stressesof intersection, to describe what isdesirable compellingapplications. As the academic for undergraduate computing studies. community tries to understand what is cohesive on and fundamental about computing andhowto In this paper the author concentrates which have received organize it for study, the various users of four diverse curricula professional computingdemand thatgreater emphasisbe national publicity in the placed on those features of the subject which literature. From these are drawn implications are immediately relevant to their individual whichguide the construction of a model The method needs. The economicsof computing justify curriculum for a small college. as important as tremendousinefficiences in software and used to derive the model is Each suggested systems designbecause the hardware is so the actual curriculum derived. powerful that the resulting output is not only curriculum represents aviewpoint, and from its curriculum is rational acceptable, it is significantly better than each'sviewpoint, what can be achieved without computers. In a and the competitors' curricula are invalid. college curriculum "results now" environment, there is little Thus, the derived small tolerance for abstractions and no motivation will be viewed as sensible by those who accept the model's derivation, and to take a little more time and do it right. the premises of whodo In a rapidly changing technological will be viewed as inadequate by those environment, it is difficult to identify not. If, however, the method for building the fundamentals. The computing fieldis surely model is hel.!:111, each may build his own ideal This is not unique, for it is an experimental field whose (for him) .-mputing curriculum. subject changes faster than it can be to say that computing curricula are arbitrary, formalizuJ. Just as we begin to understand but to emphasize the present lack of accepted sequential algorithms and automata, we find generalizations. ourselves in the age of parallel processing.

56 The author can claim both knowledge and in business practice in order that the student experience in the process of curriclum design be properly oriented and motivated. [8,9], and a special expertise in the area of small college computing curricula [17].Many The curriculum of the PittsburghLarge of the ideas presented here have emerged from Users Group [13] (see figure 4) has been a graduate course on computing curricula constructed similarly to the DPMA model, taught to small colllege computer educators. utilizing a survey of professionals to The curriculumderivation process has been ascertain topics and the degreeof emphasis. employed by dozens ofcolleges in the past: In this case, the surveywas limited to five years. large-scale corporatecomputer centers, so that the result supported greater technical The Curricula depth than theDPMA survey, reflecting the greater technical sophistication of such The major undergraduate curriculum models centers. Whileencouraging the business for computing education have been proposed by corequisites of the other professional groups, each with its own bias. commercially-oriented curricula, the PLUG Three curricula are proposed by the curriculumdoes not insist on the degree of Association for Computing Machinery: integration espoused by DPMA. On the other Curriculum '78 in Computer Science [2] (see hand, while it recommends technical courses figure I), the 1982 Information Systems similar in depth to the ACM Information Curriculum [12] (see figure 2), and the Systems proposals, it does not seek the Two-Year College Curriculum of 1980 [8]. The improvement of management's utilization of first curriculum represents computing as it computer systems. Its goals is to provide the exist in the research-oriented university. technically competent employee which the two Emphasis is placed on the mathematical year curriculum attempted to define, except analysis of algorithms, on the theory of that in this environment the four year degree numerical computationand symbol processing, is necesary to achieve that competence. and on the study of the properties of abstract machines and languages. The second curriculum Another nationally publicized emphasizes the application of existing undergraduate computing curriculum was computer technology to business organizations. proposed by the IEEE Computer Society in 1977 Concern is evidenced for the effective use of [4], intended to guide the offering of computer systems, which necessitates computer engineering within engineering understandinghow information is used for schools. The similarity between this control and decision making within curricula and Curriculum '78 has been analyzed organizations. The third curricula is focused [3] . and found to be substantial, but this on the operation of computer systems in curriculum offers little else to commercial environmentswith the goal of non-engineering programs. preparing students for carrying out the activities normallyrequired in such shops Each of the four-year curriculummodels (rather than to analyze why these activities are structured with a core of required courses are desirable). Both of the later two and a range of elective courses at various curricula emphasize a strong knowledge of levels. The Computer Science core is eight business as corequisite to utilizing the courses intended to be pursued mostly in the technical content of the computer curriculum. lower division. ACM's information system's All these curriculahave been devised by curricula has an eight course core which is academics with ±e advice and counsel, to takenmostly in the upper division while the varing extents, of the practicing major is completing the AACSB common body of professionals. knowledge (out ofwhich a statistics course and a programming course are,rerequisite to Two other curricula have been derived to the computing core). The DPI;;, curricula has a represent the business data processing seven course core, four lower division and community. TheDPMA curriculum [1] (see three upper division, while the Pittsburgh figure 3) eschewstheory, even the business curriculum has an eight course core, two lower organizational theory emphasized by ACM, and division and six upper divison. The DPMA focuses instead on criteria for practice. The curriculum is unique in being designed with curriculum seeks to produce a competent the intention that its lower division courses employee who will be skilled in applying would frequently be taught in junior colleges. existing computing technology to the The author has elsewhere analyzed the DPMA present-day business environment. The highest curriculum in the context of its stated teals goal of the curriculum is to minimize the time [10] and the similarities of the DPMA and ACM it requires to assimilate the graduate into Information Systems models and the his first position. Tomeet this goal the dissimilarities of the DPMA and Curriculum '78 DPMAcurriculumdesigners were willing to models have been noted [15,6]. sacrifice "why" for "how" when that was necessary. The curriculum is implemented as a The Computer Science curriculum lists ten rolling five-year plan, so that each upperdivision electivesand assumes that a graduating class will be state-of-the art. It major will take at least four. It is also is thus deemed essential that computer expected that at least fivemathematics technology be highly integrated with courses courses will be taken (to include calculus,

57 Figure 1.ACM Curricula. 00111 Figure 4.PLUG Curriculum BUSINESS INFORMATION SYSTEMSCURRICULUM STRUCTURE

BASIC/ Structured Assembler FORTRAN COBOL Language Programing

COMMITTERSCIENCE 11100E1. CURRICULUM mpUter Spec., Topic. Counter Systems Advanced 173 Coro Courses: Analysis COBOL Systems CS I. Cornpu 14.1 P11,21.1.1*0 A. Mancomputer Laboratory Programmin Cl2. Carp. ProCiran..1 B. MiniconossIor Laboratory CS3. Introduction lo Convu4r Sy.. C. Menem,. Evetuation CS4. IntroducSon to Coe Peer 0,11.MMI D. TIMA:CrniruniresionsiNatwotksit C8ff. rsoraxtion Io Fig Procoseng &OMAN 6. Operating Systems and Computer E. SOLOS SmistiOn CS Advanced Systems Progred..1 Ant... 1 F. Data CS7. Oats Sim.. and Algorithm Anahoin CL *tic* Comb( %Meng Laboratory Communica EDP CS a.Ortiontuttion d Prot/mm.2 imatmcio K . Simulation L SancAurd MOMOmanital tions Auditing Controls Electivo COLONS: J. ToPcs In Aulonsata Theory CS0. Compress and Socany M. ToPics n Comatibity CS 10.. Operating SOM. and CornMo88 L TopesIn Forme I. am7.00 TR. Antennae I M. MM.. and MOrleang CS IL DM Bose Management Systoms Nalco Matheinalics liegulreton(4: CS 12, MIAMI IntelaPenco CS 13. Acceithrto MA 1.bueoenmy CODAe CS .4. Solt* era CmmOn and Threeolsnant MA 2.MItionlatirol AnMysts 1 liManagementi CS IS. Theory at POO amminp Longman MA 2A. Probate. nformation CS 18. 4.500545. Compulabety and Formal MA 3. Liver Algebra Systems LonPm194. MA 4.Centel. Simone* CS 17, Nmodeol Mathernalm: Analysis MA 5.Matonnsticol Anelysis CS IS Numinsi MaThernata.: Lines, Algot. MA a.probeesty and umiak* REQUIRED COURSES

PREREQUISITE

Internship

Figure 3. OPMA ComputerInformation System. Curriculum Figure2. ACM Information Svotems Curriculum

Information Systems Information Systems Technology Process

"1 IS2 593 Computer Program and Systems Concepts Data Struc Concepts tures

194 I85 Data Information Management Analysis 0.00. 111.1a.

Oma M @Nam al 138 138 8.0. 4.1.104 Data Spasms Communica, Design tions OM 4gar. h ma PIComputer Programing is prreguisite to 151 and 152.157 and xas are graduate courses.

"10 Projects

58 probability, linear algebra and descrete operations research, the automated office, DP structures), and several courses in an law or computer auditing is beyond the applications area, such as the physical resource of the small college. The average 'r business. The Information Systems small college has mini- and micro-computer 'fits no electives. The DPMA syster little or no data processing calls for three upper division expertiL, and its computing knowledge and ul, Ives tobe chosen from a list of eight, business knowledge reside in separate and suggests eight corequisite business individuals. courses in lieu of the AACSB core (including statistics). The Pittsburgh curriculum In these circumstances the small college suggests two electives from a list of five, must find a curriculum compromise between the two of whichare lower division, and it mathematical and the commercial, between the requires eight business courses, two business abstract and the applied, which will be within electives and five mathematics courses the skills of the faculty to deliver, yet will (including statistics and business calculus). effectively prepare its students for computing careers. What these students will lack in All of the curricula are characterized by technical sophistication and state-of-the art deepprerequisitechains (up to six levels) integration, will be balanced, the liberal and by the recommedation of an upper division arts college believes, by a broader software project course. A major consists of perspective on humanity, society, cultureand nine 3 semester hour computing courses for the learning. Rather than having specialization DPMA and ACM Information Systems curricula in the environment they intend to enter, the (discounting the general education course liberal arts studentwill be prepared for a CIS 1 and including the prerequisite more general context of life. The thesis is programming course P1), ten for the PLUG that while they will still have much to learn curriculum, and twelve for Curriculum '78. At about either data processing or computer least half of the major consists of upper science, theywill be prepared to learn division courses, and none of theseare quickly, and they will organize their language-oriented programming classes knowledgewithin muchbroader perspectives, (learning to program in a high level enabling them to exercise more humane programming language is universally considered judgement in their work. a lower division task, but several languages are expected to be mastered). Rather then attempt to persuade the small college that this thesis is incorrect and that The Small College Environment they should seek to emulate either the business school or the research university, we The liberal arts college finds all four present a rationale for a curriculmwhich curricula unimplementable for a variety of meets thier perceived needs. reasons. The Computer Sciencecurricula is too sophisticated in its present emphasis, and ,A Small College Curriculum. there are pressures to make it evenmore abstract [11]. Curriculum '78 strives to The small college is advised to offer inscend the practical, but to do so requires courses at various levels and in various a theoretical knoWledge which is not available tracks so as to satisfy the needs of its to the small college. Even though most of the hetrogeneous student body. The small college computing programs in the small college will must be equally concerned with the emerge from the science or mathematics availability of opportunities for computer departments, the formality and abstractness of literacy, for two and three course "minors" computer science is yet several steps removed for students from a variety of disciplines from where the small college mathematics or (principally from business and mathematics), science faculty are. Computers are relatively and for a major. The curriculum models all new in these schools and great effort is being take the narrow view of the specialized major, expended just to gain concrete experience with so they give no guidance' on how to best serve their use. This experience base is too meager this complex goal. But since each does speak to bear the weight of a theoretical to the introduction of the major to the field, curriculum.The average small college faculty we can compare their approaches. would be pressed to do a competent job presenting all of the eight core courses of The DPMA model has the weakest Curriculum '78. introductorycourse in terms of requirements, or, if you prefer, it suggests the broadest On the other hand, the three introductory course. The undergraduate business curricula vary from too commercially-oriented curriculums in ar-.eral narrow and vocational (presupposing an presuppose the existing rr-actice of using as environmentwhich cannot be approximated at the first course in the major the the small college) to too narrow and "Introduction to Data Processing" course which sophisticated (presupposing a computationaly was developed as a core course for all trained business faculty). Specialization, business students. The DPMA model further either in the direction of computer hardware suggests that thiscourse should cater to and software (telecommunications, data base, non-business students who are seeking an large scale systems) or in the direction of introduction to computing for general

59 should alter', 'e content each education. In the DMA prit;4mming topics course, with the training begins in CIS 2. :jolt, might be taken simult .2ously topics course. It is possible, therefore, for the The ACM Information Systems curriclum and a student to break the lock ;t.fi= of the PLUG curriculumassume a more rigorous, schedule if necessary. programming-oriented first course, but still include a shar with other business majors. The PLUG The senior level courses will course, a simulation course and an currt, alum X13. hies BASIC tF first data base possible, course language, but in the ACM Information app,.'-,lion project course, and, if Both the Systems model, P1 is modeled on the first a sec_nd upper division elective. course course of Curriculum '78, and requires a simulation course and the database procedure-oriented language which will support makeextensive use of thedata structures of the later programming experience. Curriclum '78 course, and more than casual use requires an in depth first course in a high computer systems material. The resulting to the level programming languagewhichstresses curriculum bears greatestsimilarity methodology and algorithm development. ACM Information Systems requirements, but in spirit it is closest to the PLUG curriculum. It is obvious that a single course in the Elective courses might be developed to provide small collegecannot meet all these thrusts. theCurriculum '78 software and hardware Therefore, at the freshman level we propose an topics which have been squeezed into the introductory programming course, an computer systems course (course content is introductory systems analysis course and a discussed below) but the presumption is that literacy course, each taught without most small colleges will never be able to computer by prerequisite. Instead of combining an offer the number of courses reauired introduction to programming with a narrative Curriculum '78 and still have the ti--.Aware and about DP, we suggest two separatecourses faculty resources to meet the computing also which will allowgreater depth in each. To education needs of the nonmajors. It is not choose the meet the needsof the poorlyprepared or presumed that students will weakly motivated student, the literacy course small college with the intention of becoming a would provide opportunity to be exposed to designer of operating systems or the builder programming and deriving algorithms, but would ofpipeline processors. More likely, the not make that aspect of computer interaction a studentwill be content to pursue the details school, or major feature of thecourse. The literacy of computer science in graduate coursemight use either a Computers and will be intending to apply his computer upon Society text or an introduction to Data knowledge vocationally immediately Processing text, but its focus is graduation. fundamentallynon-major. Ideally it should incorporate extensive interaction with The core of the small college major would computers as media for entertainment, for CAI, consist of Programming I and II, Systems for word processing, for library use, for data Analysis I and II, Computer Systems, Data and base query, and for packaged functions like File Structures, and the senior level project statistical analysis or spreadsheetmodeling. course. A major would require, in addition, a lower division elective and two upper The sophomore courses, roughly following divisions electives. In the following the three tracks of the PLUG curriculum and paragraphs we discuss each course and its role the ACM Information Systems model, should be a in the curriculum,. The reader is directed to computer systems course, a second programming the appendix to see the course descriptions of course, and some elective language courses the courses referenced from the four models. (taught as second languages). The goal here is to provide both the necessarydevelopment of pourse Descriptions- the major and second courses for various types of minors. Thebusiness minor, after une Programming I and II should be taught in systemscourse (with or without the literacy a block-structured procedure - oriented,language course), would take the first programming and should be a "no nonsense" programming course and then a business language course course akin to the Curriculum '78 courses CS1 (COBOL). The mathematics major, on the other and CS2. The first course will introduce hand, mightmovefrom the first programming decision structures, looping structures course to the second, or to FORTRAN or some (nesting), data types, lists, tables, and applied mathematics course without ever taking subroutines. Standard file processing the systems course. activities should be illustrated (such as contol break processing and sequential The junior level courses should include a update). Proper techniques for documenting data and file structures course, the second software and for developing programs should be of systems course, and a topics course which taught and consideration given tomatters would be a proving ground for new courses. style and group programming. The second programmingcoursewould be the should prerequisite for both these courses, and one The second programming course or both of the systems coursesmightbe systematize and extend the experience of the prerequisite to the topics course.One of the first course in the samelanguage. The second languages courses, which, like the student is introduced to more sophisticated

60 use of data structures and information coding Information Systems course IS2 provides an techniques. The emphasis is on the adequate topic outline for the Data and File variability of data, complex data Structures course. The course should organization, and interrelationships between introduce a file oriented language if a data structures. Preview linked lists, trees, language such as ADA or PL/I is not already in graphs and recursion. Expose the student to use. The programming exercises should the analysis of algorithm efficiency, illustrate sophisticated applications of file demonstration of correctness (assertions and strures andlay a firm foundation for the testingmethodology), and searching and data base course. sorting prnciples. Do a large group project r0ouires module design and interfacing, The data !ease course wouldbe a notch 6 or an ,iss decomposition of problems and above IS4 and several notches above topic.: .roc .uctures. the Pittsburgh model's BIS 401. Previous exposure to the data structures and file The programming sequPnce is intentionally structures of DBMS packages will permit experience-oriented, building experience greater attention to both the 'pory of data within a language and experience in problem base architectures and to the ,,ct in which solving. Theprogramming exercisesshould they are used. It is unlike: the small dominate the students course requirements and college will have access to IM uL ,.1,mS, but should be rigorous enough to discourage those inexpensive network and relational-like data who resist practicing the tools being base packages exist for micros and most minis, presented to manage complexity. High such as HP and DEC. Thus a blend of standards of internal and external theoretical exposure and practical experience documentation and user friendliness should be could reasonablybe achieved in this senior required of all software products. level elective. In some environments CS 11 would be an appropriate guide, but it is Since this sequence must meet the needs oriented toward the writing of a DBMS rather of both scientific and commercial than toward its use in solving applications. programmer/analysts, it should focus on good, coding practices and illustrate the broadest The simulation elective could bemodeled range of techniques. It should not be taught onBIS 410 of the Pittsburgh curriculum or on in FORTRAN, BASIC or COBOL if possible, though IS7 of the Information Systems curriculum. As BASIC could be utilized if extreme care were in the case of data base, simulation languages taken to simulate structured control such as NDTRAN,or SLAM are readily available sequences. (BASIC, like PL/1, hasgained on mini-computers. Withoutprejudice to adherents in both the scientific and either major, the data base coursemightbe commercial fields, hence is justifable 1J) some more attractive to the commercial extent in terms of the generalityof its programmer/analyst, and the more mathmatically potential utility). Students emerging from the .-'ented simulation course could interest the sequence will understand the practical process scientific programmer. Hopefully bothgroups of software development and will know one would see the relevance of both electives. language very we'.1. Only the first course in the sequence is requiredbefore electing to The lower division elective courses would study computer systems or a second language. normally be second languages, and would provide intensive syntax-focused introductions The systems analysis series begins with a to the problemdemains appropriate to the survey course which touches, lightlyon language. Thus the COBOL syntax covered in computer hardware technology but treats in DPMA's CIS 2 and CIS 3 would easily be covered depth the coftware life cycle, the in a single sophomore level course after organization and interaction of data experiencing Programming I. Likewise a FORTRAN processing with users, and the role and tools course oriented about numerical methodscould of the systems analyst. This coursecould be offered, or a comprehensive BASIC course, serve as an elective for business-oriented or RPGII or assembler, etc. The upper students. Because of the great amount of division topics course would deal as overlap between CIS 4 and CIS 1 in the DPMA appropriate with comparative computer curriculum, this first systems coursecould languages, with a second "advanced cover the topics of CIS 4 from their programming" course in COBOL, with data introduction in CIS 1, but not reach as high a communications, with computer architecture, or level of skill development as is published for with Management Information Systems. CIS 4. The second systems coursewouldbe modeled on CIS 5. The capstone course for the major is the senior level application project course which The Computer Systemscourse couldbe should require that the spectrum ofanalysis, modeled on the Pittsburghcourse with that design, coding, and documentation activities title, or on the Information Systems be applied to a realistic problem andusually curriculum's IS1, or on DPMA's CIS 8, but becarried out by a small team. This course should include a software project in the same will test the graduate's preparation to language used in Programming I. This course, perform in the area of software development. along with the programming sequence is CIS 7 or ISIO are appropriate models. prequisite to Data and File Structures. The

61 Some colleges are experimenting with an A students with mathematics or science upperdivisionsocial issues course which is interest could also pursue the core courses, open to either majors or nonmajors. Such a and could achieve, with an alternate set of course could have as prerequisite as little as electives, a competent exposure to the core the freshman literacy course, or as muchas concepts of Curriculum '78, even though the both of the first courses in programming and offerings would be too meager to have ventured systems. CS 9 could be helpful, though it far outside that core. envisions a still more sophisticated audience. Conclusion Assuming that the normal faculty loal is four courses a semester, this curriculum could Despite the obviousbiases present in be offered by one full-timecomputer faculty each of the curriculum recommentations, it is member (teaching the core) supported by a possible to recognize a collection oftopics part-time business faculty and a part-time which are common to those modelswhich mathematics or sciencefaculty member (or emphasize computer applications. Not several adjuncts) for a total of 2FTEs (see surprisingly, much of this intersection is figure 5). Such a load is not desirable, and also contained in the computer sciencecore. sections would have to be kept small in order Thus the courses described for a small college to keep studentcontact hours tolerable can capture a high percentage of the topics (computing is much more student intensive than advocatedby any ofthe models, and do so is mathematics or history [7]). Yet the while emphasizing the fundamentals of national facultysituation is such that multi-lingual programming. The graduate of expecting everysmall college to have access the small college, regardless of thecomputer to two well-credentialed computing instructors equipment he trained cn, can reasonably expect is also unrealistic. to be competitive with graduates of any of the other curriculummodels for the entry level A veryrespectable systems information programming position. The liberal arts major is produced if the student completes the graduatewill bear the responsibility for core, elects the data base course, the lower doing a great deal of integrating on his own, division COBOL course (assuming the and the obligation to learn on the job new programming course is Pascal), the upper hardware and softwaredetails. But since division topicscourse in advanced COBOL these hardware and software details are applications (with the COBOLcoursebeing constanty changing anyway, and since even taught by the business facultymember or large universities' are unable to maintain a adjunct), and completes a minor in the computing environment commensurate with business department. Such a student would industry, these handicaps may well prove to be compare very favorably (having greater virtues. The liberal arts student is prepared technical skills) with a DPMA graduate who had to change, and forced to imagine different elected CIS 8, CIS 11, and CIS 13 or CIS 14. environments, where the DPMA student may well Such a student would halm: programming skills be too specifically oriented. By exceeding that acquired in the Informations concentrating on technical fundamentals rather Systems or Pittsburghcurriculum, but would than, specific business applications, the small lack exposure to data communications and to college faculty can educate both the management information systems. commercial and the scientificallyoriented Figure 5. Small College Curriculum Schedule student in the same core courses, and limit to the elective courses the need for Fall Semester Spring Semester multi-disciplinary expertise.

freshman courses References

Literacy Literacy 1. Adams, David R., and Thomas H. Athey, Systems I Programming I* DPMA MODELCURRICULUM FOR UNDERGRADUATE COMPUTER INFORMATION SYTEMS EDUCATION, sophomore courses DPMAEducationFoundation, Park Ridge, Illinois, 1981. Programming II* Language elective 2. Austing, Richard, et. al., Computer Systems* "Curriculum '78, Recommendations for the Undergraduate Program in Computer junior courses Science," COMMUNICATIONS OF THE' ACM, v. 22, n. 3 (March 1979). Data and file Structures* Topics* 3. Engel, Gerald L., "AComparison of the Systems II* ACM/C3S and the IEEE/CSE Model Curriculum Subcommittee Recommendations," COMPUTER, senior courses v. 10, n. 12 (December 1977). 4. IEEE Computer Society Education Data Base Simulation Committee, Model Curricula Subcommittee, elective Project* ACURRICULUM IN COMPUTER SCIENCE AND ENGINEERING, 1977. *signifies a course taught by full-time 5. Jones, Ron, and Rich Hamilton, "Computing computer faculty (others may be adjuncts) Education, the CPMA Model,"

62 COMPUTERWORL:, v. XV, n. 38 (September 21, 1981). 6. Kroenke, David, "A Place in the Sun," INTERFACE, v. 3, n. 1 (Spring 1981). 7. Little, Joyce Currie, RECOMMENDATIONS AND GUIDELINES FOR AN ASSOCIATE LEVEL DEGREE PROGRAM INCOMPUTER PROGRAMMING, ACM 1981. 8. Mitchell, William M., THE DESIGNOF MATHEMATICSCURRICULA FOR THE SMALL COLLEGE, Ph.D. Dissertation, Georsa Peabody College, 1974. 9. Mitchell, William, and Bruce Mabis, "Implementing a Computer Science Curriculum Merging Two Curriculum Models," SIGCSE BULLETIN, v. 10, n. 3 (August 1978). 10. Mitchell, William, and James Westfall, "Critique and Evaluationof the CAL POLY/DPMA Model Curriculum for Computer Information Systems," SIGCSE BULLETIN, v. 13, n. 1 (February 1981). 11. Mulder, M. C., et. al., "Computer Science Program Requirements," a position paper of the Joint ACM/IEEE TaskForce for Computer Science Program Accreditation, February 14, 1983. 12. Nunamaker, Jay F., Jr., et. al., "Information Systems Curriculum Recommendations for the 80s: Undergradute and Graduate Programs," COMMUNICATIONS OF THE ACM, v. 25, n. 11 (November 1982) (summarized in COMPUTERWORLD, v. XV, n. 39 (September 28, 1981). 13. Schultz, Brad, "Model DP .Curriculum Welcomed by Colleges," COMPUTERWORLD, v. XV, n. 40 (October 5, 1981). 14. Traub, J. F., Editor, "Quo Vadimus: Computer Science in a Decade," COMMUNICATIONS OF THE ACM, v. 24, n. 6 (June 1981) . 15. Vanecek, Michael T., and Carl Stephen Guynes, "Business Computer Information Systems DPMA, vs ACM: Now What?" INTERFACE, v. 3, n. 4 (Winter 1981-82). 16 Zant, Robert, MichaelVanecek and Carl Guynes, "Thoughts on the Maturing Process of the Information Systems Academic Discipline," INTERFACE, v. 4, n. 2 (Summer 1982). 17. Zientara, Marguerite; "University Summer School RetrainingCollege Professors to Teach Computer Science," COMPUTERWORLD, v. XVI, n. 18 (May 3, 1982).

63 A New Source of Computer Science Teachers:

Faculty Members From Other Departments

Keith Harrow

Department of Computer and Information Science,

Brooklyn College, Brooklyn, N.Y. 11210

Abstract graduate program.

The current shortage of computer science The CIS Department has 19 full-time faculty faculty is well-known. This paper describes members, which is about half the number that a novel solution to the problem that has been would be justifiedby our enrollment asa percentage developed at Brooklyn College. Faculty mem- of the total college enrollment. Even worse, many bers from other disciplines are retrained to computer science faculty members do not have a teach introductory computer science courses. full teaching load; most people are released from In addition to fulfilling our basic need to one or more courses because of administrative staff courses, this approach has a number of duties and/or research work.We have been rela- interesting ramifications, both good and bad. tively successful in attracting new faculty, but most of them have been research-oriented, with Introduction very low teaching loads. These people contribute enormously to the department in terms of grants, In the past few years, almost all colleges publications, and prestige; but they do not con- and universities have experienced a shortage of tribute in any significant way to our teaching qualified computer science faculty. Many power. people have explored the implications this problem and offered a few long-term solutions Thus, computer science faculty members teach [1, 2, 3, 5]. At the Thirteenth Annual SIGCSE only about forty percent of all computer science Symposium on Computer Science Education, there sections, with the majority of these being gradu- were a number of papers and panel sessions on ate and advanced undergraduate courses. In the ways to attract and retain computer science past, we have used part-time or adjunct faculty faculty. members to fill the gap. However,the shortage of full-time graduate students, plus the college's The Brooklyn College Problem lack of money, have combined to limit our ability to use adjunct lecturers. At Brooklyn College, we are experiencing the same difficulties as everyone else, with a To summarize, the problem that we (and any few extra local problems. Over the past 10 to number of other schools) face is the following: 12 years, the college enrollment first almost given these constraints, how do we meet the in- doubled to 30,000 full-time students, then con- creasing demand for computer science courses? tracted to its present level of about 15,000. The drastic changes in the size of the student Possible Solutions body, plus significant shifts in enrollment from one area to another, have left Brooklyn There are a number of unattractive solutions College with a number of overstaffed departments.- to this problem. One is to limit enrollment in For example, the Chemistry Department has over computer science courses (many schools have 30 tenured faculty members, but relatively few adopted this approach). For obvious reasons, the courses for them to teach. Recently, there has Brooklyn College Administration does not like the been an increase in the need for teachers of idea of turning away hundreds of potential stu- remedial courses, especially Mathematics and dents. Another idea is to use large lecture sec- English. The Computer and Information Science tions in the introductory PL/I course.We have (CIS) Department has experienced an explosive tried this method and found it inferior to our growth, going from three or four computer sec- current format (approximately 25 sections with tions in 1970 to the current total of more than 25-45 students per section). 100 sections. The computer science program is particularly rich, including introductory courses A third idea has gradually evolved over the (in PL/1 for those interested in an intensive past few years. This solution solves the immedi- programming course, and in Basic for those ate problems of relieving the overstaffed depart- interested in just a brief introduction), a ments and Covering computer science courses. It large number of advanced electives and a growing also raises a number of provocative questions for the future.

64 The solution that we have developed uses fac- of material, problems students seemed to be having, ulty members from other depa-..-cments to teach com and differences between teaching computer science puter science courses. To some degree, this cross- and teaching their own disciplines. They were re- over from one department to another is not new. quired to complete all programming assignments for For example, a chemist might teach a course in an the course (but they did not take any exams). area of specialization (e.g., real-time systems); a mathematician might teach a course in numerical' Current Status of the Program analysis; a linguist might teach a course in formal language theory. Most computer science departments After successfully completing the Faculty probably have such informal arrangements with other Development Program (some did not complete it), a departments. However, our use of outside staff is faculty member is accepted as a candidate to teach much more extensive. CIS 1, our introductory PL/I course. The department then provides the instructor with a syllabus, In the Fall 1982 semester, 13 faculty members several different course outlines (corresponding from other departments were teaching one or more to different approaches to teaching the course), computer science courses. More than half the sec- model homework assignments and exams, etc. Each tions of the introductory programming course in new instructor is observed informally two or three PL/I (plus a few intermediate electives) were times per semester, until we are confident that the taught by these people, with adjunct faculty teach- instructor is doing a good job and needs less super- ing most of the remaining sections (mostly at night vision. The CIS Department Appointments Committee and on the weekends). Currently, we have faculty reviews all new instructors every semester, and members from the Departments of Chemistry, Educa- has the right to reject any candidate. We discuss

tional Services, English, Mathematics, Physics, . such matters as their progress in learning more and Psychology teaching computer science courses. computer science, observation reports, and other Approximately one fifth of all computer science feedback on their teaching. courses are taught by full-time faculty members frim other departments. It is important to note Faculty members are urged (but not quite rnat these people maintain their positions within ordered) to continue their education in computer their own departments, but teach computer science science by taking further computer courses. It is courses to fill out their teaching loads. suggested that they eventually learn the equivalent of at least three or four courses, including assem- Faculty Development Program bly language programming, data structures, and programming languages. Although a few people have Of course, we are unwilling to accept just been somewhat remiss, most of the new faculty mem- any faculty member from another department. The bers have been extremely conscientious. As part of Brooklyn College Administration has agreed to give the faculty union's collective bargaining agreement, the CIS Department Appointments Committee the right faculty members receive a full tuition waiver when to interview and approve all such candidates. Nat- they register for graduate or undergraduate courses. urally, we are reluctant to accept a faculty mem- Many people have taken advantage of this offer and ber's self-assessment of his or her competence to re,-..eived credit for courses in such areas as Cobol teach computer science. Therefore, we have made a programming, file processing, compiler construction, major effort to certify some of these people. simulation, and so on. (One graduate of the 1981 Faculty Development Program has made so much pro- For the past three summers, the CIS Department gress that she has shifted her department affilia- (under the auspices of the Vice President for Aca- tion - because of budget cutbacks, she was about demic Affairs) has organized a Faculty Development to let go by her old department. She is currently Program./ In 1980, Professor Pat Sterbenz conducted an instructor teaching a full load in CIS.), We a small test project to teach PL/I to those with have then asked some of the more advanced people some previous knowledge of programming (typically to teach second or third level CIS courses, and they Fortran). In 1981, the Chairman. of the CIS Depart- have done quite well. Thus, as we produce new can- ment, Professor Frank Beckman, proposed a more am- didates to teach introductory courses, we hope to bitious program, one part of which was designed to Move the more experienced teachers into intermediate- teach other faculty members how to teach computer level courses. science. With the aid of a number of my computer science colleagues, I supervised the Faculty De- Evaluation of the Program velopment Program in 1981 and then again in 1982. As with most experiments, there are both good A detailed description of the 1981 program is and bad things to report about our use of faculty given in [4]. Here is a brief overview (the 1982 from other departments. Let's start with the nega- program was similar). All participants were ex- tive points. pected to have some previous knowledge of PL/I (in practice, this was not always true). The faculty The most obvious problem is that these faculty members were asked to audit the equivalent of a members are not computer scientists. Even if they second course in PL/I, even though they would be do a good job in teaching little details, they do teaching an introductory course. Thus, they would not always see the global point of view. In addi- be exposed to more of computer science than they tion, they tend to be less sensitive to some of the were expected to teach. In addition, weekly meet- things that we consider to be important (e.g., ings were held to discuss the style and presentation style of programming).

65 This is especially true for old Fortran programmers Second, we have gained a number of new friends who have a lot of bad habits. We are trying to in other departments. Most of the retrained facul- solve these problems by requesting these people to ty members are quite happy to be teaching computer broaden their exposure to computer science, by ask- science to motivated, intelligentstudents (for ing them to study other aspects of the field (e.g., many, the altarnative is teaching remedial mathe- assembly language or data structures), and by em- matics). It is aisc good for them professionally, phasizing to them what we consider to be important. since a knowledge of computer science will almost In particular, by having them take a second or surely be helpful as a. tool in their own disci- third course after the introductory one, we hope plines. (It will be interesting to see if any that they understand more clearly what a stu- novel uses of a computer in other fields of re- dent who h as completed the first course should search are introduced as a result of our program.) know. Teaching in our department has made them more' a- ware of many of the CIS Department's specialneeds Another potential problem that we were con- - e.g., the continual need to upgmde equipment. cerned with was the development, especially in the In many, cases, they have served as our spokesmen Administration, of a false impression that anyone on college-wide committees investigatingthese can teach computer science. Fortunately, in prac- needs. Most of them have gained an increased tice this has not been a problem. We have continu- appreciation of computer science as a legitimate ally emphasized the need for special training and academic discipline.As noted above, tnese facul- for a final review of all candidates by our de- ty members have maintained their positions within partment. So far, we have been successful in re- their own departments.We hope that they will sisting the temptation to flood CIS with a horde share their increased awareness of computer science of new instructiors, and we have been able to as- with their colleagues. sert our authority by rejecting certain candidates because of a lack of preparation. Finally, we have shown the Brooklyn College Administration that we are trying to help the col- However, there are real problems with super- lege as a whole. In a time of severe budget cuts vising people once they are teaching. Our depart- and talk of dismissal of all untenured staff, we ment has a preponderance of young faculty members, have made a good-faith effort to help solve these especially in comparison to the more established problems. The college recognizes these efforts, departments. It is quite hard for an untenured and has been generous in meeting many of our other faculty member (or a graduate student or an adjunct requests, including the hiring of several new com- lecturer) to criticize a senior faculty member from puter science faculty members. It would be easy another department. Thus, our chairman and one or to insist that the CIS Department needs five new two other senior computer scientists have been full-time faculty members per year. But given the forced to serve as intermediaries in a number of current demand for computer science faculty, it delicate situations. would be quite hard for us to find five competent people each year. By retraining faculty members The problem that we are most concerned with is from other departments, we have been able to con- related to this. How can we eliminate someone centrate on finding one or two good people per year from another department who is doing a poor job? and we are able to maintain the quality of our We have one person who seems to be relatively weak faculty. as a computer science teacher (he may very well be We have weak as a teacher in his own department). Summary and Prospects for the Future made a number of suggestions to him, but with little effect. As of now, we would rate the job he is do- In summary, we are pleased with the job being ing as adequate, so the problem is not really acute. done by most of the faculty members from other de- However, we may one day be faced with a situation partments. To repeat, we do not view them as per- in which an instructor is doing an unacceptable job. manent substitutes for legitimate computer scien- Will we be able to remove such an instructor? tists; but we do accept them as short-term replacements. Many colleges and universities either al- Despite the criticisms mentioned above, we are ready face or will soon be facing similar staff- pleased with the results of the program. First, ing problems. We hope that our program can be used our introductory courses are being covered in a as a model at other institutions. reasonable way, enabling us to maintain an assortment of advanced undergraduate and graduate courses. The instructors from other departments are all experienced teachers, unlike most adjunct lecturers who are usually full-time programmers, not teachers. Thus, we do not have to worry about missed classes, unprofessional conduct, etc.These faculty members are all familiar with Brooklyn College students and academic regulations. We encourage them to attend our department meetings and they have provided a number of interesting perspectives on some important issues.

66 References

1. Peter Denning, "ACM President's Letter; Eating

Our Seed Corn" . Comm. of the ACM 24, 6 (June 1981), 341-343.

2. Peter Denning et al, "A Discipline in Crisis: The Snowbird Report", Comm. of the A".',14 24, 6 (June 1981), 370-374.

3. Gerald Engel and Bruce Barnes, "Employment Decisions by Computer Science Faculty:a Summary of the 1980-81 NSF Survey", Proc. of the Thirteenth SIGCSE Technical Symposium on CoMputer Science Education (Feb. 1982), 167-159.

4. Keith Harrow, "A Faculty Development Program", Proc. of the Thirteenth SIGCSE Technical Symposium on Computer Science Education (Feb. 1982), 170-173.

'5. J. F. Traub, "Quo Vadimus: Computer Science im a Decade", Comm. of the ACM 24, 6 (June 1981), 351-369.

67 HOBBY ROBOTS AS TEACHING/LEARNING TOOLS

Michael Moshell Charles Hughes The University of Tennessee The University of Central Florida Knoxville, TN 37996-1301 Orlando, FL 32816

Carl Gregory, Lee Wittenberg Gentleware Corporation Knoxville, TN 37919 Abstract These mental abilities are certainly necessary to the development of good programmers. However, no The development of hardware and software for single introductory course can begin to make an simple robots is proposed as a follow-on experience employable computer professional of a total beginner. for students who have completed an introductory computer programming course or tutorial. To set such goals for a course is to delude both oneself and the students. Goals and a syllabus for a course, equipment and software needs and resources are described. We raise the same objection to the view of Emphasis is on easily available, inexpensive robotics as a subject matter appropriate for resources and the use of microcomputers as control vocational, or perhaps pre-engineering, curricula. devices. Without diminishing a course's usefulness in those roles, it is possible to,structure a course that The course is intended to teach principles of serves much broader goals in developing thinking physical mechanics, cybernetics, computer science, and reasoning skins. and to develop problem solving skills. We would also object to assertions that the The course of study being designed is appro- teaching of robotics should be motivated by a com- priate for both individual and class use. The petitive desire to compete with Japanese (or anyone course is embodied in a tutorial book, and is else's) industrial automation. The goals for our supported by a software package and a hardware educational system must remain firmly based in computer interface for robotics experiments. general (though "hard-nosed") knowledge and skills. Premature "targetting" of specific technologies is Outline a dangerous form of short-sightedness foreducators.

1. Goals of the Course *Course Goals

2. Syllabus for the Course We intend to develop skill and insight in three areas: 3. Hardware for Hobby Robotics mechanics/physics, 4. Software Concepts and Tools computers,cybernetics, and planning/problem-solving.- 5. Speculations about Impact Let's explore each area briefly. 1. Goals of the Course Mechanics *"Practical" versus "Foundation" Courses One of the most popular experiments in standard It is often the case that courses in computer introductory physics is the construction of an programming are viewed as "vocational" or at least electric motor from nails and magnet wire. This of a practical, or applied nature.Math or English motor illustrates rather well the state of electro- courses are viewed more as "foundations", sobrces mechnical technology in Edison's time, and teaches of conceptual skills that can be applied to alMost a direct, hands-on sense of what magnetism actually This hazard is(odoubly all other mental activities. is. relevant to any proposal for a robotics course

We samesort of-experience,with- , -- Why shouldn't introductory programming and simple robots. The issues of force, torque, friction robotics courses be regarded as practical or and material strengths are fundamental to mechanics;

vocational? yrt most standard lab-experience provides very , limited intuition into these issues. We have long maintained (e.g. Aiken and Moshell, 1982; Moshell and Hughes, 1982) that an introductory The ,authors' eXperience with small rebots programming course is primarily an opportunity to indicates that they are high;y motivating, and develop problem-solving skills and logical thinking.. supply exactly the kind of intuition and practice

68 with mechanical design that is needed. 2. A Syllabus

The robotics course should have a physics pre- This syllabus presumes that certain equipment or co-requisite course. is available; the details of the equipment are described in the following section. Briefly, the Cybernetics following items are needed:

Programming is usually taught primarily as a -A simple remote-controlled toy car; hands-on skill. Little opportunity is provided to consider how a program is connected to its users and -Components for a three-degree of freedom hand; to the world. (an Erector Set with three motors);

Programs to control robots are conceptually -A transporter base; (we use the "Big Trak" toy, simple, yet their design is difficult. The problem ana equivalents); is to understand the delicate relationship between sensory 4nformation and control of motion. -A microcomputer with a controller capable of sensing the positions of up to 8 microswitches, This emphasis on feedback, timing, proportional and of controlling up to 8 motors; control and error correction constitute a domain called real-time programming. The issues are not - Some miscellaneous switches, wires and connectors; intellectually very deep, at the introductory level, yet there is much art in the careful design of work- - A software system designed for educationalrobotics. ing systems. If the school or individual has an Apple II This presents the course designer with the with 64K of memory and one disk drive, the additional opportunity to "sneak up" on the student. By pre- equipment and software needed for this course will senting stimulating material with important content, cast about $400. in a format that demands many iterations and re-tries, the course can build firm foundation skills in the Most of the support software and computer desired areas. hardware is not yet available for other computers. The tutorial book will include specifications for Organization its construction by dedicated experimenters.

For the individual or for a class, the design Sequence of Events and construction of a robot is a complex enough task that planning is required. Eight lesson modules are intended to span an entire year of high school coursework. The first Merely starting to build something is most four modules could be used for a half-year. The unlikely to succeed, because the parts must work timing and relative importance of modules will have together. to be determined after an experimental teaching of the course. A robot project can be factored into several components, with each being handled by a team of MODULE 1: INTRODUCTION TO CONTROL students. A typical breakdown is: Reading: General -:oncepts of robotics. Robots are -Transporter; general purpose programmable machines which manipu- -Hand and Arm; late things. They have senses and us2 feedback. . -ControlDevices; - Computer Software. Lab: Use remote controlled toy car. Then try to use it when you can't see it; another student gives The software can be further broken down by tasks, so you instructions and you try to maneuver it through that one team writes programs to move about the a course. Strategies are analyzed and written down. room, another to find something with the arm, another to assemble a structure, etc. MODULE 2: HANDS AND MECHANICS

Reading/Discussion: Degrees of freedom. Force, Once one 'or more operational robots are torque, friction, types of motors (servo, actuator). available, a different level of skill-building is possible. Contests can be held for the most effec- Lab: Assemble a simple three-degree arm, controlled tive programs and strategies to build a castle from by manual "winches". Measure forces on control blocks; to build a bridge across which the robot wires. Test motors with various gearings to see walks, etc. When these tasks generate the need to what forces are available. modify the robot, the students have the skills to do so, if they built the robot in the first place. Use motors to control the arm. Try to pick up These interactions between task and design are small objects.' exactly what engineering is all about. MODULE 3:SENSORS Let us now consider a syllabus that organizes these concepts into a series of lessons. Reading/Discussion: Binary (touch) sensors; analog

69 8 (angle) sensors. Introduction to programming with 3. Hardware these sensors. The primary obstacle to hobby robotics, and the Lab: Write a provam that displays the status of focus of most of the literature, is the design and the arm, using two angular and two touch sensors. construction of the mechanical robot itself. Using the direct motor controller, and the program display of arm position, try to pick up an unseen It is rapidly becoming easier to build satis- object whose position is known. factory robots for hobby and educational purposes, as more sophisticated components becomeavailable. Then try to "map" an object by touching it. Robotics Age magazine is a good source of inspira- Is it long or short?How high is it? tion and ideas.

MODULE 4: -GOAL SEEKING The approach we have taken is to use the most complete subsystems that can be bought, while Reading/Discussion: Feedback, positive and avoiding the expensive special-purpose modules negative. designed specifically for hobby robotics.

Lab: Add two microswitch sensors to the remote For instance, we use a large Jeep model car control car or transporter; one for "edge of table", from Radio Shack as a transporter base. It is one for "object in front". Try to find a block of equipped with a proportional servo motor for steer- wood on a table without falling off the table, ing. The "Big Trak" toy is another useful trans- working "blind". Use other students to report car's porter. X,Y position. In the book we are preparing, the construction MODULE 5: PROGRAMMED GOAL SEEKING of three robots is detailed. The first (R1) is a simple Erector set construction. The second (R2) Reading/Discussion: Introduction of program uses a toy car as its transporter base and has a features that control motors, interacting with simple home-built arm. The third (R3, or "Woody"), sensors. uses lawn mower tires and a more substantial transporter, and has the most sophisticated arm. All Lab: Try to write a program which does what you share a common control design and contru,ler unit. just did up in Module 4. First cut: "trivial": car just explores along a one dimensional "track". Robot R3 (used in examples later in this paper) One direction is fall-off, one is block. has an arm with four degrees of freedom: lateral rotation, whole-arm and forearm elevation, and hand MODULE 6: HAND WITH TRANSPORTER grasp. It has independently controlled left and right wheels, and is supported fore and aft on Reading/Discussion: Balance and Stability; trigo- caster wheels. (See Figure 1) nometry for calculation of hand position in space. Sensors for accumulated motor travel, The following sections describe hardware and software packages being developed to support the Lab: Mount the arm on the transporter. Try manual tutorial book. control of the entire system to locate and lift small objects, first visual and then blind (with 4. Software other students and position sensor program for feedback). Using an unmodified Apple II or Radio Shack computer, it is difficult to control a robot. An MODULE 7: PROGRAM HANDS interface of some kind is necessary.

Reading/Discussion: Systematic design of complex Several of the commercially available robot programs. How to search a space. arms use serial ASCII communication, scyour computer needs only a serial port such as a printer interface. Lab: Program what you did manually in Module 6. The controller under development for this pro- MODULES: ROBALL ject uses serial communication, but doe.; not require a separate ASCII interface with the Apple II : Reading /Discussion: Description of the game. Two computer. It uses the game cbntroller port and a teams, each with one or.., more robots, attempt to special software system, instead. grasp a softball in the center 'of a pingpong table and, despite all efforts by the opposing team, PASPAL for Robotics return it to their own end of the table. PASPAL is a user-friendly Pascal interpreter. It Two divisions: one under manual control, one under supports advanced animation graphics, and many computer control. features designed to help the beginning programmer.

Lab: Design of a Roball game appropriate to the The ROBOT Library Unit for PASPAL adds a particular robots available. If only one machine variety of data structures and procedures to PASPAL's exists, make it a time trial between teams of dialect of Pascal. These commands make it possible operators or programmers. to operate the motors of a robot until some

70 condition on the sensors is met. Commands and Functions

The syntax and features described below are The intention of the PASPAL/ROBOT Library preliminary, and are likely to be changed as the Unit is to supply a kind of medium-level command system is tested and refined during 1983. set for our robots.

Before we explain PASPAL/ROBOT, we need to A low-level command structure would consist briefly discuss motors and sensors. solely of commands to start and stop motors, and functions which reported the state of sensors.The Motors difficulties in using such a system are many. Every action must be placed in a REPEAT or WHILE loop, The two types of motors in common use in hobby with the program looping until a sensor condition robotics are called "servos" and "actuators". occurs. If the sensor condition was transient, such as the detection of the edge of the table(before A servo is a motor whose output is some falling off), and the loop was slowed by the inclus- restricted motion, sic as the turning of an arm ion of several motor control commands, disaster through 90 or 180 degrees. A servo can be commanded could occur. to set its output to some value, such as 50 degrees, and (if it is strong enough to overcome the load A high-level command structure ot;,. contain imposed on it), it moves to that angle and stops. commands like "GO TO LOCATION 22,45" and -Cs:'K UP A LARGE BALL". Obviously there is a high levei of An actuator is a motor which can simply be sensor, and program, sophistication required to turned on or off. Almost all actuators are revers- supply these commands as primitives. It is expected ible. Usually an actuator iS paired with some kind that the creation of these actions as procedures of sensor so that the actuator, sensor and control- will represent something close to the ultimate capa- ler together behave as a servo. That is, a given bility of the command set to be provided. angle or position is achieved and the motor is then turned off. Actuators are usually either DC motors, A high-level system would also permit several or stepping motors. independent actions to occur simultaneously.This degree of parallelism is only possible when the If the controller is so designed, a servo can hardware supports reasonably powerful interruption also maintain the position of its output under capability. Inexpensive microcomputers such as the varying loads. This is important if, for instance, Apple II and TRS-80 don't usually have this power, the servo controls the first segment of a multi- so we are prevented from considering truly concur- segment arm. Usually, however, the friction of the rent processes in simple command systems. motor and drive train is used to maintain position when the part is not moving. A concurrent robotics system for more advanced microcomputers such as the IBM and DEC machines will Small servos are available for $20 to $40 from perhaps be developed later. The current system model airplane shops. Actuator motors are available contains some limited concurrent capability, for prices from a few dollars up, at many hobby described below in the section titled "BINDING". shops. We will now describe the "medium level" command Sensors set used by PASPAL ROBOT. An example program follows. Sensors come in many forms, but the three most popular for robotics work are switches, potentio- Software and Hardware Structures meters and photocells. The software supports up to 32 motor control A "microswitch" is a very sensitive switch; channels, and 32 sensors (which are usually switches) typically the weight of a sheet of paper will The hardware is expandable in units of eight motors operate a microswitch. These'are used to sense a or sensors up to this limit. robot's touching something, or the arrival of some part at the limit of its motion. The Apple II also provides four analog inputs.

A potentiometer is a rotary (sometimes straight- Sensors may also be defined as "logical sensors" line) device whose motion is translated into a vary- which are combinations of other sensors, or functions ing resistance. The Apple II computer can detect of the state of an analog input. In effect, sensors a resistance value between 0 and 150 Kilohms on any are like boolean functions. The usefulness of this of four "game paddle" inputs, so this is a conven- will be apparent later. ient way to report the angle of an arm or shaft. The motors and sensors are referred to by A photocell is actually a variable resistor, integers 0 through 31. CONST declarations in Pascal which can be used as a game paddle input or, with make these more legible. a simple circuit, as a switch input to detect the presence or absence of light. We frequently use CONST declarations for the analog sensors attached to game paddle inputs, also.

71 For instance: The usefulness of these commands will become apparent when we begin to explore the motor control CONST LEFTMOTOR=0 commands. RIGHTMOTOR=1 ELBOW=2 Motor Control FOREARM=3 HAND=4 Motors come in two kinds, servo and actuator. WHEELS=5 There are three simple "unconditional" commands for TOUCHFORWARD=0 the two types of motor. STIFFARM=1 HOLDING=2 *SERVO(FOREARM,60) ELBOWANGLE=0 ARMANGLE=1 causes the forearm motor to be set to 60 percent of its fullscale deflection. These declarations provide us with six symbolic motor names, three symbolic sensor names, and two *SETDIR(LEFTWHEEL,FORWARD) symbolic analog input names. causes the actuator motor LEFTWHEEL to run in a for- In the following, each new command is marked ward direction when the motor is activated. Replac- with a star (*) in the left margin, to make it ing FORWARD with REVERSE would the motor to run easier to locate. backward, whenever it is activated.

Sensors *ACT(LEFTWHEEL)

*The boolean function SENSE(whlch,TOUCH) returns turns on the left wheel. The motor runs for 0.1 TRUE if sensor number "which" is touching something. second each time this command is encountered in Similarly, SENSE(which,NOTOUCH) returns TRUE if the the program. sensor is not touching something. We say that the sensor has value TOUCH or NOTOUCH in these two cases. This simple command would suffice to control some kinds of motion. For instance, if FINGER were The sensors can be combined into "pseudo-sensors a sensor that detected the hand's having closed on by several commands. For instance, if we execute something, a loop like

*ANDSENSOR(SENSE1,SENSE2,BOTH) REPEAT ACT(HANDGRASP) we have now created a new "pseudo-sensor" BOTH UNTIL SENSE(FINGER,TOUCH) (which must be declared with a CONST and a value between 0 and 31, of course.) BOTH will have the would cause the hand to close on an object. value of TOUCH when'SENSE1 and SENSE2 both have value TOUCH. However, a smoother and more rapid action can be achieved by combining the ACT and SENSE commands Similarly, into a "conditional actuate" command:

*ORSENSOR(SENSELSENSE2,0NE) *CONDACT(HANDGRASP,FINGER,TOUCH)

would cause ONE to have value TOUCH whenever at which has the same effect as the previous REPEAT least one of SENSE1 and SENSE2 has value TOUCH. loop.

*NOTSENSOR(SENSE1,N0SENSEl) Similarly, we have a "conditional servo" command which allows the servo to quit trying to achieve would cause NOSENSE1 to have value NOTOUCH whenever some angle, if a sensor de :acts a problem. SENSE1 has value TOUCH. *CONDSERVO(FOREARM,60,COLLISION,TOUCN) *ANASENSOR(PADDLE1,GREATER,100,HIGH1) would cause the servo to try and tutu the arm to 60% causes sensor HIGH1 to have value TOUCH whenever the of its full deflection, unless sensor COLLISION analog input PADDLE1 has a value greater than 100. detects a TOUCH.

*CNTSENSOR(TRIGGER,CNTFLAG,100) We also support commands that incorporate counting directly into the motor control. causes sensor CNTFLAG to have value NOTOUCH until TRIGGER has changed to TOUCH and back to NOTOUCH Assume that DISTANCE is a sensor which "toggles" 100 times. Then its value changes to TOUCH. from TOUCH to NOTOUCH and back again every time the drive wheels rotate 1/4 turn. This is used when we need to do something until a given number of counts are recorded. For instance, *COUNTACT(WHEELS,DISTANCE,40) TRIGGER might be "toggled" once for each 1/4 revolution of a driving wheel. The sensor counter would would have the same effect as the following: be used to measure distance travelled.

72 CNTSENSOR(DISTANCE,CNTFLAG,40) command must occupy a separate line. CONDACT(WHEELS,CNTFLAG) PROGRAM FINDBLOCK That, is the robot rolls forward a distance of (* ten full revolutions of its drive wheels. A Program for Robot R3, in the PASPAL/ROBOT Dialect of Pascal. *TIMEACT(WHEELS,20) would cause the robot to roll forward for 2 seconds. (Time is measured in units Moshell 12/82 of 0.1 second). This program guides R3 around the floor randomly Several other commands, particularly related until a low-lying block of wood is detected.The to arm positioning and control, are omitted here computer then signals with a "beep" that the block for reasons of brevity. has been found.

Binding SENSORS There are times when you want two motors to act *) in unison; but separate ACT or CONDACT commands would CONST OBSTACLE=1 cause one motor to act at a time. For instance, a LOWTOUCH=2 (* PHYSICAL SENSORS *) robot with separate left and right drive motors should use both motors at once to walk forward. ANYTHING=3 (* LOGICAL SENSOR *)

*BIND(LEFTWHEEL,RIGHTWHEEL,WHEELS) (* MOTORS causes any subsequent actuator commands directed to *) WHEELS to cause both LEFTWHEEL and RIGHTWHEEL to act. LEFTWHEEL=1 RIGHTWHEEL=2(* TRANSPORTER MOTORS *) It is still possible to refer to LEFTWHEEL and RIGHTWHEEL independently. Application of SETDIR to WHOLEARM=3 WHEELS would reverse the direction of both wheels FOREARM=4 (* ARM CONTROL MOTORS. *) when actions were later requested of WHEELS, but ARMANGLE=5 (* UNUSED IN THIS PROGRAM *) not of course when actions were requested of indi- GRIP=6 (* DITTO... *) vidual wheels. WHEELS=7 (* LOGICAL MOTORS, FOR BINDING *) Example Program ARMS=8

The following program contains several proce- PROCEDURE RETRACT (* PULLS THE ARM UP SHORT. *) dures that would be standard parts of any program for R3 ("Woody"). These procedures constitute the (* "how-to" for retracting the arm, turning in place, WHOLEARM retraction is positive (upward); but and initializing the bindings and counters that are FOREARM retraction requires negative (downward) to be used. motion. *) The actual main orogram which uses these procedures is very short. Its purpose is simply to look BEGIN randomly about a room until it locates a block of SETDIR(FOREARM,REVERSE) wood lying on the floor. R3's arm projects about 4 TIMEACT(ARMS,100) inches in front of the leading edge of the trans- SETDIR(FOREARM,FORWARD) porter platform, when the arm is fully retracted. END ., Anything touching the arm is considered to be an PROCED/RE INIT (* SETS UP BINDINGS AND ORSENSOR. 9

obstacle, and activates the sensor OBSTACLE. BEGIN; '

(* FOREARM RETRACTION : *) A second sensor, on the front caster, detects BIND (WHOLEARM,FOREARM,ARMS)

objects lying on the floor. . It is assumed that the

wood block is the only thing lying on the floor. (* FOR FORWARD MOTION : *) This sensor is called LOWTOUCH. BIND (LEFTWHEEL,RIGHTWHEEL,WHEELS)

When R3 encounters an obstacle, it backs away (* FOR DETECTION OF EITHER KIND OF COLLISION : *) and turns a random angle (a multiple of 45 degrees), ORSENSOR (OBSTACLE,LOWTOUCH,ANYTHING) then resumes walking. This simple a program is capable of being trapped, if the arm is by chance RETRACT (* THE ARM *) inserted into a hole from which the "back ,way" END maneuver doesn't extricate it. The OBSTACLE sensor would have to be sensitive to touches on both the PROCEDURE ROTATE(DEGREES:INTEGER) side and the front of the arm to free R3 from this trap. (* Turns the whole robot an angle of DEGREES (counterclockwise, looking down, Since these commands are imbedded in the PASPAL is positive.) dialect of Pascal, no semicolons are required. Each

73 DEGREES is rounded down to nearest multiple Programs for arm control are necessarily someof 45 degrees, because that's the accuracy what more complicated and require analog arm position of the wheel turning sensor switches. sensing. * ) 5. Speculations on Impact BEGIN IF DEGREES>0 THEN Whereas our earlier effort at designinl, a pro- SETDIR(RIGHTWHEEL,REVERSE) gramming curriculum ("Computer Power", cf. Aiken and ELSE Moshell, 1982 and other papers in bibliography) was SETDIR(LEFTWHEEL,REVERSE) explicitly designed for the median student, we do not regard the robotics tutorial as appropriate for (* that audience. Sensor ANGTURN pulses once per 45 degrees. At the current state of development of hardware *) and software, the probability of successful, working robot systems is going to be marginal for all but COUNTACT(WHEELS,ANGTURN,N DIV 45) the brightest and most highly motivated students.

SETDIR(RIGHTWHEEL,FORWARD) This curriculum will, therefore, have to accept SETDIR(LEFTWHEEL,FORWARD) the burden of a certain kind of elitism. END Teachers may be surprised, however, at just who PROCEDURE BACKOUT (* Pulls R3 back from obstacle. *) the "robot elite" turns out to be.The authors BEGIN suspect that the skills of "vocational track" SETDIR(WHEELS, REVERSE) students, with tools and materials, may make it TIMEACT(WHEELS,10) possible for them to contribute strongly to the (* RUN FOR ONE SECOND *) success of such a course. SETDIR(WHEELS, FORWARD) END If the curriculum is successful at teaching technological "basics" such as principles of physical (* mechanics, it might even turn out to be a subtle The Main Program path for upward mobility for certain students. k) The authors eagerly look forward to in-school BEGIN tests in academic 1982/3 and 83/4, and will report results in later articles. IN IT Bibliography REPEAT Aiken, R. M. and Moshell, J. M. "Computer Power". (* Run forward until you touch anything *) The Computing_ Teacher, 9:8, April 1982. CO...WACT(WHEELS,ANYTHING,TOUCH) IF SENSE(OBSTACLE) THEN DaCosta, Frank. How to Build your own Working BEGIN Robot Pet, TAB Books, Blue Ridge Summit, PA. 1979. BACKOUT Hughes, C. E. and Moshell, J. M. "Rascal and (* Rotate 45,90 135 or 180 degrees,randomly. *) INTERPAS: Graphic Programming Tools for Kids". ROTATE(RAND(4) * 45) The Computing Teacher, 9:9, May 1982. END Moshell, J. M. and Hughes, C. E. Robots and the UNTIL SENSE(LOWTOUCH) Personal Computer. John Wiley and Sons,'to appear in 1984.

NOTE(20,200) (* Beep your success ! * Robotics Age. P.O. Box 801, La Canada, CA. 91011. Fore thews END. EleaOrion

VVI4, t /VIA Roil:bn Figure 1: R3 ("Woody"), a MobileHobby Robot Lo+end Hand Rotation isk°Pen:if-ion

74 ENDING THE ISOLATION: DEAF-BLIND AND MICROCOMPUTERS

by Dan Zuckerman

Human-Interface Laboratory, Science and Technology Studies, Rensselaer Polytechnic Institute, Troy, NY 12181

Abstract participate in amateur radio. They can communicate by "speaking" with a telegraph A method to enable a deaf-blind person key and "hearing" thevibrations from a to work with a microcomputer is described. speaker cone with theirfingertips. There Morse Code is used tactilly as a general is an enormous potential for all deaf-blind interface to the screen. Techniques, people to actively participate in society. experiences, and directions for future work By harnessing the power and versatility of are discussed. today's inexpensive microcomputers with the communicationpossibilities of Morse code the intellectual andproductivepotential of the deaf-blind may be released. Microcomputers can be used tobroaden the horizons of the deaf-blind and to Mypurpose in writing this paper is to greatly expand the number of deaf-blind who describe such a Morsecode/microcomputer are able to interact effectively within the interface, and the experiences of one deaf- context of sighted hearing people. More blind user. The interface was implemented than 15,000 people in the United States are on a Radio Shack TRS-80 Model bothdeaf andblind. Thisoverwhelming microcomputer, a machine that costs about handicap, being able to neither see nor to $1000. The machine uses the Basic language, hear, was describedby Helen Keller as a has sixteen thousand characters of memory, "dark, silent imprisonment." The lack of and uses cassette tapes for long term educational and training opp)rtunities that storage. Development of this software took are concomitantwith their communicative less than thirty man-hours. The keyboard is deficit3 contribute further to the used by the deaf-blind individual in the isolation of the deaf-blind. Specialized same way most people use it. training is intensive, requires extraordinary committment from the teacher, The Morse code interface enables a deaf- and requires physical contact for tactile blind person to interpret the screenjust communication. as a seeingperson would. As each key is pressed, that character is sounded out in A minority of deaf-blind, such as Helen Morse code. These sounds are "heard" by a Keller, have demonstrated that if the deaf-blind person as vibrations. Mr. Ray communication barrier can be breiched, the Boduch, the deaf-blind user in this deaf-blind can productively and effectively project, developed a circuit to substitute interact with their world. An even smaller a doorbell buzzer for a speaker so that minorityhave found a means and a context Morse code can be felt on a sensitivepart in whichtheir handicap is completely of the body such as the neck or thigh. An eliminated so that they can effectively infrequently used key serves as an escape participate in aninternationalcommunity into screen reading mode. The user presses of seeing-hearing people. These are the the shift and right-arrow keys followed by deaf-blind who use Morse code and a'key-indicating line'is to be heard. The sixteen lines on the screen are represented by the characters 0 through 9 and A through F. Upon pressing one of these keys, the characters on that line of the *I am, especially indebted to Dr. Linnda screen arecommunicated to the userin Caporael, friend and teacher, for her help Morse code. At this point, any key will and encouragementwiththis and other serve to temporarily pause the display and projects, Josepl.i L. Hartmann Jr., who any other key will resume the discovered the Morse code connection and communication. The enter key is available involvedme with it, and Ray Boduch, a to stop in mid-line. A utility program is source of inspiration for myselfand many used to control the speed at which others. information is sent. If the "escape" key is

75 weeks. Teachers of computer skills need pressed twice, the program will tell (in their on the screen is only have enough background to start Morse code) which line students off. Once the interest issparked, presently being typed (contains the it is easy to 1parn independently from cursor.) informatiph accesssibleondiskettes for the systembeing used. If more than one This software is designed so that it is student is involved, they will learn from always availableto the user. In general, each other. it permits theuse of any software available for the machine and allows the I assume the student knows how to touch- deaf-blind to write their own computer type, has already obtained a grasp of Morse programs. Even if the machine is busy with code, and has had the mechanics of starting complex calculations the deaf-blindperson the computerdemonstrated to him or her. can watch the screen as it changes, The first skills that need to be taught to same enabling him orher to share the a new computeruser involve reading the perception of the screen as others. screen. The teacher needs to explain that there are sixteen lines of print on the Since Morse code was not designed for screen, whichmay change after each computer work, it does nothave all the keyboard entry. The studentneeds to be characters needed. AppendixAexplains shown how to systematically read the screen extensions that were developed to display by pressing simultaneously the shift and all the characters that a TRS-80 uses. Most right-arrow keys followed by the number of of these characters are currently used by a line on the screen. The screen should be ham radio operators. The software interface readfrom the top down by pressing shift for this computer is in thepublic domain right-arrow, zero to read the first line, and is available from the author. shift right-arrow, one to read the second line, shift right-arrow, two to read the The interface described abovewas third line, and so on. Emphasis should be developed and tested in collaborationwith made on the necessity to methodicallyread Ray Boduch, whohas been blind and deaf the screen so as to be aware of how it is almost since birth. By way of background, changing. The user should also be Mr. Boduch can readBraille, read lips encouraged to read the line he is presently (with his thumbfeeling the lips and a typing. Thenumber of that line may be forefinger feeling the vibrations in the found by pressing the shift and right-arrow throat), and speak. At the age of 14, he keys twice. The current line number will be was taught Morse code and successfully heard in Morse code instead of echoing the schoolmate. The taught the code to a deaf Morse code signal for the second shift significanceof Morse code is that it right-arrow. Then, this line may be read by tactile eliminates the necessity of pressing shift right-arrow followed by the contact for communication. At 23, physical character justdetermined. Characters not a neighbor, Mr. with the assistance of previously learned will have to be Boduchbecame a ham radio operator. He can the context of the computer. send and receive Morse code at 50 words per explained in ">" as a speed For example, the TRS-80 uses minute: the federally set minimum to indicate that it is awaiting the amateur license is 5 prompt competency foran user to type something. words per minute. presently, Mr. Boduch is employed assembling electronic parts. He an Although the deaf-blind effectively use hopes to become a computer programmer, trial-and-error method for learning, the is onlynow possible a opportunity that extreme flexibilty of the computermakes of the Morse code/microcomputer. because "playing" with it, a common strategy for interface. introducing novices to computing, an unwise choice for the beginning deaf-blind user. Although it is difficult to generalize deaf-blind It is very important to present simple, from the experiences of one how program user, the activity is useful for suggesting clearly shown examples of input, listing, and execution differ. The directions for further research and user side first program should consist of something development, especially for the Character strings, human-computer interface. The user like "10 PRINT 2+2". of the immediate execution side involves both a student and a teacher. input statements, and mode (a statementwith no line number) should not be demonstrated. It is critical What does an educator of the handicapped at the initial stage that the input program need to learn to be able to teach computer be completely different from the results. To start a person learning will cause good Similarity between them computers, a teachermust have a the program 10 a language such as Basic confusion. For example, understanding of PRINT "Hello,I am yourTRS-80 Computfr." and the techniques for using the available familiarity is will only confuse the very central issue of machine. SomeMorsecode how listing a program differs from using a also needed to understand the student's program. The user should also be taught, at with one feedback. This can be learned least initially, to clear the screen with month of intensive study. Mr. Boduch taught the CLEAR keybefore each command. This Morse code to adeaf17-year-old in two

76 9 will erase the screen without affecting the Nevertheless, Braille output has an computer's memory. Theuser needs to important role. It provides a less recognize that clearing the screen makes it immediate method of communicationwhich easier for him to evaluate the effects of a doesn't require remembering the entire command. contents of a line. Thecharacters of output can be easily scanned and carefully The firstmulti-step program should be observed rather than heard just once each the speed-varying utility. This program time a line is read. uses some statements to control the speed ald timing of the Morse output. In order to Ray was recently asked what he thought work with this program, the user needs to about hoy readily deaf-blindpeoplecould change the numeric values contained in the learn to interactwith the computer via program. The results of doing so and Morse code. He said that any deaf-blind executing the new programwill be very person who could learn Braille could easily apparent: the speed of the Morse code learn this. "Yes,I am very sure, as long interface will change. Though it is as they know what is going on." difficult for the deaf-blind user to understand, at this stage, how changing A significantnumber of intelligent, just a few numbers can change the way a thinkingpeople who are presently blocked computer sends Morse code, the excercise bycommunicative barriers can become will develop an understanding of justhow participants and contributors to the Technological versatile the machine is. It will also microcomputer revolution. demonstrate how it is possible to defeat developments presentlyavailable at low cost can be used by educators and others to the Morse code interface software. This is an important lesson because, as the user make a major breakthrough in the welfare of becomes more experienced, sophisticated the deaf-blind. experimentation may possibly do this. The remaining barriers to this release of intelligence are social, not computer is a very different The the Federal a deaf-blind technological. For example, educational experience for Communications Commission requires not only person because he or she must initiate a minimum speed competency in Morse code, their learningby using examples, rather deaf- but also a technical knowledge of things than by trial-and-error methods. The as Ohm's law. The requirement of this do have the experience of such blind not technical knowledge shouldbe waived for learningby following instructions and is such a. Their entire the deaf-blind. Morse code understanding examples. tremendous break-through for the deaf-blind perspective on life is experimental. Very that legislation should be passed to more rarely are they presented with instructions license these people. The world of that make sense to easily for new experiences radio made available them. In spite of this preference foran amateur should be solely on a statement of desire to experimental approach, it should be non-structured communicate viaMorse code and a test emphasized that demonstrating sufficientunderstanding to experimentationwiththe computer, rather than trying to understandand repeat the communicate v!a ham radio. examples presented, will accomplish little. Through computer "cottage industry", a difficultpoint to get This is very deaf-blind people working athome or in across to someone whose entire life experience involves tinkering with their small businesses can compete as equals for environment until the desired effect is employment rather than having no option but achieved. to work in sheltered workshops at occupations below their intellectual Presently, Mr. Boduch uses his TRS-80 to capacity. No prejudice canaffect the exchange letters with a deaf friend via personwho submits a floppy disk for sale. program tapes. The lack of Brailled With the tremendous potential at hand, once technical materials has been a shortcoming a few of the deaf-blind come out into the of using Basic on the machine. The software world, the rest will surely follow. -interface for an IBM personal computer is presentlybeing designed. One advantage of A row of retractable plastic pins above the IBM is that thedocumentation is Braille keyboard are controlled so as to be the readable as Braille. There are at least two available on disk, so it wouldbe accessible with the Morse code interface. advantages of the Morse code/microcomputer interface. One is its low cost and easy Of course, Braille terminals are access. The other is that the interface is available and are used by a growing number designed to give the user an experience as of blind programmers. The Versa-Braille is similar as possible to the usual one such device. It has a Braille keyboard communication techniques. and a paperless Braille output device that can mimic the behavior of a video terminal.

77 9 Appendix A In addition to the table, the following Morse Code for the TRS-80 charactersaredefined: BS 8 IIII This chart shows the Morse code extensions CAN 24 IIIII that have been defined for usewith the US 31 IAA TRS-80. Both the character (or control code 39 AOE name) and its decimal ASCII representation 41 NQE are qiven. Appreciation of this information 44 MIM requires a knowledge of Morse code. To find 46 AAA a particular character, combine the Morse sounds of the charactersdefiningthat column and row. Here are two examples: Appendix B < Column N, Row B, Morse NB (dah dit dah dit dit dit) Technical Description of the Morse Code/Microcomputer Interface

! Column M, Row R, Morse MR (dah dah dit dah dit) This section is provided for the benefit of the technically inclined reader who is The Blank spaces represent character sounds interested in implementing the system. available for extending this definition to software to accomplish this scheme on a TRS-80 Model I tape-based machine or a TRS- other machines. Spaces that contain boxes 80 Model III disk-based machine hasbeen are consideredsounds too complex for It practical use,. written. has been placed in the public domain and is available from the author.

A I M Q 81 The first implementation was doneon a A 95 I X 88 86 58 TRS-80 Model 1. Itotook 475 bytes of code, % 37 I < 60 SLC * B Without provision for upper/lower case. The C ESC 27 1 ; 59 ELC ** $ 8 56 entire routine resides in the keyboard D + 43 I 96 36 that E R 82 D 68' S 83 G 71 scan. A tape patchesthe keyboard vector and loads it intoprotected high F " 34 SUB 26 9 memory enables the Morse interface. G I SOH 1 & 38 57 H 72 Z 90 Upon beinginvoked, the routine calls L 76 I B 66 H I that was the J the address original.ly- in keyboardvector. The returned character is K HT 9 , LF 10 echoed in Morse code to the speaker port. I, *** The routine will. return at this point if M J 74 Y 89 I = 61 NOT right- N P 80 C 67 L F 70 the key just pressed is not shift (ASCII 25). If it is, the keyboard 0 1 49 _L 2 50 0 48 arrow vector is patched to a scan known to be P 1 123 L available inthe ROM and a character is ( 40 1 124 Q requested from this keyboard scan. R EM 25 / 47 ! 33 S SP 32 6 36 5 53 7 55 This character is not accepted until it T W 64 K 75 U. 85 0 79 right -arrow U @ 87 CR13 4 52 * 42 is a shift or a hexadecimal digit representing a line number. If a V - 45 I 125 shift right-airow, the cursor line is [ 91 j 3 51 # 35 calculated, echoed in Morse code, and the \ 92 > 62 -. 126 DEL 127 routine returns, after fixing the keyboard Y 1 93 1 vector to point to the Morse interface. 94 j 7 63 10therwise% thehexcharacter is used to calculate a starting. addressin video SLC: Start Lower CaseLUpper!casis_ . memory. The -length.. of -this- line- is__ assUMedUniiithS-CharaCter is heard. determinedwithout trailing. spaces. The line,is -sentin Morse code followed by a ELC: End-Lower Case--This-character-.:,carriFge-return.....Before_eachharacteris (as well as CR, ., !, and .7).- signals sent; the keyboard is polled with the ribm----- that uppercase is now .being heard.. keybOard scan.

. . * * * NOT: Undefined -Code. Should a code An ENTER will cause the linedisplaying not in the tablebe requested, this loop to exit so the remainder of the line code is sounded.' will not be heard. 'Any other key will pause the routine, waiting for a different key to restart the process. The Morse code keyboard routine patches itself back in the keyboard vector before it returns. The Morse code sending routine uses the cassette port to generate a tone. The most interestingpartof this routine is the table encoding the dits and dahs. This table is a bitwise representation of Morse code such that as many as 7 dahs or 14 dits can be represented in two bytes (16 bits). Dits are encoded bitwise as a single O. Dahs are encoded as a 10. End ofcharacter is indicated by a 11. For example, a Morse A (dit dah) is encoded as 0 10 11 and padded with zeroes to form the hex byte 58 (0101 1000).

The character set is encoded in two separate tables. One represents the ASCII codes from 32 to 95. This is used as a look-up table for codes in that range and uses 128 bytes. The second table contains 12 othercharacters. It uses 36 bytes to encode the ASCII code followed by two bytes of Morse. An ASCII NUL (0) indicates the end of the table. The following two bytes are the character used if the routine is asked to represent an undefined ASCII code.

Thus, it takes 164 bytes to encode 76 characters. The remaining code consumes 311 bytes. The amount of code utilized is a great concern since many TRS-80 applications use almost all of the 16K available on the machine. Implementations on machines with more memory not need be so byte efficient.

79 PLATO STAYWELL: A Microcomputer-Based Program of Health Behavior Change that Improves With Use

Murray P. Naditch, Ph.D.

Control Data Corporation

PLATO STAYWELL is a microcomputer pro- preventative medicine program focusing gram of health behavior change. It is on cardiovascular behavioral risks in highly individualized, matches people 1979. That program, called STAYWELL, to program interventions most likely to screens employees for potential risks be effective for them, and includes in the areas of smoking cessation, phys- branches so that a person's program may ical fitness, blood pressure mangement, be changed if it is not working. The cholesterol and salt consumption, weight efficacy of person to program matching control, and stress management. Em- is evaluated by a mathematical model ployees interested in changing health- that enables the program to make in- related behaviors are given opportun- creasingly accurate decisions. ities to enroll in courses and other on-site program activities related to health behavior change. This program has been described elsewhere (Naditch, In press).

In 1981, after the STAYWELL program had been implemented in approximately ten American cities, a decision was made to examine the extent to which computer technology could be used to address fundamental unsolved problems in health behavior change.

Research in the area of response to treatment for weight control is illus- Proaram Rationale trative of one of the general problems in this area. In recent years, cogni- Cardiovascular disease is the leading tive behavioral modification techniques cause of death in the United States and have been used with some success in in other Western industrialized coun- effecting weight behavior change. (For tries. More people die of cardiovascu- example, see reviews by Stunkard and lar disease in the United States than Mahoney, 1976; Jeffrey, 1976.) Although all other leading causes combined. Pre- the results of these studies are statis- vention of cardiovascular disease has tically significant, the mean group been a major focus for scientists and weight losses are often small and do not clinicians in fields of public health reach clinical significance. Consistent, and behavioral medicine because the significant individual differences among disease is primarily caused by people's patients in weight loss are observed in behavior and is to a large degree pre- studies where individual data are report- ventable. A considerable body of reed (e.g., Harris and Bruner, 1971; search and clinical practice has focus- Penwick, Filion, Fox and Stunkard, 1971; ed on the development of effective pro- Gormally, 1979). Consistent individual grams of smoking cessation, modifica- differences in response to behavior ther- tion of eating behaviors concerned with apies indicate that small mean treatment dietary cholesterol, salt ingestion, effects observed in most programs are and weight control, modification of be- misleading. This therapeutic approach haviors related to more effective con- is effective for some types of patients trol of blood pressure, as well as pro- but not for others. grams of stress management and physical fitness. The importance of identifying individual differences as predictors of success in Control Data Corporation initiated a

80 behavioral therapies for weight control 5. Using personal data to interact as well as other behavioral risk areas with each user in a familiar way. have been recognized by a number of authors (e.g., Weiss, 1977; Coates, Matching People to Programs 1977; Leon, 1976). Unfortunately, there has been very little success in finding Although there has been a significant reliable predictor variables. This body of research that has focused on lack of success may be due to the fact the effects of individual difference that most behavior studies: 1) have variables and program outcomes, there small sample sizes,2) use univariate is very little definitive work that rather than multivariate explanations, would enable ona to effectively match and 3) focus on further differentiating individual differences with the prog- the efficacy of program comnonents rams most likely to be effective for rather than the interaction of programs each person. The PLATO STAYWELL Prog- and individual responses to those pro- ram has developed a procedure to make grams. this possible. For each behavioral intervention (for example, weight control) Another major unsolved problem in the the user completes a behavioral profile field focuses on continued availability prior to beginning the program. This of social support following the end of behavioral profile contains operation- a formal program. This lack of contin- alized versions of the key variables in ued access to social support and/or per- the clinical literature that have been vayors of the program after the inter- hypothesized to relate individual dif- vention is over often results in the ferences to program outcomes. For ex- effects of the program eroding to fail- ample, in the area of weight control, ure over time. behavior profile variables include knowledge about nutrition, the degree of A final problem that resists solution social support at home, the degree of in the development of effective behav- overweight, the number of programs the ioral interventions concerns the lack person has been in previously, sex, and of significant cumulative scientific other demographic characteristics. findings in this area. Studies tend to use different definitions of dependent In the initial interation of the pro- variables, do not all follow up dropout gram, subjects are randomized across a subjects and those who complete the pro- number of intervention approaches con- grams for adequate lengths of time, tained within the program. These inter- apply varying meanings to the program vention approaches represent approaches interventions that they use, and in gen- and configurations of program interven- eral lack a sufficiently unified set of tion approaches. For example, in the definitions or unified scientific para- weight control program, some people may digm that would be required to produce either lose weight at the beginning of more cumulative scientific findings. the program using either a fixed diet, a program of avoiding certain foods and eating others, or a program of calorie A Computer-Managed Program Can Address counting. Unsolved Problems in the Field The PLATO STAYWELL Program has formula- When a sufficient sample of people have ted one set of solutions to the prob- run through the program, the individual lems of individualization, support, and difference variables are examined using cumulative scientific knowledge. These regression equations to determine their solutions are based on unique aspects efficacy in predicting outcomes at the of program users (N = 20,000). end of the program and 12 months after the program is over. Individual differ- Individualization ence variables that are useful predictors remain in the model and those that The PLATO STAYWELL Program achieves in- do not account for a significant vari- dividualization by: ance are deleted. Variables whose main 1. Matching people to the programs most or interactive effects account for sig- likely to be effective for each person, "ificant variance are then used to match individuals to program paths in the next 2. Modifying programs while people are iteration. This procedure is repeated in them as a function of their perform- with each interation, and the system is ance. gradually able to make increasingly ac- 3. Tailoring skills to each individ- curate predictions about the effects of uals' needs, natching people to program paths. 4. Tracking progress and utilizing results to engage in a dialoguelike commentary with each user,

81 Program Branching through the program. After a specific program path is selected, the user be- Each program path includes branches so gins a lesson. Lessons usually involve that individuals who are not doing well the introduction of some specific know- may move to an alternative intervention, ledge or skill area relevant to the user, have the intervention they are in en- a simulation in which the user is given riched with adjunctive material, or the opportunity to apply new information repeat certain aspects of the interven- in a life-like context, and an assign- tion they have already experienced. ment through which the individual has Each branch point is treated and tested the opportunity to try out those speci- as an alternative experimental inter- fic skills in a real world context prior vention. The efficacy of branch points to the next lesson. For example, in the are evaluated and reconsidered with each weight control program, a lesson concern- new cohort of people comprising one of ed with eating in restaurants introduces the iterations in the evaluation process. basic skills related to eating and main- In this manner, branches may be deleted, taining a low calorie diet in a restaur- new branches may be added, or branches ant, allows the person the opportunity may be kept for people with certain char- to order a low calorie meal from a simu- acteristics but not used as branches for lated restaurant menu in a social con- people who do not share those character- text in which other people are strongly istics. and tenaciously encouraging the person to eat a higher calorie meal, allows the Tailoring Skills to Individual Needs individual to track how many calories The behavioral profile is supplemented are in the meal that they have chosen, during the program with other self-re- and makes suggestions for alternative port data related to an individual's choices and for modes of handling inter- lifestyle and needs. These variables personal situations in which people are are used to suggest choices or menus of encouraging the user to eat more calor- specific lessons to users. For example, ies than he or she would like. in the weight control program, subjects who entertain clients in restaurants, When the user returns to the next lesson, eat many of their meals in restaurants, information is collected about progress or who travel frequently are offered and success over the last week. That lessons focusing on those specific is- information is used in the tracking sys- sues. In this manner, people are match- tem to present progress, and the user is ed with the skill lessons that are dir- given an opportunity to engage in a dia- ectly relevant to their situation and logue -like interchange with the computer level of knowledge, and are not exposed concerned with the success or failure of to non-relevant lessons. the assignment for that week. The dialogue is actually a simple set of options Tracking and Commentary in which the user either moves on to the next lesson, repeats some aspect of the The program tracks each user's program last lesson, modifies their goals into history, and uses that data to review smaller steps, or reconsiders their ob- progress with each user. Tracking in jectives and moves on to a new course of the weight program, for example, focuses action. on pounds lost. Tracking in the fitness program, for example, focuses on kilo- Personalization Through Familiar Inter- calories expended, resting pulse rate, action and changes in mood since the beginning of the program. This tracked informa- The computer-managed program is further tion is presented in a graphic form at individualized by having a friendly, the beginning of each lesson. This supportive tone, referring to the user's graphic tracking of progress enables name, remembering statements made by the each user to assess their own progress user earlier in the program, allowing as well as to compare their progress users a wide latitude of choice, as well with the progress of other people who as enabling the tracking of individual's have taken the program. As the program. progress and comparison with the prog- data base accumulates information, users ress of other people. will be able to compare themselves with other users who have specific demograph- Continued Social Support ic characteristics. For example, a user The program provides continuing social can ask, "How does my progress compare support by nature of its continued avail- wlth that of other white women execu- ability. Unlike conventional program tives who are my age in this company?" interventions, where the instructor or the class disperse when the course is The commentary and tracking functions over, the computer terminal continues to are key elements in the user's flow be available for the user to come back

82 93 to review, continue or reassess progress cess in the program. People interested at any time. in being an Action Team leader take a microprocessor-based lesson on how to By remembering the user's earlier data, be an Action Team leader, and then in- and by having a number of lessons that itiate Action Teams in their work site. focus on long-term maintenance, the user A special notesfiles for Action Team is able to experience the program as leaders is available. Leaders can dis- continuous, friendly, supportive, per- cuss ccmmon problems and have access to sonal, and responsive even after the expert consultants who can give them user has been away for many months. advice in initiating and maintaining these groups. The program currently can run either on a free-standing microcomputerusing Program Evolution and Theory Construc- floppy disks or a microprocessor tied tion into a central computer via a telephone The efficacy of the program in producing modem. In those instances where there behavioral change is evaluated using a is access to the central computer the multiple regression-based structural e- program provides a social support net- quation model. This structural equation work that enables people to communicate causal model can be represented pictori- with one another about their progress. ally as a causal flow diagram. The primary mode of this communication The flow diagram is one representation of a for- is through a system called PLATO notes- mal theory, and the model tests the par- files. ameters of that theory using regression equations. The theory is redefined with Notesfiles are an electronic bulletin each subsequent iteration of users, and board in which users may write state- basically serves as a paradigm for the ments that can be read by other users. evolution of cumulative scientific re- Users reading statements can respond search in this area. This scheme is with statements of their own or initiate similar to the proposal for the formal- new statements. Notesfiles can be open ization of theory developed by Blalock to all users or access can be limited (1969). In this situation, the evolu- to users with specific characteristics. tion and testing of the computer-based Users who complete the program, or users instructional model, the clinical model who are having trouble at various por- examining the efficacy of person-thera- tions within the program related to soc- peutic intervention interactions, and ial support, have the opportunity to the formal theoretical model are synon- read and write in notesfiles of users omous. who are also attempting to stop smoking, lose weight, manage their stress, con- Computer Configuration tinue in a program of physical fitness, or manage their blood pressure. Users The PLATO STAYWELL Program runs on a have the option of creating new notes- Control Data 110 microprocessor. The files that may limit group membership, program uses two floppy disks that are focus on selected topics, or be geo- run on a one-disk drive unit. In each graphically specific. For example, a course area, there is a personal disk localized notesfile can be used to match and a public disk. The personal disk partners for fitness activities. contains information about the indivi- ual's health risk profile, their prog- A variation of the notesfiles process ress in the program, their history_in used in a work site setting involves a the program, and evaluation data. The program called Action Teams. Action public disk contains specific course Teams are work site based groups who get lessons. together to effect some aspect of the work environment related to more healthy A program session is initiated when the behavior. These groups focus on chang- user inserts his or her private disk. ing some aspect of the environment such The private disk greets the person, dis- as putting healthier foods in the vend- cusses progress during the week, and ing machines or company cafeterias, putt- refers the user to a specific lesson on ing in bike racks or showers, or on some the public disk. The private disk is more general health-related activities removed, the public disk inserted, and such as starting an aerobic dance class, the user takes the lesson on the public having contests among work groups to disk. After completing a lesson on the to lose weight or stop smoking, or in- public disk, the individual returns to itiate intramural sports activities. the private disk where a specific home- Each Action Team is led by a lay leader. work assignment is determined that enables the user to apply the information The PLATO STAYWELL program recruits learned in the lesson. Action Team leaders based on their suc-

83 -BUJ The program can run either with or with- Cliffs, NJ: Prentice-Hall, 1976. out the notesfiles support group and Weiss, A. R., Characteristics of success- Action Team functions. Notesfiles and ful weight reducers: A brief review Action Team functions require a modem of predictor variables. Addictive hookup through which the microprocessor Behaviors, 1977, 2, 193-201. can have access to the central computer and to other users.

Programs for weight control, sm-,king cessation, blood pressure manag-mient, and stress management will be in use at Control Data work sites during the first quarter of 1983. Programs for physical fitness and nutrition will be available later in 1983. Bibliography

Blalock, H. M., Theory Construction: From Verbal to Mathematical Formula- tion. Englewood Cliffs, NJ: Pren- tice Hall, 1969. Coates, T. J., Theory, research and practice in treating obesity: Are they really all the same? Addictive Behaviors, 1977,2, 95-103. G'rmally, J., Correlates of weight loss and maintenance in a behavioral weight clinic. Paper presented at the Annual Meeting of the American Psychological Association, New York, 1979. Harris, M. B. and Bruner, C. G., A comparison of a self-control and a contract procedure for weight control. Pehavioral Research and Therapy, 1971, 9, 347-354. Jeffrey, D. B., Behavioral management of obesity: Learning principles and a comprehensive intervention model. In W. E. Craighead, A. E. Kazadin, and M. J. Mahoney (Eds.), Behavioral Modification: Principles and Appli- cations. New York: Houghton Mifflin, 1976. Leon, G. R., Current directions in the treatment of obesity. Psychological Bulletin, 1976, 83(2), 557-575. Naditch, M. P., The STAYWELL Program. In Matarazzo, J. D. Miller, N. E. Weise, S. M. Herd, J. A. Weise,S.M. Behavioral Health: A Handbook of Enhancement and Disease Prevention. NY: John Wiley and Sons (In Press). Penick, S. B., Filion, R., Fox, S. and Stunkard, A. J. Behavioral modification in the treatment of obesity. Psychosomatic Medicine, 1971, 33(1), 49-55. Stunkard, A. J. and Mahoney, M. J., Behavioral treatment of the eating disorders. In H. Leitenberg (Ed.), Handbook of Behavior Modification and Behavior Therapy. Englewood

84 THE NEUROSCIENCE SOFTWARE PROJECT

by Terry M. Mikiten, Ph.D. and Ronald Pyka

Department of Physiology and Graduate School of Biomedical Sciences University of Texas Health Science Center at San Antonio

Abstract There were a number of other principal objectives in the project. They were as follows: The Neuroscience Software Project was established in 1980 to provide a complete system for 1. To provide a user-friendly learning delivering computer-assisted instruction in environment that emphasized the learn- Neuroscience for medical and graduate students. ing experience, not the hardware. An operating system using Apple II microcomputers 2. To create a system that was reliable was devised for the delivery of instruction and from the standpoint of both software for data collection on student utilization of the and hardware. system. This report describes the philosophy of 3. To create a system that could be placed the Project, the component parts of the delivery in a library setting and could be con- system as viewed by the student user and by the trolled easily by library personnel. people who manage it, as well as preliminary 4. To make both hardware and software results gathered after the first year of the sys- available to all individuals who choose tem's operation. to use it. 5. To collect data on the user identity by category, time of use and frequency of use according to individual programs. Introduction

In order to understand the Neuroscience The Hardware Software Project (NSP) goals, it will be helpful to briefly describe the course of study it is -The hardware components of the Project con- designed to assist as well as the institutional sisted of four Apple II microcomputers. Each setting in which it occurs. was coupled to a 12" color televisioi. set and a 5 1/4" disk drive. Each Apple II contained 48K Neuroscience is a 98-hour course given to of RAM memory, a disk interface card and a CCS first-year medical students at The University of clock card. The monitor chip on the motherboard Texas Medical School at San Antonio,The Medical was replaced by a 2716 EPROM containing a mod- School is one of several schools in a large ified monitor. This was done to prevent software Health Science Center. Other institutions include theft. a School of Nursing, School of Dentistry, School of Allied Health Professions and Graduate School This hardware system was mounted in a study of Biomedical Sciences. Because of the subject carel designed for individual student use. In- - matter of Neuroscience, the course is taught by structions were mounted on the walls of the carel people from a variety of disciplines, mostly to remind students of the rules for system opera- clinicians and basic scientists from departments tion. Each carel had a powerstrip to which all in the Medical School and Graduate School. of the equipment was connected. A switch on the Students taking the Neuroscience Course were from carel wall turned on all the equipment at once. the medical and graduate schools. At the same time, power to each carel was controlled by the librarian who activated a numbered The main NSP objective was to provide com- switch at the main desk. This system prevented puter-assisted instruction to the 200 or so unauthorized use of the systems and gave the students taking the Neuroscience Course. All librarian an easy way to check which systems were were linked to what we have called the Primary in use. This method forced clients to come to Assumption of the Project - people using the the desk to request that a machine be turned on computers would have no background in either and so promoted good record-keeping on system the hardware or software that was involved. use. Figure 1 below shows the overall system diagramatically.

85 The Teacher's Viewpoint: From the instructor's viewpoint, the NSP operating system has a variety of features which fulfill the objectives NEUROSCIENCE SOFTWARE PROJECT related to security, data collection and ease of management.

directions Security USSR IISOLNISTS STEP DISKETTE mom CATALOG ON LISTENS Security of the system is maintained at two levels; hardware and software. Both were meant to prevent tampering with either the hardware or programs. Hardware was protected by having pad- locks on the back of the Apple II lids. This prevented the machines from being opened. In addition, each computer was fitted with a "reset re- STEP mover", a small chip inserted in the mother board socket that received the cable from the keyboard.

STIANINTNETINENU The keyboard cable was in turn connected to the DISSETTI-. reset remover. This maneuver prevented students -LISSANIAN RITINDIS DISKIETTII from activating the reset key, thereby interrupt- TO MS ing the program that was being run. AND STEP 3 TANS EN POWER Software security was achieved by having LASARIA11 DISKETTE programs specially coded on the diskettes. The NOD ENUIll, programs were decoded by a machine language pro- TUNIS ON SISCONTIL- gram placed into the system monitor on the mother-

-STUDENT board. In order to run encrypted programs, the UST* ROORANI monitor routines were required to convert them to IN CASS or ERA', e* PROBLEM a usable form. Thus it was not possible to take I.CONSULT diskettes to non-NSP computers and modify or copy 11t110111 UST them. Nor could NSP diskettes operate on a non- NOTNNIS ON UST AMMO. NSP computer. CELL O. MITTS

Data Collection Figure 1 In order to determine which programs were used most by the students and who the users were, Operation of the System special efforts were made to collect this information. Each NSP program diskette contained, in The Student's Viewpoint: To use software addition to the Neuroscience tutorial, an inter- created for the Neuroscience Course, the student rogation program which determined the user's requested a topic package of NSP materials from identity. Through a series of questions the pro- the librarian. The package consisted of both gram determined whether the user was a member of written materials and a diskette. After the the Health Science Center, then asked if the user student surrendered an ID card, the librarian was a member of the faculty, student body or other would switch on one of the carels and give the group. If the user was a student, it then asked student the materials. Once the student was at for the user's School and class level. The pro- the carel, the sequence of events was as follows: gram then entered this data, along with the present clock time (read from the hardware clock The student placed the diskette in the 1. in the computer) into a random access file on the drive and turned on the local power switch. All disk, named the Hostory File. At the end of the programs created for the Project were designed to program, when the user indicated that it was time operate automatically when the system was started to quit, the time was taken once again, ana the in this way. quitting time was added to the History File. 2. The NSP logo was displayed while the program was loaded. The History File on a diskette contained the A series of brief menu-oriented questions 3. complete history of its use by each individual. was run. Of course, to serve many students, it was necessary The main teaching program was run. At 4. to have duplicate diskettes of each NSP program, the completion of the exercise, the student was each with its own History File. asked whether he/she would like to go through the program again. If an affirmative answer was At weekly intervals all of the diskettes were given, the sequence returned to step #3. If a collected and the data from their History Files negative response was given, the system ceased was read by a Consolidator Program which gathered operation and power would have to be turned off. the data from duplicate diskettes into a Master Torun another program the student would have to History File. This file contained the complete start again at step #1.

103 history of use for all NSP programs. Extraction from a local User's group and made these available. of appropriate data from the Master files was Iddividuals from all user categories played games. used to determine information about general sys- This was not surprising. This aspect of the system utilization, such as total time spent with tem's use has increased steadily and by 10-11 the NSP programs by all users. months, it has grown to a volume that has elicited some complaints from 'legitimate' users. The com- A large number of people used the microcom- plaints most often relate to the noise-level of puters for purposes other than the Neuroscience the games. Other complaints referred to the noise Course. Because these individuals could not made by groups of children playing games. It wts readily be interrogated by an NSP program, data interesting to learn that a large number of game- on non-NSP utilization is somewhat less reliable. players were children of the faculty. This was Since these users had to sign in at the main desk probably the least-expected result of the study. and ask to be assigned to a computer, they were The NSP policy on game-playing is being carefully also asked to enter information in a logbook. reconsidered. In addition to name and sign-in time, they were asked to indicate the use to which the computer was to be put. Typical entries were "to play games" or "curiosity". Because the library acquired a significant number of utilities and pro- System Use Trends gramming tools, "programming" was also a common entry. Over the course of this study there-. was a progressive increase in the system's utilization. The data gathered from the non-NSP logbook Figure 2 below shows total system utilization, were handled separately from the data collected expressed as the time spent by all users each by the NSP program software. lc gave useful in- week. formation about additional uses to which the computer resources were put and gave a more complete picture of the user community's interests.

Results 10000-- System Utilization iainiEXLOCROM_ 9000 TOTAL TIME_ °see -- The NSP hardware was installed in the Learn- 7060-- ing Resources section of the Health Science Center 6000-.- Library in the first week of January 1982. Soft- 5000-- ware for the Neuroscience course was made avail- 4000-.- able five weeks later, just prior to the start of the course. Some of the material was related to another ongoing course, Cell Physiology, a prerequisite for the Neuroscience course. iFeliii:allkilli ii 0 1 2 3 4 5 6 7 0 9 11 13 15 17 19 21 23 25 The software remained in place and was avail- WEEK able to all students of the Health Science Center for a period of 25 weeks. The total time logged by all four machines over this time was 93,990 minutes. This includes all uses, i.e., NSP and Figure 2 non-NSP. This was accounted for by 1138 individual user visits. Some of these were repeat visits by the same individual; data were not taken to determine the actual number of individuals who The graph also shows peaks and troughs of used the machines. use. Closer inspection of the data by NSP program users shows that some of these were linked Of the total group, 3376 minutes (3.6% of to Neuroscience-related examinations that occur- the total) were used for actual use of the NSP red in weeks 9, 16, 22 and 26. The examination programs. The remaining, i.e., non-NSP uses in week 9 was the final examination in the Cell have tentatively been classified into 'game' and Physiology Course. Figure 3 gives the data on 'other' categories. Game players accounted for the NSP programs expressed in minutes. The pat- 18,264 minutes, or 19.4% of the total non-NSP tern of use related to examinations is evident use. here. This graph also shows that NSP use fell dramatically after April 11. This is perhapsnot Game-playing was not discouraged provided surprising, since the NSP tutorials only covered that it did not interfere with use of the com- material presented in the first half of the puters for the NSP programs. For the most part, course. individuals brought personally-owned games to thec, computers, although the library did acquire some

87 98

~IN (N -3376) 72

11111E1111LIZaIIIIIIL 64 1060 BI- MEEKLY 36 elm 18 648 1 2j0 8 18 12 18 12 s 'I 2 8 v v (AM) um) 41 a 0i 14 2814 28 11 26 9 23 6 20 FEB MAR APR MAY JUNE Figure 5 Use According to Student Category

Table 11 NSP Utilization by Group_ Figure 3 No. of Time Spent The drop in the utilization around May 23 (week Group Identity Individual (min) 22) was probably related to something other than Visits examinations in the Neuroscience course. The same drop in utilization was seen in the analysis of Medical Students 244 2295 off-campus persons who used the programs. Figure Graduate Students 42 428 4 shows the pattern of usage for the latter groups Dental Students 17 of individuals. Nursing Students 11 43 Faculty 31 131 Miscellaneous 47 303 Off-Campus Visitors 38 159 Totals 414 3376

16 System Failures mmeulluzatunt ALSFF-CAMPUS USERS All problems which rendered a carel inopera- 0I- MEEKLY tive were collectively classed as system failures. When a problem occurred, the librarian was asked to report the malfunction by phone. Sometimes a brief consultation with a member of the NSP team clarified the difficulty. On other occasions it was necessary for one of the NSP members to go to the library. On these occasions,-,the carel was 14 2814 28 11 26 9 23 6 28 inoperative until the repair could be made. Table FEE MAR APR MAY JUNE 2 below lists the problems that were encountered and the total computer downtime they caused. Data given are approximate values taken from written notes made about each repair.

Figure 4 Table 2: NSP Downtime

NSP Utilization by Time of Day Cause Down Time (min)

Figure 5 below shows the pattern of use of TV improperly set by user 30 NSP programs during the course of the day. It is Damaged diskette 30 evident that, on the average, there was a progres- Improper insertion of diskette sive rise in the number of people who used the into drive 60 system as the day proceeded with a peak around Disconnected video link to TV 30 4:00 p.m. It should be emphasized that this pat- Oxidized contacts on disk tern was probably not fixed over the course of the interface card 30 study. More likely, it varied according to the Dislodged Reset Remover 1080 students' academic schedules and the proximity of Total 1260 examinations. The data here show only the cumulative experience.

88 103 The total system downtime was approximately 1.3% of the total use time. Notice that Reset Removers accounted for a high proportion of the problems we had - 86% of the total downtime. This was evidently due to the propensity of these accessories to dislodge themselves from the 16-pin sockets on the motherboard. The usual repair simply involved reseating the Remover. The time lost because of these devices was partially due to slow response time in answering the initial complaint call for the librarian. In several instances the NSP team was not available and several hours elapsed before a repair could be made. Roughly 10 hours were lost for this reason. If this time (600 minutes) is subtracted from the total downtime, we get a better reflection of downtime related to system malfunction. This corrected value for total downtime then becomes 660 minutes, or 0.7% of the total operating time, a figure that compares favorably with mainframe performance. The problem with the Reset Removers was finally solved by discarding the devices and disabling the Reset key by another means; cutting the Reset line on the Keyboard circuitry.

Conclusions

The system devised for providing computer- assisted instruction operated satisfactorily. Equipment reliability was, for the most part, very high, and the software performed well. The data- gathering system operated extremely well and re- vealed a few surprising results. Perhaps the most surprising was the overwhelmingly positive user response to the Project. Second was the finding that the utilization for the NSP programs was dwarfed by the other uses to which people put the computers.

Acknowledgements

Several people played important roles in programming the NSP tutorials. They were Shawn Mikiten, Erick Mikiten, and Robert Woodward. John Finley was responsible for the data collection snd reduction. Jacqueline Mikiten did the fine artwork and computer graphics that were used throughout. Without the support and help of all of these individuals, ths s'udy could not have been done.

We are grateful to Barbara Greene and her staff in the Health Science Center Library's Learning Resource Center for the use of their facility to house the NSP effort. Their kind cooperation and enthusiasm in managing the day- to-day operation of the checkout system was vital to the positive student responses we have received.

89 COMPUTING IN A NON-CURRICULAR SUPPORT ROLE

J. Spicer Bell Alonzo D. Peters Linda L. Royster Robert W. Jackson Richard Cornelius Dr. Randall K. Spoeri

ABSTRACT: A Microcomputer Based Vocational ABSRTACT: Individualized Grade Reports: Placement and Follow-up System Motivational Aid and Teaching Tool

J. Spicer Bell, Director, Alonzo D. Peters, Linda L. Royster, Division of Counselor Education, Coordinator, Frederick County Board of Education, University of Iowa, 338N Lindquist Center, Iowa 115 East Church Street, Frederick, MD 21701 City, IA 52240

The Vocational Placemert and Follow-up service A description of the use and effect of an was initiated by the Frederick County Board of instructor-made computer grade report, using Education in the fall of 1981 with a grant from the SCRIPT, a text editing package from the University Maryland State Department of Education under Public of Waterloo. The reports were made for junior and Law 94-482. The program serves seven comprehensive senior undergraduate students in a theories of high schools and one vocational technical center counseling course. The primary pupose of tho and is designed to serve a dual function of reports was to provide students individualized facilitating initial job placement of graduates and statements of their learning accomplishments, and analyzing the success of those graduates on the the instructor a structured system of course job. Approximately nine hundred graduates from management. The reports were also used to motivate twenty two different vocational programs register students, provide diagnostic feedback, demonstrate each year and are eligible for employment placement application of instructional and counseling theory, in local businesses. The program provides a and validate course performance standards. systematic design to ensure students equal access Rogerian, behavioral, and expectency theory and to potential job openings and employers a central Bloom's Taxonomy in the Cognitive Domain were used source of potential employees. to develop student feedback. Specific student The project was developed around the the use reaction to the reports and class outcomes will be of microcomputer technology to facilitate the discussed. copies of reports and the program will handling of large numbers of records and to provide be provided. the capacity for quick responses to employer inquiries. Commercially available hardware and software systems were selected because of ABSTRACT: Using a Microcomputer for a Test affordability and the availability of back-up and Question Storage Bank service. Data collection irr,truments were designed for use with graduates as they leave the vocational Robert W. Jackson, 2 Andrews Road, Greenwich, CT training system and with employers. Student 06830 questionnaires provide identifying information as well as information used to screen their I have repeatedly been asked if a eligibility for certain jobs. Employer microcomputer could assist the classroom teacher in questionnaires identify potential employment needs the preparation of test questions. Inevitably I and job classifications represented in the would reply affirmatively and ask to see the employer's work force. material involved. Software systems provide the capacity to Most teachers described a program that would rapidly sort graduate files for students with either randomly select questions from a large data employer identified qualifications.Many times, base or present the questions for the teacher to same day responses to employer inquiries are select and then print out the test questions neatly possible. Current informantion is kept on file on one page and the correct answers on another through the use of periodic mail and phone page. Analyzing the problem I discovered that follow-up questionnaires. Employer data is kept on sequential access was limited by memory constraints a separate data bank and is used by the placement of the micros and random access used up too much coordinator to periodically canvass the community disk space as it required all files to be as big as for available jobs. In addition to immediate the largest question. Normally I would recommend information access the system is also designed to using a commercially available data base manager provide hard copy back-up in the form of client and except that almost all of them limit the material employer lists, student referral forms, mailing to 255 characters. labels and labor market statistics.

90 I finally came upon a perfect solution t3 this ABSTRACT: An Analysis of Academic Grades at the US problem. Most schools have word processor software Naval Academy, 1971-1981 and almost all word processors allow the material to be saved in ASCII format. If the teacher would Dr. Randall K. Spoeri, Major Malcolm W. Fordham, use the word processor to enter, edit, and format United Sates Naval Academy, Annapolis, MD 21402 the questions and answers and then save the large text file in ASCII on the disk, it would be In recent years, an area of interest in possible to access that text file from BASIC just academic circles has been the phenomenon of grade like a sequential file. inflation, or "grade creep". For our purposes, This approach would combine the use of maximum grade creep refers to the steady increase, over disk space and overcome the 255 limit of random time, in grades awarded in academic couses. That access. Entering and editing the file would be is, the inflation of grades over time. This is an done from the word processor and the selection and area of interest to the Office of the Academic Dean printing of the questions would be done from a of the US Naval Academy as well. short BASIC program. I selected the TRS-80 Model In consultation with the Academic Dean's III because of its file handling abilities and office, it was decided to study changes over time because the school system that ordered this of academic grades summarized by: software used that machine, but the concept can be 1) all courses translated to almost any microcomputer. 2) selected courses 3) selected majors 4) all academic departments ABSTRACT: How Easy to USe Can a Grade Management Computer data files were established from Program Be? administrative records stored on computer tapes. The data base was organized to contain all grades Richard Cornelius, Wichita State University, awarded to all students for each semester for the Wichita, KS 67208 academic years 1971 to 1981. It was decided that the standard quality point One way to introduce teachers to ratio (QPR) for academic grades would be suitable computers is to give them a program that saves them for comparing grades for the four types of summary time and effort. A grade management program is an information. FORTRAN 77 programs were prepared to obvious choice for this purpose. If a particular access the data files to develop for each area of program is to be the teacher's first exposure to interest: computers, then every precaution should be taken to 1) the course credit hours be certain that the program is both useful and 2) a count of A's, B's, C's, D's, and F's by extremely easy to use. GRADISK is a program that credit hours has been written with the overriding goal of making 3) the percent of each letter grade it the easiest to use grade management program 4) total quality points available. This presentation will focus on 5) total credit hours demonstrating how easy to use the GRADISK program 6) semester QPR is. Features that make it easy to use include: 7) semester standard deviation a) Documentation is complete and includes a The semester QPR's for each semester were sample run. placed in data files so that plots of the QPR's b) Instructions (to the detail of "press over time could be made. Simple least squares RETURN" where applicable) always appear linear regressions were applied to each file to on the screen. provide an indicator of the QPR trend over time c) All features are menuselected. The study provided useful information to US d) Previously created files are Naval Academy administrators and department menuselected. chairmen, as well as showing that there has not e) Student records can be examined or edited, been any appreciable grade creep during the period by identifying a student with a few studied. letters of the last name or the first few digits in a student number. f) Users are warned before actions that would erase information. g) Error messages are informative. h) The program remembers options that you select and streamlines itself for the same selection the next time through. 1) Weighting schemes can be changed at any point during the grading period. j) The choice of letter grades (e.g., A+, A, A, ... or pass, fail, or ..) is up to the instructor.

91 EXPERIMENTING WITH A COMPUTER LITERACY PROGRAM FOR ELEMENTARY SCHOOL GIFTED AND TALENTED STUDENTS

by W. Starnes and J. Muntner

Montgomery County Public Schools Rockville, Maryland

Abstract Behavioral Rating Scale in Chart A lists the learning characteristics of gifted students which are A computer literacy program for fifth and found in the literature. sixth grade gifted and talented students is described. Computer literacy is a unique vehicle The task of the educator of the gifted and for enhancing the learning experiences of gifted talented student is to provide experiences and students by providing varied opportunities for environments which respond to the characteristics differentiation. Because these activities are o: the students by stimulating critical thinking, developed with the characteristics of gifted stu- fostering the use of a scientific approach to prob- dents in mind, an emphasis is placed on the use lem solving, promoting self- direction and In- of the computer as a tool. Seventy elementary dependent work study skills, allowing for use of schools participated in the project which centered creative abilities, and providing opportunities for around three elements: establishing an appropriate the development of self-evaluation. Activities learning environment, building creative thinking that stress higher levels of abstraction and con- skills, and planning differentiated computer cept development are more appropriate for gifted learning experiences. Ideas and approaches for students than those that emphasize rote learning. use by gifted elementary students are suggested. The use of the computer as a tool and the Introduction development of programming ability are activities that respond to both the identified characteristics When instituting an innovation in public of the gifted and talented and the prescribed fea- schools, it is wise to choose a small population tures of differentiation. In using the computer as on which to try the new idea. Five years ago the a tool students are able to control their own edu- Department of Instruction and Program Development cational environment and build a skill that will be in Montgomery County, Maryland decided to purchase useful in all disciplines. a few microcomputers in order to explore how they might be used with gifted and talented fifth and Implementing the Program sixth grade elementary school students.The intent was not to restrict computer learning to this Seventy elementary schools participated in the population ultimately but to use this group for curricular experimentation with a single microcom- curriculum experimentation. The school district puter available to the school for a semester each wanted to determine if elementary pupils could year. Montgomery County uses the PET microcomputer learn beginning programming skills and if they because these machines are inexpensive, easily port- would be interested and motivated to do so. In able, and conveniently packaged, having one wain situations of limited computer access, Seymour body component. They are hardy with an excellent Papert advocates looking "for a small pocket of repair rate record which is especially important students--perhaps a class of learning disabled considering the number of students using the com- youths--and then give them the computers...to show puter in a variety of school environments. the strength of the computer, which is its ability When the computers were first purchased, they to help children think." In this case, the choice were placed in a few schools with some self-in- of gifted students for the program was intentional. structional materials for student use.Without careful staff orientation, the program often was Why Start With Gifted Students? not implemented. Computer literacy is a unique vehicle for en- After that aborted start, the program began hancing the learning experiences of gifted students with a half-day training workshop for two teachers because it provides many opportunities for differ- and the principal from each school. These work- Classic and current research recog- entiation. shops focused on implementing the program, teaching nizes that the characteristics of the gifted and about computers in society, and hands-on instruc- talented should determine the nature of their cur- tional time with the PET. Teachers and adminis- riculum; in fact, differentiated curriculum must trators worked together to learn their way around be an outgrowth of those characteristics or learn- the keyboards of the microcomputers. They also ing styles (Kaplan, 1979). The Renzulli-Hartman

92 1a[ learned some elementary programming and graphics Learning Environment in BASIC. Teachers were encouraged to plan some literacy activities and some initial introduction Papert envisions the school in the year 2000 to programming for class-sized groups to be fol- as "a research lab with students engaged in projects lowed by independent activities for students to 2 and adults function as consultants and counselors." pursue. This is a description of the learning environment The goal of the program was to increase the appropriate for teaching computer literacy to the This optimum environ- confidence of students in the use and control of gifted and talented student. ment will include the following: computers. The computer literacy activities; were designed as a semi-independent irstructional unit - small groups working together to solve with objectives in the following categories: understanding computers, working with computers, problems of mutual interest - open-ended tasks computers and the society, hardware/software, and attitudes and values. - development of independent ideas - student involvement and choice in The chief instructional text used by both stu- the selection of both resources and projects dents and teachers was the BASIC Manual, written by Montgomery County Public Schools teachers as an in- - the fostering of discovery learning experiences teractive approach to teaching beginning BASIC pro- - students and teachers working together gramming. This text uses flowcharts, sample programs and end of chapter exercises to present the in interchangeable roles content in a constantly reinforcing mode. Supple- - self-pacing of activities and projects menting the text were commercially produced games, by students to allow for maximum time filmstrips, manuals, magazines, and tapes. for exposure, exploration, and acquiring ownership of concepts After two years in the pilot situation some - development and use of self-evaluative elementary schools purchased additional micro- techniques by students computers and began trying parts of the program The burgeoning appearance of computers in the with all students. The school district received funding from the Human Resources Research Organi- elementary school classroom has brought with it zation to develop a more generalized K-8 computer some feelings of inadequacy and apprehension for Teachers have had little experience curriculum. Elementary teachers who had been ex- the teacher. posed to the use of the microcomputer with gifted in either using or teaching about computers.Train- ing courses have been hard pressed to meet the needs students began clamoring for additional in-service of educators. In addition, many students have ex- training. hibited remarkable computer skills in very short The official pilot of the project ended last periods of time. This scenario will only become more prevalent as more and more families purchase summer. The overall effect of the program was home computers. As teachers face the prospect of two-fold: to learn that bright fifth and sixth grade students are motivated to and capable of students with greater knowledge and skills, they mastering programming skills and to move computer will be required to rethink their role as teachers. literacy into the elementary schools for all stu- Nowhere is this more true than in working with the gifted and talented students. Teachers have had to dents. A commitment to providing differentiated computer experiences still exists as Montgomery adapt to multi-role concepts as they worked with Molly Watt, writing of County continues to refine the elements of the their young programmers. Elementary Computer Literacy Program for Gifted her experiences in teaching LOGO, mentions some of the following as appropriate teaching styles: "dem- and Talented. onstrator, teacher-lecturer, teller, time structure; What Are The Key Elements? problem setter, management solver, arbitrator, decision maker, challenger, helper) collaborator, In Figure 1 is a schematic representation of process sharer, question asker, idea extender, ob- the elements which were determined by the pilot to server, documenter, admirer, enjoyer, time provider, be important for implementing a computer literacy technician and model learner." program for gifted students. The effective teacher needs to deal with both Figure 1 the cognitive and the affective domains. Tradi- Elements of the Elementary Computer Literacy tional teaching roles are not totally abandoned. Program for Gifted and Talented Students still want and need instruction, but they The Learning Differentiated Computer also have a need for unfettered exploration. Environment Learning Experiences teacher functions as a facilitator, helping students identify and utilize a wide variety of re- The students' range of materials needs to Critical sources. Thinking be broadened as they learn to seek out ideas and solutions in magazines and books and from mentors. Skills An important ingredient in establishing this learning environment is the teacher's ability to admit that he or she does not know all of the answers,

93 but that she is willing to guide, encourage and ap- enrichment involves a student remaining at a parti- preciate the student's efforts. cular skill level but using these skills in a variety of learning activities. These two methods are While such a learning environment can be estab- not discrete, but should be interchanged in dif- lished with the availability of only one microcom- ferent proportions as appropriate for individual puter, it is much easier to achieve when several students. computers are available. PTAs and schools are purchasing additional computers. During the course of What kinds of differentiated computer learning this project, summer school courses in computer use experiences are appropriate for gifted and talented were offered using a laboratory setting. The tea- students? The answer to this question must be un- chers who taught these specialized sessions found derstood as a tentative answer for the early 1980s that the need for role flexibility was even greater when we are just beginning to explore the methods in these situations. In one case, middle school for introducing and working with computers in the teachers were concerned because they had to estab- elementary school. Several years down the road, lish the summer laboratory with microcomputers from so much more will have been learned that present several vendors. They were startled to find that designs will seem inadequate. Present teaching ap- the students leaving the sixth grade had no dif- proaches may be viewed as obsolete in the future ficulty moving from one type of computer to another. as the present computers will be.With these dis- claimers in mind the following is a list of ideas Critical Thinking Skills and approaches useful for gifted and talented upper elementary students. The second element in the design of the program is an emphasis on the critical thinking skills 1. GRAPHICS PROGRAMS. Students can explore needed in computer programming.As soon as stu- a variety of picture and design programs. dents begin to practice beginning BASIC, they are Using INPUT a program can generate an asked to predict what an operation will do and to oriental type rug design with the students hypothesize about the RUN if the program is changed. name printed in reverse in the middle. Students are challenged to continue to hypothesize Banners and greeting cards are also rela- as they become more sophisticated programmers. tively simple, but satisfying to create. Problem solving is another thinking skill which is A student created the Happy Mother's Day Sports emphasized. The logical and inalterable step-by- program reprinted in Figure 2. step nature of learning to program the computer re- team banners are particularly successful fines the student's problem solving ability.The each season. Also a banner announcing process of debugging presents a real problem, a the classroom teacher or class nickname program that isn't running as intended. It re- can not only decorate a classroom but quires that the student find what the problem is assist in class identity.Pictures of and figure out how to solve it. Very often the hamburgers, people, robots, ice cream solution involves risk-taking, trial and error, and sundaes, Snoopy, rockets, Martians and learning from one's mistakes. symbols are, just a few of the ever popular ideas that students will use to de- The use of creative thinking is an important sign their own unique graphic programs. skill for gifted students. Much emphasis is placed Graphics can, of course, be included as on students creating their own programs. Of course, an exciting part of all of the other sug- these programs while new to the students are not gested programs, to deliver messages, to unique. The student may be building on some skill provide a title, or to illustrate a parti- he has just learned or refining some previously at- cular situation. tempted idea. 2. QUIZZES. These are popular and can be The critical thinking skills will not be em- used by other students to enhance a phasized if the student is using computer assisted specific study. Elliott, a 6th grader, instructional software which emphasizes drill and designed an Egyptian vocabulary quiz entitled Pyramid Power for use in his practice. Games and canned programs can be used occasionally as motivational tools or instructional classroom. The program ended by gen- devices to help students write their own programs. erating a picture of a pyramid with the user's score on the screen. Students Differentiated Computer Learning Experiences can also be encouraged to contract with another teacher in the building who would Because of their special learning character- like to have a quiz or drill program writ- istics gifted students can learn faster, acquire ten for his or her students. In this case, in-depth knowledge of an academic area, retain the the student has a real simulation of being knowledge, and be expected to use it to create. a programmer and having to design a pro- What relevance does this have for teaching students gram to meet the teacher's specifications. programming? There are two major ways to ensure Responding to such a request, Stuart de- that this student has differentiated computer signed a program called Planet Weight. learning experiences: acceleration and horizontal The user typed in his or her name and enrichment. Acceleration involves students learn- weight. Next a chart was generated on ing more rapidly and progressing beyond what is ex- the screen listing the planets of our pected at a particular grade level. Horizontal solar system. The user selected one and

94 was then told his weight on both earth screen (ex., you struck out, stay here one and the chosen planet. Although Stuart's turn). Unfortunately, the school year ran fifth grade classmates enjoyed this pro- out before Scott had a change to figure out gram, it was an even bigger success in the how to generate the game board on the com- second grade classroom where the students puter screen. A thoughtful word is appro- had just finished a study of the universe. priate here. Commercial games and LISTs The second graders were provided with a of game programs have a definite place and beginning computer experience and meaning- value at this level as students strive to ful learning as well. Stuart received 27 create their own and are looking for models thank you letters showing pictures of him of various ways to accomplish specific and his computer. tasks. The ultimate accolade in this category goes to David, who at age 11 designed 3. MADLIBS. The ever popular game provides a Pac-man type game for the PET computer. the source of another programming idea. (Figure 5) His knowledge far exceeded his Students write their own original Madlib teacher's, but she was able to facilitate stories and then design the program to ask his work by providing him resources, en- for the appropriate words and generate a couragement and access to a machine. story at the end. Topics are endless from "The Disease" (Figure 3) to "Last 8. ELEGANCE COMPETITION. This idea can pro- Sunday at the Redskin Game." vide a great deal of stimulation and excitement within not only a school, but the 4. ADVENTURES. This category mixes together community at large. Students are assigned adventure, simulation and variations of a specific program writing task and com- create-your-own-story. These programs pete against each other to create the most can be written on a very simple level "elegant" program to solve the problem. using IF-THEN (see Figure 4) or on a Community members can be utilized as more difficult and exciting one by using judges. This type of activity requires RANDOM. Students using the latter BASIC fairly sophisticated programming ability. statement will need to graphically re-. present the program to keep track of its 9. PUBLICATIONS. Create a book, newsletter, convolutions, and analyze the steps being or a publicity flyer. Use a printer to written. Settings can vary greatly: a produce a book of student programs. Sam wagon train in 1850, a pyramid, or on and Roy borrowed the administration printer board a space ship bound for outer space. to generate the PTA flyers advertising the (Figure 4) computer demonstration and discussion. Their flyer included a graphic of a com- 5. ANIMATION. Students moving ahead in their puter and the relevant information. Stu- programming knowledge enjoy exploring PEEK dents have also produced newsletters with and POKE to generate animated pictures on a section called DEBUG ME which features the screen. A birthday cake complete with programs such as computer riddles, puzzles blinking candles and a happy birthday mes- and cartoons which contain bugs. sage greeted one teacher as she entered her classroom on the auspicious day. Mov- The Elementary Computer Literacy Program for ing robots, animals that assemble and dis- Gifted and Talented has been an important component assemble before your eyes, and barren of the overall gifted and talented program in Mont- moonscapes that suddenly grow into future gomery County. Feedback from administrators, tea- cities are other challenging ideas for chers, parents, and pupils reflect the overwhelming young programmers. success of the pilot. Eighty-nine percent of the teachers indicated that they would like to continue 6. MUSIC. This is another area of explora- refining the program and further that they believed tion for programmers who are no longer the activities were particularly appropriate for novices. Music can be generated for its gifted and talented students. Eighty-eight percent own sake by budding composers. However, of the students indicated that they would like to sound and MUSIC can also be incorporated continue in the program in order to further their into other programs as mood-setting de- knowledge of BASIC, write more original programs, vices or as a means of delivering a mes- and design computer graphics. The staff and stu- sage. dents in this pilot would agree with Papert that computers are "carriers of powerful ideas and of 7. GAMES. This is a broad-based catch-all the seeds of cultural change...that can help people category for homeless program ideas. Word form new relationships with knowledge that cuts searches based on a theme with a suitable across the traditional lines separating humanities graphic are exciting to design and debug. from sciences and knowledge of self from both of Scott designed a game called Lucky Sport these." which incorporated a paper game board. The computer simulated the toss of 2 dice and then asked which of 3 sports symbols the token had lanced on. A random event, relevant to the sport was flashed on the

95 112 References Chart A

Scales for Rating the Behavioral Characteristics "Computers Are Objects to Think With", an Instruc- tor Interview with Seymour Papert. Instruction, of Superior Students* 3:82, p.85-89.

2 Part I: Learning Characteristics Ibid., p.87.

1. Has unusually advanced vocabulary for age or Watt, Molly. "What is Logo?", Creative Computing. grade level October, 1982, p.112-129. 2. Possesses a large storehouse of information about a variety of topics Papert, Seymour. Mindstorms, Children, Computers and Powerful Ideas.New York: Basic Books, Inc., 3. Has quick mastery and recall of factual 1980, p.25. information

Kaplan, Sandra. In-service Training Manual: 4. Has rapid insight into cause-effect relation- Activities for Developing Curriculum for the ships Gifted/Talented. Ventura, California: National State Leadership Training Institute on the Gifted 5. Has a ready grasp of underlying principles and and the Talented, 1979. can quickly make valid generalizations

6. Is a keen and alert observer

7. Reads a great deal on his own

8. Tries to understand complicated material by separating it into its respective parts

* Renzulli, J.S., Smith, L.H., White, A.J., Callahan, C.M., Hartman, R.K., Creative Learning Press, Inc., 1976. 60 PRINT40$0000 LUCKi04"

2 OPEN5,410105,LIST 10 PR11ITTP8(5)161 I pimaOM MO Ir 20 PRINTTA8(5)P1 01II 1 1 1 1 a" 30, PRIWY08(5)N011tto 00881UOISI MIS 0001" 40 PRINT AB(5)111 11 1 p 0 1 r 2 OPEN 4,4:CM04:LIST 50 PRINTTR8(5)1 i0 10NI 1 8' 60 PRINTTR8(5)1"1 10NI Ir 70 PRINTI" 70 PRINTIPRINTtPRINTIPRINT 80PRINT:PRINT:PRINT:PRINT:PRINT 80 PRIrT4(12)1"111 11" PRINTTAB(6),"**410************************" 90 CRAZE *" 90 PRINIP 1111,0 81 8" TNE,DISEASE.A MOWS 100 PRINTTA8(6);"* *" 00 PRINTIA8(11) 1 LE 8" TA8(6)ei 110 PRINT *" 10 PRINTIR8(12) "1 OV I" BY JEFFREY CORDON 120 PRINT TA8(6))"* 28 P010r88(0) "1I" AND 125 PRINT TA8(6))"* *" 30 PAINTM(14) 11 8881" STEVEN SMITH 130 PRINT TA8(6))"* 40 kiNTTA8(26)11 NO 1" *" 150 PRINT TA8(6))"* 50 PRINTM(26) "8 LE 1" TA8(6)e**************000*************" 60 PRINTTR8(27) 1 OV 0" 160 PRINT 1000 70 PRINTIRB(28)1111" 11' ''OR T1 TO 80 PRINTIO8(29) "81" )1 T 90 PAINT 170 HINT"; 200 PRINTTR8(35);"111 180 INPUT"OIVC ANOUN" ;A1 210 PRINTTR8(36) 1" DISEASE ";8$ 190 INPUT "GIVE A 220 PRINTT08(31) "Ir 200 INPUNIVEA HOUNPIC$ 230 PRINTTR8(0)140 OR MOOMINIM Ma 11 0100" 210 INPUT "DIVEA NOW/ 240 PRINTTRO(8)1 "14 IN I 1 I 1 1 I N 220 INPUNIVE AFEELIN0410$ 250 PRINTTR8(0),"1 $1 I I I10110181 MN Pr 0 260 PRINTTR8(0)1 "1 81 I I IIIIII I' 230 PRINT1" 6 THERE WAS A "pA$ 270 PRINTTR8(0) "8 1 1 1 iI i . Ie r 240 PRINT"ONCE A BAD "18$ 280 PRINTTR11(0)1 "1 1 14411 Mil II WV Cn 250 PRINT"IT HAD 4' 290 PRINT FROM MILES AROUND TOHELP IT" U. 260 PRINTC$)" CAME 300 PRINTTA8(13)1101 10" A ",E$;" CAME TOHELP IT" 270 PRINT"ONE DAY 310 PR:WM(14),i 1" 280 PRINT"IT LIVED"0 320 PRINTTR8(14)1 "1 1" 330 PRINTTR8(12)1 "11 NI 11" 285 PRINT:PRINT:PRINT:PRINT WANT TO PLAY AGAIN-Y OR 117" 340 PRINTTR8(11),"1pgrI 1" 290 PRINT "DO YOU 350 PRINTM(11);"1 14 8" 295 INPUT MS 360 PRINTTRO(11),"I 11 1 300 IF MI NY" DOTO180 370 PRINTM(11)1 "I I0 I 310 END 380 PRINTIR0(12)1 "1 NS I READY. 390 PRINTTR8(13)1 "1 II I

488 PRINTTAB(12),"1 110 I 410 PRINTM(11) "11 00 I 420 PRINTTAB(11);"1 p1 1" 430 PRINTTR8(11)4 IVO A" 440 PRINTIA8(12);11 450 PRINT:PRINTORINTIPRINT

460 PR1N11A8(11),"1141 11880 I I 1" 470 PRINTTAB(11) "1 It 11 1 1" 480 PRINTTA8(11) 1 18400 11101 1' 490 PRINTTA8(11),"100 8 1 1" 500 PRINTTA8(11),"11411" 510 PRINTTR8(11) "1182 I 8 1 1" 520 PRINTIPRINT,PRINT;PRINTIPRINT 530 PRINTTR8(16),"LOVE,MIKA" 540 PRINT:PRINT,PRINT1PRINT 550 PRINTIAB(15)1 "MRV 9,1982" REALM 1 OPEN4,41CM041LIST S PRINT"3"t0IM V(2) 10 READC,R,P,X,04,0,Y,TINFJA 20 DATA20,14,33348,33348,20,12,33268,33268,87,1,3,32 30 DEFFNZ(Z2).4004.32768 40 OEFFNW(WW).40*S+0.32768 50 FORA43140T0331571POKEA,160WOKEA4280,160iFORNIT01501208,0 I C''N4,4t0104iLIST 10 FORA'38140T033420STIP40iPOKEA,160iPOKEA+17,160!FORBATOMNEXTB,11 2 PRINT"7 70 FORA433264T033272!READCOPOKEA,C0FOR8.1,015iNEXTB,A 3 PRINT'YOU ARE INFROIT OF A PYAMID." 80 FORAN33302T033315MAOCCIPOKEA,CCARUITOISINEXTB,A 4 PRINT'YOU MUST REIREIVE A PHAROAHS STAFF,' 90 DATA13,21,14,341,32,13,1,14,225,32,40,229,4,32,7,10,15,19.18 5 PRINT"WITHOUT FALLINO INTO R TRAP OR BEING CAUGHT BY A CREATURE," 100 FORAIIIY01000iNEXTA 60 PRINT".410000 LUCKIOss' 101 005084000 61 PRINT"YOU ARE IN R DARK HALL IN THE PYRAMID." 110 FORARITO2tPRINT"TIMINEIDIGINMIMI(TIS)+40 62 PRINT' YOU ALMOST FALL IN A DEEP DARK PITT" 120 PRINT"PRESS X TO BEGIN, PLAYER"A 70 PRINT"THEN YOU ENCOUNTER R JACKAL - HEADED" 130 OETAWFASOWTHEN130 80 PRINT'000." 131 TIS "000000* 85 PRINT"SHOOT OR EXIT (S/E)" 132 XISa"5"0(2111"S"0.0iNVAL(TIS)+40 90 INPUT AS 135 AS." 100 IF RS 'S' Th:PRINT'YOU KILLED HIM 1"100TO 120 137 P333480( 33348i0m33268IY.332681C201141040411121Us0 110 IF ASs"E" THEN PRINT "YOU WEREALMOSTKILLED!" 14000512200 111 PRINT'YOU LEAVE SAFELY "100/0 460 15000T0460 120 PRINT 'YOU APE WALKINO DOWN THE ',ILL YOU SEE DEEP PIT' 200 PRIHTN7"IXIse5N ix2ses" 130 PRINT' SHOULD YOU SWING ACROSS Ok JUMP ACROSSNSW/J)":11NPUT Bs 201 nE32826,81 140 IF esn"sw" THEN PRINT"YOU BARLEY MADE IT11100TO 160 205 PRINT" SCORE! 0 MUNCH MAN MEN, 3"t 150 IF Bli"J"THEN PRINT "1 °AALLIIN4001"100TO 160 220 PRINT 160 PRINT'LUCKILY,YOU ORAB AN OLD STONE IN THE HOLEAND CLIMB UP!' 230 PRINT" 050=010010410KOMMERRIMR11=1, 170 PRINT"YOU ARE WALKING DOWN A HALL AND SUDDENLY DARTS FLY AT YOU" 1.1 0211r1 et 240 PRINT" 11014 10 180 PRINT" SHOULD YOU Or : TO THEFLOOROR RUN FOR YOUR LIFE(D/R)?"II INPUT CS 250 PRINT" 111112.1.0 ...... 1.1 030"; 00.0 11R'1 li 190 IF CS "R" THEN PRINT"DART DOT YOUR RIGHT ARMI"IGOTO 210 (I) 260 PRINT" a014 I. No. I. a. 04 0. 1.1. MI. IL 200 IF CS."0" THEN PRINT"ALL THE DARTS MISSED YOUIPIOOTO 220 S. a" 2 270 PRINT" 0...4.0.10.0.0.1.0.0.1010.0 3 210 PRINT'YOU PULLED THE DART OUT OF YOUR ARM ANO SUCKED THE POISION OUT" 280 PRINT" 00 0101 0 a"; 21 CR 220 PRINT"YOU SEE A DIM LIONT ABOUT 200 FEET AHEAD OF YOU." 290 PRINT" 00....80.0.1P012.0110.14001.0.081,..,010") U. 230 PRINT'YOU RUN TOWARD THE LIOHTI" 300 PRINT"LOU 1 0111 1 000.11"; U. 240 PRINT'THEN YOU ARE IN THE ROOM ONO IS 310 PRINT" 0444.011.111411.1 1.0.1.D141.1.1"t N...1.001.0 ") 250 PRINT'FULL OF TREASURES OF THE PHARDAN'S." 320 PRINT" 1.0101.0...0 III 260 PRINT"TNE DOOR BEHIND YOU CLOSES!" 330 PRINT" 1 1 LIMTo 00 00a 0.0.3 2"t 270 FRINT'00 YOU WANT TO STAY OR LEAVE (S/L)' li INPUT OS 340 PRINT" 100811..,11.11.101( lad IM.1.1...0=11", 290 IF 0$ "S" THEN PRINT"YOU ARE BRAVEI"WOTO 310 350 PRINT" ...... 1...0.4m0"; 300 IF DSIOL" THEN PRINT"YOUMOEIT OUT,CNICKEN1'1643TO 460 360 PRINT' 0....0000.1.600.001.0m.1401.1.15011") 310 PRINT"VI! LOOK FOR THE STAFF OF THE PHNOM" 370 PRINT" 0.1010 =IX; 320 PRINT"YrJ ARE READING THE HIEROOLYPHS TO FIND THE STAFF AND A WAY TO OET OU 380 PRINT' I 1111.1.182001.1011410.041 8"3 330 PRINT"qU FOUPO THE STAFF AND A WRY TO GET OUT.' 390 PRINT" 1001.04 8.a. $0.0.0000", 340 PRINT'YOU CAN GET OUT BY(PUTTINO THE STAFF IN FRONT OF THE MUTINY CASE." 400 PRINT" ..1.11.1.1.04 RIAM,043141-' PRINT"TEN CREATURES COME OUT OF THE OUT OF TRAP DOORS IN THE WALL!' 410 PRINT" 00 1 0.0 0 Itl 01";

370 PRINT' SHOULD YOU FIOHT OR OET OUT(F/0)",IINPUT Et 420 PRINT' 00/0...0010.001,..11.1...011.01104....1111") 380 IF Ell'," THEN PRINT"THE000$ARE AOAINST YOU,BUT YOU KILL TXMOOTO 400 430 PRINT' WISSI 0 $ 0411.40") 390 IF ESs"0" THEN PRINT"THEY DOT YOU,YOU ARE DEADI"OOTO 460 440 PRINT"tratiarawitiorsagamealumar, 400 PRINT"A TRAP DOOR OPENS AND YOUWALKOUT INTO THE SUNLIGHT WITH THE STAFF' 458 RETURN 460 TISe000080" 410 PRINT" YOU BECOME FAMOUS"100TO 460 465 POKEP,01:POKEO,T 450 PRINT"THEN THE LIGHT ZAPS YOU IN THE FRONT OF THE PYRAMID WITH THE STAFF!' 470 005138500 460 ENO 480 GOTOEOO READY. ZOO POP.E3426,10:POKE32027,51POKE32328.1 510 P0KE32229,41POKE32330,25 520 FORA617010001NEKTAAIFORAA32326TLJ2830tPOKEAA.32 530 NEXTAAIRETURN 600 8 +111F8939999999THEN8.010010600 610 IFB/2INT(8 /2)THEN1500 620 °ETAS 630 IFRINVORRIeFrOP,Ata"S"ORFISNVORRWM"THFAX26114tlitialt cm+ rv.rawat INTRODUCTORY COMPUTER PROGRAMMING FOR ALL COLLEGE BOUND HIGH SCHOOL STUDENTS

by Ken Jones and Dennis Simms, S.J.

Regis Jesuit High School Denver, Colorado

Abstract that point, were able to satisfy one student at a time. But it did allow us to get rid of our tele- In this paper we will show an affordable type and the $50.00 a month time-sharing bill. method of developing programming skills for all college bound high school students that does not It was suggested to us by Dr. Ruth Hoffman and diminish any of the other skills required of these her associates at the Denver University Math Lab. students. that if we could interface a high speed optical card reader with the micro processor we might have an Many schools cannot afford to use classroom inexpensive solution to our problem. The theory space exclusively for computers. We found a solu- was to set the card reader, the computer and a rel- tion to this problem. atively high speed printer on a cart that you could roll into any classroom. Students would then mark In April of 1982, Regis Jesuit High School of cards and run their programs with almost zero turn- Denver, Colorado received a Title 4-C Mini Grant around time. for a project that would develop introductory computer programming as apart of our regular geometry Students would be able to mark cards at home course. Geometry is required of all students at or anywhere else and we would not have to dedicate Regis and is taken by most students elsewhere who a classroom to machines. The cart could be rolled attend college. This project, therefore deals with into a closet when a computer class was not in the problems of providing computer programming for session. all college bound students. It took almost a year to get everything to- THE PROBLEM gether, but in the end the system worked perfectly. The computer took no extra floor space, did not 1) Regis is a small private school with lim- require a special room and a class of 25 or more ited resourses where 98% of the graduates eventual- students could get one or two programs running each ly go to college. day. We did indeed have the equipment to reach all students. In addition it was inexpensive, about 2) Many students and teachers at Regis feel $3000 for the actual hardware. Cards cost us about that an entire semester of programming is a low $40.00 for 10,000 and it costs about $160.00 to do priority compared with the other academic demands. a semester class, which is well within our budget.

3) The Math Department Chairperson feels that Further, there are educational advantages that all of our graduates should have hands-on experience exist that should be mentioned. ;TM computers and suffiPient programming background to allow them to write and understand pro- First, it is not fun to mark cards, therefore grams that deal with topics presented in high students tend to fUrik 1 little more before they school. attempt to run a program.

With these diverse needs and attitudes how can Second, when programs do go wrong, students Regis or any other school provide: can work through the program at their desks on the hard copy they have. Once the mistakes are found, A) The equipment and floor space to teach all corrections can be made and the program can be students programming. quickly run again. That is, the cards serve as a storage medium that is quick and reliable. B) The class time necessary to have a satisfactory computer programming experience which will Our first problem was solved at this point. serve as the foundation of computer literacy for We had a cheap piece of hardware which a lot of these students. students could use at the same time. But less than one half of our graduating seniors had even taken HISTORY OF OUR SOLUTION a computer course. This was not satisfactory.

In 1978 we built a Digital Group Micro pro- Would be possible to take two weeks away cessor. Students helped with the project and they from Geometry and teach the whole class how to do were allowed to use it anytime they wanted. We, at area problems on the new computer system?

99 The answer was a resounding YES! They all During the pilot program we developed a book- learned to operate the machine, mark cards, and to let which covered the history of computing, flow- use the Basic commands LET, READ, DATA PRINT, END charting, READ, REM, LET, DATA, PRINT, GO TO, IF and FOR-NEXT. The class' response was, as you THEN, END and FOR-NEXT statements and the function might expect, enthusiastic but it was remarkable in SORT. The students are asked to write and run several other ways. twenty programs using the statements indicated above. These programs are not all related to geom- First, the students were becoming computer etry as they cover some business topics and some literate. The Micro System has all four basic topics from algebra. parts of any computer system. There are: The objectives for the pilot program included 2 Input Devices - a keyboard and a card preparation of a syllabus for a modified Geometry reader course which would include the computer unit and which would not deteriorate the skills of Geometry 2 Output Devices - a CRT and a printer as measured EY-Our final exam.

2 Storages Devices - tape storage and We tested this objective in the following way. cards Three sections or about 90 students took the four- week computer unit. The other three sections did 1 C.P.U. not take the computer unit. They served as our primary control group. This second group spent A student can see and touch it all on one cart their extra time on constructions and review of and with a little help, understand these components topics that had already been presented. as parts of any system. In May of 1982 the same final exam was given Second, the students achieved some small power to all geometry students..It was essentially the over the machine even in a short period of time. same exam given in past years. Note, that the Possibly they could become comfortable with them previous years results served as a secondary con- and not fear them. trol group. The exam given covered all the material that was common to all sections. It did not cover Third, there is a link between Geometry and computes nor was there a heavy emphasis on con- computing. One student went as far as to observe structions. that "programs are much like Proofs in Geometry". We and the National Council of Mathematics teachers On that test the sections that had taken the agree with him. computer unit scored better than the control group.

SOLUTION TO THE SECOND PROBLEM This is exactly what we anticipated. The control group had scored lower on the first, second How do you provide the class time to have a and third quarter common exams.This is due to satisfactory introductory course for all students? slightly lower ability in those classes. However, the difference between the median score for both In the results of the experiment indicated groups remained the same for all exams. We thought above, we could see the possibility of finding a that the difference between the two groups would be solution to our second problem. The theory is that smaller on the final exam because of the time taken you can take a substantial amount of time from out for programming. It was not smaller. Geometry to teach programming without hurting Geometry because the two subjects are mutually When the results of this exam were compared to enhancing. the results of previous years' exams with the same ability range students taught by the same teacher In January of 1982, we applied for and received (our secondary control group), we found that there a Title 4-C Mini Grant to test our theory. was less than a 2% difference.

The grant proposal was divided into two parts. Geometry has not suffered. We have in hand a The first part was a pilot program in which four solution to the second problem mentioned above. It weeks would be taken from Geometry in the Spring of is possible to teach a worthwhile introductory 1982 for one half of our Geometry students. Based course on computer programming in a Geometry course on the experience gained with the pilot program the while at the same time maintaining high standards second part of the proposal would develop a six- of academic excellence for Geometry. weeks unit in computer programming and would involve all of our Geometry students in the year 1982- The four week period was not long enough to 1983. allow all students to finish all of the programs. It was, however, about 75% successful. We antici- The main point of this paper will concentrate pate that the six weeks program to be presented on the results of the pilot program of 1982. The this year will take care of that problem. We will results of the full program will be presented at be very interested to see how the scores on the the National Educational Computing Conference in final exam for this year compare with those of pre- the oral presentation of this paper. vious years. These results will be presented in the oral report of this paper.

100 GEOMETRY SYLLABUS 4. Some Consequences of the Parallel Postulate Time schedule Time schedule 5. The Angles of a Triangle without with 30-day 6. Two More Ways to Prove computers computer unit Triangles Congruent

Chapter 1 THE NATURE OF DEDUCTIVE Chapter 8 QUADRILATERALS REASONING 4 days 4 days 9. days 9 days Emphasis on: If...Ther... 1. Quadrilaterals Converse, Inverse, Contra- 2. Parallelograms positive, IFF., Only If, 3. Quadrilaterals That Are Direct Proof, Indirect Parallelograms Proof (Math example only) 4. Kites and Rhombuses Some Work on Euler Diagrams 5. Rectangles and Squares 6. Trapezoids Chapter2 FUNDAMENTAL IDEAS: LINES AND ANGLES Chapter 9 AREA 10 days 10 days 1. The Distance Between Two 9 days 9 days Points 1. Polygonal Regions and Area 2. Betweenness of Points 2. Squares and Rectangles 3. Rays and Angles 3. Parallelograms and Triangles 4. Angle Measurement 4. Trapezoids 5. Complementary & 5. The Pythagorean Theorem Supplementary Angles 6. Heron's Theorem 6. Betweenness of Rays Chapter 10SIMILARITY Chapter 3 SOME BASIC POSTULATES AND THEOREMS 11 days 11 days 10 days 10 days 1. Ratio and Proportions 1. Postulates of Equality 2. More on Proportion 2. The Bisection Theorems 3. The Side-Splitter Theorem 3. Some Angle Relationship 4. Similar Triangles Theorems 5. The A.A. Similarity Theorem 4. Theorems about Right Angles 6. Proportional Line Segments 5. Some Original Proofs 7. The Angle Bisector Theorem 8. Perimeters and Areas of Similar Chapter 4 CONGRUENT TRIANGLES Triangles Chapter11 THE RIGHT TRIANGLE 14 days 14 days 1. Triangles 13 days 7 days 2. Congruent Triangles 1. Proportions in a Right 3. Some Congruence Postulates Triangle 4. Proving Triangles Congruent 2. The Pythagorean Theorem 5. More Congruence Proofs Revisited 6. The Isosceles Triangle Theorem 3. Isosceles and 30-60 Right 7. Overlapping Triangles Triangles 4. The Tangent Ratio Chapter 5 DISCUSSION OF REFLECTION AND 5. The Sine and Cosine Ratios SYMMETRY ONLY 6. Trigonometry Is Taught from 3 days 3 days Unit Circle Approach 7. Use of Hand Held Calculators Chapter 6 INEQUALITIES Chapter 12CIRCLES 7 days 7 days 1. Postulates of Inequality 13 days 13 days 2. The Exterior Angle Theorem 1. Circles, Radii and Chords 3. Triangle Side and Angle 2. Tangents Inequalities 3. Central Angles and Arcs 4. The Triangle Inequality 4. Inscribed Angles Theorem 5. Secant Angles 6. Tangent Segments Chapter7 PARALLEL LINES 7. Chord and Secant Segments 8.* Inverse Covered by a Discussion 9 days 9 days 1. Parallel Lines 2. Perpendicular Lines 3. The Parallel Postulate

101 Chapter 13 CONCURRENCE THEOREMS 30 DAY COMPUTER UNIT

5 days 3 days SYLLABUS 1. Concyclic Points 2. Cyclic Quadrilaterals 3. Incircles Time ACTIVITY 4. Ceva's Theorem 5. The Centroid of a Triangle 3 days Discuss history, flowcharting and the 6. Some Triangle Construction first program in the text covering REM, READ, LET, DATA and PRINT statements. Chapter 14 REGULAR POLYGONS AND THE CIRCLE 5 days Show how to mark cards and use the 12 days 7 days machine. Write and run 3 introductory 1. Polygons programs. Administer first quiz. 2. Regular Polygons 3. The Perimeter of a Regular 6 days Discuss Loops, arithmetic symbols and Polygon the statements GO TO, IF-THEN and END. 4. The Area of a Regular Polygon Run 3 more programs using these state- 5. Limits ments. Administer second quiz. 6. The Circumference and Area of a Circle 6 days Discuss counters and debugging. Run 5 7. Sectors and Arts programs. Administer third quiz.

Chapter 15GEOMETRIC SOLIDS 6 days Discuss FOR-NEXT Loops and assign 2 programs using them. Discuss the first 10 days 5 days 2 programs and assign 3 more. Adminis- Emphasis on '7orhlas and ter fourth quiz. Not Proofs 4 days Assign the last 2 programs. Administer Chapter 16NON-EUCLIDEAN GEOMETRIES final exam on the unit.

3 days 3 days

CONSTRUCTIONS 30 days TOTAL

11 days 0 days

Changes were made in

Chapter 11, saving 7 days Chapter 13, saving 2 days Chapter 14, saving 5 days Chapter 15, saving 5 days Construction saving 11 days 30 days TOTAL

The only topics that have been dropped are Ceva's Theorem, Concyclic Points and Limits. All other topics are taken. Less time is spent on topic development and practice as indicated above.

102

1 1 j A Programming Environment for Preliterate Children

Charles E. Hughes J. Michael Moshell The University of Central Florida The University of Tennessee, Knoxville

Both of Gentleware Corporation Knoxville, Tennessee

Abstract Since the activity of programming is so beneficial, it is desirable to introduce it to very young This paper describes a programming environment children. Unfortunately, programming environments that provides a gentle, non-textual introduction are, in general, usable only by those who can both to programming. The, system, named KIDBITS, has read and write. Exceptions to this are certain as- been used experimentally for about two years and pects of the Smalltalk environment (Goldberg 1981, is appropriate for a wide range of beginners, Gould 1982), some implementations of turtle graph- including children as young as five years old. ics (Papert 1980), programmable toys such as Big Trak, and simple artists tools such as QUILT and PAINTER (Moshell 1982).

Introduction This paper describes a programming environment that provides a gentle, non-textual introduction to pro- Many people fail to learn programming because they gramming. The system, named KIDBITS, has been used stumble over the syntax of the language. Remem- experimentally for about two years. This is, how- bering facts such as where semicolons go is not ever the first paper that describes it. an important skill that programming develops; yet, this is the first and, for many, the insurmountable Technical Details barrier to cross. KIDBITS is a programming system in which icons When we try to teach programming to young children, (pictures that represent actions, or objects upon the syntax issue is an even greater stumbling block. which actions occur) are assembled into sequences This is most unfortunate since the real benefit of to achieve some desired effects. The paradigm used learning to program comes from the problem solving here is one of a movie director (the child) putting skills that are developed. together a film. KIDBITS permits the making, editing and playing back of these movies. Among the skills developed in learning to translate a problem statement into a programmed solution are Entering a Movie the ability to: Typically, the movie's director uses a game paddle Analyse a problem statement, to determine if it is to select the components of a movie. The chil, clear enough to allow you to start designing a first sees the following display in the lower part solution. of the screen.

Break a large problem into more easily managed sub- problems. 56 ccd OP 9 Develop an algorithmic approach to solving a clear- A ly specified problem. There are six choices here. The first option,"sun- Scientifically test a proposed solution to deter- rise", means "begin a movie".You use the paddle mine if, and under what conditions, it fails to to position the cursor (") under the chosen action, solve the given problem. and then press the game paddle's button. Devise a portion of a solution in such a way as to (Cursor control keys can be used on systems having correct this part without polluting other compo- no paddles.) nents.

Prepare descriptions of your solutions so that others may understand both what you solved and how you solved it.

103

Jr) Here is a description of all the options: Here's a typical animation seqUence:

Graphic Meaning Details BEGIN TELL BIRD Select the bird Sunrise "BEGIN" Enter a sequence of commands HOP Bird hops three times to the right Movie Camera "SHOW"Replay the movie HOP HOP Broom "EDIT"Modify a movie Glasses "READ" Get an old movie from disk TELL CAT Select the cat HOP Cat prepares to pounce on bird Pencil "WRITE" SaVe current movie on disk HOP Stop Sign "BYE" End of the day's work UP UP

RIGHT Let's explore what happens when you select BEGIN. RIGHT

You receive a new menu. Here are its contents. DOWN DOWN Cat lands on bird

TELL BIRD Select the bird SHRINK Bird gets small quickly [I] SHRINK (being eaten, perhaps) SHRINK Graphic Meaning Details END End taping session for this movie Face "TELL" Select an animation figure Replay the movie Counter- SHOW clockwise Arrow "SPIN" Spin the figure 45 degrees Clockwise Any number, 1 through 9, may be typed and then be- Arrow "HOP" Make the figure hop comes the "repeat factor". This number is displayed in the lower right corner of the screen. If Four you type 3 then select HOP, the current figure hops Diverging three times. Arrows "GROW"Make the figure grow Four Selecting the repeat symbol, a left square bracket, Converging is analogous to using a Pascal or Basic FOR loop: a Arrows "SHRINK"Make the figure shrink subsequent sequence of commands is executed some fixed number of times. Prior to choosing this sym- Arrows to "LEFT", bol, you enter a repeat factor that determines how Left,Right, "RIGHT", many times the following sequence of commands is Up, Down "UP", Move the figure executed. "DOWN Brush "PAINT" Make a copy of the figure As soon as you choose a repeat group, the menu changes slightly. Left The moon, used to terminate the Square program, disappears and a right square brackets ap- Bracket "REPEAT"Repeat some commands pears in its place. This icon is selected to close out a repeat group. Moon "END" End of the movie An example of the use of the repeat is the following sequence to create a flower garden.

When you select TELL, you receive a new menu of op- BEGIN tions. This is the list of folks to whom you can "tell" something. TELL FLOWER Select flower

There is a large selection of animation figures 4 available. However, for any given execution, you REPEAT Repeat four times may have at most six. As an example, we can have a menu that looks like this. PAINT Make a copy of flower 3 Move flower over to make RIGHT room for next copy In garden A. These figures represent: END REPEAT End repeated sequence Flower Bird Cat Rocket Triangle END

104 Editing a Story appear the set of all icons that may be used at the current point of story telling. The game paddle is If you have a movie which you have entered into the all that is needed to enter the above story. The computer, or loaded from the disk, and you want to position of the paddle controls the cursor that may change it, select the EDIT option (the broom). be moved across the icons.

The movie, as currently known to the computer, will Pressing the paddle's button enters the currently be displayed at the top of your screen.This selected icon as part of the story and shows the script is read from left to right and from top to effect of this action in the cartoon being gener- bottom, just like English text. ated. Each such selection results in the display of the text of the corresponding Pascal-like state- A menu of five choices appears on the bottom of the ment. screen. The child who is already reading is lead to learn The options are: some of the syntax of a programming language. The child who does not read is helped to develop this skill by the word/action association. Graphic Meaning Explanation Some of the prominent features of this child appro- Right Arrow Forward Move forward in the script priate computer story telling system are: Left Arrow Backward Move back,lard in the script 1) Children and adults can work as a team. Adults I(nsert) Insert Splice new action into movie provide a more highly developed sense of logic plus the reading/writing skills needed for more advanced D(elete) Delete Cut out scene from movie work. Children provide imagination and enthusiasm. Q(uit) Quit Go back to outer menu Each learns to program the computer, not be programmed by it.

2) With rare exceptions, e.g., establishing communication with a figure by the TELL verb, each The cursor movement commands (left, right arrow) command requires the selection of only one icon. operate as in a normal scan of text. Moving right of the end of a line positions you at the start of 3) Syntax errors are impossible.Thus the begin- Moving left of the start of a line the next one. ner needs to be concerned only about errors of positions you at the end of the previous one. logic. This helps to decrease the frustration level of a high energy, low attention child (or, When you select the Insert option, you are given for that matter, adult). the same menu you used when building the movie in the first place. You use your game paddle to move 4) Textis always displayed to correspond with the back and forth and select an item to insert into selectedactions. This helps to initiate partici- the movie. pants tothe next stage of programming. When you select the Delete option, the item pointed 5) Only one control structure is introduced, the to is just removed from the story. repeated set of commands. This acquaints students with repetitive execution while retaining the Modes of Presenting this Material essentially imperative nature of the language. Most children, when learning their native language, 6) Movies may be edited. This early introduction master the spoken language by the time they are to editing emphasizes debugging and the important They accomplish this remark- only three years old. realization that you need not throw everything away able task by the simple techniques of mimicry and when your story line is not exactly what you want. observation of cause and effect. Summary Several very successful educators have capitalized The Suzuki meth- on this natural mode of learning. This paper presents a technical description of od of teaching music does not initially train child- KIDBITS' features, anA a scenario on how it may be ren to read music. Rather, an adult, usually a used to introduce children to the art and science parent, is brought in to provide a role model. The of programming. The environment provided by child then learns by the familiar methods of mim- KIDBITS is object-oriented, where the objects are icry and cause/effect experimentatlon. usually cartoon characters. Programming in this system is non - texts,.... This fact, together with The software we are describing provides an environ- the appealing graph). 3 it very promising for ment for children to enter, modify and play back use with preliterate simple animated stories. Reading is not a prerequisite skill. Certain aspects of KIDBITS have been intentionally designed to reinforce other educational objectives, The child, with adult assistance, but hopefully not such as reading and understanding the meaning of interference, can enter this story via a menu single digit numbers. selection scheme. On the bottom of the screen will

105 124; The present version runs on an Apple II, II+ or IIe computer with 48K of memory and one disk drive. Distribution of the software, along with a self- teaching text, is being done by John Wiley and Sons.

References

Goldberg, Adele and Joan Ross, Is the Smalltalk-80 System for Children?, Byte, Volume 6, August 1981, pp. 348-368.

Gould, Laura and William Finzer, Programming by Rehearsal, Personal Communication, December 1982.

Moshell, J. Michael et al, Computer Power, McGraw- Hill, New York, NY, 1982.

Papert, Seymour,Mindstorms: Children, Computers, and Powerful Ideas, Basic Books, New York, NY, 1981

106 12j Teacher Training in Computer Education William Wagner, Chairperson Santa Clara County Office of Education San Jose, CA 952,15

ABSTRACT the relative lack of technological We are often asked, "Howcan we ensure sophistication of the average teacher, anti that computers will not be just another fad the lack of new teachers entering the in education, like TV or flexible system, the problem of teacher training scheduling?" Since its beginning, the becomes absolutely essential to the computer education revolution has been continued progress of this particular different from other innovative episodes innovation at this particular time. because it has featured the involvement and The participants in this panel are four leadership of classroom teachers. leaders in the field of computer education Thedilemma for continued successful who have extensive experience meeting these implementation is to alloweach wave of needs in both pre-service and in-service. newly involved teacher to experience the They bring the perspective of local school sense of ownership, pride and power which district, regional center, state department is critical in the success of any of education and teacher training innovation. institution. When this need is viewed in the light of

Bobby Goodson Computer Resource Teacher Cupertino Union School District Cupertino, CA 95014 Gary Neights Bureau of School Improvement Pennsylvania State Department of Education Harrisburg, PA 17108 Nancy Roberts School of Education Lesley College Cambridge, MA 02138

107 Instituting Computer Programs within a School District John Cheyer Manchester High School Manchester, CN 06040

ABSTRACT This session will describe computer classes "Probe'' - two teachers circulate between or programs being offered by the Manchester the ten schools teaching computer School System. programming to the gifted children in the fifth and sixth grades. "Principal In-Service Program" educate principals on different ways the "Computer Literacy' - every seventh grader microcomputers may beused in the school from special education to gifted students system. receives 15 lessons on computer literacy. Seven lessons are onhistory, parts of the "Teacher Awareness Yeat" - two VIC-20's computer and usage in societyand eight with packet developed by summer curriculum lessons are hands on use of the computer.. money explaining hook-up, simple computer programming, and assorted CAI software "Introduction to Microcomputing" - a programs. Each computer stays at each semester offering to eighth and ninth school for 2 months in which teachers can graders teaching BASIC. take it home with them to learn at their own pace. "High School Projects" Thehigh school has 21 computer available to teach BASIC "Elementary Task Force Committee'' - will be programming. Another lab is available for evaluating pilot programs that are just be CAI. Four departments (English, introduced in selected elementary_schools. Mathematics, Science, and Social Studies) For the past two years, we have been have been invited to develop materials evaluating other school district programs using SUPER PILOT or to use courseware from which are using LOGO, PILOT, and CAI MECC. materials.

108 14/4,J Voice Input/Output New Directions in Instructional Technologies

Carin E. Horn, Chair Scott Instruments Corporation Denton, Texas 76201

ABSTRACT Speech input/output is adjusting the 'traditional' computer assisted learning environment. The panelists in this session will address various aspects of voice recognition and speech synthesis technologies with educational applications.

Remarks will attend to: a) historical voice I/O developments and applications; b) the impact of voice I/O on student performance; c) real-time pronunciation feedback and foreign language learning, and d) voice recognition for special eduation. A demonstration of the VBLS (trade mark) voice-based learning system will be given.

PARTICIPANTS

Richazd H. Wiggins Editor, Speech Technology Herb L. Nickles California State University Harry S. Wohlert Oklahoma State University Brian L. Scott Scott Instruments Corporation

109 Educational Use of Microcomputers by Special Needs Students

Joan Davies, Chair Lynbrook High School San Jose, California

ABSTRACT Just because onehappens to work with computer programs, as well as, the role of special educationstudents in Silicon district level and county level personnel, Valley, the heart of California electronics in providing appropriate computer services industry, it does not mean that computer for special needs students. instruction automatically becomes part of Panel members are leaders in the field the curriculum. As much exertion is of compute use for special needs students required here as anywhere else to implement and have extensive staff development computer use in special education. experiences. Theywill address concerns Sessionparticipants will hear panel from school, district, and county level members discuss exemplary school site positions.

PARTICIPANTS: Joan Davies Lynbrook High School San Jose, CA Don Clopper Coordinator of Special Education Santa Clara County, CA

110 Needs and Opportunities for Educational Software in Grades K-12

Edward Esty, Chair OERI Washington, DC 20208 Robert Tinker Technical Education Research Centers, Inc. Cambridge, MA 02138 Lawrence M. Stolurow Center for Educational Experimentation, Development and Evaluation University of Iowa Iowa City, IA 52242 Darlene Russ-Eft American Institute for Research Palo Alto, CA 94302

ABSTRACT Recognizing the urgent need for improvement in the quality and the quantity of educational software for microcomputer, theNational Institute for Education soonsored an eight month study to document the present state of microcomputer use in elementary and secondary schools, and to determine the needs expressed by educators for software which will enable the full potential of microcomputers to be realized. The results of the survey ofteachers, administrators, and software developers will be summarized. The panel will present recommendations for the effective use of microcomputers in the areas of 1) math and science, 2) reading, writing, and communication, and 3) foreign language instruction. Discussion will focus on the ability of microcomputers to improve student motivation and performance, to introduce new topics into the curriculum, to increase teacher productivity, and to decrease educational costs. The final report of this project should be available at the time of presentation and will contain an exhaustive directory of educational software, commercial and non-commercial sources of software, anda bibliography of pertinent journals, articles, hooks, and other research studies. Program Maintenance ... The Forgotten Topic

by Frank W. Connolly

Center for Technology and Administration The American University Washington, D.C. 20016

Abstract new employees to do productive work without constant supervision. Third, doing maintenance Program maintenance is a topic that is programming gives new staff members the opportunity frequently overlooked in undergraduate programming to learn the accepted norms of their new working classes. This paper offers an approach to present- environment: programming standards, library con- ing the topic. It is based on the author's class- tents, and major file formats and conventions, etc. room experience. However, maintenance programming requires knowledge of some special techniques. If students Introduction are not introduced to these techniques and requirements, they are ill prepared to effectively handle Programming courses include numerous topics -- the initial programming assignments they are likely structured design, problem solving, language gram- to receive.Thus, the employer's initial impres- mar and syntax, interfacing with operating systems, sion of both the new employee and the new employee's file construction, programming techniques, and education may be nagative. In addition, the psycho- more. Discussions with former students brought to logical impact on neophyte programmers is signifi- light a significant topic I had overlooked for cant. While adjusting to new work settings, they years. A review of several textbooks reflected a must contend with programming tasks which are similar oversight. The topic? Program maintenance. unfamiliar to them.

Many faculty use a building block approach to Environment teach languages and programming techniques. Stu- dents begin with simple programs and then expand The Program Maintenance module I developed them, adding new requirements and using new tech- was introduced in two introductory COBOL courses. niques. Such an iterative approach to program The courses are conducted using strict structured development is a valid teaching approach, but is programming techniques. For each programming not what I consider program maintenance. For pur- project, students receive specifications. There poses of this paper, Program Maintenance is defined are two walkthru's scheduled -- the first at the as modifying, correcting and extending an operating design stage, the second at clean code. The design program written by someone else. phase walkthru requires each student to prepare IPO charts and pseudo-code.The code walkthru Recently, I instituted a learning module on requires that the compilation be free of all Warn- program maintenance in my COBOL classes. That ing, Caution and Error messages. There are strict experience is the basis for this paper. coding style requirements: naming, use of literals, comments, and format. Student teams are formed Background for each project.

Many students leave the halls of academia to At American University, students use MUSIC start work at the bottom of the computer career (McGill University's System for Interactive Com- ladder. They are not given programs to create puting). All program code and execution is done from scratch, as they did in class, nor are they on an interactive basis on the University's assigned sophisticated new systems to implement. IBM 4341 via IBM 3270 terminals. From the begin- Instead they are given the task of program ning of the semester, students work in a "library- maintenance. Maintenance tasks are logical oriented" environment as program segments and con- assignments for a new programmer. First, as trol language are retrieved by students from a individuals at the lowest rung on the professional public library and incorporated into their code. ladder, they are assigned work that more senior During the eleventh week of a 15-week semester, members of the staff dislike. Second, maintenance the students receive the maintenance assignment. tasks are easily defined and controllable, enabling

112 12i Assignment thought. While the students acknowledged they had learned a great deal, they considered the amount of In prior course projects, students received a frustration they experienced as excessive. Upon detailed layout of the input, a description of the reflection, I concluded that although the same processing to be accomplished, and a layout of the tools are used for creating and maintaining pro- uutput format. For this project, they received: grams, they are used for different purposes. Therefore, I backtracked and added a new teaching 1. IPO charts of an existing program module to explain the maintenance task. 2. pseudo-code for the existing program 3. a copy of the existing program Suggestions and Findings source code 4. a description of the modifications Finding little material in the literature, I required developed the following approach for performing 5. a format of the output for the maintenance programming: updated program I. Determine where you are After the code walkthru, data and control language for running the updated program was What does the existing program do? provided. The analysis begins with IPO charts and pseudo-code, to give the main- The program to be updated was written by a tenance programmer an overview of former student at the University who understood the program. that the program was to be used by students as described above. It was a short (approximately Verify that the IPO and pseudo-code 150 lines of code), relatively simple program are in agreement both with each other, (one file in and a report file out). The code and the code. It is recommended that did not adhere to the coding standards used in the the maintenance programmer examine course and contained few comments. The IPO and the test data, predicting the output pseudo-code supplied were accurate. prior to running the program.

The major restriction placed on the students Run the program as is, using the test was that they were not to make unnecessary data. If the results agree with the changes to the existing code. Unnecessary student's predictions, they have an changes were defined as changes made for style or excellent understanding of the code. convenience purposes. All new code added was to Ii' they don't agree. they have a be in accordance with the style standards estab- road map to indicate where the dis- lished in the class. That meant that the com- crepancies are. pleted code would contain two different styles of writing. II. Evaluate the changes required

Expectations Cla71,ify the changes as "Substantive" or 'tosfietic." Cnsmetic changes do While the students appeared to hale few rot affect the logic flow of the expectations about the project,I had several. existing prog.an. These include such First,I thought students would enjoy working on things as headiLgs, layout modifi- a program which had.existing code. I anticipated cations, field sizes, etc. These they would find an advantage in being presented generally require changes to the DATA general approach to solve a problem. Having to DIVISION, not the PROCEDURE uIVISION. contend with code not as well-written as they would have ft,I hoped to reinforce the need fr,r Substantive changes change the logic good structured technique, especially in nvil!.g flow. These include addition of and documentation. I also expected that stpdents routines and changes to the sequence would easily make the transition from "cmator" of the current program. These changes of their own programs to "maintainer" of someone ._ primarily affect the PROCEDURE DIVISION. else's programs. III. Implement the Cosmetic Changes Reality Withouta walkthru, code and then Not surprisingly, reality was significantly test these changes. different from my expectations. Students did not like doing someone else's programming. They had great difficulty adjusting to the role of "fix-it person." Clearly, the transition from creating to maintaining programs was not as easy as I

113 13 0 IV. Implement the Substantive Changes Summary

Pirst, prepare IPO and pseudo-code For 13 years of teaching college-level to reflect the program when the programming courses, I overlooked a topic of substantive changes are implemented. significance to my students -- program maintenance There is a design walkthru at this techniques. It wasn't my inspiration or insight point. that brought it to light. Former students recounting their initial, on-the-job frustration, Second, code and compile the program led me to include a unit on programming mainten- with all changes included. When it's ance in my introductory programming courses. clean, conduct the code walkthru, Based on my review of texts, it is a topic that is forgotten in many courses. But, if we are to Pinally, as with a program written prepare students for the world they will face from scratch, test it. upon graduation, we need to add Program Maintenance to our programming courses. V. Document the program

Document the entire program, not just the modifications.

Student Suggestions

In addition to normal course evaluations, I conducted a debriefing session with students when the project was completed. They had two highly constructive suggestions for using this project in the future.

1. They felt the program I gave them was too easy, As they had written programs of comparable size, they were tempted to rewrite the program, rather than upgrade it, (Several confessed to rewriting it completely. After the output of their totally rewritten program was correct, they went back to their code and replaced portions with the original code where possible, so it conformed to the assignment.) Therefore, they suggested that the original program be larger -- something that would show them the size of a "real world" program. They believed such a program would eliminate the temptation to rewrite the code instead of updating it.

2. They suggested that no new techniques or code concepts be introduced as part of the modifications. The assignment I had presented required them to add control breaks to the existing program. Having worked with accumulators previously, I considered the addition of control breaks a minor new experience. The students felt otherwise, They did not like having their concentration divided between learning control breaks and learning to perform maintenance tasks.

114

f' 131 AN ENVIRONMENT TO DEVELOP AND VALIDATE PROGRAM COMPLEXITY MEASURES

ENRIQUE OVIEDO* AND ANTHONY RALSTON

Department of Computer Science State University of New York at Buffalo

Abltract further to find which ones canprovide adequate data on the complexity of programs. In this he theory and practice of Software Engineer- paper, we -argue that college and university pro- both require the objectivedefinition and gramming courses are the best and perhaps the only measurement of the complexity of programs.Numer- useful medium ,n which to carry out experiments to ous measures have been proposed for this attribute validate an compare pros va. complexity measures but little Is known about their predictive power and models. Large numtors ,'programs ' .)al many and limitations. Extensive empirical studies are different subjects and 'n arentlan

115 132 Our thesis, and that of several other language used by the program which is Incorporated researchers, is that program complexity needs to In program complexity models varies widely. Soft- be defined asa property of the program itself in ware Science's model simply distinguishes between order to avoid any subjective components. As a operators and operands whereas models based on the first step in this directionIt seems reasonable program's controlflow, data flow or data acces- to assume that the complexity of a program results sibility require extensive knowledge of the syntax from a combination of what the program does (i.e., and semantics of the language. the program's intrinsic functional complexity) and It is reasonableto assume that the more how the program works (I.e., the program's imple - knowledge used in the analysis ofprograms, the mentation complexity). It would be widely accept- more Information can be extracted about the pro- ed, forexample, that a Fast FourierTransform gram contentsto improvetheusefulness of the program Is intrinsically more complex than a Bub- diagnostics of program complexity models. For ex- ble Sort program. Similarly, from the implementa- ample, program complexity models based on the con- tion complexity point of view,it would be safe to trolflowanddataflow characteristics of pro- assume that appropriately chosen control struct- grams couldpinpoint anomalies likeunreachable ures and data structures make a program more read- sections of code, unnecessary assignments, unfact- able and understandable than an unstructured pro- oredexpressions, etc., In the program 14,131, gram for the same function 1131. Software Science's model, on the other hand, could Although the eventualaim ofresearch Into only suggest the presence ofthose anomalies In program complexity must be to define measures of the'program 171. The useof more information the Implementation complexity of a particular al- about the program,however, makes more difficult gorithm, In the absence of a generalunderlying the formulation of simple measures and the valid- theory we argue that we should be content for the ation ofthe complexity models. This trade-off present to define relative measures which canbe between the depth of the analysis of the programs used to compare a given (e.g., student's) program required by the different program complexity with a standard (e.g.,instructor's)program for models and their predictive power Is a basic prob- the same function1141. Thus, a plausible initial lemin this area which must be resolved III. approach to the study of program complexity would (4) Data Gatherin. Technl ues and Ex erlmental be to analyze measures ofprogram implementation Design complexity by comparing the corresponding measures Numerous models and measures of program com- of several alternative programs for many different plexity have been proposed butno major efforts functions. Aswe learn more about the relative have been made to find out their predictive power importance of the different program attributes and and limitations 151. Totest these models and gain confidence in the existing models and measur- measures we need methods to evaluate independently es, we will hope to achieve gradually the goalof the complexity of programs. Obviously, these having absolute measures. methods have to bedesigned in relationto the (3) Program Complexity: Approaches to Research theoretical question being investigated. However, Researchers have generally used two different a careful analysis of these methodsis essential methods to study program complexity. One approach in this research area because they can suggest how has been to quantify the performance of program- to formulate models and measures that can be more mersin order to test hypotheses about the effect easily studiedin an empiricalform. We need to of certain program factors on the comprehensibil- determinewhich (combination of) data gathering ityof programs. The other has beento develop techniques providedata which are accurate,un- models to compare programs based ontheirstyle blaseC.and representative of the complexity of the and organization. Research based on these two ap- programs as experienced or perceivedby program- proaches has faced serious methodological problems mers. This is a major task, indeed,given the andfurther research is needed toobtain models broad spectrum of programmers' skills, program that can be used successfully to Ixplain program- sizes and functions Some methods which require a mer-program interactions 121. direct observation of the behavior of programmers Studies of programmers' behavior have tested have been used to try to measure the complexity of hypotheses concerning the effects on program com- programs. As noted, one example Is quantification prehensibility of factorslike mnemonic variable of the difficulty (e.g., amount of time) with names, indentation, paragraphing, alternative con- which programmers could trace, debug, modify, trol structures and sources ofinformation exter- etc., the test programs. A wide variability In nal to the program such as flow-charts and docu- the performance ofthe subjects was found and it mentation. In this type of study, the performance would be difficult to develop tests to control or of programmers was measuredby scoring howwell compare the subjects of these experiments based on the programmers could memorize, modify, debug, their skills,training andfamiliarity withthe hand trace, answer questionnaires, etc. (8, 9,10, purpose ofthe test programs 12, 81. Although 151. tracing, debugging, etc., are processes that Program complexity models have been developed belong to the software development cycle, they do both to achieve a betterunderstanding of the not Insure that programmers have understood the factors that contribute to the complexity of pro- test programs or faced allsources of complexity grams and to obtain objective complexity measures. in them. Finally, from a pragmatic point of view, Program complexity models are Intended to comple- these methods couldnot be usedto gather large men the more developed program correctness amounts of data because they would be too tedious models. The amount of knowledge about the

116 133 and time consuming with non-trivial test pro- Programs "), It wouldbe widely agreed that the grams. appropriate choice of control and data structures We believe morecomprehensive andadequate can make programs more readable, easier to debug data can be obtained from experiments where prop- and more reliable. This general argument and the erly trained (panels of) experts Judge the rela- results of studies that suggest that the control tive complexity ofalternative programsfor the flow and data flow of programs play an important same function. A generalargument In favor of role In the ability of programmers to understand, this contentionis that reading and Judging pro- trace and debug programs id, 6, 9, II, 121 have grams Is an essentially simpler process than writ- led us to formulate a model of program complexity ing, tracing, modifying, etc., them thereby making based on the control flow and data flow character- easier the collection of data. Reliable data may istics of programs. be expected from expert Judges who havea clear The modelof program complexity we have de- understanding of the purpose of the programs, the fined Is based on the assumption that,in order to language used and thetechniques used In their understand, trace and debug aprogram,a program- implementation (e.g., abstract data types, recurs- mer must,at least,be able to determine a)the ion, etc.). Additionally, a relatively large num- statements that precede andthestatements that ber of alternative Implementations should help the follow each statement in a program, and b) the set Judges give a wellInformed opinion of the rela- of variable references affected by each variable tive complexity of the programs. assignment and theset ofassignments by which Program complexity Isan elusive and multi- each referenced variable could have been defined. faceted concept. Therefore, extreme care must be Based on this assumption we have defined two taken In the design ofcase studies and experi- program attributes called control flow complexity ments. For example,comprehensive and unbiased (CF) and data flow complexity (DF). These attri- guidelines should be given to the experts Judging butes (CF and DF) have been defined Ina language the complexity of programs. These guidelines independent form and we have implemented a system should enhance the panelists' understanding of the that measures automatically the CF and DF of Pas- concept of program complexity,carefullydefine cal programs. In order to define control flow the program attributes they should (not) consider complexity, we view a computer program asa flow In the evaluation of the programs, how finely they graph having a single entry and a single exit node should classify the programs accordingto their anda number of nodesin between where each node complexities, etc. Similarly, test programs have represents a sequence of statements without to be carefully selected In order to control those branches intoIt or out ofit and the nodes are program attributes not takenInto account by the connected by edges thatrepresent pathsthrough particular model or hypotheses being tested (i.e., which the dynamic execution of the program flows. Insure the Internal validityofthe experiment) We assume that the difficulty in understanding the and to testthe models or hypotheses with a sequences of node executions (i.e., Its control variety of program types, sizes,and programming flow complexity CF)can be measured directly by languages (i.e.,insure the external validity of the number of edges of the program flow graph. the experiment'. Several studies indicate that the size of the (5) An Exemplary Experiment to Test a Model set of variable references affected by each vari- of Program Complexity able assignment and the set of assignments where We have developed a model of program complex- each referenced variable could have been defined ity whichIs based on the control flowand data have a significant effect on the ability of a pro- flow characteristics of programs (see below). The grammer to understand, trace and debug a program. case study we present here Is part ofan ongoing We surmisethat the difficulty in understanding effort to test this model. The test programs for the definition-reference relationships in a pro- this experiment were written by students ofan gram (i.e., Itsdata flow complexity DF) canbe introductory computer science course and a prelim- quantified by counting the number of variable inary analysis ofthe behavior ofthe model is definitions (references) associated with each done by comparingthe complexitiesassignedto variable reference (definition)In that program. these programs by the model and the graders of the The fifty-nine programs usedIn this experi- course. We do not regard the results of this ex- ment were the first programming assignment of an periment as a definitive test of the model but as introductory computer science course. Only cor- possible insights Into the data gathering techni- rect programs were usedin order that the assess- que discussed in this paper and the behavior of ments by graders of the implementation complexity the model that should be further explored. not be affected by syntactic or semantic errors in We firstgive a brief description of the the programs. These programs were written in Pas- model whichIs sufficient to understand the data cal and they ail compute the cube root of a set of gathering method we have developed to validate It. numbers providedin an Input file. Their output A detailed definition and analysis ofthe model was alist of the numbers read from the input file can be found in 1131. and the cube root corresponding to each number. All Control structuresand data structures are programs consisted of a single main program with no fundamental and closely Intertwined program procedures or functions. components (as exemplified by the title of Mirth's Two computer science graduate students book 1181, "Algorithms + Data Structures

117 134 assessed the complexity of the programs as excel- We have used the data obtained from this ex- lent (A), good (B) or poor (C) (with "excellent" periment to compare how well the control flow com- corresponding to "less complex"). One student plexity (CF), data flow complexity (DF), cycioma- graded 33 programs and the other graded 26 differ- tic complexity (V)1111 and a combination of con- ent programs. The graders were told only to Judge trol flow complexity (CF) and data flow complexity complexity In terms of the choice and use of con- (OF) correlated with the complexity grades assign- trol structures and data structures in the pro- ed by the graders. The grader of the 33 programs grams but they were not told what the details of assigned 13 A's,15 B's and5 C's. The grader of the model being tested were. In order to enhance the 26 programs assigned 4 A's,16 B's and 6 C's. the graders' understanding of the concept of pro- In Figures 2 and 4 wehave plotted the control gram complexity, they were asked to consider the flow complexity (CF), data flowcomplexity (DF) set of questions concerning some general and spec- and complexitygrades (A,B,C) assigned by the ific program characteristics shown in Figure I. graders corresponding to each test program for the It was our intent that these questions not give 33-set and 26-set, respectively. Figures 3 and 5 clues about the model being tested. Additionally, show tallies of the complexity grades assigned to to ensure a consistent classification of the pro- the programsand their corresponding cyciomatic grams, we instructed the graders to scan all the complexity (V) andcontrol flowcomplexity (CF) programs before grading any of them to form a for the two sets, respectively. clear notion of what constituted an excellent, In order to compare roul;':iiy these measures of good and poor implementation, and to ignore pro- program complexity, we have defined areas for A, B gram attributes not considered in the experiment and C-programs In Figures 2-5 and then have count- such as mnemonic variable names, indentation, com- ed how many programs have appropriately and Inap- ments, etc. propriately fallen Into each area. Our criterion to obtain these areas has been to take the maximum DATA FLOW (DF) RELATED_QUEST1ONS V, CF and DF of the A, B and C-programs asthe limits for the maximum complexities that A, B and I. Are the data structures appropriately chosen? C-programs can have. The maxima for V,CF and DF 2. Are the data structures properly used? for the programs from the two sets areshown In

3. Are the boolean conditions properly used? Tables I and 2,respectively. The zones defined 4. Are there unnecessary variables? by these limits are shown as dashed lines In 5. Are there duplicated sections of code? Figures 2-5. 6. Are there variables that should have been

declared as constants? TABLE 1

CONTROL FLOW (CF) RELATED QUESTIONS Max. V Max. CF Max. DF

1. Are control structures appropriately chosen? A-programs 4 13 10 2. Is the structure of the algorithm reflected In B-programs 6 19 16 the control structures used In the program? .C-programs 7 22 28 3. Can nested if's be comprised in simpler conditional statements, as TABLE 2 if bI then ifb2 then X A versus If bl and b2 then X = A? Max. V Max. CF Max. D.F 4. I; the level of nesting of control structures not so deep as to make the program unintellig- A-programs 4 13 9 ible? B-programs 5 16 17 C-programs 7 30 29 CF AND OF RELATED QUESTIONt

1. Is It easy to find and understand the conditions under which a section of code would be executed? Tables 3 and 4 show the number of A, B and C- 2. isit easy to follow the sequence of statements programs contained In the A, B and C-zones of Fig- In each path of the program ? ures 2 and 3, and 4 and 5, respectively. Based on 3. Is it necessary to backtrack often In order to theinformationprovidedby Tables 3 and 4, we follow the implementation of the algorithm? might conjecture thata combination of CF and DF 4. Could certain sections of the program be can classify the programs better than V,CF or DF programmed In a simpler way? alone, but obviously this hypothesis remains to be proved. Weemphasize that thiscase study was FtGURE 1 presented solely to Illustrate amethod for gathering data and validating models ofprogram complexity.

118 135 FIG. 2

C C 20

BB C B C B B

15 B

B A A B B B

10 AAB AAAAJ It BB AAA

P 11111111i 104 15 20 25 A ZONE B ZONE C ZONE IDF

F I G. 4 CF 30 C

C C 20 C

B B

BBB A

I° AAALB

20 25 A Z011E i B ZONE C ZONE DF

119

136 TABLE 3 TABLE 4

V CF V CF (see Fig.3) (see Flg.2) (see Fig. 5) (see Fig. 4)

A-Zone 13Aos, 9B's 13A's,8Bos A-Zone 4 Aos,14 Bos 4 A's, 10 B's B-Zone 6B's,4 Cos 7Bos,2Cos B-Zone 2 B's 6 B's

C-Zone 1 C 3Cos C-Zone 6 Cos 6 Cos

DF CF EDF DF CF d DF (see Fig.2) (see Flg.2) (see Fig. 4) (see Fig. 4)

A-Zone 13A's,5B's 13Aos, 3B's A-Zone 4 A's, 5 Bos 4 Aos, 4 Bos B-Zone 9B's,3Cos 11 Bos,2Cos B-Zone 11 B's, 1 C 12 B's

C-Zone 2Cos, 1 B 3Cos, 1 B C-Zone 5 Cos 6 Cos

FIG.3 FIG.5 CE CF 30 30

91 i25i 9-CCC zs

1Z 10 IN CI 8 tE 8 o, 1 CC N C;

201 C -J 0 _7 ci 7-C

'0 - - CCC -CC a l6 BB 6 BB-

15 11 0 BBBB Eo -J BBBB C BBB= 5 AAAAA -BB BUBB B,_5 AAABBBB1 BE Li All) s A 10 IN BMUS BBBBBB :Z E EBBBB BBBBBBB A, 4 AAAAAA 10:1?; Al 4AAAA 10 I z, 0' o, N. 41- E

120 137 SUMMARY 121 Brooks, Ruven, Studying Programmer Behavior Programming courses are not only an approp- Experimentally: The problems of Proper Meth- riate environment to perform experiments to test odology, CACM, Vol 23,1980, 207-213, hypothesesand models ofprogram complexitybut perhaps the only available environment in which to 131 Brooks, Ruven, Using a Behavioral Theory of do meaningful, experiments involving the collection Program Comprehension in Software Engineer- of large amounts of data. From such courses we Ing, Proc. 3rd Int. Conf. on Soft. Eng., can'obtain many programs for the same purpose, In IEEE,1978,196-201. many different computer science subject areas (e.g., numerical analysis, data structures, etc.) [41 Chapin, Ned, Data Accessibility in Structured and In a variety of programminglanguages. Many Programming, NCC, 1978, 597-603. experiments with different models can be performed to rank these programs according to their complex- 151 Curtis, Bill,In Search of Software Complex- ity, and thegrade assigned by the instructor ity, Workshop on Quantitative Software would be a readily available Independent evaivat- Models, IEEE catalog #THOO 67-9, 1979, Ion of the programs' complexity. 95-106. Given the methodological difficulties for studying program complexity that we have mentioned 161 Dunsmore, H., and Gannon, J.,Data Referenc- in this paper, we believe we should Initially con- ing: An Empirical Investigation, Computer, centrate our efforts on the development and valid- Vol. 12, 1979, 50-59. ation of models of program complexity for elementary programs and gradually develop more sophistl- 171 Gordon, Ronald, Measuring Improvements In cated models for larger programs. A large number .Program Clarity,CSD-TR-268,Purdue Univer- of elementary programs can be obtained from intro- sity, 1978. ductory computer science courses for testing simple models of program complexity which are based, 181 Gould, J. D., and Drongowski, P., An explore- for example, on the complexity of each module of tory study of computerprogram debugging, the program. Similarly a fairnumber of large Human Factors 16,1974, 258-276. programs rom advanced courses(although smaller thanlarge production systems) can beused for 191 Green, T., Conditional Program Statements and testing models that,for example,not only take Their Comprehensibility to Professional Pro- into account the complexity of each module but also grammars,J. Occup, Psychol., Vol.50, 1977, module interactions, appropriateness of data 93-109. structures, clarity ofdesign, etc. We believe that accomplishing these steps In the study of 1101 Love,T.,An Experimental Investigationon program complexity are essential before we attempt the Effect ofProgram Structure onProgram to develop absolute measures for the complexity of Understanding, Proc.Reliable Soft, Conf., large production systems. SIGPLAN, March, 1977,105-113. From a didactic pointof view, theuseof objectiveprogram complexitymeasures that can 1111 McCabe, T., A Complexity Measure, IEEE Trans. complement program correctness models is of Inter- on Soft. Eng., Vol. SE 2, 1976, 309-320. est Initself. The guidelines or feedback provid- edby good measurements together with good diag- 1121 Miller, L., BehavioralStudies ofthe Pro- nostic messages could provide more meaningful com- gramming Process, IBM Research Report RC 7367 mentary for students on their programs than they (#31711), 1978. are accustomed to getting from instructors or teaching assistants. This point suggests one 1131 Oviedo, Enrique, Control Flow, Data Flow and final--and potentially major--benefit of a soft- Program Complexity,Proc. Computer Software were quality measurement system based on the and Applications Conference, IEEE,1980, 146 - notion of quality relative to an instructor's 152. solution and/or the other students solutions to a programming problem 1171. With the rapid Increase 1141 Ralston,Anthony,Mathematics and Computer in demand on computer science faculty coupled with Science, Research Directions in Software a decreasing supply ofboth faculty and graduate Technology, The MIT Press,1979. students(i.e., teaching assistants),any system with the potential to supply automatic grading not 115) Shnelderman, Ben, Software Psychology, Just of a program's correctness but also of Its Winthrop, 1980. quality could be a major boon to'computer science departments everywhere. 1161 Simon, Herbert, The Sciences of the Artific- ial, MIT Press, 1969. 9,1979, 107-116. REFERENCES 111 Basil!, Victor and Reiter, Robert, Evaluating [171 VanVerth,P.B., andRalston, A., A System Automatable Measures of Software Development, for the Automatic Grading of Programming Workshop on Quantitative Software Models, Style, Proc. NECC, June 1983. IEEE catalog #THOO 67-9,1979, 107-116. 1181 Wirth, N., Algorithms + Data Structures = Programs, Prentice-Hall, 1976.

121 138 TEACHING A SOFTWARE ENGINEERING CLASS USING AN IBM PERSONAL COMPUTER (tm)

RONALD I. FRANK (Visiting Lecturer)

Framingham State College Computer Science Department Framingham, MA 01701 The equipment we use includes a large CDC Twin We outline the successful use of an inexpensive Cyber 172 System located in Boston, 30 miles personal computer to support the programming away. This had been our traditional support project portion of a standard Software machine for the Software Engineering course. Engineering Course. First. as background, we We access it via local terminals running at 300 specify the environment, i.e. our institution and 1200 baud. It was reconfigured to 1200/300 and the equipment we usually use. Second, we from 300 only, and was quite unreliable during outline the course, its purpose, its level, the the semester this course was given. This en- type of students we have, and their background. couraged us to use the IBM PC. We thus provide a reference for the following discussion of the student projects. Third, we We have a DEC PDP-11, a PRIME 400, and a Wang discuss the projects themselves as actually VS-80, locally. The Wang is restricted to accomplished on the IBM PC. Fourth, we discuss administrative use and word processing. We the pros and cons of using such a "small' recently purchased an IBM Personal Computer (tm) machine as a project development vehicle. with 128K (64K on a QUADRAM QUAD (tm) Board) and Fifth, and last, we review the lessons we have an 80 cps matrix printer. We use the color/ learned from this educational experiment. graphics adapter and an Electrohome green phos- phor monitor as our standard display (we would A cost estimate sheet is provided showing a cost now opt for the more expenseive but higher of computing of about $40 per student_per quality IBM monochrome monitor). We have two semester for the complete machine service for single-sided 160KB disk drives.Our hardware in- this Software Engineering course, assuming that cludes an FSC audio-visual department (shared) the entire cost of the PC was borne by this one color TV for color graphics output and an Electro- course. This cost includes an all-points- home green monochrome projection TV for classroom addressable color graphics configuration presentations (projecting the PC display screen supporting extended Pascal with separate onto a regular flat suspended movie screen). The compilation, a full macro assembler, color TV has "video-in" so we drive it without an 128KB of main memory, and an 80 cps matrix printer. "rf" modulator thus getting better quality color I). The Environment (School and Equipment) in "medium" (320 by 200) resolution.All three Framingham State College (FSC), Computer TV's are driven by the color/graphics adapter. Science (CS) Department. Our software includes the PC Macro Assembler (1), FSC is a small (3100 students) four year liberal arts Pascal (2) BASIC (3), and the BASIC compiler (4), college which is part of the Massachusetts State all running under PC-DOS 1.1 (5). Advanced, College System (9 colleges). Massachusetts also graphics, BASIC comes with DOS.All programming "supports a University System. At the time of was done with EDLIN (5) which comes on the$40 this course (Fall 1982), the FSC CS Department DOS disk. DEBUG (5), a single-step debugger and has one full time faculty member (the chair- unassembler, also comes on the DOS disk. person) and 27 part-time adjunct faculty members from industry who teach one or two course sections II). THE SOFTWARE ENGINEERING COURSE per semester. The department supports 206 CS majors and other students as well. As part of the ?our year CS curriculum, we teach a contemporary Software Engineering Course. The The home offices of Digital Equipment Corp., class is comprised of third and fourth year Data General, and Prime Computer are within a students, and is taught by a practitioner with the ten mile radius of FSC. Most of the Route 128 purpose of introducing them to the Software Life complex is less than twenty miles away. Wang is Cycle, the rationale for the various documents about 40 miles away. IBM has a strong marketing required, and the literature of the field. The presence in this area, along with its Cambridge students get some practical advice from "war Engineering and Scientific Narket Support Center stories" and also experience the problems of a and Cambridge Scientific Center. From these team project. They write the documents and per- demographic facts derives the unusually strong form all the functions required in a software and experienced faculty we enjoy.To accomodate development project including generating require- this faculty, many "day" sections are taught at ments, generating time and cost estimates, meeting night, as was this course. deadlines, and giving oral presentjtions.

122 139 Structured modular programming is taught through- reporting system, with data file encryption for out the four year curriculum. This course is a security. It had a menu ariven user interface. review and application of what the students have been taught, but in the context of Structured The other two projects, reported here, were Design. similar to each other. They provide to PC Pascal most of the extensive macro-level color Walk-throughs and reviews occur in class and graphics capability which is inherent in PC during team meetings. Team meetings occur outside BASIC. In both projects, the screen driver is of class. More time is spent in teamwork than in the BAS.COM file from the PC BASIC Compiler class or in class preparation. which provides the color graphics macro library to compiled BASIC. The projects bridge PASCAL The class used Brooks (6), and Pressman (7), as to the same code - they enable a PC Pascal pro- required texts with Myers (8) as a recommended gram to generate color or monochrome graphic reference. A number of class handouts were used images. This is not a capability inherent in including a MIL-STD for software development (9), PC Pascal. the life cycle document outline from a computer manufacturer, and various reprints from the IEEE The BAS.COM file is proprietary to MICROSOFT "Transactions on Software Engineering". A reserve (tm). That implies that absolutely no documen- reading list was maintained at the library. The tation is available to find out how it works. books included many alternate Software Engineering It is a library of machine language graphics texts by authors such as Constantine, Myers, macros with undocumented functions and calling Jensen, and Orr. A number of the IEEE tutorials sequences. It is not our purpose here to were also on the list. document this information - doing so might be a copyright violation. The course met for two hours, twice a week, for one four-month semester. There were eighteen Our purpose is to document the successful use of students, including two professionals and 3-5 the IBM PC in the Software Engineering course. others currently working part-time in programming The two PC-based projects, totaling twelve in the Route 128 area. The students had all people, shared access to our one PC without long previously taken one year of programming including waits. Informal self-scheduling sufficed. Our Pascal and FORTRAN. They had taken a Data Struc- PC is in a controlled, but public, location. It tures course and had experience using files. Many is available M-F, 8:00 AM - 11:00 PM, and on had had Programming Languages (covering the theory Saturday 9:00 AM - 6:00 PM. Each team has avail- of computer languages) which included a large able a set of diskettes. Each member maintains individual programming project. All had completed their own work diskette, but not system diskettes an assembler based Machine Architecture Course. such as DOS, Assembler, or Pascal. The compiler, Various other courses were shared by the students the assembler, all work diskettes and all docu- as a common background, but not a microcomputer mentation are checked out from an operator as if course per se. Our CS students have college they were a restricted-use library reference ,entrance examination SAT averages of about 1075 book. Diskettes are kept in the Computer Center. (math + verbal). This means a student gets his diskette and his team's copy of the system software from an III). THE SOFTWARE ENGINEERING PROJECTS installation operator and returns all of it after the session. The students would have significant The eighteen students were divided into three difficulty making and keeping copies of the fee teams of six each. A team was comprised of a software. It can be done, but it wasn't. leader. a leader back-up person, a recorder, a chief programmer, and two editors. The leader Since most of the word processing was done on the represented the group in class, the recorder kept PRIME, Wang, and Cyber Systems, there was little the group meeting records, and everyone did word use of the PC for report writing. Next semester processing and coding. Also, everyone kept their we will plan an experiment where we use the PC own Programmer's Notebook. All documents were for report writing also. There is more than worked on together, but a single person naa main sufficient time available on the PC to allow for responsibility for each document. There were six this extended use. We plan to use Easywriter 1.1 documents: (11) because it is not expensive but provides sufficient functions for coding and documenting. 1. A Proposal (Software Plan) 2. Requirements Specification No special classes were held on the PC. The 3. Architecture/Functional Specification students had to learn PC-DOS, PC Macro Assembler, 4. Preliminary Design the use of the PC BASIC Compiler, and the Pascal 5. Detailed Design and Test Plan (and Error Compiler. we did hold two tutorial sessions on Reporting Form) 8088 architecture, the assembler, and the overall 6. User's Manual and Documented code structure of the PC Pascal.At times we had to advise some team members on 8088 register use. The three groups, after two weeks of investigation, decided on three projects. One, not discussed here Importantly, the PC Pascal is powerful. It is was a traditional large machine Pascal-based pro- more powerful than the mainframe CYBER Pascal we ject on the CYBER System. It was a general pur- use. As such, it is a significant item to learn. pose grades maintenance, grades computing, and

123 1 40 The PC Macro Assembler is also a mainframe level system. The teams divided up the labor of learn- V). LESSONS LEARNED IN THIS LIMITED EXPERIMENT ing into individual efforts and then held cross- tutorials for each other. Certain advanced indi- Even the smallest schools can now provide their viduals were clearly individually responsible for CS students with all of the hardware and soft- the major breakthroughs in understanding and pro- ware tools for real-world projects.This new 16 blem resolution. Insight in the face of zero bit micro, for example, provides large systems documentation clearly varies among individuals by capacity, capabilities, problems, and opportuni- at least a decimal order of magnitude, as ties at a very low per-student cost. Very in- measured by the time needed to solve a problem. expensive, technically advanced bit mapped graphics is a major innovation. IV). PROS AND CONS OF THE PC USE Physical and media security is not a major pro- There are no significant differences between the blem. Student enthusiasm is a major plus. capabilities of the IRM PC with its languages, and large systems with their languages, except The Programmers' Notebooks that the students that the PC has wider capabilities, such as color maintained became so valuable to their work that graphics, at a much lower per student cost of the students didn't want to hand in the notebooks computation. for instructor review until the coding and documenting were completed! Cost is the main factor. We supported twelve students doing a major software effort with no VI). ACKNOWLEDGEMENT strain on the PC schedule. All this for less than $4700, as a one-time-charge. (See the cost I want to thank Anita Goldner, FSC CS Department sheet at the end of this article.) Chairperson, for supporting this experiment and participating in it.Also, I want to thank Paul The Pascal reference manual needs a tutorial Ferguson, FSC Computing Center Director, for manual. The assembler also needs a tutorial but handling all the administrative details needed to to a lesser degree than Pascal. We can't expect order, set up, and run the open-shop PC. My thanks BAS.COM documentation, but it would be helpful. to the class for their help also. The portable Electrochrome TV unit, driven off of COST SHEET the graphics adapter, enabled us to have a large screen for in-class presentations. The port- For Computer Components Only ability of the PC enabled in-class presentations. The fact that the PC color graphics could use an $4700 Purchase (Including Software) minus cost of ordinary A-V color TV set meant that we could capital and maintenance which comes to: afford to do a color graphics project. $1175/year for 4 years The completely open (i.e. fully documented) $118/month for a ten month year nature of the IBM PC system means that we can $.33/hour for 84 hours a week availability assign such "real-world" system design problems and get them done within the confines of a single Twelve students, not doing word processing on the semester. the Technical Reference Manual (10) PC itself, had little conflict using one PC.With for the PC explains the hardware in complete de- word processing there would be a need to schedule tail. The fact that native BASICA and the BASIC usage, but the system (1 PC)would still not be Compiler are essentially identical means that we saturated. We estimate that fifteen students. could run color graphics tests interpretively, each needing five hours per week on the system and then compile to find links to BAS.COM in (75 hours out of 84 available) would saturate the order to isolate its functional routines. system. If the Center hours were extended to include late night and Sunday use, we estimate that Development would have been faster if we had twenty students could be supported on one machine bought the double-sided (320KB) disk drives. That for this course. would have reduced the "Floppy Flipping" necessary to do a Pascal compilation or an assembly. It At fifteen students per machine, the total cost is would be ideal if our students had these machines $7.80/student/month. This includes all system at home. That would make it easier and more pro- components, but not room rent or security super ductive for them. One student did have an IBM PC vision. By opening the center for more hours, at home_ However, by chance, he was on the enabling 20 students to use the PC, the cost drops CYBER project! to about $5.90/student/month. This comes to $40 per student or $30 per student,.respectively,for The reliability of the hardware was perfect. The a semester (1/2 year).This estimate does not in- software was also reliable, but we suspect certain clude media, supplies, or a formal maintenance difficulties came from either poor documentation contract. Media and supplies come to much less in the Pascal manual or from true bugs. We than $50/team/semester. The 4 year life time couldn't spend the time to isolate the problems. assumed for the PC is reasonable due to techno- The students found fixes or bypasses that worked. logical progress calling for its replacement. None of the problems encountered left "time bombs" The machine will last and be useful much longer in the codes. The codes came out "clean". than that. A maintenance contract costs approximately 10% of system cost/year.

124 I 1. 141 BIBLIOGRAPHY

(1) IBM Personal Computer Macro Assembler. #6024002.

(2) IBM Personal Computer Pascal Compiler. #6024010.

(3) IBM Personal Computer BASIC. #6025010.

(4) IBM Personal Computer BASIC Compiler. #36024003.

(5) IBM Personal Computer DOS. #6024001.

(6) Brooks, F., The Mythical Man Month, Addison-Wesley, 1975.

(7) Pressman, R. S., Software Engineering, McGraw-Hill, 1982.

(8) Myers, G., Software Reliability, John Wiley, 1976.

(9) MIL-STD 1679 (Navy), Military Standard Weapon System Development, Chief of Naval Material, NAVMAT 09Y Wash., DC 20360, 1 December 1978.

(10) IBM Personal Computer Technical Reference Manual. #6025005.

(11) IBM Personal Computer Easywriter (1.1). #6024005.

125 142 CRISIS IN PROGRAMMING, OR HISTORY DOES REPEAT ITSELF

Jacques LaFrance

Department of Mathematical Science Oral Roberts University, Tulsa, Oklahoma.

Abstract between two or more alteratives; and an iteration is the repetition of something until some The increasing cost, unreliability, complexity and terminating condition is met. unmaintainability ofprogramming effortsof the 1960s gave riseto the discipline of structured More complex logic is created by including these programming. We face a similiar crisis in structures within each other. For example, microcomputer software today because of a new consider an algorithm in which the mainlogic generation unfamiliar with the past. New structure is a sequence of three things, the first approaches to introducing programming are needed to of which is a selection and the last an iteration. solve the problem. A programming language called The selection chooses between two sequences, each Antfarm has been developed to introduce structured of whichcontainsan iteration, which in turn programming logic in such a simple and entertaining containsa sequence, one element, of which is a format that even young children can begin to learn selectioa, etc. At eachlevel, the structure the currently best accepted principles of program conta:n,- only these' three basic forms. Each of design. The program has been used with kids from these contains one entry and one exit,the logic age 5 to college students and adults who are either cannot wander off in some hard to follow path. The new to programming or have only worked with BASIC. discipline of limiting one's logic design to these simple structures is called "structured programming," and the programs whichresult are Historical Background called "structured programs." Other concepts that have developed in conjunction with structured The Software Crisis of the 1960s. programming are "top-down design", "structured design", "Chief Programmer Team", "structured As computersystems grew from the early days of walk-throughs", and "proving program correctness," computing, a crisis was reached sometime in the mid some of which are often included in the meaning of 1960s when the cost of software became noticably structured programming. For a more complete larger than the cost of hardware and the logical discussion of these and examples of how they have complexity of software systems became unmanagable. improved software development, see N and In fact some systemr became so complex and Kelly, Top-Down Structured Programming. interconnected that someone once calculated that any changes introduced to correct errors had such a These new programming disciplines have become large probability of introducing other errors that generally widely accepted in industry and higher it was very unlikely that the system would ever be education. The Pascal programminglanguage was correct. developed by Niklaus Wirth around 1970 to give a tool forbetterteaching of modern programming Leaders in the industry began desparately seeking a concepts than was possible' with existing solution to this crisis, a solution that would give programming languags. In particular, it was the field of software design a precision similar to designed to be able to teach structured programming what was being achieved in the hardware area. What concepts in a natural way. was needed was a methodology for software development which would allow it to be accomplished History Repeats Itself efficiently and would enable the logic of complex systems to remain manageable. The New Software Crisis.

A Solution to the Software Crisis. Now weare entering the software crisis of the 1960s afresh. The problem we now face is that the A paper by Bohm and Jacopini in the Communications widespread multiplication of personal computers in of the A.C.M. in1966L provided afoundation the last 5 years, all of which come equipped with for a new discipline of programming. They proved some version of BASIC asthe default (or in some that any logicalsystem can be reduced to a cases,the only) programming language, is leading combination ofthree basic logic forms. These to a new generation of software developers forms are called sequence, :election, and 'unfamiliarwith the abovehistory. BASIC was iteration. A sequence isone operationafter developed in the early 1960s prior to the advent of another; a selection isthe making ofa choice modern structured programming principles, and

126 143 therefore these principles had no influence on its graphics. The processor has been written in UCSD design. Although one can do structured programming Pascal and therefore will run on the widest variety with any language, BASIC does not naturally of processors and machines. A Forth version is encourage structured programming and naturally also being developed for use on smaller machines. leads the notice programmer away from structured The screen isa farm on which an "ant" lives and programming principles. Its design structure raises food. Everything the ant does uses energy, encourages' obtuse linearlow-level coding with and therefore it must eat some of its crops to stay arbitrary branching. alive. It can plant interestingly shaped fields, dance, march, explore, find its way around, and so Since the distribution of personal computers has has proven to be a highly interesting environment exceeded the computer professicaal community's in which the children can work. Whereever we have ability to provide adequate training, many of the used it, the children have been very intrigued and new users are being self taught or are being taught highly motivated. Goals 4 and 5 hayed been by others who are largely self-taught. Thisis achieved successfully. The following section shows leading to a generation of new programmers who are how the first three goals were achieved. not aware ofthe progress the field has made in software design principles in the last twenty years. As George Santayana said once, "Those who The Antfarm Language cannot remember the past are condemned torepeat it." The Antfarm Setting.

A Solution to the New Crisis. With Antfarm, the screen is a farm on which an ant colony lives and raises food. The antsconsume One solution to this crisis is to develop tools for energy and must be fed to stay alive. The ants and teaching structured programming concepts on the other symbols are all made of typewriter microcomputersto all age levels, especially the characters and do not restrict the language to use younger children who will soon become thenext on special graphic hardware. All movement isin generation of programmers. The concepts of terms of the rows and columns of the text screen. structured programming or structured programs are The ants can dance, march, explore, have lunch, and notdifficult and could easily be understood by draw rudimentary pictures by planting seeds. One children if given the right tools. The basic logic student even programmed a hurdle race with three structures of sequence, selection,and iteration ants, and in another program an ant finds its way are understood by children before they reach through a maze. elementary school. Consider the following sentences: "Put your toys away, put on your The ant looks like this, depending on orientation, pajamas, and then we will read a story," a where the "*" is its head: sequence; "If the weather is nice on Saturday, we will have a picnic," a selection; and, "Write your \*/ *1 \ / spelling words 10 times each," an iteration. -x *X0 - 0 0- / \ Over the past three years we have been working on /1\ such tools.6 Aspecial languagefor usein teaching structured programming concepts to Figure 1 children has been developed. This language, called Antfarm, has been designed according to the Thebody characters represent the ant's weight, following criteria, which, if met, could make it a energy supply, and change with each command. "XO" very useful tool to introduce structured 4.s the maximum, andit represents 400 units of programming concepts: energy, or enough to do 400 commands. The seeds grow into plants as follows: 1. The language wouldonly include the basic structures of sequence, selection, and iteration, seed i.e. no unconditional transfer ("GOTO"), but would germinated seed include a complete set of these basic structures. starting to come up 2. The language would be as close as possible to young sprout the English with which the children were already Y tall plant with branches familiar, and the vocabulary would consist of short P flower simple words learned early in school. full grown mature food 3. The language would focus on program logic structures, which are deemed harder to understand, Figure 2 by eliminating any consideration of data structures. The Antfarm Commands. 4. The situation of the language would be fun and imaginative, providing high motivation to work with Basic Commands. The basic commands in Antfarm are the it. MOVE, one row or column forward, BACKUP, one row or 5. The language processor would be implemented in column backward, TURN LEFT, rotate 45 degrees to as transportable a way as possible. theleft, TURN RIGHT,rotate 45 degrees to the right, PLANT, put a seed down where the head is, The Antfarm languageuses the text screenand EAT, consume whatever is under the head,and WAIT therefore is not limited to hardware with special or NOTHING, which has the ant do nothing but use up

127 .t 144 one energy unit and spend one time unit. commands. The format ofthis command is "LEARN command-name definition". An example is: These commands or any others can be listedone after another making a sequence, such as MOVE PLANT LEARN REVERSE DO TURN RIGHT 4 TIMES MOVE TURN LEFT. The command the ant is to learn is named REVERSE. Program Logic Control: Iteration. Two forms of After the ant is given the command REVERSE, it will repetition are allowed,the common counting loop perform DO TURN RIGHT 4 TIMES. and the conditional loop. Both begin with the word DO or the word REPEAT and end with the condition Top-down Design in Antfarm. Since words do not for loop termination. In -he case of the counting have tobe definedbefore they are usedin a loop, this is a natural followed by the word definition, the ant can be taught a complex program TIMES. In the case of the conditional loop,this by themethodology of top-down design. The is either the word TO or the word UNTIL followed by following isan example ofteaching the ant to the condition. The ant has a rich set of plant a gardan, which has the shape of the outline conditions to chose from, which contributes to the of a square, by the principle of top-down design: captivationof the languageand itspower in illustrating program logic structures. These LEARN SQUARE DO SIDE 4 TIMES conditions include the ability to check position, LEARN SIDE ROW -OF- PLANTS CORNER e.g. ROW number, PAST COLUMN number, NOT PAST ROW LEARN ROW-OF-PLANTS DO PLANT MOVE 5 TIMES number, orientation, e.g. FACING N, NOT FACING SE, LEARN CORNER TURN RIGHT TURN RIGHT appetite,e.g. STUFFED, NOT HUNGRY,smell,e.g. SMELL FOOD, NOT SMELL DIRT,and sight,e.g.SEE After teaching the ant all of these commands, the PLANTLEFT, NOTSEE JUNK AHEAD. In addition ant may be instructed to to make a square simply by conditions may be combined with AND or OR. typing the command SQUARE.

Program Logic Control: Selection Antfarm Examples.

Selection is implemented with the same word most The next program is an example in which the ant common in English and other programming languages, makes itself oriented toward the north, pointed IF. Thereare twoforms, "IF condition THEN toward the top of the screen: commands END-IF", and "IF condition THEN commands IF-NOT commands END-IF". Inthelatter, ELSE, LEARN FACE-NORTH IF NOT FACING N THEN IFNOT, and OTHERWISE are accepted as synonyms for TURN-NORTH ; IF-NOT. In either form a comma is a synonym for LEARN TURN-NORTH IF FACING-EASTERLY THEN THEN and a semicolon or REGARDLESS is a synonym for TURN-TO-LEFT IF-NOT TURN-TO-RIGHT ; END-IF. The condition is exactly the same as the LEARN TURN-TO-LEFT DO TURN LEFT UNTIL condition allowed for loop termination. Two FACING N examples of selection commands are LEARN TURN-TO-RIGHT DO TURN RIGHT UNTIL FACING N IF FACING N ,TURN LEFT ELSE TURN RIGHT; MOVE MOVE LEARN FACING-EASTERLY FACING NE OR IF HUNGRY AND SMELL FOOD THEN EAT REGARDLESS MOVE FACING E OR FACING SE

Program Logic Control: Loop Exit. Another flow of In this example, both the "IF conditionTHEN controlstructure is STOP,which terminates the command" and the "If condition THEN command IF-NOT loop if it is used in one. An example of use is command" formswere used. Also this command illustrates that named conditions can be learned,

DO MOVE IF NOT SMELL FOOD, STOP ; EAT 10 TIMES FACING-EASTERLY inthe example. This allows a powerful hierarchy of logical conditions for which is a counting loop with the possibility of program structure control and provides a basis for premature exit due to running out of food to eat introducing the concept of a function in the before the count is exhausted. student's subsequent education.

Subprogram Definition. Another major consideration The following is another example in which the ant in modern program development is that of building finds its way through a maze using the algorithm of the total program out of relatively small modules. always staying against the left wall. The maze is The main program is simply a set of logic made of plants just two columns apart so the ant structures to put the major modules togetherin can use its rather near-sighted vision to see the their logical order. Each of these modules is also walls. There is food at the end to signal just the set of logic structures needed to combine completion of the maze. the modules out of which it is made, and so forth, until the structuresbeingcombined aresimply fundamental commands in the language.

Antfarm forces the development of modular programs by restricting each module to no morethan80 characters, one line on a crt or two lines on an Apple II. Thefacility provided for building program modules is that of teaching the ant new

128145 LEARN FOLLOW-LEFT-WALL DO GOLEFT UNTIL OUT-OF-MAZE The students all responded very enthusiastically to LEARN GOLEFT LOOK-FOR-LEFT-GAP this project and began tograsp the concepts of MOVE MOVE AGAINST-WALL? program structure, even though the computer time LEARN LOOK-FOR-LEFT-GAP IF NOT was severely limited. The best work was done by a WALL-ON-LEFT THEN TURN-TO-LEFT ; student who created a program of about 19 modules LEARN TURN-TO-LEFT DO MOVE MOVE TURN which made the ant draw (plant) theshapeof a LEFT TURN LEFT UNTIL WALL-ON-LEFT Christmas tree and then wait for it to grow up. He LEARN AGAINST-WALL? IF WALL, DO TURN only needed help on the latter :art and two simple

RIGHT TURN RIGHT UNTIL NOT WALL ; errors. It clearly showedtop-downstructure. Also one student compared noteswith hisolder This example shows powerful uses of selection and brother who was taking a high school BASIC class. iteration in expressing program logic. Notice the They agreed that the younger one had learned more form "IF X THEN DO SOMETHING UNTIL NOT X" in the about programming with Antfarm than the older one definition of AGAINST-WALL? This is equivalent to had with BASIC. the form, "WHILE NOT X DO SOMETHING", which illustrates the ability tobuild different logic Use of Antfarm in Summer Camps. forms froma fewsimpleones. Theformer is probably more common in natural language, as in, A Two-week Camp at ORU. During the summer of 1982, "When the bell rings, if you haven't finished, keep the Oral Roberts University School of Education ran working until you are finished," and, "After two two-week camps for kids from 5 to 17 years of school, if it is raining, wait there until it stops age using their new microcomputer laboratory. Each raining." This example also shows the benefit of student had 15 hours in the camp, half of which was mnemonicnamesbeing used,especiallythe names spent on the computer and the other half receiving used for the learned conditions, OUT-OF-MAZE, instruction. All students began withAntfarm; WALL-ON-LEFT, and WALL. Theseare definedas however,for the second week, the older students follows: switched to a Pascal system called Computer Power, produced at the University of Tennessee under the LEARN OUT-OF-MAZE SEE FOOD AHEAD direction of Prof. Mike Moshell in order to provide LEARN WALL-ON-LEFT SEE PLANT LEFT materials for teaching Pascal in high schools. LEARN WALL SMELL PLANT The two five-year-olds that attended learned This illustratesthatnot onlycan theantbe something about the preciseness of computers. They taught command sequences as modules,butit can were able to give the ant a sequence of commands to also be taught conditions as named modules as wel). make it do what they wanted and quickly saw the This in a sense gives both procedures and consequence ofwrong commands when the antdid functions, although functions are limited to type something different from what they hadplanned. Boolean. However, they were not able to get to the point of being able to compose learned programs. This level Antfarm for teaching structured programming. ofabstraction requiresa more mature mind. We found that at about eight years of age the children These few examplesillustrate the potential of seem to have developed' enough of an idea of Antfarm to teach the basic concepts of structured abstraction to be able to define simple programs programming. As we have used Antfarm, we have for the ant. found it to be very effective in teaching concepts of structured programming, not only to children but All students, no matter how old, definitely learned also to college students and adults. Other some goodfoundationalprincipleswithAntfarm. professionals familiar with structured programming The older ones seemed to be able to carry the ideas have also been impressed with its potential when from Antfarm over to their work with Pascal quite they have had the opportunity to observe Antfarm. easily, although, we believe more time spent with Antfarm would have benefitted them. The results of this experienceconfirmed the previous results: Experience with Antfarm they enjoyed it very much and were abletoput together logical program sequences of varying An Elementary School Antfarm Class. complexity.

In the fall of 1981, Antfarm was used one day a A Summer Camp Conducted at OU. The University of week for eight weeks at Grissom Elementary School Oklahoma College of Education alsoconducted a in Tulsa, Oklahoma.There were eight classes, each computer campin the summer of1982,under the limited to 10 students on a first come first served direction of Evelyn Gatewood.The students in this basis as part of an enrichment program available to camp attended 3 hours each day for twoweeks. all students. The classes were about 35 minutes During this time they were given experience with long. One computer was available and that was one BASIC, Antfarm, and LOGO. In the Antfarm segment, the author brought to the school for the day and the students, especially the younger ones, seemed took back home after Cie last class. Because of to grasp the concepts more quickly than with either the severely limited computer time available, the of the other two languages. The vocabulary and instructor did all thetyping and went over the language structure wassimpler and moreeasily students' programs verbally when there were serious understood. It was noted alsothat t'he children shortcomings the students needed to correct. were more fascinated by having an "animal" to take

129

146 care of than by just making:pictures. Some of the the key to enable it to contribute to the solution children becameveryprotective of their ant, of the new software crisis. making sure it didn!t starve, sometimes by overfeeding it,-:f1 perhaps7they:mre_apt to do with Conclusions imnsehoId pets well. Th.-,,Ite:n.7w---,sonal interest their ant =0:_the ct-LLOrm=ora-motivated and It is possible to teach structured programming to seemed , better=-oommr-rarton between the children in a way that they enjoy and which lays a cLild zr , .-muter. good foundation for future programming. We have seen Antfarm do this. Antfarm does indeed achieve cults _ College S Ildents. the goals set out for its design. It is fun; it allows only structured programs; itis based on We have also used Antfarm for a period of about one simple English words; and it is simple enough for week in our Introduction to Computing class at Oral young children. Itis our hope that Antfarm, or as RobertsUniversity. Wehave found that this things like it,can achieve widespread use experience helpsthe new students understand the introductions to programming andforestall the concepts of program structurebetter prior to present crisis in programming by making it possible studying it in Pascal. It also seems to serve as a for potential future programmers to begin their good tool to wean students who have programmed in computer careers learning to make modular BASIC prior to entering the course away from the structured programs. unmodular, low-level, arbitrary branching style of programmingto which they had become accustomed. References In both cases the students were more quickly able to begin to develop modular structured approaches 1. Bohm, C., and Jacopini,G., "Flow diagrams, to program design than without the brief experience Turing machines and languages with only two with Antfarm. formation rules," Communications of the ACM, vol. 9, no. 5 (MAY 1966), pp. 366-371. Other Languages for Teaching Structured Programming. 2. Cohen, Harvey A., "Oznaki and Beyond," There are several other languages that can be used Proceedings of the National Educational Computing to teach structured programming, including Pascal, Conference, Iowa City, IA, 1979, pp 23-24. Karel the Robot, and LOGO. Pascal was designed by Niklaus Wirth5 specifically to teach 3. Dahl, O. J., E. W. Dijkstra, and C. A. R. Hoare, programming. It remains the primary tool for this Structured Programming, Academic Press, New York, in higher education. ComputerPower, developed by 1972. 1011 Mike Moshell, Robert Aiken, ' and a team of others at the University of Tennessee, makes Pascal 4. Dijkstra, Edsgar, "GoTo.Statement Considered suitable for teachingat the high school and Harmful,"Communications ofthe ACM,Volume 11, possibly junior high school level. Number 3 (March 1968), pp. 52-54.

Karel the Robot was developed by Richard 5. Jensen, Kathleen and Niklaus Wirth, Pascal User 14 Pattis to introduce structured programming to Manual and Repo, Springer-Verlag, New York, 1973, his university classes in very much the same way as 1976. wehave used Antfarm. Like Antfarm, Karel the Robot leaves out the concept of data structures in 6. LaFrance, Jacques, "Shall We Teach Structured order to focus attention on the logic structures, Programming to Children?", Proceedings of the and the language consists of commands to a robot to NationalEducationalComputing Conference 1980, move around it's environment. It implements only Norfolk, VA, 1980, pp. 261-265. structured programming constructs, and gives a graphic situation in which the student can see the 7. LaFrance, Jacques, "ReorientingStudents to effect of his commands on the the Robot. Structured Programming with Antfarm", Proceedings of the 1982 Western Educational Computing 1213 LOGO was developed by Seymour Papert ' as a Conference, San Diego, CA, 1982, pp. 35-42. toolto use with children to study their patterns of problem-solving. Although it was not designed 8. Lieberman, Henry, "The TV Turtle: A LOGO for teaching programming,it does contain much of Graphics System for Raster Displays," MIT, A. I. the structures needed to teach structured Memo 361, June 1976. programming. 9. McGowan, Clement L., and John R. Kelly, Top-Down In comparing Antfarm with these, thefollowing Structured Programming Techniques, points can be made. Antfarm is simpler than any of Petrocelli/ Charter, New York, 1975. these, yet gives more powerful program logic structures. Also, it is more English-like and 10. Moshell, J. M., G. W. Amann, and W. E. Baird, gives the student a more imaginative environment "Structured Gaming: Play and Work in High School in which to work. The simplicity and ComputerScience," Proceedings of the National immaginativeness makes it more suitable with Educational Computing Conference 1980, Norfolk, VA, younger children than any of the others, and the 1980, pp. 266-270. powerfulness ofit's logic structures makeit a challenge to older minds. The fact that it forces the learner to make modular, structured programs is

130 141

1 11. Moshell, Michael, Proj. Dir., Computer Power: A First Course in Using the Computer, Gregg Division, McGraw Hill, 7ew York, 1982.

12. Papert, Seymour, "Teaching Children Thinking," MIT, A. I. Memo 247, October 1971.

13. Papert, Seymour, Mindstorms: Children, Computers, and Powerful Ideaa, Basic Books, Inc., New York, 1980.

14. Pattis, Richard E., Karel the Robot: A Gentle Introduction to the Art of Programming, John Wiley & Sons, New York, 1981.

15. Santayana, George, The Life of Reason, pg 284.

16. Solomon, Cynthia J. and Seymour Papert, "A Case Study ofa Young Child Doing Turtle Graphics in LOGO," MIT, A. I. Memo 375, July 1976.

17. Tomek, Ivan, "Josef, Programming for Everybody,", ACM SIGCSE Bulletin, Volume 14, Number 1 (February 1982),The Papers ofthe Thirteenth SIGCSETechnicalSymposium on Computer Science Education,Indianapolis, IN, February,1982, pp. 188-192.

18. Watt, Daniel H., "A Comparison of the Problem-Solving Styles of Two Students Learning LOGO: A Computer Language for Children," Proceedings of the National Educational CoLputing Conference, Iowa City, IA, 1979, pp. 255-260.

131 148 AN EVALUATION OF A LOGO TRAINING PROGRAM

M. Elizabeth Badger

Massachusetts Department of Education

Logo into the ordinary class? This paper describes the evaluation of a computer-based school program which util- This paper represents an attempt to answer this ized the turtle geometry and sprite fun- question, it describes a program that was ctions of the Logo language. The study carried out in two Cambridge, Mass. schools and used both observational and quantitative was conducted by a group of individuals who were methods to measure the effect of the not experts in the Logo language. The evaluation program on: 1) pupils' familiarity with itself is primarily illuminative, with some basic geometric concepts; 2) pupils' reliance upon test results to confirm its more understanding of Logo commands; and subjective findings. 3) the relationship between the two. Attitudinal and organizational factors The Program were also considered. Although generally positive results were noted, the The program took place in the winter of 1982, author expresses some reservation about when the Cambridge School Volunteers proposed to the cognitive benefits of unstructured teach a 5-week course in Logo to the 6th grade activities. She suggests that the classes of two Cambridge schools. The majority strong visual appeal of the Logo proof the volunteers were Harvard students who were grams may obscure their potential use experienced in programming languages; however, as "problem-solving environments". few had had previous experience with Logo. In preparation for teaching, they were given a short training course in Logo, as well as vartuus Logo materials that had been produced by the MIT- sponsored project in the Brookline Public Schools In Mindstorms, Seymour Papert 1. argues for anew They were also asked to keep a journal of class conception of education and mental growth. Given progress. This was to be used, not only as a the opportunity to use computers and, more parti- vehicle for suggestions and cautions, but to cularly, to use the MIT-developed system called insure a continuity during the program. logo, children could develop an intuitive sense of geometry, physics and the procedural thinking The Schools that underlies problem solving. Not only would the graphical nature of the system appeal to School A (Classes Al and A2). children's imagination, but it would serve as visual reinforcement and feedback to their mental The two schools differed in population, c4rriculum explorations. and environment. In large part, the pupils in School A were newly arrived in the United States Among educators, there is general agreement that and Cambridge. All but 7 of the 32 pupils in the Logo has features that make it especially appro- 6th grade were receiving some kind of extra help priate as a conceptual tool. In comparison with duq to language or learning problems. Because of other readily available languages, it is inter- this, teachers tended to stress the basics.. Most active and well-structured, allowing the user to of the mathematical teaching concentrated on define and manipulate sub-procedures. It contains computation, with little reference to more general good graphics, with very simple body-centered and mathematical "concepts". The children stayed in Cartesian coordinates; and it is designed to be their own self-contained classrooms and were conceptually easy. 2 On the other hand, there is taught most subjects by their classroom teacher. little specific evidence for its effectiveness in As a result, except for the occasional deliverer the class. The evidence that has been offered of a message, there was little movement in the has been the product of a special environment, corridors. At School A, the 8 Apple microcompu- i.e., programs sponsored by the Artificial Intell- ters were arranged in a rectangular array in one igence laboratories of MIT and Edinburgh. There corner of a large library. Each class was brought exists little guidance for the teacher or admin- to the library by its teacher and, for the most istrator who asks the question:What intellectual part, each teacher stayed to help supervise the benefits can I expect to derive from introducing work done.

132 149 School B (Class 81) In addition to these differences in mathematical experience, there was a definite progression in School B is surrounded by Harvard and Lesley their standardized mathematical achievement scores. College. Part of its student body came from the When grade equivalent scores were ranked (a total immediate neighborhood and tended to be the of 45 were available), the pupils from School B children of professionals; part was brought by bus had, on the average, higher scores than those from from other areas of the city. Each teacher among School A. Within School A, the two classes also the older children is considered a subject special- differed. The average of the ranks of the three ist, has covered a wide range of topics with his classes are listed below (1indicates the highest pupils and was eager to house the 6 Texas instru- score achieved; 45 the lowest): ments microcomputers at the back of the classroom. The 24 6th grade pupils in the class were divided Class B 16 into two groups, and the afternoon of each day was Class Al 31 devoted to Logo. During this time, their teacher Class A2 23 met his assigned mathematics classes in another How the Program Went room in the building. The 6th grade group that was "off the computer" would either go elsewhere The program at School A took place every morning for instruction or would remain in the room with from 11to 12:30. Each class was allocated 45 pre-assigned seat work. minutes, with 2 pupils per computer. There were usually at least 2 tutors present, as well as the The Microcomputers class teacher.

The two types of microcomputers also differed in Using the model of computer ds "tutee", the pupils ways which had consequences for the teaching in were encouraged to explore the computer capabili- the two schools. The Logo language is built into ties. They seemed to learn the basic commands very the TI microcomputers. As a results, pupils at quickly and on the first day the tutors' journal School B had access, not only to turtle geometry noted, "High optimism.A few know Back. All which allowed them to draw figures and patterns know Left, Right, Forward."Within a few days, on the screen, but to the sprite program which however, some of the tutors seemed perplexed at permits the programming of motion and speed. On the amount of raw energy that the computers seem the other hand, they could not easily save their to have released. The pupils themselves were programs. The tape device proved to be a cumber- delighted in their new-found power when they some procedure and was barely used. Consequently, found that the computer would respond immediately, the procedures for individual programs were and in marvelously unexpected ways. Type in some usually typed anew at the beginning of each numbers and commands, and the turtle would zip session. around the screen, building up a constantly changing pattern that could be changed even more In contrast, Logo is not a feature of the Apples, by varying the color. but must be read in through the use of a disk. This limited the scope of activities for the The tutors on the other hand, worried about the children in School A because the disk contained extent of their understanding. It was the only turtle geometry. However, the use of the holistic impression that seemed to capture the disk drive had certain advantages. It allowed pupils' attention, not the component parts. children to save their own programs on personal Sensing their lack of focus, one of the class- disks and allowed for later printout. This room teachers mimeographed sheets listing the feature was used extensively by all the pupils basic primitives and some nested repeats. The at School A. pupils copied these into their notebooks and on to the computer, but tutors continued to wonder The Pre-Test if the REPEAT instruction was understood. The children loved large numbers and used them for In order to assess the pupil's familiarity with everything---REPEAT 765, RT 675, FD 786,etc. various mathematical concepts, a test was admin- It was unclear whether or not they had any con- istered at the start of the program. It was com- ception of the role of each command in this posed of a series of questions concerning area, grand design. Their encitement and energy set angles, coordinates and permutations. Standardized up so much static that it was hard to see that mathematics achievement test results were also they were actually progressing intellectually. collected. In an attempt to focus their attention on the From the pretest responses, it was evident that procedures involved, cards giving instructions the pupils in the two schools differed greatly in for making designs were introduced in the third their mathematical knowledge. For example, al- week. These seemed to work fairly well. The most none of the pupils in School A knew what an children liked them and enjoyed constructing the angle was. When asked to estimate size, many small designs. Perhaps as a result of the hased their Judgements on the length of the arms. design cards, the tutors noted that about 50% Most left these questions blank or wrote "don't started using procedures and almost all had understand". In contrast, most (91%) from School settled down. Also, during this period each B correctly drew 900 and 450 angles, and over pupil was given a disk on which to record in- half were able to recognize a 900 angle that was drawn at an unfamiliar tilt. dividual programs. These provided a focus for

133 151 activity because designs could now be saved. They of turtle commands to produce a given series of enjoyed this and began to share programs with each figures; 21 draw the figures that would be pro- other. Many of their programs were copies of or duced by a given procedure: 3) draw and estimate modeled after the design cards, but some children the size of a series of (same as pretest); used the design cards as a take-off point for their 4) list all possible permutations of 4-colors own explorations. (similar to pretest).

However, on the Monday of the fifth and last week, Part 1: Given a Figure, Write a Procedure the tutors noted that some were getting bored. "Several didn't even want :heir disks today. There were 4 variations of this type of question. Others seem to need new ideas, to be encouraged Pupils were askeu to write procedures that would Someone in the second group to explore further." produce a set of irregular stairs, a triangle, a This found out how to make the computer "beep". rectangle and an angular figure. Answers were spread and led to intermittent beepings through- judged in terms of the correctness of 1) direction out the sessions. The tutors' comments grew of angle; 2) size of angle; 3) length of line. plaintive: "Lots seemed bored. We need some- Results are as follows: thing new. Perhaps more guided instruction at this point? A lot of use of background color Percentage Correct change. Some kids still don't have a grasp on the basic turtle commands. Telling left from direction size length right is a problem for at least one of the girls." Class B 79 79 79 School B Class Al 31 31 Class A2 47 67 8o The program at School B is less easy to describe. Possibly because the program itself was Class B 75 08 71 less structured, it was harder to see where it Class Al 56 06 50 was going. Membership in the two groups working Class A2 73 07 67 on the computer was not fixed but changed according to outside scheduling. This caused con- Class B 8o 75 79 fusion at the beginning of the sessions, and Class Al 56 44 25 the tutors noted that the classes tended to Class A2 73 73 27 be "unruly". At any time during the computer period there were pupils coming in and out, Class B 17 54 some working at their desks on other projects, Class Al 06 44 In fact, some with no discernible work to do. Class A2 06 25 at times some children did not work on the computer at all during the scheduled period. Two interesting findings emerge from these re- However, this did not visibly bother them, sults.' One is the extent to which Class A2 possibly because there was the opportunity to use resembles Class B in its percentage of correct the computer at leisure and in relative privacy responses. With little previous experience in after the rest of the school had left for the muring and rotation, a large percentage of In fact, a kind of computer club grew up day. Class A2 was able to give the correct values for for projects after school and some of the boys the procedures. In contrast, the pupils in used their knowledge to teach Logo to the 7th Class Al showed a lack of understanding of many and 8th grades. Possibly because of this, set of the computer commands, despite the fact that time was less important to the children. they had used these figures often in their designs. Their responses to the first question, in parti- The Post-Test Results cular, showed that a large number of children had no clear sense of turtle commands (e.g. FD 3, As the program developed, an insistent question RT 5, LT 180, FD 4 ...). Secondly, the large That was: how much did the began to emerge. percentage of errors in regard to size of angle children understand whet they were doing? At on items 2 and 4 represents a problem inherent School A, motivated by the promise of printed in the turtle language. Although each internal results, the children produced a multitude of angle within an equilateral triangle is 60°, the designs. turtle must be programmed to go around the exterior of the triangle. Its turn, then,is not 60° At School B, without a means of recording their but 180° minus 60°. Few children realized this work, they seemed to have been captivated by and nerhaps would have come to an intuitive. the activity of sprites. At the end of 5 weeks rec ton only by actually walking around an of daily work on the computer, did the pupils equ. aral triangle. have a sense of computer programming in turtle language, or had they become mesmerized by the Part 2: Given a Procedure, Draw a Figure visual potential of the computer? This group of items was desigrld to measure These considerations led to the construction of pupils' ability to differentiate between a post-test which was composed of four parts. different commands. If reflected the evaluator's In it pupils were asked to: 1) write a series concern that children were not aware of the

134

15-j 1 . effects of the component parts of a procedure. taught to recognize and draw angles of various The three procedures given were: sizes, while the children in School A had no experience with angles before the training pro- a) REPEAT 4 FD 40 RT 90 FD 10 LT 90 gram. During the program, the pupils in Class A2 b) REPEAT 4 FD 40 RT 90 FD 40 LT 90 had received some instruction from their teacher, c) REPEAT 4 FD 40 RT 45 FD 40 LT 45 and this appeared to affect their performance. On the post-test, 19% of Class Al and 33% of Class As in the case of the previous set, responses were A2 could draw a 90°angle. The same percentage in judged in terms of the correctness of direction Class A2 could recognize a right angle in an un- of angle; size of angle, and length of line, with familiar tilted position, although only one the following results: pupil in the other class could do so. Not surprisingly, correct responses to the angles Percentage Correct questions were closely correlated to the successful description of procedures. Of the five direction size length highest scores on the procedures questions, four were produced by pupils who recognized and drew Class B 71 88 92 90° and 456 angles. Class Al 25 69 25 Class A2 33 87 67 Part 4: Permutations.

Class B 50 79 79 A permutation question had been given on the pre- Class Al 13 63 19 test, so it was possible to compare results. Class A2 33 80 53 Success rate rose among the pupils in School A, with 6 giving the complete set of permuations Class B 42 42 46 (in contrast to 1 on the pre-test).A corres- Class Al 06 13 31 ponding increase appeared at School B (from 2 Class A2 33 27 40 to 11). All pupils who gave the entire set of permutations showed evidence of an orderly pro- Most pupils recognized RT 90 as the command for a gression through the range of possible combina- right angle; far fewer were successful with a 45' tions. angle. Most of the attempts had no relation to the instructions given and seemed to indicate that Teachers' Appraisals the pupils had little idea of how to incorporate change in direction with an angle other than 90°. As the programs differed in the two schools, so In fact, there were very few "near misses" on the did teachers' attitude toward it. The teacher last qUestion. It seems that children either of Class Al thought that some children, par- understood the gestalt or did not. ticularly those who were receiving learning disability tutoring, got some positive feelings When correctness for angle size, direction and and a sense of accomplishment. However, she length were totaled throughout the questions as a saw little carryover to their classroom work. whole, the following average scores were ob- Pupils saw it as an "isolated program."With tained: few exceptions, their attitude and achievement in the program were predictable on the carryover direction size length of positive attitudes. Noting that some of the (1-6) (1-7) (1-7) children who were not scholastically tops did very well on the computer, she saw this new con- Class A1(n=18) 1.9 2.4 2.3 fidence carried over into the class. She cited Class A2(n=18) 2.8 3.7 3.1 two Spanish-speaking boys who were below average Class B(n=24) 4.0 3.9 4.9 in general academic work. On the computer, possibly unhampered by the problems of verbal The generally low scores of the pupils in Class Al communication, they seemed to flower. Working reflects the specificity of their learning. For in a purposeful, competent way, they showed an example, although approximately 2/3 drew a aspect of themselves that could not have been right angle in response to RT 90, only half of predicted. Another boy, who was usually anxious those pupils were able to respond correctly to and impulsive in the normal course of school the inverse operation (i.e., a command of RT 90 work, found a new self-confidence. In general, in response to a right-angled figure). These she believed that pupils felt more positive. children also showed little understanding of relative length, and their sense of direction Structure also concerned the class teacher at was particularly poor in the interpretation of School B, but his criticisms were directed at the procedures. However, because these concepts were physical set-up rather than at the teaching embedded in command fundtions, their confusion methods. As discussed previously, the split of about the effect of the functions undoubtedly the class into two groups resulted in half the contributed to the results. class remaining in the room doing project work while the others were at the computer. This Part 3: Angles. caused a feeling of resentment among those who were scheduled to remain at their desks. They As noted previously, pupils in School B had been were interested in seeing what was happening and

135 152 found it hard to settle down. The idea developed What was most fun about learining to use the that some had to do tedious work while others were computer? having fun. (This was not strictly true, since each half was scheduled for daily computer time). . You feel like an inventor.

In spite of this, their teacher was very en- . You find out things, like if you told the thusiastic about having the computers within the computer to go LEFT 20, you won't know what classroom. As a math specialist he could see their would come out. potential for mathematical learning, and he used . When I make a real neat procedure and the more enthusiastic of the 6th graders to teach people say---Ohl Can I have that pro- Logo to the 7th and 8th grades. cedure on my disk?

. When'the turtle keeps going off the com- Asked if he could see such a course as a permanent puter and makes funny lines. feature in the curriculum, he answered,"Definitely yes to turtle geometry." On the whole , the class In what ways do you feel smarter? felt great enthusiascl for the computers. "They couldn't get enough of it It gave them some- . When I make or learn something new by thing to look forward to. The things that myself. went before were less of a chore." Fascination . Being able to do something that is pretty. with the computer appeared to bring about a new . Because I did the difficult one, the seriousness of purpose to school work in general. EXPLOTION.

Tutors' Appraisal Summary and Conclusions

The tutor organizer, who was himself a computer What was accomplished?What could have been enthusiast, was more cautious in his appraisal. improved? Does Logo make a significant contri- Indeed, he questioned a fundamental premise of the bution to children's understanding of mathematics? training, that computer programming could be used There is no doubt that the schools differed in as an introduction to problem solving activities. fundamental ways---the preparation of the children, He believed that, in the program, the children not the atmosphere to which they were accustomed, the only failed to reach the level of programming expectations that teachers held for them, and ability that would allow them to attempt problems, the computer set-up itself. Conclusions that one but they were presented with no strong incentive might draw on the basis of one group were often to develop the necessary attitudes. He attributed contradicted by evidence from the other. Further- this to the "bang per buck" effect of Logo. more, the organization of the program, relying as Problems involving recursive designs and line it did on the enthusiasm of untrained tutors, drawings can be seen as tame and difficult when could not be called idealin any sense. However, compared with the startling visual effects so some points emerge: easily achieved with random input and massive repeats. Milder, more esoteric goals that in- 1. Work with Logo did affect some basic volved frustration, could be looked upon as a learning. For example, pupils improved case of diminishing returns. There is little mo- in their ability to recognize and describe tivation to persist in the face of difficulty length and a basic angle (900). However, when immediate success can be obtained almost there is evidence that this learning was effortlessly. not generalized. For instance, although 69% of Class Ai (who had received no formal He also suggested that there are certain features training in angles) were able to write of Logo that are unneccessarily confusing. For a Logo procedure to tell the turtle to example, in Logo both length and rotation are draw a 90° angle, only 19% could draw it signified by a combination of letters (FD, RT, etc.) directly with pencil and paper. Further- and numbers. There are no aids to distinguish more, pupils who had not had previous ex- between the two, although they signify qualita- perience in measuring a variety of lengths tively different operations. If symbolic represen- failed to notice changes in the FD command tation were used as commands, children might be from one procedure to another. It seems more aware of the fundamental differences between that their understanding of these concepts the two. was embedded in a particular context, namely turtle geometry. Pupils' Appraisals 2. Attitudes toward Logo material may be The pupils were asked for their reactions to the affected by previous learning experience. program. These were generally favorable, leaving Children who had a history of independent little doubt that they enjoyed it. Aside from work and explicit teaching of the math specifies inherent in the two langauages (turtles concepts involved in turtle geometry vs. sprites), the comments from the two schools and sprites showed more sustained intellec- were similar and are amalgamated below. tual involvement than those who had not. On the other hand, some low achievers and children with reading disability also gained new confidence that was justified by performance.

136 153 3. Although the pupils enjoyed their work with computers, it is unclear whether or not they felt a sense of personal control in terms of being able to program the computer to do a variety of tasks or designs. For many children their "computer experience" was essentially affective and aesthetic in character rather than intellectual. This is not necessarily invalid; the disadvantage to this approach is its limitations. Since the computer is not essentially an aesthetic tool,in order to explore its visual capabilities one must engage in a cognitive effort. An understanding of the processes is an essential condition for exploration. Otherwise one is stuck at th' affective level which, because it depends upon visual excitement, loses its appeal with repetition. By the last week, it seemed that the computer had lost its appeal for some children, and this appeared to be related to their lack of cognitive involvement.

References:

1. Papert, Seymour. Mindstorms.1980. New York: Basic Books.

2. Howe, J., Ross, F., Johnson, K., Plane, F., Inglis, R. Learning Math Through LOGO Programming: The Transition from Labratory to Ciassroom. 19-82. Department of Artificial Intelligence, University of Edinburgh.

137 EDUCATIONAL COMPUTING POST HASTE.A CASE STUDY

Deborah E. Blank

Electronic Learning Facilitators 9510 Linden Avenue, Bethesda, MD 20814

Enthusiasm for educational computing of each school year. can lead to the purchase of computers 2) The computer lab would be staffed before plans for their integration into by regular faculty, and computer literacy the curriculum have been considered. This activities would be carried out in every case study details the evolution of a classroom. computer education curriculum that was 3) Students were to use computers designed to accommodate two factors. for as many purposes as possible. 1) the computers were already in place, 4) The curriculum would be imple- and 2) the principal of the school wanted mented in September. every faculty member and student involved in hands-on computing as soon as possible. DEVELOPING THE CURRICULUM Design of the curriculum proceeded according to these basic principles: 1) Knowledge of the cognitive INTRODUCTION development of children must be reflected in the curriculum goals and activities. How quickly can an elementary school 2) Computers are tools, not a sub- devise a plan for hands-on computer ject to be studied, and should be integra- education for every student, particularly ted into the existing curriculum. when the faculty is untrained and the 3) Computer programming is a computers have already arrived? An valuable activity for children because it independent group of computer educators encourages systematic problem-solv-ng and was asked in May, 1982, to design a logical thinking. kindergarten through eighth grade computer curriculum and prepare teachers to help The next step in designing the students learn about computers. This paper curriculum was to identify goals for the will describe the curriculum that was These area produced, how a faculty was trained to entire project. teach it, and how its implementation is 1) To help students, teachers, and proceeding. staff develop a sense of control over The setting is a parochial school computers. located in a medium-sized northeastern 2) To help students, teachers, and city. Most of the 500 students come from staff discover computer applications that middle to upper-middle class homes. The are useful to them. principal is totally committed to compu- 3) To encourage students to develop ter literacy for all of her students, and an appreciation of computers as tools is bolstered by parental support that for life-long learning. guarantees that all wishes will come true. 4) To encourage the development of The school's board of directors had problem-solving skills that can be applied purchased eight Apple II Plus microcomputer to all facets of living. systems; two of them were being used occa- 5) To stimulate thought and discus- sionally tc run software, and the remain- sion 'VI(mg students, teachers, and staff ing six were sitting idle.Another about: i;orre ridvantages and disadvantagi component of this idyllic situation was a of use in school and in soWtY faculty that was eager to learn about ) tO generate interest in the computing - even if it meant learning field of co:.puter science and related orb with, or from, their students. careers. The principal set these requirements for the computer curriculum: Once these goals were established, the concepts that were to be the basis for 1) All students (K-8) had to have hands-on computer time for at least half the classroom and lab activities were

138 155 defined. The curriculum was divided into a minimal amount of training before they two parts: Computer Literacy and Problem- began instructing students, each activity solving. The programming language LOGO was was written as a complete lesson plan. An selected for grades kindergarten through important feature of each plan was the four, and BASIC for grades five through identification of skills from other subject eight. Concepts were written for each areas that are practiced during the activi- grade level for both sections; many of ty. These listings helped alleviate the concepts were repeated several times concern that other subjects were not throughout the total structure so that receiving sufficient attention because of knowledge and skills would become increas- the new emphasis upon computer education. ingly sophisticated as students progressed. The goals, concepts, and lesson plans were Samples of the concepts are listed below: combined into a curriculum guide that was provided for each teacher. Computer Literacy Concepts: Divel Ks Computers can do things over SELECTING SOFTWARE and over again without getting tired. Since the scKool had a rather loosely People need to share compu- Level 1: organized general curriculUmcourseware ters. selected could not be made to match spe- Level 2. Programs are sets of instru- cific objectives. Teachers requested '.f.e ctions that tell the computer what to do. purchase of math drill and practice pro- The Level 3: People can learn the same grams and some educational games. things in different ways; learning from a remaining software was selected to provide examples of the varied usot. of computers computer is one way. for instructional purposes and to F;ery Level 4: Vocabulary to describe as the basis for some critic: thinking computer parts is extensive and constantly activities (such as software evaluation) growing. by students in the upper grades. A word processing program and a typing tutorial Level 5: Word processing requires were purchased for use by the upper gm:es, the use of a program. and the LOGO language for the younger Level 6: Commercial software can children. It was suggested that teachers be improved by thoughtful development select their own courseware as the year and programming. progressed, since the full software budget had not been spent. In audition, Level 7: Many electronic games are single-purpose computers. several magazines were ordered, and an assortment of books to provide technical Level 8: Computers are significantly information and ideas for using computers altering the way we live. in the school. Problem-solving Concepts: SCHEDULING Level Ks Computers follow instructions given by people. Rather than expecting newly trained teachers 'co program in both LOGO and BASIC, Level 1: People must use special the decision was made to teach LOGO to words to give instructions to computers. the younger children in the fall, and Programs are made up of devote the spring to teaching BASIC to Level 2: the older students. one or more procedures. Each kindergarten through fourth grade class was scheduled Level 3: A computer screen is a to spend 45 minutes per week in the lab grid - all drawings are actually made by with a maximum of two children per compu- connecting straight lines. ter. Fridays were reserved for lab use by students who needed to use drill and Level 4: Some programming languages can be used to write interactive programs. practice programs or for enrichment. All students were to participate in classroom Level 5: Each programming language computer literacy activities throughout has a special vocabulary and syntax. the school year. BASIC is a highly inter- Both the principal and the teachers Level 6: agreed that more computers would allow active language that is user-friendly for year-round use by all students, and and good for writing educational software. that their purchase should be a priority Level 7: Problems to be solved using budget item. Because the principal felt computers can be broken into smaller sub- that she could not afford to employ a problems that operate under the control of full-time computer lab teacher, she the main program. decided that the lab would be staffed by faculty members who would spend one hour Level 8: Solutions to computer problems are arrived at using algorithms. per day teaching hands-on computer uoe. The remaining portion of their school day Classroom and computer lab activities would be spent with their regular classes. had to be planned to correspond with This meant that the lab teachers, as a concepts. Because teachers would have

139 15G rule, were not instructing their own 1-7,me computer to purchase to supplement the school's curriculum, and upon the students in the lab. widespread concern that time spent learn- STAFF DEVELOPMENT ing about computers would, by necessity, mean less time learning about other The faculty returned to school a few subjects. days prior to the opening of the fall semester. Every teacher was expected to 2VALUATING THE-CURRICULUM participate in two days of computer training. All but a few were enthusiastic, During early October, approximately and all were somewhat intimidated by the one month after the curriculum was imple- equipment and scope of the curriculum. mented, one of the trainers visited the Two faculty members had had computers in school for an informal meeting with the their classrooms for several months, and "Intensive 8."At that point there was had begun learning some BASIC. The remain- both good and bad news. The good news ing 23 teachers had no experience with was that the teachers had exceededtheir computers. own expectations and were able toinstruct Four trainers worked with 25 teachers, LOGO with a fair degree of confidence. four administrators, and several parents The children were very excited, and were for two seven-hour days. Eight of the progressing rapidly. The unhappy news teachers had been selected to staff the was that the schedule that called for lab, including the two who knew some a different teacher in the lab each hour BASIC. Since these teachers would soon was a disaster. There were two major be teaching programming one hour per day, problems: the teachers had no time to four days per week, they spent almost two prepare for their programming classes full days with one trainer learning LOGO. (their one-hour preparation time was used These eight teachers became known as the to plan for their regular subjects), and "Intensive 8," and met for short periods their arrival at the lab was always rushed with the larger group to be introduced to and disorienting. (Several parent volun- the curriculum guide and to preview soft- teers were assisting in the lab,but their ware. The remaining teachers spent their schedules were erratic.) Teachers were days as follows: also concerned about having to teach 1/2 day learning terminology and how BASIC in January when they were just to operate the computer to run mastering LOGO. courseware To help address these problems, the principal agreed to: 1/2 day previewing and evaluating courseware 1) Hire (on a tuition-for-services- rendered basis) one of the parent volun- 1/2 day (primary teachers) learning teers who had a programming background. LCGO in the lab (taught by the Provide training in BASIC prior "Intensive 8") 2) to the Christmas break, and allow teachers 1/2 day (upper grade teachers) learn- to take the computers home during that ing BASIC and word processing period. 1/2 day planning "what do I do on Two complete, formal evaluations will Monday?" reviewing those por- The tions of the guide that they be done during the next six months. would begin teaching the next week focus of the evaluations will be upon the proficiency of the teachers, and the value By the end of the second day, both of the curriculum guide in helping them teachers and trainers were tired but satis- achieve their computer education goals. fied. Several teachers remarked that the The results of these evaluations will be specificity of the activities in the guide described at the NECC conference. would help "pull them through" the first weeks of teaching.

INFORMING PARENTS During the evening following the first day of training, parents were invited to interact with the computers and learn about the new computer curriculum. Many parents came, and many stood with arms folded while a courageous few played educational games and moved the LCGO turtle around she screen. The questions that ware asked after the presentation of the curriculum focused on the type of

140 157 LOGO

Carolyn Markuson Joyce Tobias Martin Saltz James Gottlieb Bobbie Gibson Roy Moxley Steve Tipps Hal Evans Glen Bull Terry Schwartz Mary King Steve Taylor Susan Walker Pete Davidson Leah Rampy Rochelle Swensson Barbara S. Hilberg

ABSTRACT: LOGO A Three Year Sequence, Grades Education) decided to develop a program introducing LOGO and a Floor Turtle to a group of nursery school youngsters. The Director and others at the Carolyn Markuson, Joyce Tobias, The Public Schools College of Human Resources and Education became of Brookline; Brookline, MA 02146 involved in developing a curriculum. This presentation is designed to: Seymour Papert created a learning environment, 1) discuss the process used by a group of where learning becomes as natural a process as educators while developing a school walking and talking, all with a new technology: program involving LOGO and Terrapin the microcomputer. Why is LOGO such a unique Turtle in a nursery school environment; language for elemantary grade children? What 2) present an account of the program as used characteristics of LOGO enhance the total learning by teachers and children in the environment challenging and honing the problem classroom; solving capabilities of young children teaching 3) report on the behavior of adults and logical thinking skills and creativity at the same children as the program was developed and time encourages children to test their ideas and used; and receive immediate response? How can spatial and 4) evaluate and suggest modifications to the number relationships, as well as geometric program based upon observations made proportions, become part of a child's during the developmental phase. consciousness? The two presenters were among those initially LOGO has been taught in Brookline schools involved in the developmental phase of this course. since the 1978 pilot program sponsored by MIT and They will present the basic information through a the Nat:tonal Science Foundation. Today, it has lecture. This lecture will be augmented by slides evolved into a threeyear sequence for all children which will show faculty, staff, children, and in grades 4, 5, and 6. The curriculum developed to parents involved in the project. Audience support this program, the inservice training interaction will be invited and an annotated program developed for classroom teachers, and the bibliography of reference and research material decisions required to select a version of LOGO best used during the course development will be suited to the needs of this yearlong educational available. experience will be presented.

ABSTRACT: LOGO Instructional Development Project ABSTRACT: Development of a Program Designed to Use LOGO and a Floor Turtle in a Nursery School Steve Tipps, Hal Evans, Glen Bull, University of Environment: Trials, Tribulations, and Triumphs Virginia School of Education, 405 Emmet St., Charlottesville, VA 22901, Terry Schwartz, Mary Martin Saltz, James Gottlieb, Bobbie Gibson, Roy King, Steve Taylor, Susan Walker, Pete Davidson, Moxley, West Virginia University, Morgantown, WV Albemarle County Schools 26506 The University of Virgins School of Education With some knowledge of microcomputers in the and Albemarle County Schools joined in a schools and funds available to buy the needed cooperative effort to develop and monitor LOGO in et./64L(Olt, the frector of the West Virginia fifteen fourth grade classrooms. The major Nursery Schofil (College of Human Reso0r47,3 an components of the project were:

141 158 Inservice training conducted by two Because of the growing concern about a University professors over the fall year possible gender gap in computer literacy, a final of the project.The goal of inservice focus of this study was a comparison and contrast was proficiency with LOGO. of the programming styles of boys and girls. Implementation--carried out by the teachers The NECC presentation will present a brief with records of progress and change in summary of the findings of project with the actual classroom. Journals and regards to 1) the programming styles of fifth weekly reports were kept. graders using LOGO; 2) the relationship of field Evaluation--cooperatively done with both dependenceindependence to programming styles; and standardized testing for possible 3) the relationship of gender to, programming cogni.ive and attitude change and specific testing for LOGO proficiency. The project began in the summer of 1982 and has gone throughout the year. Teacher had two ABSTRACT: Modifying Papert's Vision: LOGO Lessons months of training before taking the computers into the class. Inservice continued throughout the Barbara S. Milberg, Electronic Learning year with teachers discussing implementation Facilitators, 9510 Linden Avenue, Bethesda, questions and learning new language skills as they Maryland 20814 worked on programming projects. Testing was done early with the children Seymour Papert e:Ivisions children discovering involved in the project. Specific instruments were LOGO in a total LOGO environment, which is freely the Cognitive Abilities Test, the Fennema accessible to them throughout the day.While this Mathematics Attitude Test, and Test of is the ideal, very few schools at present have both IntrinsicExtrinsic Classroom Motivation. Post the commitment and the funds necessary to provide tests will consist of the same tests and exercises this. A reasonable alternative is the LOGO class in LOGO and problem solving. held for a specified time either within a school From the records and experience of teachers in program or as an after school activity. this year's project, instructional suggestions will The project presented here describes a series be compiled for expansion of the project into fifth of LOGO classes, offered by Electronic Learning grade. Facilitators to children, ages 5 through 12, both during the summer and after school during the school year. The success of the program was ABSTRACT: The Programming Styles of Fifth Graders dependent on two factors: 1) the format, which Using LOGO provided for a balance of structure and discovery; and 2) supplemental materials, which allowed Leah Rampy, Rochelle Swensson, School of Education children without computers at home to continue to #337, Indiana University, Third & Jordan, explore LOGO concepts outside the classroom Bloomington, IN 47405 setting. The children were divided roughly into younger Twelve fifth grade students from a local and older groups, working two to a computer. All elementary school participated in a sixweek long groups in the introductory class met for a total of class at Indiana University designed primarily 1) 10 hours, the summer groups meeting on consecutive to introduce students to the LOGO language and 2) days, the fall and winter groups once a week. to study the programming styles exhibited by Followup classes were also available. students learning LOGO. The classes involved large Because of time limitations, teachers made group instruction, time to work on the computer on brief presentations of new concepts to the class instructorsuggested as well as studentinitiated followed by time for children to experiment projects, and opportunities for students to individually. A series of verbal, physical, and demonstrate their programs and to exchange ideas. spatial relationship activities preceded and One Apple IIplus microcomputer was available for accompanied computer use. There was a strong every student. Undergraduate elementary education emphasis on writing procedures and problem solving. majors, trained in LOGO, served as tutorobservers Supplemental materials, such as Program Puzzlers for the students. The observations suggested and Playing Turtle, were developed for distribution similarities and diTferences among the to children to be used outside of class.These and participating students' approaches to programming. other materials developed for the program will be A secondary concern of this study was to described and demonstrated. examine the effect of cognitive style on programming style. Researchers in the Brookline LOGO Project (1979) had argued that students' cognitive styles were reflected in their approach to programming but apparently no attempt was made to measure the cognitive style of the students involved in that project. In this project, students were selected on the basis of extreme scores on the Children's Embedded Figures Test, a measure of field dependenceindependence.

142 15%i ALTERNATIVE APPROACHES TO PROVIDING COMPUTING FACILITIES

Dr, Mary Lucy Sennett Richard V. Murdach Pat Kelly Richard W. Evans Mary M. De Boer

ABSTRACT: CompuShare: A SchoolCommunity Project their number one need. The Western Illinois Center for Educational Improvement served as a Dr. Mary Lucy Sennett, Deer Creek School, Box coordinating agency to provide districts resources 2086, Arcola, MS 38722 to serve the growing need for microcomputer information and inservice. It soon became Over the past few years the message "no funds apparent that individual districts would not be available" has echoed through the halls of all able to solve this problem independently without levels of academia. Unfortunately, the trends seem wasteful duplication of efforts and financial to predict an even more dismal financial future. resources. However, if a number of school In the midst of this period of financial chaos, we districts were to pool their collective resources, are also faced with volumes written on the these needs could met be more effectively at less significant difference microcomputers can make in cost. Through the WICEI, a committee was formed to education. However, there is one simple fact that gather data and information concerning ways in must be faced. It takes money to purchase which districts could collectively meet their microcomputers, If the funds are not available, individual needs for microcomputer support. The the current literature becomes lost rhetoric. result of that effort is the formation of the Project CompuShare is a program designed to Central Illinois Computing Consortium. develop a cooperative venture between small rural The Central Illinois Computing Consortium is schools and small local business enterprises. The designed to provide vitally needed educational major objective of the project is to obtain free services and technical assistance to school microcomputers for the schools from donations from districts, nonpublic schools, vocational and small local businesses. The businesses involved special education cooperatives, community colleges, will also realize both short and longrange and other educational agencies in the benfits from participating in the project. fourteencounty region of the Western Illinois The priorities of Project CompuShare include: Center for Educational Improvement. The computer (1) To establish a workstudy program where a consortium will provide districts with inservice member of CompuShare would pay a student and staff development opportunities, technical minimum wages to do comuter work for assistance, software evaluation, membership in the their business. The work could be done Minnesota Educational Computing Consortium, on the computer that the business donated collective bidding for hardware ands software, and to the school. access to a library of educational software in a (2) To develop an overall training program highly efficient and costeffective manner. Any or for students in the use of microcomputers all of these services wiould be available to in business applications. districts who join the Central Illinois Computing (3) To provide an interchange of ideas and Consortium. support between the school and its local The presentation will concentrate on the business community. benefits, development, and organization of an Tne presenter will discuss guidelines of the educational comuting consortium. project and will elaborate in detail on the benefits and problems that result for both the school and business by participation in Project ABSTRACT: ARelocatah': Computer Laboratory CompuShare. Pat Kelly, Computer Resource Teacher/Coordinator, Carroll County Board of Education, Westminster, MD ABSTRACT: The Central Illinois Computing 21157 Consortium Federal funds have made it possible for Richard V. Murdach, Director, Central Illinois Carroll County to establish the position of Computing Consortium, 1444 Maine Street, Quincy, IL Computer Resource Teacher/Coordinator in order to 62301 create a Relocatable Computer Laboratory. The goal is to travel around to every Middle school in the During the 1981-82 school year an overwhelming county and provide some "hands -on" experience for majority of school districts in central Illinois each seventh and eighth grade student. It is one identified the use of microcomputers for way of beginning to meet the rapidly developing instructional and administrative applications as need to provide "computer literacy" to our

143 160 students. In addition, instruction is provided for its courseware (these results are all highly teachers and administrators in special workshops. significant with p<.001). While this may not be the ideal solution to In addition to students' independent use of meeting this need, it offers an alternative that CALL for basic skills and tutorial purposes, the helps close the gap between the need to raise facility operates in a number of classroom modes to computer literacy levels and the high cost of support instruction. Special classes in Physics, purchasing sufficient hardware to do so.Many Spanish, and English composition (utilizing the unanticipated difficulties were encountered in APPLEWRITER word processing program) make organized preparing schedules and purchase orders, purchasing use of the facility with very positive results. It hardware, and developing the curriculum. The is anticipated that this use will soon expand to presentation will share with others what has been include laboratory exercises in Chemistry, Biology, learned, to assist them in avoiding pitfalls and and Psychology. Another role played by the oversights. facility 1.3 that of providing computer literacy workshops to area educators and to the college's faculty.Workshops as well as individual ABSTRACT: CALL: A Multipurpose Educational visitations are continually held, and it is Computer Facility expected that the facility will have a generative effect in increasing the college's use of Richard W. Evans, Associate Director, The Learning educational computing. Center, SUNY College at Farmingdale, Farmingdale, NY 11735 ABSTRACT: Cost Effective Implementation of a The ComputerAided Learning Laboratory (CALL) Microcomputer Program in the Elementary School at the State University of New York College at Farmingdale is emerging as a very cost effective Mary M. De Boer, Assistant Director, The Learning support mechanism for students and faculty at the Center, SUNY at Farmingdale, Farmingdale, NY 11735 college. CALL utilizes twelve Apple II+ microcomputers linked with a ten megabyte Corvus Seven alternative microcomputer implementation hard disk and Constellation network.This advanced plans for a New England School District were hardware configuration provides educational examined to determine the most cost effective computing support to the college with great approach. Areas of major concern in choosing a flexibility at low staffing and maintenance costs. program were centralized vs. decentralized labs, The staff has created a turnkey instructional curriculum specific vs.general CAI software, and program which allows students new to the facility age/grade level of students. The key factors used to quickly learn its operation and gain access to to determine overall.program cost were training its courseware. Since the ten megabyte disk holds time, hiring of new staff, hardware, and any the equivalent of 64 mini diskettes of software, additional costs (such as busing) resulting from there is no need to maintain and handle a library plan requirements. of floppy disks. This allows the facility to This particular school district had 13 maintain minimal staffing generally a workstudy elementary schools with a total enrollment of student to maintain records and assist students in 5,583.. They had made a recent purchase of 20 the selection of courseware. In addition to microcomputers and had available some limited reducing staffing costs, the network configuration software programs but had no implementation plan. also reduces maintenance costs:the weakest links In addition they were greatly affected by budget in our hardware configuration are the floppy drives cutbacks yet wanted to provide a quality required to mount courseware on the storage disk. educational program utilizing the microcomputers The ability to avoid the use, of floppy drives for for as many of their students as they could. student use has made it possible to operate all The following seven plans with their projected twelve microcomputer stations with zero downtime costs, exclusive of hardware,costs, were proposed since the assumption of fulltime operations in and examined: Spring, 1982. PLAN 1: To place microcomputers in the Resource The facility offers a wide variety of Rooms, implement computer literacy/general educational computer support to the campus. As CAI for grades 4-8, 9 weeks, 4 cycles/year part of the College's Learning Center, CALL costs: yr 1 $18,900, yr 2 $17,500 provides students with both basic skills support PLAN 2: Place microcomputers in classrooms, (reading, writing, and mathematics) and tutorial scattered throughout the schools at assistance in content areas. Courseware is various grade levels available in Physics, Chemistry, Biology, costs: yr 1 $19,200, yr 2 $17,500 Sociology, Psychology, Mathematics, Secretarial PLAN 3: Set up one or more centralized computer Science, and Spanish, as well as in basic English labs providing general CAI structure, reading skills, and mathematics. costs: 1 lab, yr 1 $34,000, yr 2 Student use of the facility for these purposes has $33,500; 2 labs, yr 1 $60,000, yr 2 grown dramatically since the facility began $59,500 operation. A continuing program of evaluation \PLAN 4: Place microcomputers in libraries. This reveals that students react very favorably to the plan was immediately rejected since the important aspects of CALL's operation such as its costs were enormously prohibitive, as the General Functioning, Potential, Proctoring school district had no library staff Assistance, and the Interest and Effectiveness of members.

144 161 PLAN 5: Implement curriculum relevant CAI programs with microcomputers in the classrooms covering various grades costs: yr 1 $35,000, yr 2 $30,500 PLAN 6: Curriculum relevant CAI with microcomputers in centralized labs

costs:1 lab, yr 1 $34.500, yr 2 $33,500; 2 lass, yr 1 $60,520, yr 2 $59,520 PLAN7:Curriculum relevant CAI with microcomputers in the Resource Rooms costs: yr 1 $35,000, yr 2 $30,500 General CAI and computer literacy were chosen over a curriculum relevant approach because of the history of difficult and unsuccessful implementation,of curriculum relevant CAI in elementary schools. Plan 1 was chosen because it provided quality computer literacy/general CAI exposure to a large number of students, while it did not strain the school district's budget. The higher cost in year one was incurred due to training time of staff during the initial implementation phase; year two includes maintenance costs.The grades 4-8 were chosen on the basis of a minimum of nine hours exposure for each student, with a priority for older students to better aid in their preparation for high school.

145 162 DISTANCE TEACHING OF SOFTWARE ENGINEERING

Darrel Ince The Open University United Kingdom w. S. Matheson The Open University United Kingdom

As part of amajor programme of scientific described and the structure and aims of the and technological updating, the Open course outlined. University is developing anovel Masters Course which deals with the industrial The second part ofthe paper will describe applications of computers. The course is one of the major modules of the course, intended for practising programmers, software engineering. This module, which engineers and technical managers who work will be taken by themajority of students In a real-time environment and who are who enter the course, takes the view that finding their work transformed by the coding represents only a small part of the micro-computer. program development process, and that other activities (analysis, specification, The course is intended to alleviate a major design, maintenance) are, at least, equally problemwhich is currently facing British important. The modulewill consist of a industry over the retraining of existing number of components: personnel. There are now a large number of personnel working inareas such as process (i) Course texts which dealwith parts of control, avionics and command and control the software life cycle. systems, who require updating on the software and hardware aspects of (ii) A number of industrial, real-time microcomputer systems. Unfortunately, software case studies. these personnel often occupy critical positions in their company and cannot be (iii) A home experiment kit which consists released for the 12-18 monthperiod of a stand alone micro-computerwhich can necessary for the study of a conventional act as a terminal to the Open University Master degree. Wehope that the Open mainframe network, for 'hands-on' University, an institution set up to experimental work. provide degree-level education for the house-based student, will be able to (iv) a video tapewhich shows the progress respond to this challenge, and provide a of an actual software project using a large model which can be used in other shortage number of personnel. areas where there is a need for scientific and technological updating. The paper will conclude with a discussion of some of the difficulties encountered in The first part of the paper will describe providing such a module to the postgraduate the work of the Open University. The student studying at a distance. background to the course will also be

146 163 District Planning for Computer Use in K-12

Glenn Fisher Alameda County Superintendent of Schools Office Hayward, CA 94541

ABSTRACT

An outline for a K-12 computer literacy . For which students? staff? scope and sequencemodel will be discussed . What courseware is appropriate? by Gary Bitter. Included will be computer . Do we need to develop our own software? awareness and computer programming details. . Can we afford to?

A brief implementation model for curriculum . Which brand(s) should we buy? will be discussed. . What monies will be committed?

Don Rawitsch will discuss the following: . How will we train staff? 1) In creating an instructional computing The chair will discuss the role of the plan, instructional expertis'.! is more administrator in this process. District important that computing expertise. planning is a political process and 2) Although implementation factors such requires support of staff, Board, as equipment, materials, and training take administration, and parents. It should up themost time in the planning process, parallel other curriculum planning in the these items should not be discussed until a district. Staff development is an rationale is developed for why the district important part of a plan. thinks it should be involved in computing. The plan should extend over time to 3) The district must determine the demonstrate commitment and the possibility relationship of computing to its of change and input. There should be room curriculum, considering both how computing for serendipity and learning at school can be integrated into the present sites within the district plan. curriculum, and what new curriculum areas The administrator is the crucial element might be suggested by computing. in successful implementation. 4) Computer literacy is much more effectivelydeveloped in s ddents when it PARTICIPANTS: is planned as a cumulative effect of activities throughout the curriculum, as Gary Bitter opposed to being set aside as a single Arizona State University course. Tempe, AZ 85287 5) The purchase ofcomputing equipment should be determined based on educational Pristen Bird goals more than by hardware features and Instructional Computing Consultant cost. Department of Education A variety of questions to consider will Tallahassee, FL 32301 be addressed by Pristen Bird. Don Rawitsch Among these are: Director of User Service, MECC What do we want to do with computers? St. Paul, MN 55113

147 164 Information Technology and Its Impact on the United States - Overview and Implications

Sponsor: ICCE

Linda Garcia Fredrick Weingarten Office of Technology Assessment U. S. Congress Washington, D.C. 20510

Linda Roberts Office of Library and Learning Technologies U. S. Department of Education Washington, D.C. 20202

The "information revolutionbis profoundly affecting American education and training -- creating new demands for instructional services and, at the same time, providing new opportunities for improvement and delivery of such services. The new information technologies can help all educational institutions meet these new demands. Many are already being effectively used in education and training. However, OTA has ientified a number of barriers to their use -- their high initial cost, the lack of high quality programming, and the shortage of local personnel with adequate training. Whether or not new information technologies will fulfill their potential will depend, in part, on the kinds of actions that the Federal Government takes. What is needed is a broad approach that takes into account the changing needs for education and training, considerations (IF equity, and changing institutional roles.

148 THE ELECTRONIC BLACKBOARD USING A MICROCOMPUTER AND LARGE-SCREEN TELEVISION AS A LECTURE AID

JAMES E. CLARK Department of Economics Wichita State University Wichita, Kansas 67208 Telephone 316-689-3220

ABSTRACT

This paper will describe and demonstrate the to enhance instruction, particularly in the large- development, use, and benefits of programs that use section classroom. The Electronic Blackboard is a an Apple microcomputer and large-screen projector step further in integrating the microcomputer into television as the primary lecture aid (replacing the the classroom. traditional chalkboard or overhead projector) in large lecture sections of college Principles of ADVANTAGES OF THE ELECTRONIC BLACKBOARD Economics classes. Some of the benefits of using the Electronic Blackboard are:1) its attractiveness Compared to chalkboards and overhead projectors to students; 2) its ability to interweave text with (the traditional lecture aids for large classes), the graphics and simulations; 3) its legibility, even Electronic Blackboard has several advantages that in the back of large (250+ seat) classrooms; 4) its make using it worth the necessary investment in ability to use the Apple's high-resolution colors equipment and preparation time. Among these are: to highlight and tie together key concepts. The Electronic Blackboard can be used to prepare lec- Attractiveness to students ture materials before class, and also can be used in a "live" mode to display text and graphics The current generation of students are accus- created during class; examples of both typus of tomed by a lifetime of experience to watching the presentations will be shown. movement and color of television shows; the black - and- white, static pictures produced by chalkboards ORIGINS OF PROGRAM and overhead projectors are, by contrast, very boring and unexciting. The Electronic Blackboard The impetus for the Electronic Blackboard can promote student attentiveness and learning by program came from a National Science Foundation providing the color and motion to which students CAUSE (Comprehensive Assistance to Undergraduate have become accustomed and attracted. 'In addition, Science Education) grant entitled "Interactive since the Electronic Blackboard embodies up-to-date Microcomputing in the Classroom." The purpose of the technology and resembles (faintly) an arcade game, grant is to encourage the use of microcomputers in students are more likely to pay attention in class classroom settings for creating demonstrations, sim- to material presented on the Electronic Blackboard. ulations, etc., that can be easily manipulated to answer "what-if" questions from students.Funds from Le 1.1)lit the grant have been used to purchase 19 Apple II+ microcomputers and seven Kloss Novabeam projector When properly prepared, text and graphics pre- televisions with large (4' x 6') screens. On a com- sented on the Electronic Blackboard are much more petitive basis, 22 faculty members at Wichita State legible to students than are the same materials University have been selected to receive summer written on a chalkboard or overhead projector.This stipends for the development of software to imple- is especially true for students who are forced by ment the objectives of the grant. To date, the large class sections to sit in the far reaches of following disciplines have been involved, either several-hundred-seat lecture rooms. Proper presen- through summer sub-grants or voluntary participation: tation does require that attention be given to clear astronomy, biology, chemistry, economics, electrical and unconfusing screen layouts (this is also neces- engineering, industrial education, industrial engin- sary for clear presentations on chalkboards and over- eering, mathematics, and physics. head projectors). To insure legibility, it is also Many excellent modules have been produced by necessary to use non-standard character sets. The participants in this grant, ranging from simulations Electronic Blackboard uses both modified, all-white, of queuing and other random processes to the graph- standard-size upper and lower case characters, as ical depiction of the evolution of the universe; some well as slightly larger upper and lower case letters 1,2 with true descenders; this larger character set of these programs were demonstrated at NECC 1982. greatly improves legibility, especially in large While these modules, and similar ones created else- classrooms. Details on the character sets used may where, have been valuable to students,3 they have bn found in the section below on Program Description not made full use of the microcomputer's ability and Capabilities.

1 6 149 Color arisen with the use of the Electronic Blackboard. Due to the design of the screen, the intensity of the The ability of the Apple microcomputer to create image from the projector is diminished when the colored lines and images has been harnessed in the screen is viewed from a large angle; with the present Electronic Blackboard to make presentations both equipment, the Electronic Blackboard requires clearer and more interesting. In particular, complex several projectors and screens for classrooms that presentations involving graphical analysis (common in are wide relative to their depth. Also, prepared economics, and_in many other disciplines) can be made text materials appear on the Electronic Blackboard much clearer, and associated elements tied together, much faster than they can be written on a chalkboard with a careful use of color. While colored diagrams or overhead projector, creating a tendency for the can be created on chalkboards and overhead projec- instructor to present materials faster than students' tors, such diagrams are typically quite messy, and note-taking can keep up. very difficult to modify to show the effects of For lecture materials prepared in advance of changes. classes, creating materials for the Electronic Black- board takes several times as long as does preparing Animation the same material for writing on a chalkboard or overhead projector. At present, preparing one class The one area where the Electronic Blackboard hour's worth of material takes between two and three represents the greatest improvement over its tradi- hours of time for the Electronic Blackboard, with tional counterparts is the microcomputer's ability to considerably more time necessary for complex manipulate rapidly large amounts of information, and graphics; this compares to the approximately half- to clearly display the results.While this ability hour usually required to prepare an hour's worth of is most obviously applied to purely numerical infor- chalkboard presentation. The benefits of the Elec- mation, it can also be very effectively applied in tronic Blackboard seem (to the author) to be worth classrooms where the relevant information concerns the time involved, especially since the materials the location of a line on a graph, or the position of created can be reused in future classes, with any an image on the screen.The microcomputer can be needed modifications made fairly rapidly. An easily used to create, display, and manipulate numer- authoring system for the Electronic Blackboard is at ical and graphical examples, simulations, and so on, present being developed that should substantially in ways and at speeds that are simply not possible reduce preparation time. with the traditional tools. This is especially true in cases where students ask questions which need PROGRAM DESCRIPTION AND CAPABILITIES substantial calculations to answer. The Electronic Blackboard is designed to allow easy insertion of The core program of the Electronic Blackboard is modules representing graphical and numerical written in Basic and is based on the concept of examples, simulations, and other similar types of organizing the material to be presented into "pages;" materials; several examples of such modules will be one page is normally one full screen of material. presented, along with examples demonstrating the Within each page, the program writes a section of text other advantages of the Electronic Blackboard. onto the screen, then waits for the space bar (or firing button on paddles or joystick) to be pressed STUDENT REACTION TO THE ELECTRONIC BLACKBOARD before writing the next section of text. Section length can be anywhere from one character to an The author has used the Electronic Blackboard as entire page, and can be varied for each section. the primary lecture aid in teaching two large class "Turning pages"' (clearing the screen and starting a sections of Principles of Economics during both the new section of text) is also accomplished by pressing Fall 1982 and Spring 1983 semesters at Wichita State the space bar. The instructor is thus always in University. While final course evaluations are not control of how fast the material is presented. If yet back, some preliminary evaluations are available. desired, it is possible to have more than one Student attitudes toward the use of the Electronic screenful of text a particular "page;" the screen Blackboard, as expressed both in class discussions can be scrolle? to make room for new text without and in outside conversations with the author and erasing the entire screen. It is also possible to go other faculty members, have been almost unanimously back to devious pages or to move forward by skipping favorable. Comments show that students feel that pages at any time. text materials presented on the Electronic Black- Text and graphics are displayed on the Apple's board are more interesting and easier to learn from high-resolution screen through the use of Image than would be the case if the same materials were Printer (a machine-language graphics program soon to presented using traditional methods; the reaction to be available from C & C Software).As used in the graphical materials has been even more favorable, Electronic Blackboard, Image Printer can very rapidly with stress placed on the increased understanding write upper and lower case letters, numbers, and that is generated by the used of color and motion. symbols anywhere on the Apple's high-resolution In addition to classroom use, demonstrations of screen (regardless of byte boundaries). Character materials prepared on the Electronic Blackboard have sets available include a standard-size (7x8 dots) been presented to conferences of persons interested all-white character set that is much clearer than the 4 Apple's standard character set, an enlarged (9x12 in computer-based education and to economIt7 dots) white character set with true descenders, and a tors5, with highly enthusiastic and positive recep- small (6x6 dots) character set with only upper case tions. letters and numbers. Other character sets in other Along with its benefits, several problems have sizes can be created if the user desires.For use in

150 161 large classrooms, legibility is best with the large students, its legibility, and its use of color and character set, although the standard-size characters animation to clarify complex materials and maintain can also be used, and are appropriate for tables, student attentiveness.The Electronic Blackboard labels, and so on. program, created to implement the concept, has been The maximum amount of material contained in an described and demonstrated, and experience so far Electronic Blackboard program is limited by the with using the program to teach Principles of Econ- amount of meuory available. The Image Printer omics has been reviewed. The Electronic Blackboard routines and character sets fill the space below the represents another step forward in utilizing the Apple's high-t:esolution graphics page 1; if both abilities of the computer, and the microcomputer in high-resolution graphics pages are to be used, the particular, to improve the quality of education. space available for programs is from 24576 to 39590 for a 48K Apple with MAXFILESI set. This is enough ACKNOWLEDGEMENTS space for 15-20 average pages of text with colored underlining. Longer presentatioas can be created by My initial work on developing the Electronic having one program end by running the next program. Blackboard, and on modules for teaching Principles This can be done without disturbing the text or of Economics, was supported by NSF CAUSE grant #SER- graphics being displayed, so that the instructor can 80-04784. I would like to thank Dick Cornelius be explaining the present display while the next (project director), Mel Zandler, and the other program is loading. faculty at Wichita State University who participated The organization of the program into "pages" in the CAUSE grant, for help, encouragement, and makes it easy to add modules of graphics displays, good examples. simulations, etc., into the program (Image Printer Apple, Apple II+, and Apple Graphics Tablet are is also an excellent way co create graphics displays trademarks of Apple, Inc. Image Printer is a trade- and animation). Each page begins with an even- mark of C & C Software. thousand line number (page 1 starts at line 1000, page 13 starts at line 13000, etc., so that by REFERENCES appropriate numbering a module ca be executed at the appropriate time in the program. Modules 1. Cornelius, Richard, "User Power; A Discussion created outside the Electronic Blackboard can be and Real-Time Demonstration of Valuable Programming renumbered and merged into the Electronic Blackboard Features," NECC 1982 Proceedings, pp. 95-96. using any good commercial program editor. 2. Cornelius, Richard, "Microcomputers and Large- Graphics and text can be very effectively inter- Screen Projectors in Science Lecture Halls," NECC woven by putting graphics on one of the Apple's high- 1982 Proceedings, p. 231. resolution pages and the corresponding text on the 3. Bowman, Barbara, and Randy Ellsworth, "Micro- other high-resolution page. At any time, the Elec- computing in the College Classroom and the Effects on tronic. Blackboard can flip between the two pages by Student Attitudes Toward Computers," presented at the pressing the "1" and "2" keys. Annual Meeting of the Association of Psychological Also available is a nondestructive pointar (an and Educational Research in Kansas, Emporia, Kansas, arrow shape) that can be placed on the screen to 1982. point to, key terms and illustrations. The position 4. Clark, James. E., "The Electronic' Blackboard," of the arrow can be controlled from the keybord, or presented at the Third Microcomputers in the Class- by a joystick or paddles. room Conference, Wichita, Kansas, 1982; presentation The Electronic Blackboard can be used not only to IEEE Student Branch, 'achita State University, to present materials that have been prepared and 1982. stored in advance, but also can be used "live" to put 5. Clark, James E., "Using Microcomputers and text and graphics on the screen. At any time, the Large-Screen Projectors for Teaching Principles of instructor can move from a prepared page in the Economics," presented at the Annual Meeting of the Electronic Blackboard to a "blank page" routine that Joint Council on Economic Education, Kansas City, permits the instructor to put upper and lower case Missouri, 1982. text, numbers, and symbols anywhere on the screen by typing at the keyboard. Key terms can he underlined in color on the blank page by using keyboard commands to specify the color and the endpoints of the line. More complex graphics can be drawn "live" through the Electronic Blackboard's interface with the Apple Graphics Tablet. All of the features of the Electrotic Black- board discussed above will be demonstrated during the paper presentation.

CONCLUSION

This paper has discussed ani demonstrated the advantages of using an Apple II+ microcomputer and large- screen projector television as a replacement for the traditional chalkboard and overhead projector as the primary lecture aid for large classes. The advantages of this approach are it,. att-ictiveness to

151 168 RESULTS AID LESSCNS FPCM ? STUDY OF PFArERS' CCNTPCL OF PATE CF TEXT PRESEPTATION OrCONFUTER SCREENS

Werner Feibel

Educational Technology Center University of California Irvine, Calif. 92717

There is a growing body of research literature on AEFTRACT variables to be considered in formatting text on the pages of a book; these variables range from comparisons of different type styles to the most In an effort to begin eNploring the factors effective ways of laying out material on the page -- important for effectively using the capabilities e.g., whether to justify margins, and how to use of the computer to present text, two ;E,Idies have indentations and spacing to provide the reader with been carried nut.Results from the first one are additional information anti help for most effectively reported. In t4is study, students in introductory identifying major concepts and points (see, e.g., college pnysics courses read tnree texts, presented Partley, 1980). While this research has not provided via computer screen. After ea.h session, students any clear-cut results to date, it has served to 'all were tested on the text content, and were into question a number of conventions and interviewed regarding their reactions to the texts assumptions, and also to emphasize the complexity of and the presentation on screens. During one of the sessions, each student was given control over the the issues involved. rate at wnich text appeared on the screen, If we expect to get on the right track quickly including en option to stop the text at any point. with the new, electronic medium, such factors must While no significant differences were found be considered in the context of presenting text and between tne control and no-control conditions, pictorial information with computers. Furthermore. several tendencies were observed, and several additional considerationicenter the picture -- many useful lessons for research in this area were of which have been discussed by Alfred Pork in his gained. These are discussed. "Textual Taxonomy" (1981). In particular, the consequences of certain differe between the

computer and the book must ' tai, -1 irRo account. Introduction Two types of differences, te.oral and spatial will be considered here. As the computer comes co play an increasing role in schools and other learning situations, issues In contrast to the information on a book page, relating to the most effective pedagogical uses of a which is static and completely present when the book new tool and medium must be dealt with. To date. leaves the printer's, the computer screen can be most of the research on these issues has concerned filled at a rate that the reader can select. This the.use of the computer as a tool for testing would give a reader the option of interacting with learners' understanding of material presented, for the material in a manner very different from a book. providing them with additional opportunities to work In addition to such peripheral possibilities as with the material, -- e.g., in the form of drill and using this option to exercise and perhaps increase practice -- or for presenting tutorials on certain her reading rate, controlling tne rate of topics in the curriculum (although this use of the information presentation could enable the reader to computer is still much less prevalent than its use view material in a much more active manner -- e.g., for drill and practice). by stopping the text presentation and trying to anticipate where a certain argument is leading, then Cuestions relating to the computer as a medium for testing her expectations (and thus, to some degree, presenting curricular material. -- i.e., as an her understanding of the material) by letting the alternative to media such as books or films, subsequent text appear on the screen. Related to particularly the former -- appear to be getting much this is the possibility of animating diagrams and less attention. To some extent this may be due to an illustrations; the student could formulate implicit assumption that the resolution of most of hypotheses about a given sit :zstion and could these questions will take the same form it did in actually view a simplified esentation of books. Several considerations, however, suggest that experimental results relati to ',tie situation and this may not be appropriate; moreover, it is by no discussed in the reading. means clear that ventions and tendencies common for the printed page are the most effective ones possible, even for that medium.

152 16J resides tnese temporal factors, a second set of Procedure spatially-based differences between book page and computer screen as media concerns the cost, and paterial3; therefore the use, of empty space on the Portions of two chapters from a widely used presentation field. Plank space on a book page costs introductory physics textbook were used in the money, and is tnerefore generally avoided in this studies. The goal was to find texts that were medium. One consequence of this is the relatively relatively self-contained, that wouldbe of dense packing of print on the page. Wnile approximately equivalent difficulty, and that were economically advantageous, this may prove to be considered equally interesting by readers. The pedagogically detrimental. plank space on computer actual texts were selected after pretesting them on screens, however, is free; this opens up many students in physics classes and selecting thn two possibilities for using space on the presentation that students regarded as most comparable in field to provide supplementary information about the difficulty and level of interest. Certain, very relative importance of different parts of the minor, modifications were necessary to make the two material, as well as facilitating the formation of a readings sufficiently self-contained and visual image of the probable content stylistically acceptable; however, these changes interrelationsnips because one can easily scan a were limited to the insertion of a few clauses, and body of information repeatedly. the deletion of some short discussions of material related to other sections of the text. Cne reading Thus, the luxury ci`free space can make various dealt with electromagnetic induction, and the other presentation possibilities more attractive. Dividing covered the kinetic theory of gases. In addition, a text into natural phrasing, thereby providing what third text from a different textbook, on the law of is likely to be a more natural pace and chunk size gravitation, was used in a practice session. for the intake of information, is one way of using space to possible pedagogical advantage. Formatting Tests designed to assess students' comprehension text in a way tnat makes clearer the hierarchical of various aspects of the material were constructed relationsnips in the material also opens up a large - - for the practice text and for the two readings set of possibilities. Vnile the latter, used in the study. Questions ranged from those hierarchical, formatting depends directly on the designed to determine the reader's recall of very actual content for its effectiveness, natural specific information -- e.g., whether a particular pnrasing simply attempts to capitalize on processes phrase was used as a section heading, or where on and tendencies implicit in various informal learning the screen something appeared -- through contexts, but conspicuously absent in our understanding of various concepts -- e.g., defining educational system teyond the first few grades. I the basic ideas and terms presented in the reading refer here to the use of meter for passing on oral - - to grasp of the conceptual consequences of the literature and information, the timing that makes material -- as reflected in students' ability to some speakers much more memorable and effective than others, and tne use of motoric activity and rhythm solve problems about tie concepts and to discuss the to teach various skills and concepts -- (see, e.g., significance of particular experiments for the topic under discussion. the work of Furth & Wachs, 1576, and that of Creenfield & Childs, 1(376). We explored the Finally, a set of questionnaires was developed to possibilities of using such a resource to fuller obtain information about students' background with advantage in learning settings. physics and computers, their attitudes toward computers, their study strategies and habits, and As one step toward helping to clarify these issues their reactions to being able to control rate of and making decisions based on as mu:i. information as text presentation. possible, we are carrying out two studies, in a project funded by the rational Science Foundation, Equipment and Programs: to investigate som4-'&f the variables relevant to the The texts were presented on a Terak F51CA personal use of the computer as a medium for presenting computer. To allow the reader to adjust the rate of textual material. Tne first study investigated the text presentation, Michael Potter, one of our effects of giving students control over the rate of student coders built several speed control boxes, text presentation on tneir comprehension of the which could be connected to the computer. The boxes material and on their attitudes toward the learning operated by moving a knob to a position situation. In the second study we investigated the corresponding to the desired rate of text consequences of natural phrasing on these same presentation; in addition, a button could be pre!ised variables. I will present the results of the first to stop tne text at any point -- until the reader of these studies -- discussing hoth the data and restarted it by pressing the button again. some of the lessons we have learned from this. initial investigation. some of these results and These boxes worked through a rrogram written by lessons have methodological implications for Adam reneschan, another student coder. Tne program research in this area, and I hope tney will be of translated settings on the box into timing delays or use in guiding others in their own studies. interrupts in the main program; these controlled the presentation of the text on the screen. resides presenting the text on the screen, and drawing the accompanying diagrams, the main program also allowed the user to flip around to earlier or later sections in the text. This flipping could be done easily, and

153 170 students reported having little difficulty learning generally they did not use it beyond an occasional now to move around in the text. Finally, a program attempt -- generally to simply try it out. The that recorded any changes in presentation rate and control boxes permitted the reader to change the wrote this information to a data file for later rate from a very slow, letter by letter, analysis, became available during the study, and presentation of the text at one extreme, to a rate this was used to record the actual changes for some comparable to a reading speed slightly faster than of the participants. the average reading rate for nontechnical material. While this rate was regarded as sufficient, the Participants; students' reactions to this provided one of the Students were recruited from three introductory lessons to be considered in research of this nature. physics courses, and were paid for participating in tnree sessions. Yost of the students were majoring In the no learner control condition, the text was in some field of science, although this was not a simply thrown onto the screen, so that the entire criterion for selection. ef the33students (23 screen appeared in a couple of seconds. This rate male's,10 females) who partiCipated beyond the was much too rapid for anyone to read it as it was practice session, 31 completed the study. Cnly two coming out. males dropped out.

Method_ pesults Students were tested over a six week period in a repeated measures design, with three sessions per While test scores in a sufficiently large class student. When they arrived for their session, can generally be approximated quite well by a normal students were placed in a room with the computer, distribution, such an assumption was considered too and the program was started for them. They were told risky in the present study -- because students to read the material, spending as much time as they actually came from three different classes, so that wished on it. They were told that they would be the underlying distribution would more probably tested on the material at the end of the session, consist of a combination of normal distributions. and were permitted to take notes on the material. Consequently, analyses were done using nonparametric When they completed the reading, they were asked to tests -- particularly normal deviates tests (see, give up their notes, and were given a test on the e.g., Marascullo & McSweeney, 1977). material. After this was completed, they were interviewed briefly about their reactions to the Cn the whole, the major analyses identified no reading -- both as content and with respect to its significant differences on the comprehension tests differences presentation via computer. as a function of learner control. While in performance were generally in the expected The first, practice, session was not included in direction, these differences were not sufficiently any comprehension analyses. rather, this session was large to permit rejection of the null hypothesis. about used to obtain background information on the Thus, mean proportion correct on the text electromagnetic induction was C.565 (s.d. = 0.163) students, and to give them practice reading text from a computer screen -- including developing for the learner control students, and 0.529 (s.d. = familiarity with the means for flipping around in 0.188) for the no learner control students; tne text. Furthermore, the practice test similarly, for the reading on the kinetic theory of administered et the end of this session informed gases, the figures were 0.654 (s.d. = 0.180) and students about the range of detail tney were 0.593 (s.d. = 0.150) for the learner control and no Since expected to attend to in subsequent readings. learner control students, respectively. Winsorization -- truncating the range of scores by discarding a certain-number of scores at each tail After the first session, students were randomly assigned to either a learner control or no learner of the sample distribution, in order to get rid of control condition for the next saession. Pecause of outliers -- made no substantial difference, the full equipment difficulties, the division for the second set of available scores was used in analyses, to session was IP no learner control and 15 learner take advantage of the additional degree's of freedom control. For the final session, students were put in available When working with the full sample. whichever of the two conditions they had not done in their previous session. Pecause of attrition, the The fact that the learner control individuals did third session consisted of 14 no control and 17 score higher on both texts is encouraging, and learner control participants. Crder of texts was speaks in favor of investigating this variable fixed; all students read about electromagnetic further -- particularly since the same students read induction in the second and about the kinetic theory both texts. This latter point makes it unlikely that of gases in the third sessions. Thus, roughly half the mean differences are attributable to the the people had control over rate of presentation of individuals or to the texts, since the relative the electromagnetic induction text, and the rest standing of the two groups switched for the texts, better over the kinetic theory of gases text. and since the common denominator in the performances was the learner control. Nonetheless, In the learner control condition, the student since the results only approached significance, they could adjust the rate of text presentation, or stop must be considered merely suggestive. the text, at any point. Unfortunately, problems with the control boxes made the hold button somewhat' unreliable, and most students reported that

154 In the learner control conditions, students who both because the number of people who considered checked more positive adjectives to describe their themselves slow readers was too small, and because exrr' -"- ',mputers tended to score slightly English as a second language was not explicitly hi ,Y1 students less positively inclo as a ffIc'-- in the present study. in rs. The difference for the t w-7 > aLput 11, but again in favor of students witri positive experiences with computers. Discussion The samples in the learner control conditions were too small to make it feasible to use this attitude Overall the results of the first study were information as a covariate in analyses. inconclusive regarding the effects of learner Nevertheless, these results make intuitive sense, control of text presentation rate on subsequent and it will be helpful to explore the more general recall and understanding of the material. questions of the conditions under which computers Nonetheless, several indications and lessons did facilitate learning or understaning in people with result from the study. different experiences with computers. First, for our purposes, the finding that Physics background did make a difference, as is to performance on the two texts was comparable is be expected, with those in the more advanced encouraging for us, since it supports our intention introductory course and those who had taken more of using the same texts in the second study. The physics courses performing better on the average; average performance scores do not prove that the however, these differences again failed to reach texts are comparable, but they do make such a claim significance. Moreover, this factor did not play a plausible. role in the grouping, since there were no differences in the make-up of the two groups (i.e., Other issues arising from the first study will control with first reading versus control with help us in the details of running the second study. second reading). The greater heterogeneity as well as providing some guides for other introduced through this broader range of physics researchers in the area. First, the trends reinforce background may have contributed to the overall the importance of avoiding samples that are too results, however, since the increased range of heterogeneous. Possible solutions for this performance attributable to theme differences in difficulty include either ensuring that all familiarity with the subject matter would be participants are drawn from the same classes, or expected to increase the variance of the performance designing the study to permit separate analyses for figures. different groups within the sample. Post hoc examination of subsets of comprehension Tne indications of differential effects of learner test items (i.e., specific memory, definitions, control as a function of attitudes about computers concept memory, and problem solving) did not permit. suggests that such information be obtained and an unequivocal identification of the particular either built into a multiple factor design as a aspects responsible for observed differences. In grouping variable, or that the study be planned so part this was due to the much larger proportions of as to permit the use of such information as a ties in scores on these item subsets -- because of covariate in the analyses. the r-h smaller possible range of values. There did be any consistency in the types of A more important and more general issue arises iLems where the two groups differed -- something from students' reactions to the consequences of that might possibly be examined more systematically controlling presentation rate. College students in a design where interactions can be studied belong to a group whose reading and study habits are directly. largely developed, and likely to be relatively entrenched. In particular, this is a group whose Two other data are relevant here, since they also members have all been reading for over a decade -- provide hints for subsequent research. First, many almost exclusively (if not entirely) from a printed of the students in the learner control conditions medium. Thus, these people have invested a great said they found the continuous appearance of text on deal of effort and practice in mastering the process the screen distracting; most said they simply set of obtaining information from the printed page. Most the box to maximum and then waited for the screen to college students are relatively proficient at fill before actually reading the text. Related to reading, and they are quite likely to consider this this is the fact that many students mentioned the aspect of their education as more or less completed. novelty of dealing with text presented in such a Therefore, hindsight suggests it should not be sequential and temporally variable manner. This surprising to find them somewhat disconcerted and novelty was not always perceived favorably. A more negative when required to deal with a situation that lue raised by such reactions is considered demands adjustment in some of these habits and in bel the expectations they have built up about the "behavior" of text. A second datum of interest concerns the finding that several of the students who considered The individuals who responded more favorably to themselves slower than average readers -- most of this new situation -- those who considered these being students with English as a second themselves slow readers (and thus presumablyfeel a language -- found the learner control box useful. certain dissatisfaction on this score) andthose who Again, however, theme findings provide only have been reading English for a shorter suggestions for future studies. They cannot be time than native speakers -- are students Whosehabits can be considered primary results of the present study -- regarded as still changing, or at leastmore

155 172 amenable to change. These are people more likely to be receptive to the possibility of modifying their Acknowledgments reading habits or of taking advantage of situations tn-can facilitate the process of gaining Tnis research was funded by the rational Science iniormation for t . , Indation, through their Research in Science P.n important implication of this consideration -- Education progr,,'1TSE Grant I SED - 8112378) if correct -- is that such factors' as control over rate of text presentation might be more profitably any thanks to Kristina Hooper, Cynthia Powell, studied in individuals whose habits are still in the and Ruth Von Plum for their invaluable suggestions, process of being perfected -- e.g., younger students insights and efforts. Thanks also to Michael Potter or learners for whom particular reading situations and Adam Feneschan for their efforts in making the are still relatively novel. In a sense, this speed-control box and writing the programs for the implication can be considered a special case of the study. Without the efforts of these people, the more general issue of the diffusion of a new tool or project would never have come to fruition. context -- something generally much easier with Pibliography people who have less of a stake in what was previously available. Pork, A.Tex tual Tax nomy . Unpublished paper, Such habits should be less influential in the Educational Technology Center, University of second study, however, since there even those who California at: Irvine, 1981. consider themselves sufficiently competent readers will be dealing with a process very similar to Furth, H.C. & V'achs, 11 Tninking Goes to School: activities in which they commonly engage --grouping Piaget's Tneory in Practice. Mew York & London : information in a manner that seems natural. Oxford University Press, 1974. The present study -- despite its inconclusive Greenfield, & Childs, C.P. Weaving, color results -- does indicate several questions and terms, and pattern representation: Cultural directions in need of research. Before we make hasty influences and cognitive development among the and ultimately untenable decisions about the manner Zinacantecos of Southern Mexico.Paper presented in which computers should be used in learning at the First International Conference of the situations, we must be certain we can make those International Association for Cross-Cultural decisions from a reasonable and sufficient knowledge Psychology.Hong Kong, 1973. base. While I nave been unable to add much specific information to that base, I hope I have provided Hartley, James(Ed.)The Psychology of Written some insight into some of the priorities and Communication: Selected readings, New York: difficulties that must be considered as we move Fichols Publishing Co., 1980. toward the computer age. Marascuilo, L.A. & McSweeney, M.;!on- Parametric and Distribution-Free Fg.thosis for the Social Sciences. Monterey, Calif. :['rooks/Cole, 1977.

156 173 An Experimental Comparison of Discovery and Didactic Computerized Instructional Strategies in the Learning of Computer Program:4,6=g

Brian McLaughlin

Aion of Computer Research and Technology ::-Eitional Institutes of Health Bethesda, Maryland 20205

Abstract "is still in need of testing. But it This study compared the effectiveness of hypothesis of suchimportanthuman implicE.iut, instructional strategies for teaching computer thatwecannotafford not to test it...". Twenty programming. College students, pretested for Locus years of research have demonstrated the of Control and cognitive ability, were assigned to difficulties in testing Bruner's "hypothesis". Yet use short instructional computer programs many important issues raisedby the discovery - characterized by either a "discovery" approach or didactic controversy remain unresolved. By an expository "programmed instruction" sequence. examining and improving upon methodological All instruction and testing was administered by the difficulties encountered in previous discovery computer. In general, expositoryinstructionled learning research, this study aims to demonstrate to betterposttest recallof basic programming the utility of a rigorous experimental approach to factswhile discovery instructionresulted in this intractable area of educational research. better performance on extrapolation tasks and Most experimental research comparing discovery actual programming. Discovery instructionled to and expository instruction has used some variant of higher self-confidence about newly learned "example-rule" methodology. The learner is given information and a greater willingness to continue examples and must infer or "discover" the instruction. Students withan Internal locus of underlying rule. The narrow focus of this control performed better under discovery artificiallaboratory implementation of discovery instruction, while External locus of control leRrning has unreasonably constricted the students did better underexpository instruction. nperationalization of discovery processes. In this This interaction was remarkably consistent across 'tudy, a computer program was used to generate a eleven cognitive andaffective outcome measures. discovery 'Z'-.7.-ning environment whichprovided An instructional program based on this research has and had been implemented on an Apple micrwromputer. ol,portUnities for "messing about "7 sufficient openness to accommodate the fuller exercise of intellectual and cognitive capabilities Rationale implied in broader conceptions of discovery The psychologicaltheories of Piaget and Bruner suggest the value of exploratory learning11'14'6. discovery-oriented instructional experience. The Anothermethodological weakness hampering theories ofSkinner and Aue.,bel emphasize more previous studies has been the all-too-common focus systematically controlled expoeitory instruction. on a single instructional outcome, typically an Ambiguityand controversy hft?e long surrounded the achievement posttest. Proponents of the discovery comparison of instructional methods based onthese approach have suggested a tantalizingarrayof two schools of thought. A core epistemological resulting educational benefits. In this study, the rationale for the discovery approach is the effectiveness ofinstruction was assessed on a Piagetiannotion of a learner interacting with the variety of cognitive, affective-and motivational environment, .actively constrcting knowledge outcomes. through continuous application and reorganization A third methodological difficulty has been the of personal cognitive structures. In contrast, the need for more sophisticated hypothetical models of notionmore cftenassociatedwith the didactic complex instructional situations. As Cronbach3, 4 approach suggests a more passive, dutifullearner who incrementallycopies or absorbs new knowledge has repeatedly advised, it ie crucial to carefully from an outside source. These two approaches might describe the specific instructional elements being investigated, the type of material being presented, be broadly characterized .as differingin their the characteristics of the individual learners, and relative emphasis on "learning bY doing' versus "learning by being told." the nature of the outcome variables. In this study, specificinstructional strategies were Despite its far ranging implications for both implemented as specific algorithms and routines in educational theory and classroompractice, the an instructional computer program. Focusing discovery versus didactic controversy has been only on a single subject matter area (computer programming), minimally illuminatedbydecades of empirical an Aptitude by Treatment Interaction (ATI) design research. Bruner1 asserted his broad claims for a was used to assess the relative effectiveness of discovery learningapproach inthe spirit of an instruction on learners With differring individual hypothesis. In 1961 he stated that thehypothesis characteristics.

157 174 Method Aptitude by Treatment Interactions were A minicomputer was used to teach introductory interpretedusinggraphic representations and the computer programming to 81 college undergraduates Johnson-Neymantechnique8 '12. An exampleof such instructional using either a discovery'or didactic an interpretation is illustrated in Figure 1. strategy. Learners had no previous computer Posttest Effort represents the subject'sresponse The didactic strategy used programming experience. to a final rating scale: "How seriously did you The an expository programmed learning sequence. try to answer all the previous questions?". After question-answer-feedback format of this approach is beingresidualized on other significant predictors instructional representative of many current (in this case, a better(faster) training score instruction applications of computer assisted predictedhigher reported effort), Posttest Effort (CAI). The discovery strategy used a computerized was separately regressed on IE andAbility. The discovery learning environment incorporating resulting disordinal interaction obtained for IE is epistemological instructional heuristics and sketched in Panel 4 (bottomright quadrant) of of the discovery approach. The assumptions Figure 1. Construction of a Johnson-Neyman (p<.05 shortmodel discovery environment consisted of level) region of significance indicated that for which could be learner-modified and programs subjects with IE scores less than 8.9, discovery executed. During(deliberately slowed) program instruction was significantly more effective while execution, the computes screen displayed simulated for subjects with IE scores greater than 16.2, dynamic computertransactions. Theamount of Programmed Learning was more effective. learners personal initiative required of discovery Interpretationof significant three way was variedby prohibiting or encouraging learners interactions (Ability by IE by Treatment) was aided model to initiate their own changes in the by three dimensional figures. For example, the programs. In all discovery treatments, the visible results for Confidence in Right Answers (confidence executing programs immediate feedback provided by ratings after each posttest item were averaged variations provided the opportunities and program separately for items answered correctly and for learning. incorrectly) are, illustrated in Figure 2. approximately Instructional treatments (each Comparison of correspondingcorner points of the were compared for one hour nufstion) two surfaces in Figure 2 helps summarize the effectiveness on eleven outcome dimensions (derived character of the interaction: posttest items, from these: multiple choice 1. Discovery was more effective for Internal generation ratings, program tasks, confidence low Ability subjectsand External high Ability Continuing Motivation measures9, risk-taking subjects. measures, and general rating scales.) Two 2. Expository instruction was more effective individual differencevariables were examined for for Internal high Ability subjects. the possible Aptitude by TreatmentInteractions: (1 ) 3.The most strikingdifferencebetween 13 two surfaces was the exceptionally poor performance , ( IF " ); and Internal-External Locus of Control of Internal low Ability subjects underexpository ability ("Ability", (2) a measure ofcognitive instruction. based on the SAT Verbal score and an algebra word problem test.) Summary and Conclusions The differences among discovery treatments Data Analysis and Results were more suggestive than dramatic. However, these stepwise Data analysis was performed using within discovery comparisons did demonstrate the regression as suggested by multiple techniques practicality of comparing variations in the Cronbach and Snow4. There were four experimental discovery approach. Experimental control over treatments (N=20 for each treatment group): instructional elements ofdiscovery instruction 1. Discovery High Initiative (learnersmade offers numerous departures for further research and own changes) promises continuingrefinement of a computerized 2. Discovery Low Initiative (learners made discovery approach. only suggested changes) In contra& to withindiscovery comparisons, 3. Discovery Optional Initiative (learners the differencesbetweendiscovery and expository made own and/or suggested changes) programmed learning were substantial. These 4. Expository ProgrammedLearning (branching results can be summarized in five general programmed instruction sequence) conclusions. Treatment effects were partitioned into the 1. Expositoryinstructionled to superior three orthogonal treatmentcontrasts of most reception and recall of basic programming facts and theoretical interest. Independent variables were rules. Discovery instruction led to better entered in the prediction equation in a integrationof newly learned material as evidenced predetermined order: Ability, IE Locus of Control, by superior performancewriting programs and on pretest covariates (deri-edfrom an identical multiple-choice posttest items requiring training session given all subjects), treatment interpretation or extrapolation. It is interesting vectors, then second order interaction vectors, and to note that if all cognitive outcome items had finally third order interactions. Using this been lumped together as a single achievement "step-up" regression strategy, the increase in R posttest, the differential performanceoneach squared due to each successive term was tested for subskill would have gone undetected. The failure statistical significance with F ratios constructed to separate posttest items of different types may using the error term of the final full model. The beone contributingreason to the dearth of results are summarized in Table 1.

158 .17, Table 1 Increase in Percent of Outcome Variance Accountedfor by Each Term (Values Below 1.0% Not Displayed)

Outcome Full Measure P N Model A I P TI T2 T3 AT1AT2 AT3 IT1 1T2 IT3 AI AIT1 AIT2 AIT3

Neat 8130.6P16.74 XXX 4.1P Transfer 1.7 3.1 XXX XXXXXXXXX 1 Far 8132.3q02.7q5.7p5.7P XXX Transfer 8.3q2.6 XXX XXX XXXXXX

Gen 32.1q125.7q XXX 4.5P 81 XXX XXXXXXXXX

24.5 7.0P onCiPageMa 78 9.1q 1.7 1.7 1.9 1.2

;A b oaihteiss 7641.7q 16.6q 7.24 6.94 2.1 4.6P

34.0p 11.84 3.4 inVOge. 76 2.0 9.51; 2.8 XXX XXX XXXXXX

Risk Taking 8138.8q ;17.2q XXX 2.8 1.6 6.4P 2.9 3.7 a PffliFit 7429.81)1 4.7P 5.4P 1.1 2.5 12.4qXXX XXXXXXXXX c CM 78 14.2g 32.4P 4.6P 1.44.2 2.5 2.3

7825.6 2.1 2.9 4.6P Inforlig2Rgi 9.841.7 3.5 XXXXXXXXXXXX b TakUppellar 7631.1p 5.6P 1.64.8P2.33.9 1.73.6 4.8PXXXXXXXXXXXX

8130.6P 1.2 XXX 1.0 2.6 Pcate 5.5P6.7P 14.0 1.9 1.8 5.9P q Instruction 6034.IP15.5 XXX 5.2 1, lime 1.35.63.21 2.2XXXXXX XXXXXX

Significance of F Test : p=p<.05 qr.v.01 ma = Not in Model A = Ability I = Internal-External Locus of Control (IE) P = Predictor: a=Training Score b:Pace c:Key Count T1 t. Nigh Initiatve Discovery versus LowInitiatve Discovery T2 Optional Initiative versus Mean of High and Low InitiativeDiscovery T3 t Programmed Learning versus All Discovery Figure 1

Separate Regressions of Ability and IE on Posttest Effort.

Within Discovery Prgrm'd Lrning vs Discovry 6.0 t + i

i 5.3 E 4 1 F DO 1 1 PL D Fo 4.5 pg+ NI o'

R T PLI 3.8 4 t

1 1 ++ + + + + + + + 2 1 X +1 +2 2 1 X +1 +2 Ability Ability

Within Discovery Prgrm'd Lrning vs Discovry 6.0 t

5.3 t F DH t PL DI F BE o 4. . N .

1 3.8 pLf

. , .

. + + + + 1 + + + + 2 6 10 14 18 2 6 10 14 18 (Int) IE !(Ext) (Int) IE (Ext)

Posttest Effort: high="tried hard to answer all questions correctly" low=udid not care about answering questions correctly"

Each X and Y axis unit = 1 standard deviation Int = Internal Locus of Control Ext = External Locus of Control DH = Discovery High Initiative DO = Discovery Optional Initiative DL = Discovery Low Initiative PL = Expository Programmed Learning D = Combined Discovery (DH, DO & DL)

160 Figure 2 Regression of Ability and IE on Confidence on Right Answers

FroBrammed Learning Combined Discovery

6.5 6.5 65 3 0

5.5,t 55

4.50 45

43.6 C 0 U U 2.6

k,11110,410

,4$tot°

Sot'

Resealed regression equations: Programmed LearningY' 4,118g +.85Ability +.191E -.44Ability x IE Combined Discovery Yis4.63 +,27Ability -.O1IE +.17Ability x IE

Note: each scale unit represents 0.5 standard deviations.

178 consistent findings in previous dtudies of 2. Cronbach, L.J. The logic of experiments on discovery learning. The results of this study discovery. In L. Shulman & E. Keislar (Eds.) 10 Learning by discovery: a critical appraisal. strongly support the work of Mayer in emphasizing the differential structure of cognitive outcome. Chicago: Rand McNally, 1966. 2. Diucoveryinstruction led to higher self-confidence ratings, but only on correctly 3. Cronbach, L.J. Beyond the two disciplines of scientific psychology. answered posttest items. This is interpreted as The American Psychologist, indicatingsuperior cognitive discrimination after 1975, 30, 116-127. discovery instruction. Thiscritical ability to 4. Cronbach, L.J. & Snow, R.E. discriminate levels of certainty aboutnew Aptitudes and instructional methods. knowledge is seen as an important (and neglected) New York: educational outcome. Irvington, 1977. 3. Discovery instruction led to higherlevels 5. DeCharms, R. of expressed interest in continuingcomputer programming instruction inadditional educational Enhancing motivation: change in the classroom. New York: Irvington, 1976. settings. This supports Maehr's9 conceptualization of Continuing Motivation as generalizing beyond the 6. DiVincenzo, R.M. Forming a theoretical synthesi3 liking of instructional method to increased for viewing discovery learning instruction. interest in instructional content. This finding School Science and Mathematics, 1980, 80, 218-226. also supports the motivationalimplications of discovery learning as expressed by other 7. Hawkins, D. Messing about in science. 1,5 researchers Discovery instructionappeared Science and Children, 1965, 2, 2-6. more capable of "turning-on" learners than did expository instruction. 8. Kerlinger, F.N. & Pedhazur, E.J. 4. The results on affective and motivational Multiple regression in behavioral research. New York: outcomes indicate that lowerAbilitylearners Holt, Rinehart, & Winston, 1973. benefited more from discovery instruction than from expository instruction. Additionally, in contrast 9. Maehr, M.L. Continuing motivation: an analysis of a seldom considered educational outcome. to many previous findings, lowerAbility learners showed the same level of cognitive achievement Review of Educational Research, 1976, 46, 443-462. (relative to higher Ability learners) regardless of 10. Mayer, R.E. Different problem-solving discovery or expository instruction. Thus this study suggests that discovery can beaneffective competencies established in learning computer and promising strategy. for teachingcomputer programming with and without meaningful models. programming to lower Ability learners. Journal of Educational Psychology, 1975, 67, 5. The results for IE Locus of Control showed 725-734. that Internals did better under discovery 11. Papert, S. instructionwhile Externals did better under expository instruction. This interactionwas Mindstorms: childrent_computers, and powerful ideas, New York: particularly evident on motivational andaffective Basic Books, 1980. measures but was remarkably consistent across all 12. Rogosa, D. Comparing nonparallel regression eleven outcomes. These findings confirm the lines. relevance of Internal-External Locus of Control as Psychological Bulletin, 1980, 88, 307-321. an important variable in educational research. 13. In summary, this findings lend definite Rotter, J.B. Generalized expectancies for internal versus external control of reinforcement empirical support to important expectations of and decision time. discovery learning theory.Despite the relatively Psychological Monographs, 1966, short duration of instruction, a substantial number 80, (No 1, Whole No. 609). of significant treatment differences were observed. The use of computerized learning environments is a 14. Strike, K.A. The logic of learning by -- discoVery. flexible and promising research approach for the Review of Educational Research, 1975, controlled investigation- "of instructional 5, 461-483. processes. An instructional, program based on this research has been implemented on anApple microcomputer. The program, named Sherman, provides self-instruction in the Pilot programming Note: this article is a summary of the author's language using a discovery oriented approach. It doctoral research at The Catholic University of requires an Apple He computer or.an Apple II with America. The dissertation based on this research 64K and an 80 column board. was among five finalists for theAmerican Educational Research. Association's Outstanding DissertationAward for 1982. The instructional References computerprograms used in this study were originally 'developed by the author to run on a 1. Bruner, J.S. The act of discovery. PDP-11/34 minicomputer. Harvard Educational Review, 1961, 31, 21-32.

162 17J CHECKING LAB CALCULATIONS

William F. Pelham

Physics Department, Towson State University, Baltimore, Maryland 21204

Abstract mind. A commitment was made to develop systematic procedures and computer programs that would Computer programs were written, as APL func- check lab calculations. The remainder of this tions, that checked every single calculation made paper will describe the general plan for doing as part of the required laboratory work in a two this, pedagogical decisions made, several results semester General Physics course. Both the values of a six semester trial, and some conclusions. and the roundings of the results of the calculations were enamined. Students reported individu- General Plan ally, in an interactive mode, and received immediate evaluations; records were kept of their Every single calculation performed by every entries and successes. A description of the work, single student would be checked for numerical the bases of the programming, and some results of correctness and proper expression. Proper a six semester trial are given. expression means the correct number of significant digits in the result of a computation as determined by the rules of significant digits and the Students attending a required laboratory uncertainties of the contributors to the computa- session in General Physics courses are presumed tion. to know the purpose of the work to be performed, to have read the directions, and to understand In an interactive mode, then, a student would the theory behind any calculations to be made. enter raw data, carrying only one uncertain digit, Typically, some instructions are given them and the results of the requested computations. before starting work, particularly on the use of Thn computer would perform the same computations equipment, and the experiments are then conducted. and compare its results with those of the student. After the work is completed, a written statement,_ Questions requiring discursive answers could be following the raw data, giving the "results" of askai at appropriate points. When all questions the experimental work is prepared in the labora- were answered, or the student elected to stop, the tory notebook. The notebooks are periodically computer would prepare a summary of the student's submitted for evaluation, and the contribution to success and it, along with everything the student the course grade of the laboratory work is entered, would be stored. The instructor may largely based on these evaluations. examine the stored material for grading purposes and for helping students see where they went Almost all the laboratory exercises involve wrong. using the data in calculations; these vary in complexity, but are usually quite simple in form. Pedagogical Decisions It is in doing the physics, relating the data to the physical concepts being used, that students Before programming could start, certain ques- have trouble. Using the wrong force, or the tionsneeded answers. The list that follows gives wrong current, for example, will give numerically most of the quesOons and the answers used in wrong results that may not look suspiciously dif,. guiding the construction of the programs: ferent from the right ones. 1. When in the computing session, if ever, When the notebooks are evaluated, often by should a student be told whether his and the com- student assistants, every calculation is not puter's results agree? Immediately? At the con- repeated and there is, thus, the possibility that clusion of the session? erroneous results, derivinr,, from faulty physics, Answer: Immediately; first, whether the will remain undetected.The power of the labora- numerical result is corrert and then whether there tory to reenforce and explicate the lecture can is a significant digit error. be, clearly, compromised. A way to minimize undetected calculation errors is to have a com- 2. Should a student be given a chance to puter perform the same calculations, using the correct an erroneous result, and if so, how many student's own data, and compare results. This times? gives an immediate evaluation and has the advan- Answer: Yes, unlimited times. tage of doing it while the work is still fresh in

163 18o 3. Should the student be given the computer's functions, a kind of dense pack of instructions, results? albeit sometimes hard to read in detail, but rela- Answer: No for the numerical part, yes tively easy to follow as far as the flow of thought for the rounding, but only on request and with a is concerned. concomitant loss of credit. As hinted above, each student entry was 4. May a student redo all or part of the examined for the number of significant digits and report during the current session or a later one? the place value of the uncertain digit.Up to 5 Answer: Either way. A count will be items (an arbitrary decision) were entered on one kept of the number of recordings of the report. line, so entries were stored, temporarily and in a file, as 6 column matrices. Each calculation got 5. May a student stop before a report is its own row. One matrix was used to store data completed and finish it at a later date? and results, one to store numbers of digits, and Answer: Yes. one for place values; these were catenated into a 3 dimensional array before storage in an APL file. 6. Should there be a penalty for late reports? Security was maintained by having the func- Answer: Yes. tions that read from and wrote into 'the file locked. Locked API. functions cannot be listed by 7. Should the computer attempt to provide anyone including the locker. The file name, con- instruction? sequently, remained unknown to the students. To Answer: No. know what was in the locked functions meant that unlocked copies had to exist somewhere out of stu- Programming dents reach, and.they did. To get at them required a knowledge of 4 different keys and the The entire operation was to be run on the instructor's personal ID. If one was not experi- available UNIVAC 1100/10 computer through dial-up enced in APL, it also meant the ability to decipher access with DECWRITER II terminals; the terminals the UNIVAC APL Users Guide. were not, and could not be, located in the laboratory. APL was to be used to write all the pro- The master functions, one for each lab exer- grams (henceforth called APL functions) because it cise, were stored as elements in a UNIVAC program seemed especially convenient for the task being file and written with digraphs rather than the considered. Doubtlessly, the special features of special APL symbols. This was done originally to APL affected the programming strategies used, but avoid filling an APL workspace, but turned out to that is probably the case whatever the programming be advantageous in that the UNIVAC Editor could be language. used to correct and update functions, a much more efficient method than using the available APL Some features of APL and how they were function editing procedures. When it came time for employed will now be given: a particular master function to be used, it would be added to the runstream, inside the APL pro- a) The APL environment itself, with its cessor, as a file element with the UNIVAC ADD com- ability to have numerous functions, sitting in an mand. Since an APL command was also in the file easily retrievable workspace, that can be called element, the reporting process was initiated auto- as utility functions by a master function. All matically upon completion of the ADD. the digit counting, rounding and file usage were handled this way. A sample of a student report together with a few annotated lines of programming will be found b) APL random access files into and from in the APPENDIX. which multi-dimensional arrays can be written and read as a single variable. Each student's report Results of Trials was stored and retrieved, with one instruction, as a 3 dimensional array containing 3-2 dimensional As the reporting system began to be used by matrices of, typically, 26 rows and 6 columns each. the first of the total of 350 students, unsuspected problems arose, some of which could be corrected c) The ease of locating and selecting data and some of which await solution. Perhaps the elements from both character and numeric vectors most frustrating occurrence for students was to and matrices. Each student entry was a character make a typing error of a sort that would cause the vector and it was examined for errors, and the execution of the function to be suspended; this digits of each number counted before conversion, happened often in the beginning. In APL, if a via the execute operator, to numeric data. The statement cannot be executed, an error message is summary of the student's success was also simpli- printed and the number and content of the unexecut- fied by these features of APL. able line follow. The system then waits. Things are not stopped, however, merely suspended, and d) The ability of APL to perform parallel the function can be restarted at any line with an processing facilitates the entering and handling appropriate command.All variables remain intact. of repetitious data. When 5 runs were made, all 5 Attempts to instruct the students how to restart calculations were done simultaneously with one the function were not completely successful. (They instruction, greatly shortening the function. were to type $GGOBACK whereupon the responss would Indeed, the compactness of APL results in short be "To Where?" Their reply was supposed to be the

164 181 question number to which they wished to return.) Conclusions The main problem seemed to be a lack of recognition that the function had been suspended. None Some of the answers to the pedagogical ques- of the students was familiar with APL.A special tions given in an earlier portion of the paper may error detection function was written to help this have to be changed. Unlimited ability to make problem and it has somewhat. Other solutions will changes, in particular, seems to be in need of be attempted. revision. This is what ultimately uses the students' time; perhaps an automatic review of Another unexpected problem was that many stu- entries or a referral to the instructor after a dents refused to believe the computer. They were few tries is what is needed. right and the computer was wrong.They would repeat the same entries over and over sometimes Writing the functions and continuously with slight changes. They would claim that their improving them is sometimes tedious, but is not lab partners had gotten the exact sequence of unpleasant and is rec3mmended. A rethinking of entries declared correct that, for them, were the entire scheme seems now due, particularly with declared wrong; and reams of paper, a view to making it easier to add and subtract they would finally slap the print-out down on the questions; if other faculty members, not adept in instructor's desk and essentially dare him to APL, are to use the reporting system, it is essen- prove them wrong. He virtually always would and tial that tinkering with the questions be as easy it was almost always because the students had made as possible. physics or digit mistakes. Merely checking the results of calculations The above circumstances inevitably arose with is, obviously, not the whole of evaluating a stu- the less able students. The good students, who dent's learning in required laboratory exercises. had everything right before they started, breezed Other aspects of laboratory work (preparedness, through the reports. The unsure students would quality of work, follow-up) can be part of a com- start skipping about, calculator in hand, changing puter managed laboratory. The insistent nature of earlier entries until, after a while, they lost the computer's responses and the motivation to suc- track of what data the computer was using to make ceed that it induces suggest that learning will be its calculations and consequently accused it of improved through computerization of laboratory being in error. By the end of the second semester work. To have a "go-no go" test of preparedness of the course, the general level of the discipline to perform a lab would, however, require on-line necessary to track through the print-outs had computation in the laboratory and, most likely, a improved, but the full acceptance of the need to so-called "open" laboratory schedule; i.e., a be completely logical about a sequence of steps schedule that permits students tc work in the was not universal. laboratory at virtually any time. The work being reported can easily be adapted to the open lab con- Another unwholesome situation developed when cept and was undertaken with that possibility in % differences were being calculated; students were mind. instructed to express the results of such calculations to only one digit (the fact that they would Appendix round 0.62% to 1% and throw up their hands at 13% didn't help matters). When the computer would First, some printout of a typical student tell them for, say, the fourth time that their 5% interaction. Q1 is straight data input, Q2 is the was wrong, they would resort to trial and error: result of a calculation, correct, in this case, Q6 first 4, then 6, then 3, then 7, and so on. Some- shows some student errors and Q7 shows the reintro- thing will have to be done about this. duction of previously entered information.

Good things happened, too. Misunderstandings > @ADD CML.REPORT20306 were brought to light, even with better students, ap1/1100 level 7r1bv2 tue 03/01/83 16:20:39 that the reading of a lab notebook would not have clear ws detected. There was a drive to get it right, good day garrett , if you are all instilled by interacting with the computer, that a set to report a lab, type ok, if not, type stop. red X in a notebook would not create. Consulta- >OK tion with the instructor to see why the computer it is 2 days after the lab was performed, and you called them wrong, and the subsequent acceptance are: of the correct result provided considerable satis- on time in reporting. faction to many students. this is to report lab 6 type done to stop, delta to change place. Upon informal inquiry, it seemed that the type y to have sigdigs evaluated, n not to (with better students liked reporting their results on lower grade). the computer. For students poor in physics and/or >Y for those few who seemed permanently bewildered by Q1:first, we'll check the calculation of the the process of using the computer, computerized friction lab reporti;ly cost them hours of time; a reduction force. enter fup, fdown (in newtons) and theta, in of this waste is needed and will be attempted. radians. >1.62 .539 .524 Q2541:give the friction force in newtons. >.

165 182 calculations all ok right and returns the rounded number. Lines 86 sig digs all ok through 92 are doing the same calculations on two Q6: give the change in potential energy, in j, as numbers simultaneously. the block moves between the two beams, run 1, run 2. >.2320. .2020 incorrect values are 2nd do you want to reenter it (them) (y or n)? >Y >.222 there is a sig dig error in the 2nd entry(ies) do you want to replace it(them) y or n? >N do you want to know the correct form(s), y or n? >Y 2nd entry should have 4 digits. Q7: and now the ke change + the friction work, in j, run 1, runt. >OLO print old entry? y or n.

2.200,1 2.20e&l calculations all ok sig digs all ok last chance to make a change. type done or delta. >DONE done--your work has been recorded you will now be automatically signed off apl. sign off system or @add report another lab. apl terminated

Next are the master function lines associated with Q7.

80:Q7:'Q7: AND NOW THE KE CHANGE + THE FRICTION WORK, IN J, RUN 1, RUN2. 81:PL$S'7' 82:NUM$S2 83:ENTER 84:$QACT 85:ANS1$SKF$SENT 86:ANS$S-/PQ$S.5$XSLM[3]$X(SLM[2]%1[2 1])*2 87:ANS$SANS,-/WR$S.5$XSLM[3]$X(SLM[2]%T[4 3])*2 88:ANS$S ANS+FF$XSTAR 89:i$S$C /PQ RNDNP $L/MT[4;1 2],MT[3;2] 90:N$SN,$C/WR RNDNP $L/M1[4;3 4],MT[3;2] 91:WW$S(FF$XSTAR) RNDNP MT[2;1]$LMT[5;1 2]

92:ANS$SANS ROUNOP N$SWW$CN . 93:ITEM$S8 94:RETEST 95:N SIGCHECK 'AA' 96:KF$SANS1

Bear in mind,that APL goes from right to left. Line 85 is the student's entry called KF and ANSI. Lines 86, B7 and 88 calculate the "correct" result from data entered earlier; it is a 2 element vector, unrounded, called ANS.

Lines 89, 90 and 91 compute rounding information and line 92 rounds ANS correctly. The function RNDNP takes the quantity to the left of its name and rounds it to the number of digits to the right of ics name ($L/ selects the smallest number of digits from the group of 3 to its right); the function returns the place value of the rightmost digit of the rounded quantity. ROUNDP rounds the left quantity to the place value given on the

166 183 TEACHING UNDERGRADUATES TO THEORIZE THROUGH THE USE OF A COMPUTER SIMULATION OF KIDNEY FUNCTION.*

by David L. Wilcox

Biology Department, Eastern College 3t. Davids, Pennsylvania

Abstract A partial solution to this problem may be the use of simulations in investi- Using a simulation of osmoregulat- gative labs. Simulations provide a power- ion, ninety seven physiology students in ful substitute for certain investigations four different classes engaged in student for which one does not have the time, equip- research projects, deducing model struct- ment, expertise or moral right. Due to the ure through open ended experimentatidn. speed of the process, and to the limited They each developed theories, designed number of variables which can interfere experiments, interpreted results, and with the experimental process, the computer reported to the class in journal format. simulation of biological systems may be Each class made considerable progress in used to give the student experience in the deducing model structure, developing such process of scientific discovery within the typical traits of the scientist as doing time constrants of a single course. Such extra (no credit) experiments and part- open-ended experimentation introduces stun isanship of certain theories. They de- dents to scientific thought at a level not bated their opinions concerning kidney usually experienced before graduate school. control mechanisms with great enthusiasm, In a time of limited educational funds, and they learned a considerable amount simulation may be the only affordable way about kidney function. to allow students to repeat an experiment and learn from their errors3. Simulations may be used as laboratory Introduction: Creativity and Simulation exersizes in various ways, ranging from simple demonstration to the design of new Often we are victims of our own mathematical models. A few notable examples success in the teaching of science. Be- include: the demonstration of system con- cause the current "state of the art" in trols, the deduction of system relationships, science changes so rapidly, we spend most an effort to control the system, the deduct- of our time explaining current formulation of parameter values, the identification ions, and our students know little of the of unknowns, the upgrading of present models, thrill of discovery. However, a number and the design of new models. of writersl have demonstrated that the quality of independent student investi- The Design of the Kidney Simulation gation such as publications, senior theses, or even science fair projects, are dis- The project being reported in this pa- tinctly better predictors of students' per used a simulation of mammalian fluid future productivity in their discipline, balance as a system for open ended labor- than such, traditional measures of student atories. The program was originally de- ability as the SAT, GRE, or college grades. signed to run on an Alpha-Micro mini-com- Directed student research is, of puter with time sharing capacity, using course, costly both in institutional funds five CRT terminals and a TI 810 printer. and in instructional time, and it may re- Each CRT is assigned 32,000 bytes of dynamic duce the content which students will mast- memory. In writing the "Kidney Project" . er. It seems likely, however, that net- glecting of the needs of our more creative simulation, I used the MENTOR simulation students may be even more costly, and may system, designed at Eastern College. (A reduce the number of really creative short version of MENTOR has recently been students who choose to enter or remain in written for the Apple II.) MENTOR consistsfq science2. of a modular base upon which models using L sets of difference equations may be sim- * Under partial support of N.S.F. LOCI ulated. Besides its availability to us, its grant: SER 79-00115. most important advantage was its very high

167 flexibility in experimental design, nec- independently designed, ran, and "published" essary if students are to effectively form an experiment in an in-house journal,"The and test their own hypotheses. In MENTOR Kidney Project Record". Each paper thus simulations, the user (student) sets the became primary literature for subsequent length of the experiment and the time in- student research. Students could also terval between observations, chooses the write "letters to the editor" for extra observations, changes the number and the credit. Following each set of papers, the size of the experimental groups, sets class discussed their current understanding times at which values may be changed, and of the model. assigns the values for any and/or all of the simulated system's parameters. This Results of the Use of the Kidney Program flexibility allows an almost infinite number of possible experiments. In addition, Variation in and between classes the MENTOR system supports its simulations usually makes the evaluation of an educat- with a package of 'utility' programs to ional innovation very difficult. However, analyse and graph data, advise on exper- this project produced products which could imental design, and store or copy data be analysed and evaluated as concrete files. Additionally, the interactive entities, student scientific papers. The nature of the system enables a student with results clearly demonstrate the scientific ten minutes of introduction to work on a process going on in the class. simulation without supervision. Fig. 3 shows three figures from a The computer program, "KIDNEY", sim- student's papers (Brauch, 1979), summariz- ulates a terrestrial mammal's homeostatic ing the current thinking of 'the class at controls of fluid and salt levels. As the 3 points during the semester. Her diagrams compartment model in Fig.1 shows, the clearly show the development of theory in model has four interacting "flows" of the group of students. There was an aver- material: the two substances controlled, age of eight new relationships"discovered" fluid volumes and amounts of sodium; and with each new set of papers. Although the two controlling hormones, anti-diuretic total knowledge increased, it was not by hormone and aldosterone. I assumed that a smooth accumulation of data. Even in the water reabsorbed from the nephron this simple system, students sometimes comes from two locations, the convoluted disagreed in their interpretations of ,tubules (including the aldosterone sen- data, and there were some "contradictory" sitive site in the distal convoluted results.. Even "test" experiments might tubule) and the collecting duct. Sodium not resolve debates, and new parameters and reabsorption I assumed to be a function complexities were discovered. only of the first location. A good illustration is the changing Fig.2 shows the major variables of view of the relationship between intake the model linked in a signal flow diagram. rate and the blood volume. The student Flows and levels of sodium are given in researchers initially thought this a percentage (meq.). Cardiovascular con- simple, direct relationship (Fig. 3a). ditions control hormone synthesis, and the By the second paper this consensus had hormonei in turn control reabsorption. broken down, and for most of the rest of Clomerular filtration oas a function of the semester my class had two "schools blood volume and several other cardiovas- of thought" about the effect of fluid cular parameters. intake rate on blood volume (Figs. 3b The model has 41 parameters, and can and 3c). The confusion came from neglect be reduced to eight differential equations of the indirect effects of salt intake vs. (Appendix 1) representing the blood levels, water intake. Instead of immediately the amounts in the intake and urine col- correcting them,I let them argue. Fig. 4, lection "jars", and the hormonal levels. which shows the final understanding of Observable variables given students include several of the students, demonstrates the the sodium concentrations and fluid volumes results. Eventually, most of the class of the intake, the blood, the glomerular was convinced by the "negative feedback" filtration, the flow in the convoluted school (which was correct). The speed of tubules, site specific reabsorption and simulated studies allowed the experimental the urine. The model also includes di- process to continue long enough to be self urnal changes in CFR, and tolerance limits correcting. This pattern of changing for blood volume. thought is remarkably similar to the ideas KIDNEY has been used to allow my of Thomas Kuhn on the process of science4. physiology classes to pursue a semester By the end of the study when the ion, class research project. Each class illustration in Fig. 4 were prepared, most had the goal of deducing the structure and of the students had learned to weigh data the control loops of the kidney model. and reject useless hypotheses, as seen Starting from scratch, every conclusion in the reduction of total complexity be- had to be supported directly from their tween Fig. 3c and Fig. 4a, b, and c. A data. At monthly intervals, each student few students, however, seemed unable to do

168 185 B -GFR-->TV-->UVkj...

PTR-

CDR

BNA-1GFNA-tAR2NAtii

ItPTR-

-SADR*ADH-LADH* -SALDO)ALDO-LALDO-*

Fig, 1 A model of Kidney Function: Fig. 2A Model of Kidney Function: Material Flow Signal Flow

BV total blood volume BV = total blood volume IV = rate of water intake IV = rate of water intake GFR glomerular filtration TV = tubular flow rate UV urine production rate B% = percent blood sodium BNA = total blood sodium I% = percent sodium intake INA . rate of sodium intake ALDO = aldosterone level GFNA tubular load of sodium JAR = volume in intake jar UNA a sodium excretion GFR = glomerular filtration ALDO aldosterone level UV . urine production rate SALDO . ALDO synthesis rate T% = percent tubular sodium LALDO . ALDO degradation rate U% . percent urine sodium JAR V volume in intake jar ADH = antidiuretic hormone level JAR V . volume in collection jar TV . tubular flow, rate PTR isotonic reabsorption CDR . water reabsorption JAR NA = sodium in intake jar JAR NA . sodium in collection jar PTR = reabsorbed sodium ADH . antidiuretic hormone level SADH = ADH synthesis rate LADH = ADH degradation rate

iG (.0 Stan,r,Nary ePCkee

NQIns} 3 et-fF 11. J: \o- :$ it / / I 'Bloodvet. 0--) -1.. - - - / d;adireerrAni i I TZA....ia.- I / I V I Fit t...k. Or.e 0k...rear< kg r I /4. + -41--- I I th;nc No.t 4 i I r- ft . . , \ .. \ Wievican Ira uck ,.\ A.,,:.!env.,I .\ % r / Adosvona t' .1 ADR Fig. 3 Changing Views of Kidney Function in a Physiology Class Using Open- Ended Simulation to Deduce System Function. (Figuresfrom student papers)

a. Summaryof10studentpapers: Oct.15 b. Summaryof10studentpapers: Nov. 1 c. Summaryof10studentpapers: Nov.21

170 (a)

Fig.4 Four Final Views of Kidney Function iu a Physiology Class Using Open- Ended Simulation to Deduce System Function. (Figures from student papers)

171 188 this (Fig. 4d), retaining most of the increases exponentially, it becomes in- postulated relationships of Fig. 3c. creasingly difficult to train new scientists. Although the class developed a generrl Equipment costs more, the "core" information consen,:us concerning the model kidney, is far greater, more connections beiween areas of disagreement still existed at the disciplines are evidently necessary, and time we terminated the project with a dis- our educational resources are reduced. As cussion of the real model (Fig. 2). As a result, students today are unlikely to be an example, consider the four views of able to use the tools of their discipline the direct control of ADH expressed in at their own discretion until they complete Fig.4. One student (Fig. 4b) was correct, their terminal degree. one added two non-existant or remote re- These problems can be met in part by lationships (Fig. 4a), and two of them the judicious use of computer simulation. (Fig. 4c and d) missed the direct link Computers may simulate expensive equipment, entirely. If the project had continued, both training students in its use and sim- this disagreement would have led to debate ulating its output for student analysis. and research aimed at discovering the roles Computers may also effectively demonstrate of blood volume and blood sodium in trigg- complex phenomena, including cross dis- ering ADH release. ciplinary topics, which include processes The last observation is of great im- happening in time dimensions impossible for portance for evaluating the use of sim- the laboratory. In addition, computers ulations as student research systems. By may give the student looking for a career the end of the semester, more than forty a taste of the investigative process, and individual investigations had been re- perhaps attract them into science. Last, ported without completely explicating the but surely not least in the America of dynamics of the model system. In fact, the 1980's, all this may be done at a the class never moved beyond signal flow reasonable cost. There is no need to use diagrams to the next logical step, writ- a large main-frame computer for effective ing their own model equations. student oriented simulation. With care, Model systems of the complexity of even the "home" microcomputers can produce the kidney model (six differential surprisingly sophisticated results. equations) are very unlikely to be exhausted by student research. Such complex models will be as opaque to the Bibliography student's understanding as the biological systems on which they are based, and if 1. Postlethwait, S. N. "Improvement of accurate, will be equally beneficial in Science Teaching." Bioscience 30:601- teaching the dynamics of the system. In 604; September 1980. fact, since the student will be able to 2. Anderson, David E. "Computer Simulations observe at least ten fold more events, in the Psychology Laboratory." many of them impossible in a student Simulation & Games 13:13-36; March 1982. laboratory due to constraints of budget, 3. Wallach, Michael A. "Tests Tell Us time, equipment, and technique, a well de- Little About Talent." American signed and accurate simulation could be Scientist 64:57-63; January 1976. an even more effective experimental system 4. Kuhn, Thomas S. The Structure of than the real thing. Scientific Revolutions The University of Chicago Press, Chicago: 1970. Conclusions This report demonstrates how a simulation of a biological system, such as osmoregulation, may be effectively used in the teaching of physiology as a system for open ended student research projects. Students were able to conduct significant research as part of the course laboratory, and they were able to analyse their data and synthesize their new experimental knowledge into significant scientific models. In the process they evaluated experimental results which would have been impossible for undergraduates under traditional laboratory conditions. Beyond the limited use of computers in specific disciplines such as Physiology, I believe this study may also have significance to a wider frameof reference, i.e., the whole process of science education. As our knowledge of the universe

172 Appendix Parameters of Osmoregulation Model

Model of The Control of System Osmolarity Symbol ValueDefinition

1. Equations of water balance IV 20 Intake rate Jar - Jar - IV Mi 150 Maximum intake rate IV - Mi /(1 + K1 / ADH + BV / K2) ADH 100 Unite of ADH in blood GFR - TPR / BV BV 1000 Blood Volume TV = GFR /(1 + ALDO / K3) x 0.8 Jar 1000 Amount in the water jar BV = BV + IV - UV + Met TV 120 D.C.T. flow rate UV = TV (ADH / Cl + K4) / (ADH + K4) TPR 0.2 Total peripherial resistance ALDO 100 Blood level of Aldersterone 2. Equations of salt balance GFR 160 Glomerular filtration rate INa - 1% x IV UV 20 Urine production rate BNa = BNa + INa - UNa U% 0.05 Meq. Na of urine UNa - TNa = T% x TV B% 0.1 Meq. Na of blood T% = B% / C2 T% 0.03 Meq. Na of tubular fluid U% = UNa / UV I% 0.05 Meq. Na of intake fluid

B% .., BNa /BV K1 150 ADH level at which IV=M1/2 K2 500 BV level at which IV -M1 /2 3. Equations of hormone level K3 17.65 ALDO level:GFR/2 returns to BV ADH - ADH + Sadh - Ladh C2 12 Conc. effect: collecting duct ALDO = ALDO + Saldo - Laldo K4 175 ADH level: TV/2 returns to BV Sadh - Mh /(1 + (K6/ B%) E10) Cl 3 Dilution: Distal cony. tubule Saldo = M1 /(1 + (BV / K5)E10) Ll 0.75 Z loss of ALDO per hour Ladh = ADH x Lh M1 150 Maximum release of ALDO Laldo = ALDO x L) K5 1000 BV level: ALDO release=M1/2 El 10 Intensifier of ALAO response Lh 0.75 % loss of ADH per hour Physiology Simulation Usage Mh 150 Maximum release of ADH KG 0.1 B% level: ADH.release = Mh/2 Semester Course Stdt. Hrs. Hrs/Stdt. E2 10 Intensifier of ADH response Fall 1979 Gen. Phys. 10 120 12 INa 2 Total meq of Na in intake Fall 1979 Human P&AI 22 110 5 BNa 100 Total meq of Na in blood Spring 1980 Human P&AII 20 140 7 GNa 16 Total meq of Na in GFR TNa 3.6 Total meq of Na in tubule 1979-1980 3 courses 52 370 7 UNa 2 Total meq of Na in urine Riso Reabsorption: P.C.T. RH2O Reabscrption. Collecting duct Fall 1980 Human P&A 117 54 3.2 RNa Reabsorption of Na (total) 8.4 Spring1981 HumanP.SAII 17 142 BP 100 Mean blood pr. ssure 25.3 Spring 1981 Gen. Phys. 11 278 UJar 0 Volume: urine collection jar 1980-1981 3 courses 45 474 10.5

1979-1981 6 courses 97 844 8.7

173 MICROCOMPUTER-BASED DATA ACQUISITION '21i NEUROBIOLOGY

by Richard F. Olivo

Department of Biological Sciences, Smith College Northampton, Massachusetts 01063

Abstract neurophysiology course [1]; since that time, we have used the system for three A microcomputer system for capturing years, the analog interface and the transient analog signals and displaying software have each been redesigned, and I them repetitively on at oscilloscope has can report now on what I regard as a been developed for student use. The tested and effective system for hardware 13 based on Rockwell's AIM-65, computerized data-acquisition. Although supplemental memory, and an analog-digital our system was developed for interface that is described. The neurophysiology, it can be used for any software, in ROM, uses single keystroke data-acquisition task in which the commands, any of whic..1 may be entered at sampling interval is between 100 any time. Input modes include continuous microseconds and 65 milliseconds; this input, triggered input, averaging, and encompasses a wide range of laboratory rate histograms. Display modes include applications. I would be pleased to make scrolling and jumping,' both of which can the software available to colleges or be frozen or have their direction universities that wish to duplicate our reversed, and slow output to a chart data-aquisition system. recorder. Intervals between samples are resettable in the range from 100 usec to Hardware 65 msec. The system has been used for three years; students have found it At the time I designed our system, helpful, consistent, and easy to control. Rockwell's AIM-65 microcomputer was the most cost-effective choice for our introduction purposes. We wanted six set-ups, one for each student group in laboratory, and thus Many traditional laboratory courses the cost of each one was very important. can be made slightly easier or a bit more Although a number of low-cost productive through the use of micro- microcomputers are now available, the computers, but a neurobiology laboratory AIM-65 remains a good choice because of can be substantially transformed by its particular combination of features. computerization. Neurobiology students It includes a full keyboard for entering typically record sequences of action commands, and a 20-character display for potentials from nerve cells. Each action prompts to the user; it has two Potential lasts only 1 or 2 milliseconds, input-output ports, for connecting the although the sequence of events ray extend analog interface; and it includes a over several seconds. Thu action crystal-controlled 1-microsecond clock and potentials, amplified and displayed on an 16-bit interrupt timer, for accurate oscilloscope screen, seem to flash by; timing of data-acquisition intervals. In they can be glimpsed but not studied. addition, the AIM is compact and does not They are too fast to be written by chart require a video monitor, which makes it recorders, and although photography can easy to integrate into an existing capture them, photography is too laboratory set-up. The AIM's major inconvenient and too expensive for routine disadvantage by current standards is that use in a student lab. A microcomputei' it has only 4K bytes of en -board memory, with an analog/digital interface, however, and expansion memory boards frod various can record these transient events for vendors are slightly more expensive than repeated playback to an ostilloscope, they ought to be. Nevertheless, a freezing the signal on the screen, or for complete system (AIM, 32K expansion slow playback to a chart recorder, to memory, and analog interface) costs about provide a permanent record. Several years_ $1000, which is much less than systems ago, I reporte' on our preliminary plans based on other popular microcomputers and to use microcomputers in an, undergraduate very much less than commercially available data-loggers or digital oscilloscopes. The output (digital-to-analog) conversion is provided by an AD558 Although the AIM, enclosures, power integrated circuit, which produces supplies and expansion memory all are voltages from 0 to 10 V and settles to its available from commercial vendors (some of new output value within 2 microseconds whom will even assemble and test the after it receives the data. Note that the package of components before shipping), input circuit's range is -5 to +5V, while there is not (so far as I know) a the output range is 0 to +10V. For each, commercially available analog/digital the span is 10V, but the midpoint (the interface board that is inexpensive, equivalent of 0 V at the input) differs by provides preamplification, and is capable 5 V. This means that the analog .output of fast analog/digital conversions. We will be stepped up on the oscilloscope have built our, own interface, based on a screen compared to the input. This has handful of integrated circuits that are not proven to be a problem for students, relatively easy to use. My design for but one could add a lx amplification stage that interface, which incorporates that offsets the output signal by 5V to improvements suggested by Artner Chace of give the output the same midpoint as the Mount. Holyoke College, is shown in Figure input signal. The remaining components, analog/digital the trigger-in and pulse-out lines, are 1. , The task of an interface is to convert voltages to their protected from inadvertent connection to digital equivalent, which is sent to an large signals by a comparator (LM 311) and input port of the microcomputer, and to by a resistor diode network, convert digital information from the respectively. The microcomputer expects computer back into voltages for display on to see (and produces) TTL-compatible an oscilloscope screen. In addition, pulses (0 or +5V), and if these are becaus'e voltage signals from available from the other laboratory neurophysiological experiments are very equipment -- and if the students.are small (even after preamplification, they always careful -- the protection circuitry are in he range of 10-100 millivolts), could be omitted. our interface includes- a stage of preamplification to bring the input Although it generally is good advice signals\ up to the size that the to buy rather than build equipment, that analog-tO,digital converter expects (10 option was not available for this'system. V); als, because we need to take in However, since the' two converters from trigger signals from stimulators and emit Analog Devices are so easy to use, there pulses to synchronize oscilloscopes, the is little difference between wiring a interface board includes connections and cable between the ATM's application protection \circuitry for trigger and pulse connector and a board containing these signals. circuits, and wiring a cable to an equivalent commercial board (if one were For analog/digital conversion, available). The additional components chose integrated circuits from Analog (preamplifier and trigger protection Devices, Inc. (Box 280, Norwood MA 02062) circuitry) are necessary for our that were relatively inexpensive, application, but they may not be for performed well,\ and required a minimum of yours; if they were omitted, the external components. The analog-to- interface would be extremely simple to digital converter (AD 570) accepts signals construct. The total cont of the in the +5 V to -5V range, and takes 25 integrated circuits is less than $40; our microseconds to perform a conversion. To complete interface, which is mounted on a boost input signals to the required range, rack panel under the oscilloscope and I added a two-stage amplifier based on the includes the.various features shown in the commonly available 1458 (dual 741) op amp'. diagram, costs well under $100. The amplifier has a variable -gain control (lx to 1000x in 1,2,5 steps) using a Software 10-position switch and a set of common resistors (these and other parts can be The software for the data-acquisition purchased from suppliers such as Radio system had to meet two principal-criteria: Shack or Jameco). If calibrated gain is it had to b3 fast, and it had to make the not needed, a 1-megohm potentiometer may sy3tem easy to use for students who had no be substituted for the selector switch. experience with computers. For ,ease of The preamplifier also provides a choice of direct- or capacitor- coupling (AC/DC switch), and it has an offset (zero) control; again, these may be omitted for Figure 1 (next page). Analog/digital ease of construction if they are not interface based on integrated circuits for needed. analog-to-digital (AD570) and digital-to- analog (AD558) conversion.

175

1 9 '"? SPDTe..."0 INPUT

AC /DC 0.1 1K uF x2 ioaNA/LfTr10K x10ed x20 .__,vv4_2j__.(x50 100Kx100 100Kx200 ANL--...x1000x500

our 10K

+IN 6

-12V V- 1458

20 CAl

NC NC 1 18 DATA, 14 PA0 LSB2 17 Ig 4 PA1 DIG COMMON 3 16 PA2 BIPOLAR OFF 3 4 , 15 PA ANALOG COM 52 ---=-1' 5 14 PA4 ANALOG IN 6 13 5 PA5 7 12 PA6 7 8 11 O PA7 B U 8 9 AD570 10 NCA2 H(-- 21 .4 SPDT PBO DATA, TSB V OUT OUTPUT 1 16 N 9 VSENSE 10 PB1 2 15 _ PB2 V SELECT 11 3 14 GND 12 P 4 13 I 4- PB4 GND 13 5 12 OUTPUT 16 PB5 6 11 +12V ro (CHART) PB6 CS 17 7 10 -t- 0.1 15 PB7 MSB AD558 9CE uF 18 CB1

GND 1 8 +12 10K +IN OUT 7 +5v

-IN NC +12V 3 6 10K TRIG:1ER '12V- NC 30K LEVEL 114311 5 -12V 10K +12V

10K...TRIIN -12V +5V 4 -I-- 19 CB2 PULSE 1N914 47K _E*0 UT GND 1 A POSER +12V N _12V CONNECTIONS

176 193. use, the program is stored in permanent The data-acquisition program is memory (ROM), so that no loading from disk- writtenk in assembly language, and is or tape is necessary, and no accidental container' in a 4K ROM that plugs into a over-writing of the program can occur. socket on the AIM intended for Rockwell's Program start-up and all subsequent assembler. Although writing an assembly commands are by s±ngle keystrokes, and any language program of this size a command may be executed at any time. The substantial task, it was clear to me that command set is listed in the Appendix, and the data-acquisition routines would have will be discussed further below. Each to run at machine speed (eliminating the student receives a User's Guide, which possibility of working in BASIC), and it explains each command. The cover of that seemed simpler to do everything in Guide (Figure 2) also illustrates the assembly language rather .than trying to performance of the system; the photograph mix high-level and assembly language. have described elsewhere [2] an example of an assembly language routine for acquiring one byte of data. In retrospect, now that FORTH is available for the AIM, I would probably write the program in FORTH (which makes it easy to include assembly language USER'S GUIDE routines) if I were to write it again.

The program has three major input modes: continuous input, triggered input, and averaging. These place data into mv-)H\ilvir three separate buffers in memory. At start-up, the available memory is automatically allocated into three non-overlapping buffers, but the allocation may be changed by using the VARIABLES command, and any buffer may be assigned to any part of physical memory. 10 ms I ms Our systems have 36K of memory, which means any buffer could store up to 36,000 samples. In continuous INPUT mode, the input buffer is filled repeatedly (with the newest data over-writing the oldest) until INPUT is halted or another mode is selected. Each digitized data point is echoed to the oscilloscope as it is taken. NEUROPHYSIOLOGY DATA INPUT mode is useful for capturing spontaneous neural activity and responses ACQUISITION SYSTEM to hand-delivered stimuli (such as mechanical prods or the application of drugs). When an event of interest has occurred, one halts INPUT mode and displays the buffer, which contains the last few seconds of data. When electrically triggered stimuli are used, DEPARTMENT OF RIOLOGICAL SCIENCES SMITH COLLEGE NORTHAMPTONMA 01063 however,' TRIGGER ,mode provides a more efficient way of capturing the response evoked by the stimulus. In TRIGGER mode, Figure 2 the computer waits for a trigger pulse; when the pulse is received, a pre-set number of data samples is taken and stored, after which the computer awaits the next trigger pulse. The amount of at the top left shows an example of analog data stored in each sweep is determined by data (upper ..trace; the brief downwa2d a variable, which can be reset in units of spikes are action potentials) and the 256 samples (1 page of memory); the digitized echo of the same data from the default value is 2 pages (512 samples). computer (lower trace). At the right, a similar photograph of the oscilloscope The entire orientation of the system screen shows the original and digitized is toward temporary storage of data, with data at a faster sweep speed. The chart the expectation that data of-interest will record below was obtained by playing back be written out to a chart recorder for a the same data from memory at a slow rate; permanent copy. This gives students their not the accurate reproduction of the data in its most useful form for further analog signal. analysis, and it also avoids the expense

177 194 fast and complexity of disk or tape (either of our work, fast acquisition, whinh, from my point of view, only, oscilloscope display,.and slow output to a postpones the moment when a hard copy must chart recorder make the-most sense, but be made). Data that have been captured by the software is wric.ten in such a way that the system can be played back by typing it can also be used for acquiring slow the DISPLAY command, D. DISPLAY data (approximately 16 samples per second, automatically selects SCROLL mode if Minimum), where it might be appropriate to continuous input was used, and JUMP mode use a display rate much faster than the if 'triggered input or averaging was used. input rata, and a chart writing rate that In addition, SCROLL or JUMP may be entered is equal to the input rate. directly (and any buffer examined) by typing S or J. Finally, the two special input modes deserve some comment. AVERAGING resembles SCROLL is a particularly useful triggered input in waiting for a trigger display mode. The signal appears pulse before acquiring a sweep of data, continuously on the oscilloscope screen but the data are summed with previous but seems to slide across- it, like a sweeps rather than stored separately. The moving chart. The direction of travel can average is calculated whenever- the number be changed by,typing R, REVERSE (this is a of sweeps is an exact power of 2 (1, 2, 4, feature that was added to the system at 8,... 256). At each sweep, the last student request); the image .can be frozen calculated average (rather than the input (or scrolling resumed) by pressing the signal) is displayed, while the AIM shows space bar; and the rate of scrolling can the number of the current sweep and the be increased or decreased by pressing the number of sweeps in the last average. + or keys. As a result, one can examine Averaging is useful for extracting an data with very high temporal resolution evoked response that is hidden in a noisy (our oscilloscope sweep speed is typically baseline, and students have used it more 2 ms /div), move forward or backward in often than I first expected. During time at one's convenience, freeze time, averaging, DISPLAY can be entered to show expand the oscilloscope sweep to examine the current average frozen on the screen, events of interest more closely, and and averaging can then be re-entered handle a large data-sample. without losing the accumulated data. The second special input mode, HISTOGRAM, is JUMP mode is the display mode probably of more interest to selected automatioally if triggered input neuroscientists than to scientists in or averaging was the last input mode used. general. HISTOGRAM sends to a chart One sweep appears frozen on the screen, recorder a voltage proportional to the and then the system JUMPS to the next number of events per seczmd, updated in sweep. Jumping can be halted (or resumed) 0.1-sec bins. It is uses': for_connting the by Pressing the space bar, and the number of action potentials per second, an direction of advance can be reversed by important measure of neural activity. In typing R. The current sweep number is this implementation, events are detected shown on the AIM's display, an asterisk in software: any signal that crosses the appears if jumping has been halted, and an 0-volt baseline (either in a positive or R appears if jumping is reversed. Once negative direction, as set by the user) is again, by freezing a sweep of interest on counted as an event, and the count is the screen, experimental data can be updated every 100 msec. By using the zero examined at length, the sweep can be offset control, it is usually possible to -expanded, one can back up to the previous adjust an input signal so that only the sweep, and so fortn. biggest events trigger the counter. A pulse is emitted whenever an event is In addition to these display modes detected; one can trigger the for an oscilloscope, two other display oscilloscope from the pulse to observe the modes produce slow output for a chart input signal being counted, while the recorder. EXCERPT (E) plays out the chart recorder simultaneously writes a segment of memory currently displayed on record of the histogram. the screen. WRITE plays out the whole buffer, starting from the beginning. Evaluation Either of these can be used for making a chart'record of data obtained from any The data-acquisition system has been input mode. For convenience, the rate of used in my neurophysiology course for sending out data samples is slower for three years by undergraduates whose prior EXCERPT than for WRITE, so that one chart experience with oomputers ranged from speed can be used to produce a high- extensive to none. Even the least resolution (E) record or a temporally experienced students found the computer no compressed (W) overview. However, the harder to use than the other equipment default va,,e of an; timer interval can be they encountered, and all of them, were changed using the VARIABLE command. For eager to use the system since it let them

178 19'j 538 their data clearly and gave them hard Output modep copies to show others. The command set seemed consistent and easy to understand; D Display. Automatically selects the the students controlled the system appropriate output mode and buffer, properly right from the start, and readily determined by the last input mode made it do what.they wanted it to do. As that was used. a result, while previously students inthe course had at best obtained a few polaroid S Scroll. Data from the buffer, photographs from some experiments, they selected scroll across the screen, as now routinely capture as much of their on a chart recorder. Usually used data as they wish. They work more with the input buffer. quantitatively, and the level of the laboratory now is better matched to their J Jump. Successive sweeps from the other work in the course. buffer selected appear one after another. Usually used with the References trigger buffer.

W Write. Contents of a buffer are 1. Olivo, R.F. (1980) Microcomputers as output slowly, for writing by a chart laboratory_ instruments: two recorder. Outputs the entire buffer. applications in neurobiology. Proc. National Educ. Computing Conf. Z: E Excerpt. The part of a buffer 81-85. currently displayed on the oscilloscope screen is output 'very slowly, 2. Olivo, R.F. (1981) An efficient A/D for writing by a chart recorder. interface. Compute! a (9): 140-142 (September 1981). C Continue. Resume last display mode from the current position; used after Excerpt.

System control

Appendix Q Quit. Exit from the current mode; this command is available at all times.

Whem_the computer is first trged on: (space) Pause. On input modes, suspend data collection. On output modes, N starts the data-acquistion program suspend advancing through the buffer, and. sets default values. All freezing the current data on the Subsequent operations are controlled screen. Typing (space) again ends by the ,following commands: the pause.

Input modes R Reverse. Change the current direction of scrolling or jumping. I Input.'Continuous storage of data in New samples over- the input buffer. +/- Increase or decrease the rate of write the oldest previous samples. scrolling or jumping. May be typed more than once. T Triggered input. A single sweep of data is stored after each trigger Variables. Print current values for pulse is received. Data are stored memory allocations, timer intervals, in the trigger buffer. or constants, and permit their alteration. A Averaging. Successive sweeps of data, initiated by trigger pulses, are averaged. The last average calculated is displayed.

Histogram. Events that cross ground level (with or - slope, as ,specified) arecounted, and the count is output as a voltage every 100 msec. No data are stored.

179 COMPUTER LITERACY

Margaret Christensen Dr.,Carla Thompson Dr. Joyce Friske Wiliam H. Pritchard, Jr. Donald Z. Spicer Ronald R. Bearwald David J. Lewis

ABSTRACT: AlgehrS,_BASIC, and Computers: The This presentation will focus on the ABC's for NonScience Majors development and integration of a computer literacy discipline area (CLT) into the twoyear college Margaret Christensen, Widener University curriculum. 1'r6S-enterswill describe the role of the CLT discipline ,area within the overall Seven realistic problems lead students through curriculum and discuss the focus of CLT courses a semester of elementary algebra and computer with respect to specific needs-of twoyear college science in a new course we have developed. students. Concerns associated with integrating As computers become less the exclusive domain computer literacy couses into a noncomputer of the scientist or specialist and more the daily science division will also be discussed. companion of educators, social scientists, In order to meet the computer literacy needs humanists, and businessmen alike, the pressure to of all students regardless of background or major, become computer literate increases for all the CLT curriculum was designedi- elative to three students, including those who lack the mathematics areas: liberal arts students, proficiency which is prerequisite to successful technical/occupational students, and education completion of computer courses. Rather than majors. Course descriptions,,, aboratory subject these students either to a computer course activities, and class projeCts will be suggested which is devoid of math (if such is possible!) or for all computer literacy courses within the CLT to a traditional, remedial college math course curriculum. taught in the same way that lost them before, we The session will conclude with an overview of have developed a course which combines computer future perspectives for the twoyear college science and elementary algebra in a symbiotic computer literacy scope and sequence. relationship. Central to the course are seven problems which require some thought, the use of algebra, and ABSTRACT: The Vassar College Computer Literacy computer programming in BASIC for their solution. Program In the course manual the problems sections follow a logical order of progression, in terms of material William FL Pritchard, Coordinator, Computer covered and difficulty, and have solutions which Literacy Program, Donald Z. Spicer, Associate Dean require only the use of material taught in that of the College, Vassar College, Poughkeepsie, NY section or in previous Sections of the course. 12601 Students have responded enthusiastically; they have clearly learned a lot; and the In January 1982, Vassar College with support: instructor found the course a lot more fun to teach from the Fund for the Improvement ofd Postsecondary than standard remedial algebra or computers without Education (FIPSE) initiated a computer literacy math. program that has a number of unique features that resolve many problems associated with computer literacy in the context of ,a relatively small ABSTRACT: Computer Literacy in the TwoYear liberal arts college. The thrust of the program is College Curriculum to infiltrate computer usage into the curriculum of the College as a whole. Therefore faculty Dr. Carla Thompson, Dr. Joyce Friske, Mathematics development, which provides sustained impact, is an Instructors, Tulsa Junior College N.E. Campus, important component., In a core course a group of 3727 E. Apache, Tulsa, OK 74115 participating faculty, together with students sharing common interests, are. introduced to the As the belief in the need to educate future availability and use of computer resources that citizens in the operation, use, and impact of support their discipline. To reinforce this computers gains support in educational circles, it initial experience, participating faculty agree to is becoming increasingly clear that all students introduce computer usage into designated courses should be provided with educational opportunities taught in later semesters, and students who take that allow them to become computer literate. This the core course are expected to also take at least need can be appropriately met at the junior college one of those' subsequent courses. While many_ level by offering computer literacy courses aspects of educational computing can be managed on designed for noncomputer science majors. micro and minicomputers, there are also many uses

180 197 required to support diverse faculty interests that consecutive weeks during the first cannot be cost efectively acquired 1-5, a small semester and two during the second. college. Since the Vassar Computer Literacy Classes will be twentyJfive minutes in Program is intended to support a broad duration. The computer language used cross-section of the faculty, the program makes use will be LOGO. of the varied software and hardware resources Grade 4: Students will come to the micro learning available remotely through EDUNET. Data base and center in groups of twelve.\\Each student information services such as DIALOG, Compuserve, will receive four consecutive "weeks of and The Source are also used. instruction. Classes will be forty A major aspect of the Vassar program is that minutes in duration. The computer it is designed to be a model program that is language used will be PILOT. intended to be readily transportable to other Grade 6: Students will come to the micro learning colleges. Therefore, the program is designed to center in groups of\sixteen. Each have relatively modest start-up costs. The major student will receive\four consecutive continuing costs are for personnel and remote weeks of instruction. Classes will be teleprocessing. forty-five minutes in duration. The Beginning Febraury 1893, the program will be computer language used will-be-BASIC. expaned to include more courses, an increased Students will be grouped heterogeneously for number of faculty development workshops, and a instruction. Additional opportunities will be unique program of faculty ownership of provided to students in the non-target grades on, an microcomputers. Under this grogram faculty will be individual, small group, and special-unit basis. encouraged to purchase their own microcomputer All students will have equal access to through joint support from a grant from the Sloan participation regardless of ability 'eve?. Foundation, a College low-interest loan, and a negotiated group buying plan with a vendor. It is hoped that within the first year, approximately 25% ABSTRACT:. Machine Language in Computer Literacy: of the faculty will own a microcomputer. Strategy and Supporting Software The presentation in this session will elaborate on the initial conception of the program, David J. Lewis, Department of Mathematics, Ithaca explain the design of the core course, and discuss College, Ithaca, NY 14850 the experience learned during the first year and a half of implementation. A grasp of fundamental principles of machine language is an important subject for computer literacy in its own right. It can also serve as a ABSTRACT: A Microcomputer Literacy Program concrete introduction to other, topics such as variables, the CPU, main memory, hardware vs. Ronald R. Bearwald, Assistant Superintendent, software, high-level language processors, the Lincolnwood Scool District Number 74, 6950 East notion of a program itself, and many others. Price, Lincolnwood, IL 60645 Further, machine language experience illuminates the otherwiCe murky gap between logic design and .- The Lincolnwood Schools have developed a plan high-level language. For these reasons computer for mocrocomputer literacy which provides for the literacy courses at Ithaca College generally begin instruction of every child in grades two (primary), with one to two weeks on Slic, a highly four (intermediate), and six (upper). This interactive, full-screen, instructional simulator instruction will be based on a sequential for a simplified machine language we have developed curriculum which includes general cognitive, and for the purpose to run on the Apple II. -affective goals as well as specific operational and Slic is a simple decimal machine like those in programming objectives. many computer literacy and data processing texts. General goal of the microcomputer literacy What is important about the Slic processor is its program are: totally visible memory and animation of program - To develop computer literacy by teaching execution under student control. The accumulator, importnnt computer related concepts, program counter, all seventy memory locations, and increasing the awareness of the values a display/reply "screen" remain on the Apple and applications of computers in our screen, reflecting all events in the execution of a world, and providing opportunities to program in "real-time". attain a certain level of competency in Slic renders many concepts graphic that performing fundamentaL computer students must ordinarliy infer, even with the best operations. of the available high-level language processors. - To improve the overall cognitive and The, Slic intruction being executed is highlighted affective abilities of each child by in reverse video, and one can SEE a loop in action developing structured and logical as the program runs and a sum grows in the thinking 'skills, positive attitudes, and accumulator. The instructor can POINT AT a innate creativity. variable (that is, a location) on the screen and Each student in the specifieed grades will change it by executing an instruction.Students receive a minimum of four weeks instruction which can discover the principles of destructive read-in will take place in the following manner: and non-destructive read-out simply by observation. Grade 2: Students will come to the micro learning The sharing of main memory by program and data area in groups of six. Each student will appears natural when seen in real-time action. receive four weeks of instruction; two The major technical features of Slic are: (1)

181

19tS target machine instructions and data are stored in fivedigit decimal locations and asIngle accumulator; (2) the instruction repertoire includes laod and store, arithmetic, jumps, and input/output; (3) the command structure is based on the UCSD Psystem and Apple Pascal; (4).a memory "editor" allows direct manipulationof "main memory" and easy entry of programs; (5) a filer 'provides saving, retrieval, and printing of programs; (6) a program may be run at a specified rate, stepped one instruction at a time, or "crawled _thnough all fetches and stores; (7) any machine instruction may be executebInimmediate mode, simply by typing it as a command.

182 19J COMPUTER EDUCATION FOR ELEMENTARY SCHOOL TEACHERS

Joyce Currie Little Robert Wall Harold R. Strang Ann Booker Loper Dr. AliceAnn Winner E. Muriel J. Wright Helen V. Coulson

ABSTRACT: Computer Literacy for'Elementary and' Hardware component,consist of an OSI 48k Middle School Teacher's A Report on a MAEUC/AEDS microcomputer equipped with two 8inch disk drives, Project a real time clock; a video terminal, a second independent video display, and a serially connected Joyce Currie Little, College of Natural and printer Mathematical Sciences, Robert Wall, College of er loading the simutlation's BASIC program, Education and Instructional Technology, Towson an operator assigns, via the terminal's keyboi.rd, State University, Towson, MD 21204 the following to each of the class's four simulated, students: a probability for knowledge (the A pilot project in computer literacy, likelihood of answering correctly) and a sponsored by the Maryland Association for probability for initiative (the likelihood of Educational Uses of Computers, Inc. (MAEUC), and volunteering to participate). the Maryland affiliate of the A ssociation for The teacher, who is situated in a booth Educational Data Systems (REDS), was done to test a physically isolated from the operator, next strategy for increasing literacy among teachers in receives a brief familiarization with the elementary and middle schools. The primary simulation's graphic feedback characteristics and purposes of the project were to provide information then actually conducts a lesson with the four to the schools for awareness of computer literacy; students whose names are displayed on the experiment and evaluate the use of certain independent video monitor in front of the teacher. exercises, materials, and activities; determine The teacher may verbally instruct', ask questions, computer literacy levels; and determine the solicit questions or answers, or just respond to effects of exposure to computer literacy materials. students whose graphically defined hands pop up due It was decided that a minicourse format would be to programmed initiative. Each teacherstudent used. interaction is coded by the operator and keyed into Three elementary schools and their feeder the terminal. The terminal's display directs the middle school in an inner city area were chosen for operator, who functions as the voice of the the minicourse. All teachers and administrators, students, what to say to the teacher. After the and some parents, were invited to participate.The teacher has solicited an answer, for example,-the course was organized, managed, and taught by operator'-.. screen might present WRONG ANSWER 1, the members of MAEUC/AEDS, and equipment was provided 1 designating the type of wrong "answer to'be by their institutions. relayed to the teacher. Rapid computer feedback to The Computer Literacy and Assessment the operator coupled with careful operator training questionnaire, developed by the Minnesota insure a pace in the teacherstudent verbal Educational Computing Consortium (MECC), was used exchange not appreciably slower than that found in for pre and post testing of knowledge and attitudes an actual live classroom. in order to assess changes in computer literacy. During the teaching session, the computer also This presentation will describe the project collects time and frequency data on over 110 and its management, give an overview of the specific types of cycles that the teacher has minicourse, present the results of analybis of employed in interacting with each student. As change, and summarize the local MAEUCkAEDS group's cycles emerge, the printer fuiibtions as an event recommendations. recorder, displaying immediatelyfol,lowing each cycle the student involved, the time elapsed, and the cycle type. Cycles are defined according to ABSTRACT: Microcomputer Simulation: An Aid in their antecedent characteristics (e.g., teacher Training Elementary School Teachers ----soliciting an answer, a question, or responding to a 'hands ups), student responses (e.g., right Harold R. Strang, Ann Booker Loper, 405 Emmet answer, wrong answer, or question), and teacher Street, Ruffner Hall, Unversity of Virginia, consequation (e.g., inclusion of feedback, Charlottesville, VA 22903 instruction, and/or emotional affect). A complete hard copy teaching profile is This micocomputer project is being developed obtained at the conclusion of the teaching sesion. to assist in the preparation of elementary school This record displays over 500 separate measures teachers. Our primary goal is to help bridge the pertaining to the events that occurred during that gap between academicteacher preparation and actual session. teacher:student classroom contact. Two projects have addressed one of the most

183 2ou basic questions pertaining to the simulation - sense of shared responsibility. - namely, whether it truly creates an environment Five assumptions were articulated and functionally similar enough to an actual classroom supported by the results of the study. Briefly to be of value in teacher training. Initial stated they are: 1) workshops conducted by outside results are very encouraging. When, for example, agents are less effective in promoting innovation class or individual characteristics were adoption in elementary education; 2) a single manipulated so as to produce initiative and/or school, which possesses a positive attitude toward knowledge contrasts, participating teachers change and the requisite supportive mechanisms, can responded in ways paralleling those that would be be a decisive instrument of change; 3) an inside expected in actual classrooms. change agent can initiate a successful change In addition to pursuing the validation process in a single school environment; 4) research, the authors are also exploring the computer literacy at the introductory level needs possibility of developing a pre-teaching diagnostic to include careful analysis of "wise use of tool based on the system's extensive data microcomputers" to avoid an over-emphasis on collecting, sorting, and display. features. computer assisted instruction as the prominent mode of implementation; 5) computer literacy at the elementary level should provide an awareness of the ABSTRACT: Toward Curriculum Development: A Case implications of process learning to develop Study in Computer Inservice Training curricula for optimal implementation of the computing potential. Dr. Alice-Ann Winner, United Nations International The data, collocted from a variety of School, 24-50 East River Drive, New York, NY 10010 evaluatory procedures, support an additional arlumption undefined at the outset - curriculum Teacher training in appropriate use of the revision should follow rather than preceed teacher microcomputer in the elementary classroom presents training. The results also indicated apparent a problem for educators.Many of the computer stages in the computer implementation process, and training programs are conducted by "experts" who illustrated some of the problems inherent in that have little understanding of elementary education implementation as well as general ones relating to and/or the specific needs of its teachers, and inservice training. The data underscored the consequently curriculum development for wise necessity of integrating developmental inservice computer use has not been forthcoming. This training into the regularities of tbe,elementary project report describes the design, school. inplementation, and evaluation processes of an inservice training program especially created, for elementary teachers. ABSTRACT: Incorporating -the Microcomputer into the Three other differences separate this training Department of Mathematics Program for the program from most of the current computer Prospective Elementary School Teacher-at California workshops, even though similarities in content k.nd State University, Northridge structure exist: 1) the use of an inside change agent as the inservice instructor; 2) the extended E. Muriel J. Wright, Helen V. Coulson; California duration of the workshop series over the entire. State University, Northridge, No thridge, CA 91330 school year; 3)the use of formative evaluation in the design process. Computer literacy is an essential outcome of The development of this program started with contemporary education.To keep pace with the reviews of the literature on change and computing inevitable sophisti-cation of their students, it is in education with regard to the implications on imperative that-1C-8 elementary teachers - perhaps curriculum development in elementary education and the single most influential group in the inservice training. The training took place at a mathematical competence of the country - acquire an single school with an international population. understanding of,the versatility and limitations of While the demographics of the study were to some the computer through a working knowledge of how one extent idiosyncratic, the course content is interacts with computers and how one uses their applicable to other school environments as many capacities. elementary teachers share common goals. The The Department of Mathematics at California materials and experiences designed and implemented State University, Northridge, has just completed in this study reflect this commonality of purpose. the PILOT MODE of a two year project for (1) the The objectives of, the program were to increase development of a laboratory of eight the awareness, exploration, and experimentation microcomputers, (2) incorporation of interactive levels of the participating teachers, and to lay a compute( experiences within an existing strong_two foundation for permanent computer implementation semester mathematics program for prospective and curriculum revision. Designed by the author, elementary school teaschers involving'700 students who as a fellow teacher was aware of the annually (90% women, 20-40% ethnic minorities, 10% regularities of the school environment, it re-entry women), (3) selection-and_design-of reflected the specific needs of the participants. software using the computer's special capabilities These needs were addressed in the workshop sessions (iteration, recursion, graphics,'number-crunching) and in the intervals between sessions. Use of to enhance the treatment of the mathematical topics formative evaluation permitted the teachers an within the available time, (4) bringing the novice active role in the planning process, transmitting a to a beginning but sound programming capability.

184 201 We will (1) build mathematical and computer skills - by hands-on drill, testing, and assignment correction, (2) teach new mathematical and computer concepts - by tutorial programs which will also invite student modification, (3) provide problem solving experience - by interactive programs in simulation of probability, statistical, and transformation geometry problems, (4) illustrate by examples of our programs;and format the variety of uses that an elementary teacher may make of the microcomputer in the classrOom - in mathematics, but also in other aspects of the curriculum, and (5) develop programming skills - by teaching computer commands, by requiring modification of some existing programs, and by student-written programs suitable for elementary classroom use. During the PILOT MODE, using an Apple microcomputer and peripherals purchased under a grant, we were able to select and develop software,/ for drill, tutorials, and simulations; make extensive student testing of the materials in bot one -unit electives and seminars; and determine th values of individual and paired use of the terminals for student -worker 7--A-room for setting up \ a model microcomputer laboratory housing eight machines was obtained adjacent to the classroom/tutorial complex used for the methematics courses. Proposals were submitted to various agencies within and outside the University. Six Apples with color monitors and disk drives have been granted and will be installed during January 1983. Peripherals such as a graphics tablet and two printers will also be in use. A-voice-over \attachment for use with haddicappedstudents is attached to one computer. Beginning with the Spning Semester 1983, full implementation of the project is planned. Using tested materials - mathematics text, programming manual, computer film, and computer programs and materials selected and designed by the two project directors ----three hundred students under ten instructors will develop programming capability along with the required mathmatical knowledge. For years, a standardized common final-has been administered to all sections of this course so that well established norms are available for comparison of the effect of incorporating the microcomputer into the program without reduction of mathematical content. Level of Computer Literacy will "be determined using a test and norms developed by the Minnesota Educational Computing Consortium. Finally, instructor and student attitude will be determined by self-report. The goal is preparation of 350 students per semester trained in the mathematics essential for the elementary school and with a working knowledge of computers. An important spin-off will be the replacement of the now superfluous one-unit elective in introductory computer programming with a course in Graphics and Turtle Geometry.

185

.1, COMPUTERS IN EDUCATION

C. Michael Levy Dr. John R. Pancella Edward Zeidman Melvin H. Wolf

to enhance their visual quality but to serve asa ABSTRACT: RealTime Microcomputer Programs for signal or cue. More importantly, the keyboard was Teaching Statistics used as an asynchronous device, permitting us to create dynamically changing displays and, in C. Michael Levy, Department of Psychology, University of Florida, Gainesville, -FL 32611 general, to give these programs many of the characteristics of some of the exhilirating arcade While many instructors believe that the key to games. The obvious appeal of this utilization is mastery of statistics is the successful completion related to the fact that students who often resent of many sample problems, it is often infeasiblefor using "canned" data feel that they (rather than the them to implement instructional programs that computer) are in primary control in these provide students with a rich corpus of problems. situations. Partly in resonse to this implied need, I have In short these programs provide students with developed a 100 kilobyte ensemble of interactive powerful opportunities to discover for themselves CAI exercises for the Apple, IBM, Atari, and DEC the answers to important substantive-questions, thereby making the abstract theoretical ideas of microcomputers. The materials are solely statistics more concrete and memorable. The instructional; they provide no means to perform statistical calculations of usersupplied data. presentation will include both a discussion and These programs are used in conjunction with a demonstration of these materials. workbook that portrays the rationale and demonstrates the use of each statistic with a detailed example, and presents scenarios describing ABSTRACT: High School Science Microcomputer the problems that the student completes. Project The design of our programs was heavily influenced by insights that emerged as I developed Dr. John R. Pancella, Coordinator, Secondary other courseware for undergraduates during thelast Science, Mr. John Entwistle, Teacher Specialist, decade. For example, the software contains Science, Mathematics, and Microcomputers, routines whose sole function is to reduce the Mrs. Carol Muscara, Specialist, Computer Related opportunities for students to make entry errors Instruction, Montgomery County Public Schools, that result in abnormal program termination_ or Rockville, MD 20850 evocation of an obscure error message. The ensemble of programs contains two broad A project was begun during the 1981-82 school categories of exercises. One includes multiple year to implement microcomputer technology in Feedback from computational problems for each statistic that is senior high school science courses. in the family of problems found at the end of the three school tryout centers was used to develop chapter of many statisticslbooks. These were teaching and training plans to incorporate designed to provide the guided practice many microcomputers.in biology, chemistry, earth students need to appreciate the process of science, and physics classes of 12 senior high Training determining the correct solution. schools during the 1982-83 school year. I have dubbed the second category of exercises for 36 science teachers was conducted during summer Use of microcomputers will be phased into "Exploratory Problems". These were designed 1982. expressly to give students an intuitive grasp of the remaining 10 senior high schools by September each statistical technique. Great effort was made 1983. to incorporate special features thatwould make 2'heproject gopls are: moderately deep levels of cognitive processing 1) Reinforcement or review of information gained in the classroom for students to nearly unavoidable. This was accomplished by grasp subject matter that is perceived as directing students to "play with" the techniques and "manipulate" the raw data in order to answer difficult for them. Development of problem solving skills "what if" sorts of questions for example, "'What 2) will happen to the variance if I add a constant to using programming algorithms or microcomputer capabilities to increase each score?""What happens to the standard error student ability to solve if I increase or decrease N?" "What will happen to the value of t or F if I reduce to zero the multidisciplinary problems. of simulation software for laboratory variance of one group?" 3) Graphics were used whenever possible to investigations that would otherwise be too enliven the displays, and color was used not merely timeconsuming, hazardous, or expensive

186 203 for high school investigation (e.g., to ones which were previously not possible and for longterm genetic investigations, ammonia which the computer is especially well suited. production, electric charge on atomic The graphing of functions receives a great particles). deal of attention in the community college 4) Interfacing (connecting laboratory mathematics courses. However, student performance equipment with the microcomputer) for in this area has been poor. This is particularly investigation and analysis of data distressing in light of the importanze of functions previously not readily available to and graphing, both to applications and to further students (e.g., continuous pendulumperiod work in pure mathematics. analysis, rapid coolingcurve plotting, The microcomputer, with its unprecedented genetic statistical analysis). ability to do tedious calculations almost Each science department of the 12 high schools instantly, makes it a natural choice for material received four Apple II microcomputers, four intended to improve students' understanding of monitors, four disk drives, one printer, and functions and graphs. This paper explains a miscellaneous accessories. Each school received computer activity designed to provide students a about 30 software programs for immediate use. fertile, mathematically accurate environment and Other software was loaned from a central the motivation to manipulate that environment to collection. Schools purchased evaluated and learn about graphs ofequations. This activity is approved software to meet their local program called the Function game. needs. The Function Game is a computer educational Four 3hour followup meetings of the summer game designed to aid the-student in learning to participants were held during 1982-83 to exchange recognize functions of a single variable a ideas and programs and describe diffusion and necessary skill for mathematical modelling. The dissemination activities done at the school level. Function Game challenges the student by effective Products of the project included: use of computer graphics, immediate performance 1) A model 10day inservice program for feedback, scoring, and competitive skill ratings. training teachers. The student is able to explore and discover the 2) Design and implementation of several association between graphs of functions and their labisratory investigations using equations, with the computer giving the student thermistors and a teachermade interface immediate feedback. module to connect the experiment to the The student starts the game by selecting microcomputer through the paddle port functions he/she wants to study. (The game package analogtodigital converter. comes with a set of over seventy functions, and 3) Implementation of an evaluation form and more can be added.) A 'function is randomly chosen selection process for approving software from this set; the graph is drawn;the student for school use, and a data base storage then examines it. Ths student can imagine that system of the information. he/she is a detective, whose job it is to solve the 4) Data collection on microcomputer use case of the "mystery" function. The student must during school and nonschool hours. guess the name of the function and determine the 5) Lesson and unit plans with descriptions of value of its unknown constants. Then the graph of approved software coded to each science the guess is drawn, overlaying it on the graph of course by topic and lesson objective. the actual function. The student can visually 6) A bimonthly newsletter to schools on new ascertain the correctness of her/his answer. (This products, new applications, and school provides the student with immediate performance implementation activities. feedback.) In keeping with the game aspect, the Data on the results of the project showed an student is scored on the proximity of the actual unusually high rate of diffusion of the training function to her/his guess. among other teachers in the schools, students, and An important feature is that hints are even parents. There was a rapid increase in provided when the student wants help. A set of microcomputer use 1:), students. Many microcomputer hints comes with the package, but as with programs were writ'in by students and teachers and functions, these can be modified. shared among,schoris. A special editor is provided to modify and add to the list of functions and hints. The instructor, can even design entirely different sets of ABSTRACT: The Function Game: Using a functions and hints, to apply to a physics, Microcomputer to Improve Graphing Skills chemistry, or economics course.The editor is designed so that the teacher can tailor the program Edward Zeidman, Division of Science and to fit his/her needs without knowing anything about Mathematics, Essex Community 'College, Baltimore programming. County, MD 21237 The Function Game has been classtested in precalculus avid calculus courses at Essex Community Computers are gaining widespread use as College, Baltimore County, Maryland.We found that instructional tools in the classroom.They have the program worked most effectively with groups of shown promise in such traditional tasks as drill three or four students. The interaction among the and practice, question and answer formats, and data students was very important in, their. learning manageLent. These are important applications of process. Inotructors remarked on how quickly they the computer, but in order to realize the full were able to identify the different functions. instructional potential of this relatively new The program, written to work on the Apple II medium, we must look beyond the above applications microcomputer, will be demonstrated.

187 . r 204 ABSTRACT: Computer Chronicon Pro_ect

Melvin H. Wolf, Professor of Humanities and English, Pennsylvania State. University, Middletown, PA 17057

In their endeavors to achieve a better understanding of the past as a step toward developing clearer views of the present and future, scholars and students in the humanities have long made use of two standard forms of chronological tables; the first provides chronological lists of items, and the second horizontal or vertical timelines. Both forms are designed to facilitate perceptions of relationships between various events and persons, but the utility of both forms as they are now generally available is diminished by their being set on printed pages by the choices of persons other than the users themselves. With computer terminals now readily available to investigators in all fields, it is time for the development of a machinereadable data bank of chronological informantion which a user can query with an eye to his particular needs and interests. Just as library users can now broaden or limit the range of responses to their bibliographic questions at library computer terminals, they should be able to broaden or limit the range of responses to their chronological questions at similar library, university, or classroom terminals. The purpose of this project is to develop a Computer Chronicon capable of generating customized chronologies to users' specifications. The specJ.fic objectives of the pilot project now in progress are to: 1) compile a sample machinereadable data bank of chronological information, 2) develop a working set of computer programs which will a) generate both cathode ray terminal (CRT) and hardcopy output b) of both list and timeline chronological tables c)using both data bank and usersupplied data d) for both interactive and batch processing, and 3) provide full and readable documentation describing the system and its use.

188 205 WHERE ARE WE GOING IN THE USE OF COMPUTERS IN PUBLIC EDUCATION

Sylvia Charp Editor-in-Chief Technical Horizons in Education

The use of computers in education covers a wide range of activities and applications and its use in schools has increased tremendously. We are currently 'being swept along with the tide, using buzz words such as computer awareness,. computer literacy, computer based instruction, etc., but without sufficient thinking concerning the direction we wish to go, profiting from experiences of others and then developing a plan for implementation. There is no doubt that-Many aspects of our lives will be changed, adapted, or modified by,' computer technology and that education ust prepare students to live in a technological world. However, "to get on the bandwagon" should not be the prevailing desire. We -now have sufficient accumulated experience and research tohelp determine, our direction, though what should be taught and to whom' is still being debated. We do have alternatives, and we can determine our needs and wants and better evaluate our options.

189 20G Computers in the Undergraduate Mathematics Curriculum Sheldon P. Gordon Suffolk Community College

ABSTRACT This session is designed to explore some division undergraduate mathematics courses. of the most, sophisticated and novel uses of The modules are used for problem solving, the computer in the undergraduate tutorial and modeling in differential mathematics curriculum. equations, calculus and precalculus At Brooklyn College, CUNY, a group of mathematics courses. are using many faculty have developed and explored a The mathematics faculty multifaceted approach to the use of of these modules for class demonstrations, the functiongraphing and computer to perform , the clerical and particularly routine work involved in producing drills, modeling programs. Several modules are laboratory worked examples and related instructional designed to be used as materials. The main direct and immediate supplements to classroom lectures. These impact on education at the College has been modules include suggested laboratory to the establishment of \a math workshop, exercises. Students haveaccess the laboratory staffed by, faculty and student tutors, to materials in a departmental Apple computers provide a major component of the math equipped with several remediationneeded by many students. The networked to a hard disk. workshop is used by apprOximately 1000 per At che University of Pennsylvania, the year and is based on the use of drill faculty has adopted the philosophy that materialsgenerated byprograms developed computing can be used in mathematics within the project, largely by Professor instruction to motivate the mathematics and Kovacic. to enable the student to solidify the grasp an Other major areas of development and of mathematical concepts. As exploration at Brooklyn Co lege include; 1) illustration of such usage, two versions of in syntax driven authoring languages and a course entitled Senior Seminar programs to support them; 2? the production Computational Mathematics will be of -materials for computer literacy and described. This course is a reguired computer scienceeducation\that illustrate course for computer math majors at the the workings of various ,algorithms by University. printing traces of theiroperations; 3) the The, first version of the course is An use of computer graphics on large and small Introduction toComputer Simulation. This machines to produce instructional materials course involves mathematical modeling, for distribution and use in classes; 4) the differential equations and probability utput of material designed this way to theory. The second version of the course ary high qualitytypesetting devices for is the Mathematical Foundations of Computer ucational publishing. Graphics, This version requires At SUNY College at Brockport, a group of proficiency in linear algebra and some fa ulty has developed a number of geometry. instructional modules for use in lower

ARTICIPANTS: M"chael P. Barnett B ooklyn College Ci y University of New York The on D. Rockhill SUNY College at Brockport Gerald Porter University of Pennsylvania

190

2117 Simulation: A K-College Teaching Strategy Beverly Hunter Targeted Learning Corporation Amissville, VA 22002

SPONSOR: Society for Computer Simulation

ABSTRACT Computer simulation is a widely used simulations or teaching students to problem solving aid inmany areas of life. be model builders. A variety of simulations, from board games The panel represents some of today's to role playing, hage been traditionally leaders in both the academic and commercial used by teachers. With the increasing community of model and simulation availability of computer resources, it is advocates. Thepanelists experiences span likely that computer simulations will all grade levels, from primary through become widely used in classrooms. college. As with every pedagogical strategy, educators should be cognizant of its benefits and drawbacks. The purpose of PANELISTS: this panel is to clarify some of the issues surrounding the use of computer simulation Alfred Bork as a teaching/learning tool. Discussion Education Technology Center topics'will include: UniVersity of.California, Irvine I. a recapping of the history of computer simulation models as classroom aids; Ludwig Braun

. 2. an overview and assessment of New York Institute of Technology commercial simulation models available Old Westbury, NY 11568 today; 3. specific examples of integrating Jonathan Choate simulations into the curricula; Groton.School 4. a look at-the development processes of Groton, MA .01450 simulation models; 5. the inherent dangers of using canned Tom Snyder simulations; i.e., the black box Tom Snyder Productions, Inc. phenomenon; Cambridge, MA 02138 6. alternatives to using canned

191 1)ULs Considering the Lack of Instruction Computing in Higher Education: Why?

Lincoln Fletcher, moderator MECC State University Instructional Coordinator St. Paul, MN 55119

ABSTRACT There is far lesshappening in the area question. Members will represent different of instructional computing - using perspectives on the situation: Computing' computers as an instructional tool - in Center Director, faculty member, higher educational institutions than in administrator, and computer,, coordinator. elementary and secondary school across the Throuall discussion and sharing of concerns country. Why is this the case? perhaps some new understanding of the This panel will discuss some of the problems and some possible solutions can be issues and problems related to this found.

PANELISTS

James W. Johnson Directorof Information Technology Univer ty of Iowa Iowa City, IA George Culp Assistant Director of Instructional Computing University of Texas Austin, TX.

Charles Parson Assistant Professor, Geography Department Bemidji State University Bemidji, MN ...

3PONSOR: ACM SIGCUE

192 2 0 The Funding Game: Playing to Win John T. Thompson Barker Central School Barker, NY 14012

ABSTRACT This. session describes the grantsmanship and area of funding priorities. Audience process inherent in applying for private is informed of the regional collections of monies for microcomputers. The audience the Foundation Center in public libraries, will leave the workshop equipped with a specific books onprivate foundations and broad background in applying for such funds addresses to write to for more detailed with detailed references to monies information on their individual needs. available through the large computer Emphasis is placed upon how to establish companies (e.g., Apple, Radio Shack, initial contact and rappdrt with foundation Atari). Contact persons, telephone numbers trustees. and extended bibliographies in the Attention then will focus upon the grant. grantsmanship area will be given. application procedures for several of the Fast-paced and_packed full of useful large computer companies that have insights into the grantsmanship process, corporate foundations. The majority of the this session will appeal to beginning grant workshop will be devoted to these specific writers, plus established money 'seekers Procedures. looing for a quick update and new Throughout the presentation, handouts information. will be distributed to the audience. The Attention is first given to the private method of presentation will be a lecture foundations which are depicted by type format with questions from the audience. (e.g., independent, corporate, community)

193 210 DESIGNING A PROGRAMMING COURSE FOR MBA STUDENTS

*David V. Cossey, Director The Wharton Computer Center The University of Pennsylvania - The Wharton School Philadelphia, Pennsylvania 19104 (215) 898-6422

David Rossien EXXON Corporation New York, New York

This paper discusSes the redesign of a was hoped that knowledge of APL would non-credit programming course for MBA students. provide a useful tool that students At Wharton this course is entitled "Problem would utilize throughout their time at Solving Using the CompUter" (BA814), and all MBA Wharton. students must either waive the course by credentials (based on prior computer COURSE GOALS experience), waive the course by exam, or pass the course as offered by the Wharton Computer Faculty and administrators have described Center. Wharton has offered such a course for the primary purpose of the course as several years, and recently the course was "facilitation of a better understanding of the redesigned. The study'described in this report nature of computers and their capabilities". and the redesign took place in 1980-81. The The course instructors and the Wharton principal issues raised during the redesign administration believe that such an period concerned the goals of such a course, understanding is important to a manager, whether the existing course met those goals, and especially. given the rapid influx'of computer if not, was it possible to design a course technology into the business world. The better suited to meet the-goals- consensus is that an active, hands-on approach to computer programming is essential for a INTRODUCTION foundation of computer operation and utilization. In the course of our research we For several years, BA814 used-the spoke with faculty and administrators from programming,language APL. The course consisted numerous graduate business schools throughout of approximately 22 hours of lecture presented the country. Most stated that'developing an by members of the Wharton Computer Center staff understanding of computers is important; their (generally MBA candidates), and was offered foUr respective schools concurred by requiring times each year: in August, during the Fall and students to complete a course similar to BA814. Spring semesters, and during the January inter-semester period.The August class met for A secondary, but noteworthy, goal of the three weeks, the semester classes for seven BA814 course is to provide the student with a weeks, and the January class.for two weeks. tool, the knowledge of a programming language, Which may be used while at Wharton. All MBA candidates must either pass or waive the course in order to graduate. During Finally, a tertiary but relatively less the past two years the waiver requirements have important goal is to provide 'a tool which might been relaxed, so that "substantial computer--- be usefultoi the student after graduation. experience" is sufficient to receive a waiver by While it is not expected that any Wharton MBA credentials. About one -third of the recent student will be doing a significant amount of Matriculants currently waive the course by programming, it is felt that with 'more and more credentials. businesses gaining access to computing power, the knowledge of a high-level programming APL was chosen for two principal reasons:%i,.., language might be useful to graduating MBAs. The widespread purchase and use of personal 1. It is an interactive language. BASIC, computers for home and office use might, in the the only other interactive language near future, make this a more important goal. available, was not well implemented on the DECsystem-10., DESCRIPTION OF THE FORMER (APL) COURSE

2. It is a very powerful language, and it The 22 hours of class time that comprised

211 194 BA814 were designed to instruct the student in 4. Students often found themselves the fundamentals of the API, programming ill-equipped to work with the computer language. A simplified outline, indicating the packages which they do use while at approximate number of class hours spent on each Wharton, such as SPSS, EMPIRE, IFPS and -subject, follows: TSP. This was because they had only (2) Introduction to Using APL on the been taught concepts relevant to the APT, DECsystem-10 language. __.(2)__APL Data Structures --- (6) Using APL as a Calculator (4) .Writing Simple Programs (Functions) in APL is fundamentally different from almost APL all other computer languages. Though it is true (3) More Complex Programs in APL that those differences make APL a powerful tool, (5) Manipulating Matrices in APL they also set it apart from languages such as (4) Formatting, Flowcharting, Discussion and BASIC, FORTRAN, PL/I, COBOL, Pascal and Ada. Review APL often makes the computer transparent to the user, hiding what the user should really see. There were three 1-2 hour 'quizzes', 8 In this way it may distort both the "nature of assignments, each of which required one to three computers," and their capabilities. hours of computer terminal use, and a final project that generally took 8 to 10 hours of There is without a doubt an APL "mind-set," terminal time and an equal amount of outside a way of viewing problems in a different fashion preparation. than one would if trained in one of the other languages mentioned above. The fact that APL is To pass the course, a student had to .not used extensively in the business community, accumulate a satisfactory average on the exams nor is it the primary teaching language of any (generally about 70%), and complete the other business school in our survey, suggests assignments-and the final project. Since there that emphasizing the APL "mind-set" does not was no penalty for dropping the course, most necessarily_facilitate_the_understanding of students did so when it appeared they were not computers and their usage in the business world. passing. Therefore, the number of students who receive grades of NC was quite low, around 5-10 Proponents of the APL language view it as percent. However, the course was sometimes the language of the future. However, computor subject to a high attrition rate; in January scientists have been hearing this claim for ten this was sometimes as high as 40 percent, and years. A recent survey, taken of the readers of ranged betueed.15120 percent in August, and BYTE magazine (BYTE, October 1980), shows the 20-40 percent.during the semester. "literacy" rate of APL to have dropped faster in the past two years than that of any other Students did not generally find the course computer language. The rate, according to this conceptually difficult. There is no question survey, is now 22. percent. The literacy rate that most students were learning what we have (i.e., percentage people who "know" the asked them to learn, though their responses on language) of Pascal, on the other hand, had course evaluation forms suggested they did so increased by over 150 percent, and is now at 40 grudgingly:. percent, up from 14 percent two years ago. This suggests that 10 years time should be sufficient to detezmire whether APL is the "wave of the PROBLEMS WITH THE FORMER COURSE future". Admittedly, more and more companies are using APL, just as more and more companies The critical problems with BA814 as it was are using computers, but the widespread taught are outlined below: utilization of APL that had been envisioned appears not to have taken place.

1. The course did not teach concepts which Virtually every BA814 course evaluation are basic to an understanding of the form used to come back with a comment about "the role of computers in the business heavy workload". With three exams, eight environment. Concepts such as assignments, and a long final project, BA814 structured programming, files, records, compressed the work of a semester-long course databases, and text editors were among into a few weeks. A significant part of the those not covered in the course. problem stemmed from the inherent nature of the APL language. There are so many different 2. Students reported the course unduly operators in APL that complex expressions must detailed and time-consuming, be taught before the student oan really use its particularly in light of its non-credit full power. It became apparent that it was not status. possible to teach APL in a less detailed manner,

3. The skill(s) taught in the course (e:g. the APL language) were not used again by the vast majority of students while at Wharton.

195 since we reit that we had already pared the MBA curriculum. The three that do not have such language down to its smallest possible subset, a programming course, have a formal course which while still demonstrating the power of the utilizes computer packages, such as those previously mentioned, in their "core" courses. language. We felt that it was not possible to further reduce the amount of time necessary to This suggests that a majority of these business teach APL. schools believe that a course in computer programming is relevant to the MBA degree. Experience over the past several years suggests that while APL may have been a useful Each school that requires a course utilizes tool at some point in time, the average Wharton a different format. In about half of the student was not using his or her knowledge of schools, the course combines Management APL in other courses while at Wharton. This was Information Systems and computer programming to a large extent directly traceable to the into one course. These courses are always for advent of the hand-held financial calculator and credit, and typically last a full semester. the development of special-purpose high-level Some schools, such as NYU and Columbia, had software packages. course structures similar to that offered at Wharton; non-credit short courses which teach When the computer is used in a Wharton only programming. Other schools give one-half course, it is usually via a pre-written package or one-third credit for their courses. No 'which requires only that the user know how to course lasts less than two and one-half weeks, log in and initiate the program, and not any and most short courses last for six weeks. knowledge of a computer "language". The By far the most common language taught is packages which are most frequently used are . LINDO, IFPS, EMPIRE, SPSS, BMDP, Minitab and THE BASIC. Nine of the twelve schools with MANAGEMENT GAME. programming courses teach BASIC, one teaches PL/I, one FORTRAN, and one a combination of PL/I In ai lition to the use of the computer in and APL. regularly-scheduled Wharton courses, many students use the computer to assist them in For our study, we requested syllabi from their Advanced Study Project (ASP) courses. In' each school, and received about half to examine. almost all eases it is a package, and not a Our findings indicate that most courses cover general-purpose language, which is used. Many the rudiments of BASIC programming.A few marketing students analyze their questionnaires simple programs are assigned, and some courses with SPSS, finance students use TSP or IFPS, and have,a mid-term or final. ASPS have also made use of EMPIRE, BMDP, and As far as programming as a tool for a IDA. In some cases, databases such as COMPUSTAT are accessed, and since APL can only read student's use while in graduate school, most specially constructed databases, a consultant professors admitted that it was not used very must write a program in FORTRAN or Pascal for much. Stanford noted that it had an extensive BASIC library (much like Wharton's APL Library), the student. and that several programs were used by students, Many students did not know what facilities particularly the plotting and regression (other than APL) were available on the Wharton routines, but that a knowledge of BASIC was not essential for this use. Some courses, such as DECSystem-10. BA814 did not attempt to teach the students what tools were available on the that at Carnegie-Mellon, spent a lecture on Wharton DECSystem-10, and because of time canned packages such as SPSS and text processing constraints, it was not possible to conveniently programs such as SCRIBE. The instructors at include this knowledge in BA814. The experience Carnegie-Mellon said many MBA students used the of one of the authors of this paper (Rossien) text editing facilities to write their papers, . suggests that the tools which students want to as well as to generate cover letters and resumes use now are not those which mimic calculators, for job searches. but rather packages such as those mentioned above- and programs with access to databases, which calculators cannot provide. SUMMARY

, 1: Learning a high level computer EXPERIENCE OF OTHER BUSINESS SCHOOLS programming language (in contrast to a "package" such as SPSS or TSP), is quite Part of this study included conversations useful, and-perhaps essential, to with administrators'and faculty at fifteen facilitating a good understanding of graduate business schools. Of the fifteen computers and their capabilities. surveyed, twelve offered some sort of required course in computer programming as part of their 2. The advantages of teaching the APL

196 213 programming language have not been \ or Pascal realized. For many reasons, APL was not used as a tool by the vast majority of 3. Instruction in a package of the MBA students, before or after they student's choosing graduate. Students found the previous course to be much more time - consuming It was envisioned that each of these parts would and difficult than they felt the be viewed as an independent "mini-course", and situation warranted. Yet for all this would normally be taken in the above order. work, students did not have a good understanding of what computers can do The goal of the first course segment was and how one might expect computer threefold: technology to affect the business environment in the future. 1. Live the computer novice an overview of the very basic concepts of computer 3. APL instructors agreed that 113cauee of systems. These include: its power and breadth it takes more time to teach the fundamentals of the APL -How to log on, log off, get help language than it would to teach many other higher level languages. -What are files, and how they are used 4. The concepts and principles behind APL, as well as the language itself, are not -What is the significance of terms at all widely utilized in the business such as CPU, memory; disk, terminal, world. Concepts such as Mee, records, batch, interactiveon-line, and program structure were dot really prograMming language and covered in BA814, as they are not an canned-package. important part of APL. These concepts are considered important and integral 2. Give the students some idea of what parts of EDP and should be included in tasks can be easily accomplished on the the MBA curriculum. Wharton DECeyetem-10, and generalizing from this, on any large computer system 5. Packages such as EMPIRE, IFPS, SPSS and in the business world. This means TSP are not as readily understandable to giving a description of the packages, persons' knowing APL as to those knowing databanks, hardware, and libraries "FORTRAN-like" languages such as BASIC available on this system as well as or Pascal. presenting a summary of what i8 available in today's business world, and 6. Most of the business schools we where trends in such fields as office contacted have a required course in automation, computer modeling, and computer programming as part of their database management are headed. MBA curriculum. Most schools teach the BASIC language in one of its many forms. 3. Teach the student how to use a simple Usually the BASIC that is used is an text editor enhanced version of BASIC.

The above material could be covered in four INITIAL RECOMMENDATIONS to five class hours.

It was clear that the goals of BA814 were We concluded that some form of high level not being met in an effective manner within the programming language instruction facilitates a structure of the APL course. After studying the better understanding of computers and their course outlines and syllabi of other business capabilities. The following 18 a net of schools, and speaking with Wharton faculty, concepts that should probably be covered in the staff, and students, we concluded that-it would introductory programming module: be possible to design a course to.better meet our goals. Our initial design of such a course 1. Variables, numeric and character-' consisted of three parts: 2. One and two dimeneional arrays 3. Conditional statement execution 1. An introduction to the Wharton Computer (1f-Then-Else) Center Facility 4. Looping and iteration 5. Subroutines and functione 2. Instruction in a language such as BASIC 6. Input and output to the terminal and to

197 214 files As originally proposed, the new course 7. Construction of a "Conversational' would have required slightly more class time program than previously. 8. Program modularity, block structured code, hierarchial problem solving As described above, each segment of the course attempts to meet specific goals. The The language of instruction should be one students would have learned the fundamentals of which is comparatively easy to learn, and allows computer systems and thinking, as well as the students to study the concepts suggested above capabilities of the Wharton Computer Center. without being constrained by artificialities They would have been presented with three major inherent in the language. A study of language areas of computer technology in business: text choices suggests that the possible languages are editing and word processing, data processing via all "FORTRAN-like", but there are several a high level language, and computer packages alternative languages to choose from. We felt which can be used to solve business-oriented that the best choices would be BASIC, Pascal, or problems." PL/I. PL/I was not implemented on a DECsystem-10. When we began our study, we were not aware of an enhanced version of BASIC for FINAL IMPLEMENTATION the DECSystem 10, so our initial Choice was Pascal. In -ourse of discussions about the After the initial recommendations were

proposed r, . of the course, we discovered distributed and discussed, a revised course an enhanced Jn of BASIC, which contained began to emerge that was a modified version of advanced control structures (IF-THEN-ELSE, the initial plans. We had initially recommended etc:), long variable names, expanded the use of Pascal for the Programming Language user-defined sub-programs, and many other module of the course. After we discovered and advanced features. Upon investigating this evaluated the MaxBASIC language from National version of BASIC (MaxBASIC from National Information Systems, we recommended its use as Information Systems), we decided to use BASIC in the language to teach. the ProgramMing Language module of our redesigned BA814. We decided that having three separate modules would create administrative problems, Topics in this section include the but we thought that the content of the three following (the numbers in parentheses are the modules should somehow be provided, perhaps by number of hours estimated): Introduction to alternative means. To this end the first module programming (1.5), structure cf a computer (Introduction to the Computer Center) was program (1.5), variables (1.0), expressions and combined with the second module (Introduction to assignment statements (1.5), input/output (0.5), a Programming Language) in a single course that conditional execution (1.5), iterative-looping would be required of all MBA students without (0.5), flow charts (0.5), one-dimensional arrays previous experience with computer programming. .0),.stri,Igs (0.5), two-dimensional arrays (1.5), subroutines and program modularity (1.5), The content of the third module reading from files (0.5) and 'Putting It All (Introduction to a Package) is provided in Together' (1.5). optional short courses offered by the Wharton Computer Center. Packages taught to meet the In the third section of the course the objective of the propcsed third module include: student would choose which ofa\half dozen or so SPSS, BMDP, IFPS, EMPIRE and TSP. packages he/she wishes to learn.\This section of the course would have certain prerequisites, Officially, BA814 now consists of the first since most of the packages assume he user and second originally proposed modules in a understands the nature of the problems he/she single non-credit course. This course is wishes.to solve. A partial list of the packages offered four times a year:August, Fall, which would be offered includes: SPSS, TSP, January and Spring. In August the course runs IFPS, EMPIRE, IDA, Minitab, Runoff (text for 7-8 class days, during the Fall and Spring processing). the course runs for about 7 weeks for 2-3 hours per week, and in January there are six 3-hour The course would have one individuat final sessions (Morday-Saturday).The January session project which would require the student t'use is run the week before classes begin for the the package to solve a real-life problem. Spring semester. Enrollments for this course----- Experience at the Wharton Computer Center during the first year were as follows: suggests that is is definitely possible for the student to become quite facile in any of the August, 1981 300 above packages after at most five class hour Fall, 1981 50 January, 1982 60

198 Spring, 1982 30 August, 1982 200

The course has been received very well, and we have found much less resistance to the revised course than we did to the previous course. We attribute this to two primary reasons:

1. The use of BASIC rather than APL 2. A more modest workload for the course

Many students either are buying or expect to buy a personal computer system within the next five years. A recent questionnaire' of Wharton MBA students indicates that 46% of the respondents expect to buy a personal system in the next five years, and 40% of the respondents indicated that they have used a personal computer system. The proliferation of personal computers has made the teaching of BASIC more 'palatable' and logical for many of the MBA students who take BA814.

As a result of this redesign of BA814, and 6supplement BA814, and to assist MBA students, faculty members, and other interested parties in selecting a personal computer system, the Wharton Computer-Center-sponsoreda-Computer Fair in September, 1981. The fair brought twelve local vendors to in an exhibit area. There were as many different systems on display as vendors. There were also lectures on various topics, including 'Personal Computer Systems: What They Are and Whatto Buy'. The 1982 Fair had twenty vendors and 1,500 attendees. We feel that the Fair has become an important part of the education prograMs provided by the Computer Center, and although not an official part of BA814, the Fair is in many ways an extension to BA814.

199 2 G A CURRICULUM FOR A MASTERS PROGRAM IN COMPUTERI'ZE'D MATERIALS MANAGEMENT

by Daniel G. Shimshak and Dean J. Saluti

Department of ManageMent Sciences University of Massachusetts/Boston-

Abstract materials management evolved from two sources- - minicomputer vendors and manufacturing practi- In this paper, a curriculum in computer= tioners. Digital Equipment Corporation (DEC), ized materials management is presented. The the largest AmeriCan manufacturer of minicomput- curriculum is designed for a graduate program ers, had recognized the fact that their PDP 11 and developed as the result of the combined family and the new VAX models supported a maior efforts of (1) Cambridge College. (2) American proportion of materials management' installations. Production and Inventory Control Society, (3) DEC's field service representatives and sales Digital Equipment Corporation. (4) Bay State personnel had come to the realization that the Skills Corporation, and *(5) the Boston Acade- users at these installations had skill level in- mic Community. The program is comprised of adequacies which prohibited proper system apPli.r. six modules of training, each lasting three cations. For example, users with materials/man- months. Courses are designed to be taught in agement training lacked computer skillOrhile the evening and weekends to allow students to technical computer persOnnel failed to understand maintain employment or to enter optional mater- materials management tecVniques. FUkhermore, the ials management internships. users had begun to speak outand/sought any and all available training sources( DEC's formal recognitpm of this problem was Introduction matched by APICS, the American Production and In- ventory Control Society. APICS, a professional This paper presents a curriculum in com- association for' materials management practition- puterized materials management which currently ers, clearly voied'a need for education in com- serves as a pilot graduate program in Cambridge, puterized materials management. An informal task de- Massachusetts. Such a curriculum although group of Bost6n APICS members. DEC representa- signed for a graduate program, can be modified tives and,University professors began to work to- fOr an undergraduate Bachelor of Science curric- gether -to develop a curriculum to meet the needs ulum. In addition, the actual courses within in the field. What evolved was a pilot program the curriculum. linked in such a way as to ad- with support from several sources: dress.various.training deficiencies can then be (1) Cambridge College--a small private college used in specific industrial applications. accredited to grant. masters degrees in management, The field of materials management is one of- contributed institutional resources which will growing importance. Industries involved in the lead to a graduate program in computerized ma- production of equipment, components andmater- terials Management. ials comprise a major element of the economy and (2) APICS--the'Boston Chapter membership donated their productivity strongly influences domestic time and resources to the development of a cur- living standards and competitive positions in riculum which reflects speCific skill levels rec- international trade. The materials manager must ognized by practitioners in the field. be involved with three mayor functions--systems (3) DEC--contributed a PDP 11/34 minicomputer design, "systems operation, and systems control system to be housed at Cambridge College to 2 (see Stevenson and Monks ). These encompass a support all program training. In addition DEC wide range of management activities related to provided technical inputto the curriculum de- produetion,.inventory, purchasing, stores, dis- velopment and made a heavy commitment toprogram 3 delivery. The computer has had a tribution and quality. (4) Bay State Skills Corporation-processes significant tmpact on recent advances inthe grants for the development of highteChnology field. COmputers are used not only for solving . oriented training through funds allocated by the problems dealing with materials management, but Governor of Massachusetts: and the State Legisla- also for generating reports, automating manufacture. They provided funding for this pilot pro- tUring processes, controlling operations and in gram. 4' designing products. Now, more than ever; is Boston academic community--representing Uni- the computer and materials management linked to- versity professors throughout the state from gether. academic disCiplines such as computer sciences, From a historical perspective, the need for management, management sciences, and management a masters degree program in computerized information systems. They played an integral part.

200 217 in the development of this curriculum. These pro- minocomputer system. Although generic minicomput- fessors held positions at institutions such as ers operating system's functions are presented, the University of Massachusetts/Boston, South- emphasis is placed on the knowledge of the-DEC eastern Massachusetts University, Suffolk Univer- PDP 11 due to its popularity and applicability sity, Bentley College, Northeastern University in the manufacturing environment. and Boston University. Finally the Program attempts to prepare the

. The current economic status of Massachusetts student for the interpersonal dynamics off' the supports thedemand for positions in computer- manufacturing work environment. Students must 5 learn to work in a project task group setting ized materials management. Massachusetts has where technical computer and manufacturingtasks the second lowest unemployment rate of the industrialized states. With'regard to manufacturing must be shared and accomplished by a team of pro-. activity and employment, MassachusettS statistics fessionals. Various interpersonal skills and man- are much healthier than mosi; of the remainder-of agement principles.; essential-to success in this the nation., Also it is critical to note that the area, are taught. most important piece of tax limiting legislation in the state, Proposition 2 1/2 which was approv- Curriculum Tracks ed in November 1980. had had its major impact on The program curriculum can be found in the norunanufacturing employment. Thus economic in- Appendix. This program is designed to be taught dicators reveal a maintenance of employment rates in the evening and weekends to allow students to and economic stability in Massachusetts which is maintain employment or to'enter optional mater- essential for continued growth in this area of ials management internships. There are six mod- computerized materials management. ules of training, each lasting three months; thus the student may acquire a masters degree in Program Objectives computerized materials management in one and a half years. The admissions criteria provide an The first objective of this program is to opportunity for individuals seeking a career build a foundation of technical skills in ma- change as well as those who have a background in terials management. These are primarily mathe- .business or manufacturing. Various courses such matical skills with manufacturing applications. as accounting techniques or statistics can be For example, inventory control draws upon alge- waived or required given the student's previous bra Lad probability theory, quality control re- educational background. lies on statistical analysis, and shop floor Track I -- Materials Management Techniques- - scheduling applies matrix algebra. Whereas includes the following courses: mathematical analysis of these problems has un- Basic Production and Manufacturing Princi- til recently been performed by hand, today reliance on the computer has become mandatory. ples Bill of Materials Techniques A second major objective is to develop the Capacity Requirements Planning ability to write 'programs to perform materials Purchasing management functions. Teaching' emphasis is placed Introduction to CAD/CAM on the development of the students' ability to de- (Computer Assisted Design/Computer Assisted sign systems in a structured logical manner in Manufacturing) virtually any and all applications and systems Introduction to Robotics languages. This is accomplished by providing an Forecasting Techniques environment'which will require students to build Shop Floor Control systems (such as Materials Requirements Planning Inventory Control Systems --MRP) in the following applications languages: Manufacturing Production Scheduling BaLic. Fortran IV, Fortran 77,,Watbol. Cobol, Basic Principles of MRP ,PL/1, and PL/C. The student completes the pro- Advanced Principles of MRP. gram with a portfolio of materials management programs, each materials management technique This track provides a technical expertise in being programmed in as many as five different nearly all areas of materials management as en- languages. In addition these same techniques will dorsed by practitioners and professional associ-

be programmed in PDP 11 Assembly in order that ations . as APICS. The traditional educational they may acquire systems programming expertise envirom_ these courses would allow the and familiarization with macros. Proper program- student t, thematical techniques to ming techniques are emphasized such as structured applications. Program moves forward to pro- design and eodingas well as systems documenta- vide'the tools for the design of computer systems tion. In meeting this objective students will for the specific materials management techniques. gain the ability to effectively link materials The logical nrogression of the courseware begins management skills end techniques to computer with the basic materials management principles, applications and programs. moves through specific techniques, and concludes The third objective is to develop in:the with the all-encompassing systems environment of student an expertise in minicomputer operations. MRP. It is important that the students can configure a minicomputer system within a manufacturing en- Track II--Applications Programming for vironment; can plan a conversion in hardware in- Materials Management includes the following stallation; and can actually operate and manage a courses:

201 218 Introduction to Computer Programming Logic of this track is to acquaint students with impor- for Manufacturing (Basic) tant principles of the business environment that Intermediate Computer Programming Logic for they will undoubtedly encounter in the perform- Manufacturing (Fortran IV) ance of their materials management functions. Intermediate Computer Programming Logic for Track VI--Behavioral Dynamicsincludes Manufacturing II (PL/1) Management Style for Manufacturing, Job Oearch Advanced Computer Programming Logic for Techniques and six human relatiOns, skill build- Manufacturing (Cobol) ing courses taught within a workshop environment Systems Programming Techniques (PDP 11 in each module. The topics covered-in the Human Assembly). Relations courses include interpersonal communications. team building, group processes, motiva- The intention of this track is to familiar- tional techniques, leadership, conflict inter- ize.the student with various applications lang- vention, organizational effectiveness and inte- uages while integrating materials management gration techniques. A course in management style programming assignments with programming tech- is included because it is perceived that grad- niques. The student will learn Fortran IV by uates of this program will quickly move into writing Bill of Materials programs, for example. entry-level management positions. Finally the The number of languages taught will depend upon course dealing with job search techniques will the.availability of compilers for.the PDP 11/34; assist graduates in moving into immediate employ- thus languages such, as Watbol, PL/C and Fortran ment. 77might also be introduced to the student. Mac- ros and systems Programming skills will be ac- Summary quired through PDP 11 Assembly language. Track III--Computer Systems Skills for Mini- A consortium of contributors from the menu- computer Applications--includes the following 'featuring industry, computer industry and aca- courses: demia have worked together to develop a graduate Introduction to Computer Applications in program in computerized materials management. The ManufactUri:Ig intent of the contributors was to develop a pro- Computer Law gram curriculum with the following primary ob- Computer MRP Systems I. jectives: Computer MRP Systems II. (1) to build a foundation of technical skills in Within courses such as Introduction to materials management, Computer Applications in Manufacturing and Com- (2) to develop the ability to 'write programs to puter MRP Systems I and II students will become perform materials management, familiar with commands for various software (3) to develop an expertise in minicomputer oper- packages that are a7ailable from vendors ations, and (Original Equipmer Manufacturers or OEMs) for (4) to in-et/are for the interpersonal dynamics of the DEC PDP 11. Computer Law is an important part the manufacturing work environment. of the curriculum given that practitioners must The courses'are offered in six modules and continually negotiate contracts with hardware assembled within the following tracks: and software vendors /OEM's who offer materials ,(1) Materials Management Techniques, management processing resources. In addition DEC (2) Applications Programming for Materials Man- RSTS Operating Systems:. and DEC Systems Manager agement, topics are included within the content of speci- (3) Computer Systems Skills for Minicomputer fic courses in this track. Applications, Track IV--Mathematics Workshops--includes (4) Mathematics Workshops,

six optional, noncredit mathematics workshops (5) Manufacturing Business Environment, and , and a ,Statistics Applications course. AS quanti- (6)Behavioral Dynamics. tative methods are taught within each course in The pilot class will begin in July 1983, the Materials Management Techniques track, the with an expected enrollment of 30 students. Con- student is offered support ina workshop en- tinuous input will be solicited from program de- , vironment. Thus students with an innate fear of velopers and Program graduates. In this way the mathematicsnan,be assisted. A basic statistics courseware will remain within the state of the course in included which can be waived by stuart. dents who have achievecha grade of B or better in Appendix a college statistics course. Most materials man- Computerized Materials Management agement techniques incorporate statistics prin-, Proposed Curriculum cioles thus justifying the statistics reauirement. Track V-- Manufacturing Business Environment Module I Contact Hours --includes the following courses: Introduction to Manufacturing for IntrOduction to Manufacturing for the Busi- the Business Environment 20 ness Environment Basic Production and Manufacturing Accounting Techniques for Materials Manage- Principles 140 ment Introduction to CoMputer Applications Financial Techniques for Materials Manage- in Manufacturing 20 Introduction to Computer Programming Againment. the opportunity exists for students to Logic for Manufacturing, 8o waive these courses if they have achieved a B or (Human Relations Course) 140 better in similar college courses. The purpose

202 Math Workshop I (Optional) 40 Outlook on MatistchUsetts, Commonwealth Books, 240 Palisades,NJ. 1982. Module II Accounting Techniques for Materials Management 40 (Human Relations Course) 40 Intermediate Computer Programming Logic for Manufacturing 40 Bill of Materials Techniques 40 Capacity Requirements Planning 40 Math Workshop II (Optional) 40

Module III Financial Techniques for Materials Management 40 Statistics Applicatiohs 40 InterMediate Computer Programming Logic for Manufacturing II 40 Purchasing 40 (Human Relations Course) 40 Math Workshop III (Optional) 40

240 . Module IV Introduction to CAD/CAM 20 Introduction to Robotics. 20 Forecasting Techniques 40 Advanced Computer Programming Logic for Manufacturing 40 Computer Law 20 Shop Floor Control 20 (Human Relations Course) 40 Math Workshop IV (Optional) 40 240 Module V Inventory Control Systems 40 Master Production Scheduling 20, Basic Principle's of MRP lib SysteMs Programming Techniques; 40 (Human Relations Course) 40 Compuier MRP Systems I 20 Math Workshop V (Optional) .40 24o Module VI Advanced Principles of MRP 40 Management Style for Manufacturing 4o Computer MRP Systems II 60 (Human Relations .Course) lip Job Search Techniques 20 Math Workshop VI (Optional) 40 24-6-

References

(1) William J. Stevenson, Production/Operations Management, Richard D. Irwin, Homewood, IL. 1982. (2) Joseph G.-Monks, Operations Management, McGraw Hill, New York, 1982. (3) Jack N. Durben, "Materials Management Online System," Proceedings of the 24th Annual International Conference of APICS, Boston, MA. October 6-9, 1982, pp. 27-30. (4) Donna Hussain and K.M. Hussain, Information Processing Systems for Management, Richard D. Irwin, Homewood, IL, 1981. (5) Roger J. Deveau, Dean J. Saluti and Daniel G. Shimshak, "Economic. OpportUnity--An Economic Profile of Massachusetts" in The

203 22u Information Literacy Course: A Recommended Approach

Eileen M. Trauth

School of Management, Boston University Boston, MA 02215

information in whatever form it arrives and by Abstract whatever means it is processed - via computer, Given the pervasiveness of computers in our or otherwise. society,much recent attention has been focused Fundamental to this view is the notion that on the development ofliteracy courses to prepare students for this new era. The typical information is a phenomenon that has Thus, one could study approach is a computer literacy course which independent existence. information in much the same way, as one could introduces the students to some programming study energy, or music, or basket weaving. The language. This paper presents an alternative The goal of emphasis would be on the issues associated with course on information literacy.. Some of these information literacy is to be able to respond the existence of the phenomenon. behavioral. to'the demands of an information-intensive are technological, others are Thus, the spectrum of possible study ranges society. from systems design and programming to human information processing to societal impacts of Introduction information. These areas of study are, The arguments supporting the need for a however, shaped by an "information This perspective means literacy course in the computing area havebeen perspective." considering the phenomenon to be distinct from made and are widely accepted. What is not so widely accepted, however, is the form thatsuch the media used for storage, processing or transmission. Since the focus is on the a course would,take. The rationalen for computer literacy stems primarily fromthe phenomenon and not the-medium, a broader Thus, the emphasis for pervasiveness of information processing outlook is possible. problem solving would be on the satisfaction of technology in our society.Most fields of the information need rather than on the study require interaction with the computer As such, attention Upon graduation, virtually all areas manipulation of technology. resource. for of employment will bring the graduatein can be given to appropriOte tools information processing and transfer- whether it contact with information processingtechndlogy. be print, electronic, video, etc. An additional aspect of this perspective is that The response to this need ranges from a Since service course in the computer science it allows for the behavioralcomponent. the various media involved with information are department to specialized courses within the seen as being only the tools, the- focalpoint various disciplines. Literacy in this regard can become the people - thoseaffected by these is generally labeled "computerliteracy" and aspires, to make the students proficient in the tools. manipulation of a computer via some highlevel The underlying assumption, it seems,' Such an approach to literacy will enable languge. the student, in this author's view, to cope is that learning how to successfullymanipulate the technology is the best preparation for an with the issues surrounding an 'information-intensive society that go beyond information- intensive society. This paper suggests another approach, one based uponthe manipulation of the tools. The student would develop a methodology for "learning to learn." goal of information literacyrather than Additionally, the student would develop an .computer literacy. According to this view, orientation toward information as anexplicit what should be the focus of attention is the - and valuable organizational resource. Finally, information itself. The computer is viewed as assuming that most people will be users of 'a tool - albeit the major tool -used in the. information rather than developers of its processing of information. Thus, it is studied systems and technology, they will,ke able to - but from the viewpoint ofits relationship to develop their ability to articulate'their the information. Information literacy, then, information needs and communicate them either is a broader concept than computer literacy. to another person or to a machine. The goal is not competenceinmanipulating the computer, as the latter termimplies. Rather, the goal is the capabilityof working with

204 221

1 Recommended Undergraduate Literacy Course understanding it and is motivated to do so. Some of the distinctive properties of There are certain assumptions underlying information that can then be noted are the the recommendations that will follow. First, following: Information has subjective it is assumed that this is a required course existence, while data has objective existence. for allstudents in the academic unit the Information is not depleted with use. department,' school, or college. Second, for Information is intangible. The overall intent the majority, of the students, this literacy of the treatment of information as a distinct course might be the only computerrelated entity with distinct properties is to separate course.that they will take. Third, it is the information content from the media or assumed that one of-the goals of such a course technology by means of which it is conveyed or would be its integration into the subject processed. matter of the Major discipline.That is, this course would become an integral part of the A second major topic that should be student's educational experience. This addressed is that of systems. Hany people use reflects\the attitude that the study of the terminology of information systems without informatiOn should not be separated from the ever having stopped to think about the setting in which it is to be used. A final implications of the term. Recommendations assumpton is that the study of information is regarding the treatment of the term not synonomous with, the study of the technology "information" have just been presented. used to process it. The latter is an aspect of Recommendations about the term "system" follow. the former. Students should first be introduced to the In keeping with the emphasis on information notion of a system in general (and accordingly, literacy this course should, among other systems thinking). They should then probe the things, familiarize the student with the interaction of systems and information. Part information environment. That is, the student of doing so involves examination of systems should become comfortable with the process of used for the processing and communication of articulating theinformation requirements of a information. Thesesinclude software packages, 0.-Ven situation. In addition, he/she should'be new telecommunications offerings and manual able to identify and understand the role of procedures. Another apsect of the study of information flows within the organization. For systems involves the study of the this reason the course should include organizational systems within which information discussion of notions associated with flows. If an information system is to automate information as an independent entity. The the flows of information within the origins of informaton should be examined organization, then it is encumbent upon those relative to-the situations in which the working with such systems to be able to students would be working with information. understand the existing framework that holds The properties of this phenomenon should be the information. examined as well. In order to this, it is necessary to understand the distinction between An overriding orientation that should data and F.-.formation. While this may seem influence the preceding treatments is the fairly obvious, there appears to be much particular disciplinary setting. It has confusion over the terms as evidenced in already been argued that information only has students' perceptions about the topic. A recent real existence within a given context. For discussion in an introductory course_yielded this reason it is crucial to the introductory. such responses as: understanding of information and its systems that different scenarios of information "Data is what you have when generation, and processing be used.Business you use computers; information students will encounter a different type of

is what you have when people information than that for library science - do the work." students. Liberal Arts students interact with information in ways much different from those Or of science majors.

"You take information and put A third major topic should be the study of it into the computer and do information processing. In an era when the things like statistics on it limits of computer processing continue to be and. then you get data as the expanded, students who may never take another output." formal course in this area need to develop a view of the entire srectrum. This involves The viewpoint held for this proposed course treatment of three major areas. is that data should be understood to be the raw material out of which information is created. If it is accepted that information is a It is, therefore, information in potential. phenomenon whose existence depends upon the Data, when processed, does not necessariy characteristics of the recipient then it is result in information. It only does so when necessary for the students to have some the recipient' of the data is capable of understanding of human information processing.

205 222 This includes treatment of both the ways in should examine the ways in which information, which humans process information in general and systems, and organizations interact. Again, the ways that the particular people involved in such a treatment should be geared to the disci- the student's discipline process and use plinary setting. In a business curriculum the information. A secondary benefit is that this students would consider the information understanding is helpful when learning about problems of industry and the available ways of machine processing. solving them. A literacy course for a liberal arts curriculum might emphasize societal impacts Since commercially available software of new technology, coping with the "information packages are increasingly the norm, students explosion," etc. One way that such topics can should have exposure to the types that they be woven into the fabric of the course is to would be likely to use. They will then not have each student give a five minute presenta- only have experience manipulating "real world" tion at the beginning of class meetings. The software, they will also have the opportunity students are told to report on a recent journal in an introductory course, to examine examples or newspaper article. They typically focus on of more sophisticated software than they are leading-edge applications, societal issues, and capable of writing themselves. the new types of technology that are emerging. These talks often lead to stimulating The final aspect of information processing discussions. In addition, the students are is to experience developing programs made to feel that a portion of the course is themselves. This is often the easiest segment governed by their particular interests. of such a course. The students usually feel that they are working with something concrete The Goals of an Information Literacy Course that has specific outcomes. Despite the fears that some might have upon entering, by the end A very important outcome of such a literacy of the course the majority of students are course is the demystification of the usually quite excited about working with tecnonology. Students with no particular computers and are pleased with the acquisition interest in or inclination toward computers of a new and valued skill. The' challenge when develop a sense of self confidence. Through incorporating a programming segment into such a successful experiences with machine processing course, however, is to maintain the proper they are reinforced that with some knowlege perspective. The goal of the programming humans can be in control of the technology. segment is not one hundred percent proficiency in some high level language. Rather, it is to Closely related is the second intent, that understand how to process data into information the technology be placed in the proper by means of a computer program. The emphasis perspective: secondary to an understanding of is placed on learning the input-process-output the need for and the uses of information. Tne sequence. The students learn how programs work course should convey a user-oriented in general through writing specific programs. perspective. Since the majority of the The particular language used is treated as a students will eventually be users, this outlook vehicle for conveying such concepts. With is appropriate. Because majors should develop thirty to forty percent of the coursework a sensitivity to the people for whom they will devoted to computer processing only a subset of be working, this perspective is fitting. the language can be taught. The students are \told this. They are taught enough to be able The third objective can be inferred from to see the way in which computers process the comments just made. This course should data. Suggested constructs would be: simple include both majors and nonmajors. Both groups input and output, sequential file processing, will benefit from exposure to the other. calculations, transfer of control and looping. Having a range of capabilities and interests By writing programs in a given language the poses a challenge to the teacher-. This author students should develop a generalized believes, however, that the benefits are worth understanding or how computers and software the extra effort. This type of literacy course function. If some attention has also been provides nonmajors with an exposure to the given to human information processing then the entire field. By having majors take such a students are able to contrast their own mental course as their introduction to the field they operations with the operations of the receive an overview of the range of issues computer. This understanding has proven to be rather' than in-depth exposure to a narrow area valuable when students encounter difficulties. (uhic:h is usually the programming dimension). with theirprograms. A typical example is the holistic or gestalt approach'to problem solving The final objective of such a course is 'to taken by, humans as opposed to the linear provide the students with some tools for coping approach taken by the computer. Another is the with the information-intensive society in contrast between the amount of ambiguity each which they will be working and living. By can tolerate. learning about information and how it is processed they will develop skills.in "learning The final topic to be covered in such a to learn."With increasing obsolescence of course should bring together the previous knowledge, this is perhaps the best that can be three. Through problems and cases the students hoped for from an education.

206 223 Conclusion

This type of literacy course has been taught in a variety of institutions to diverse types of students. Student reaction is consistently positive. In many ways this type of course is more difficult than one. which treats a narrower cut of the subject area. The body of knowledge is large and is constantly growing. Most textbooks are not orientecito such a course. But given the pervasiveness of information technology and the need for all students to learn how to cope with it, such a literacy course appears to be a reasonable response.

207 224 A System for the Automatic Grading of Programming Style

Patricia B. Van Verth and Anthony Ralston

'DepartMent of Computer Science State University of New York at Buffalo Amherst, New York 14226

Abstract in the program. Quality in this sense is taken to mean the antithesis of the complexity of control With current emphasis on programming and data structures. The research described in methodology in introductory computer science this paper is concerned primarily with automating courses, automatic grading systems which merely the evaluation of program quality in student check whether programs produce correct answers programs, and, in particular, program quality as have become obsolete. This paper describes a it appears in introductory Pascal programs. method for automatically grading student programs for style using a system which implements a Benefits of Grading Style Automatically mathematical model of program quality. The quality metrics obtained from this system relate Automatic grading of student programs places to the choice -a use of data structures and the process of evaluation on an objective basis control structures,- and to t eIvision_of_the since the automation itself requires that program into procedures. A database of programs standards for judging have been defined and

and grades is being created in order to test the implemented. Often in the case of human graders . system and provide material for further research. lack of objectivity occurs when a set of programs is graded by several different persons or even when processed by the same person. This usually happens due to the absence afclear guidelines for Introduction style, with graders relying on personal opinion or intuition. In the case of several graders, these Relief from the burden of grading student opinions often 'vary, thus, producing programs is high on the wish lists of all computer inconsistencies in the grades. In the case of an science instructors. Student programs are a individual grader it is difficult to maintain the necessary evil in introductory programming Classes same level of consistency in the presence of large and other computer science courses in order for numbers of programs and lack of clear-cut students to demonstrate in,a practical manner standards. Moreover, time pressure on graders their mastery of programming skills. However, it results in evaluations of program quality and is possible that performing the grading task could appearance which are often too cursory. With an be relegated to the computer itself. This in fact automated system, the definition of measurements has been done in the past for evaluating whether implemented in a computer program means that those programs run correctly or Compile correctly. With measurements will be applied in a manner which is the current emphasis on programming methodology in impartial, effective., and consistent. This in elementary level courses, these automatic grading turn should instill confidence in students that systems have become obsolete by virtue of the the assessment of programs is meaningful and fair. deemphasis on correctness: not only are programs It is our opinion that an automatic grading system expected to work correctly, they must also be with well-defined measurements of program quality well-written, thus demonstrating good programming can do as well as expert human graders who take style. considerable care in their grading, and will do better than almost all human graders in practice. Programming style in the context of this _paper is taken to be a combination of program As well as achieving objectivity, an appearance and program quality. Program appearance automatic grading system would considerably is that component of style that includes the alleviate the shortage of personnel to conduct format and internal documentation of a program. computer science courses and to do the grading Program quality refers to the way in which the associated with these courses. Grading projects implementation of the algorithm has been is a lengthy, time-consuming and boring task. It accomplished, i.e. the way in which data is beneficial to an instructor when the instructor structures and control structures have been used obtains feedback_on the effectiveness of

208 225 instructio' ;however, most times the instructor just simple counts of structures since a count' does not do the grading and gets only a essentially reflects which structures have been superficial impression of how students are selected. To gauge the use of these structures, performing in this aspect of the course. Indeed, the measurements should also incorporate the data from an automatic grading scheme could give effect these structures have on the program an instructor far better feedback than is normally itself. Thus, measurements ofprogram quality achieved. On the other hand the student derives should include both syntactic and semantic the maximum benefit if the grading has included analysis. E.g., it is not only important to note constructive criticism and praise. Given the size that a WHILE loop has been used in a program but of classes and the number of persons available to also to recognize the effect that loop structure do the grading, this is an almost unreachable goal has on the flow of control of the program. within reasonable time constraints. An automated system would be capable of providing similar An Approach to Program Qualitr/Complexity analysis in greater depth and, with more accuracy. The approach to grading program quality Background of Automatic Program Graders discussed here originated in program complexity studies done by Enrique Oviedo at Programs for grading student programs have SUNY/Buffalo[5,6]. In that research Oviedo been in existence almost since the time when proposed a mathematical model of programs and from computer programming courses were first taught. that model derived a set of measurements for The earliest known such system was developed in program complexity. Prdgram complexity is an area 1958 by Jack Hollingsworth at RPI and used to of software metrics which seeks to measure in a grade assembly language programs[4]. This program program the degree of difficulty a person has in evaluated programs for correctness of the answers. understanding a program in order to debug, modify Later versions of program graders shared the early or maintain the program. Its importance arises version's concern with correctness, included from the couderable amounts of time and money efficiency checks for time and memory usage, and which are spent in performing these three tasks in were extended to grade programs in a variety of large computer systems.To.date there are no high level languages[1,3]. In the early 1970's, absMute standards of program complexity; in fact emphasis in computer programming courses shifted there is no.generally accepted model or toward programming methodology. This shift measurement of that attribute[2]. Nevertheless, resulted in a relative deemphasis on correctness we see in the work of Oviedo the po ential for in evaluating programs. Other factors such as developing an accepted model of program quality readability, quality and clarity became equal in which can be used is the context of automatic importance to program correctness. Since a grn4'.ng for programming style. 'correct program in this context is one which produces correct answers for a standard data set Description1EComplexity Measures or student-supplied data set, correctness can be readily checked by computer. However, evaluating a In his work, Oviedo models programs using the prograM for its style is a much more difficult control flow graph ci the program.Measurements task by virtue of lack of either a generally for the complexity of control flow are made by accepted definition of style or standard-3 by counting the edges on the g.-aph which relate to which to judge style. the number of potential branches in a program. Data flow measurements are made by computing data Some attempts have been made io relate flow equations over the same control flow graph; program style to resource usage_ program these are related to the data complexity in a appearance18,91. These attempt:: have not gained program. These measurements combine syntactic wide acceptance since it seem,., clear that style analysis, a parse of the control and data encompasses more than is implied by these terms. structures in a program, with semantic analysis, Criteria used in grading style in programs usually the effect each structure has on the control flow includes the appropriate choice and use of control and data flow of the program, to arrive at the structures and data structures, and the division complexity measures.While no absolute standards of the program into procedures. These criteria exist, measures of this kind permit comparisons remain sufficiently ambiguous (note the word between programs implementing the same task "appropriate") but can be rigorously defined by involving essentially the same algorithm, i.e. one embodying them in a mathematical model of can ascertain whether one program is less complex programs. Then the quality of the use of data than another by comparing complexity measures. If structures, control structures, and procedures in the model is appropriate, the results wi.11 a program can be measured using that model. This identify those programs which perform the permits the comparison of programs performing the programming task in a less complex manner than same task on the basis of these measurements, others. This then provides the basis for an ultimately allowing one to determine that one automatic program grader for programming style. program is better than another if the measures See Figure 1. have been suitably chosen and the measurements are consistent with this conclusion. Programs implementing the complexity measures were used as part of early experiments conducted To support such conclusions the measures by Oviedo. Based on these early experiments and derived from the model should include more than reported elsewhere at this conference[7], the

209 22G PROGRAM COINT(INPUT,OUTPUT); PROGRAM TEXTCOUNT (1NPUT,OUTPUT); CONST BLANK = ' '; CONST B="; VAR VAR WHAR, NWORD, NLINE: INTEGER; NCHAR, NWORD, NLINE INTEGER; POS: BOOLEAN;

OLDCH,NEWCH : CHAR; CH: CHAR; BEGIN BEGIN NCHAR := 0; READ(CH); NWORD := 0; IF CHOB NLINE := 0; THEN POS:=TRUE WHILE NOT EOF DO ELSE POS:=FALSE; BEGIN IF TRUE OLDCH := BLANK; THEN NWORD:=I WHILE NOT EOLN DO ELSE NWORD:=0; BEGIN NCHAR:=I; READ(NEWCH); NL1Nr:=0; NCHAR := NCHAR +1; WHILE NOT EOF DO IF (OLDCH = BLANK) AND BEGIN (NEWCH 0 BLANK) READ(CH); THEN NWORD := NWORD + 1; IF EOLN OLDCH := NEWCH; THEN END; BEGIN NCHAR :m NCHAR + 1; NCHAR:=NCHAR+2; NLINE := NLINE + 1; NLINE:=NLINE+1; READLN; READLN; END; CH := B; WRITELN(' THE NUMBER OF WORDS IS ',NWORD); END WRITELN(' THE NUMBER OF LINES IS ',NLINE); ELSE IF (POS=FALSE) AND (CH=B) WRITELN(' THE NUMBER OF CHARACTERS IS ',NCHAR); THEN NCHAR:=NCHAR+1 B4D. ELSE IF (POS=FALSE) AND (CHOB) THEN Data Flow Complexity = 20 BEGIN Control Flow Complexity = 14 NCHAR:=NCHAR+1; NWORD:mNWORD+1; Example of Good Program END ELSE IF (POS=TRUE) AND,(CH=B) THEN NCHAR:=NCHAR+1 ELSE IF (POS=TRUE) AND (CHOB) THEN NCHAR:=NCHAR+I; IF CHOB THEN POS:=TRUE ELSE POS:=FALSE; END; WRITELN(' THE NUMBER OF WORDS IS ',NWORD); Figure 1. Example Programs WRITELN(' THE NUMBER OF LINES IS ',NLINE); WRITELN(' THE NUMBER OF CHARACTERS IS ',NCHAR); END.

Data Flow Complexity = 69 Control Flow Complexity m 34

Example of Poor Program

The programs included in this figure are whereas the poor one has one WHILE loop and eight programs which count the number of characters, IF-THEN-ELSE statements. The data flow complexity words and lines in a text file. Characters mean is lower in the good program since there are ten all characters including the end-of-line mark. assignments or variable definitions with ten Words are strings of characters excluding blanks referentes to those definitions vs. eighteen which are separated by blanks and end-of-lines. assignments and twenty references in the poor Lines are signified by end-of-line marks.The program. The large difcerence in data flow most difficult part of the program is determining complexity (20 vs. 69) between the two programs the word count, i.e. detecting where one word arises from the effect assignments have on ends and the next begins. references in data flow complexity measurements. At each reference, all of the previous assignments The good program has lower control flow which might affect the value of a variable at the complexity than the poor one since it uses two point of reference are taken into account, WHILE loops and one IF-THEN-ELSE statement, producing a multiplicative effect.

210 22 measures obtained using the complexity model defined rules for indentation and commenting; the compared favorably with the results obtained when automatic grader then measures how well the rules programs were graded for complexity by expert have been applied.' Output format reflects the graders. These early results imply that, in manner in which the output is presented. Note that practice, when program quality grades are assigned program correctness does not include the by human graders, the automated complexity appearance of the output. External documentation measures will compare favorably with these. Thus, includes any separate documentation required as the automated results seem to be capable of part of the programming assignment such as a _discriminating quality programs as well as human discussion of the stepwise development of the graders. These results have encouraged the program. Thegoal of such categorization of grades current work in expanding the model and in is to make students more aware of the various developing a comprehensive automatic grading aspects of producing "good" programs. Hopefully, system. results from human grading will be sufficiently accurate to permit comparisons with automated Assignment of Grades measures in the various categories.

Since no absolute standards for program Implementer'.. of the System quality exist, i.e. a quality measure of 50 has no ° meaning by itself, part of the process of using Implement:ti, , the autr cic the grading programs will be to calibrate the analysis portion of the system nas bee. 'tiall; grades for the particular programs at hand. An completed using a front-end parser for Lae Pascal assumption that must be made with this system is language to derive control flow graphs and sets of that programs being compared are programs which defined and referenced variables from source implement essentially the same algorithm and are programs. These graphs and sets are then thus capable of being compared. One approach to processed by a complexity-measuring program that calibrating the system is to provide a "perfect" counts edges on the control flow graphs and program against which to compare student calculates data flow equations. Programs are solutions. To ensure fairness, the "perfect" (i.e. processed on a procedure-by-procedure basis for instructor's) program could be awarded a 95% grade both types of complexity measures and the total to give the students the opportunity to arrive at for each type is obtained by summing over all a better solution. Student grades forquality procedures. This yields two components in a will be assigned using the model values relative program complexity vector which is subsequently to those of the instructor's program. used for comparing programs.

Testing of the System In addition to the programs for measuring complexity,. implementation of the grading system The grading system described here is unique has progressed in three other areas: the since it not only 'automates the process of grading collection of programs, the.collectiou of grades, program quality, but it also tests this process the construction of a database. Programs are against human graders to determine system collected by requiring all students to use a performance. The testing portion of this research special job control language (JCL) command when has been started by collecting the programs from running versions of programs to be handed in for several sections of introductory computei science grading. This JCL is very easy to use; it is no courses using the Pascal language. Along with the more complex than the normal system calls to the sources for the programs, grades assigned to the compiler and the linker/loader. The JCL programs are also being collected for later' controls the compilation and execution of the testing of the system. These are being entered program, and it collects a copy of the source and into a database constructed for the purpose of output of the program for entry into the database. providing sets of programs and grades for a The student receives a -compilation listiLg, output variety of experimental' purposes. and two grading forms attached to the listing by the JCL. The grading forms have a,coded ID number In order to make grades more meaningful for which permits later correlation with the collected experiments, the recorded grades are sub-divided copy. When the-program is graded by the grader, into grades for program correctness, program the second form is copied from the first and quality,_program readability, internal retained by the grader before returning the documentation, output-format,-and external listing to the student. The grades on the form documentation. Program correctness includes the which has been kept are then entered into the compilation and running of the program; full marks database and correlated with the copied source are awarded only, if the program produces correct programs. To ensure consistency in the database, answers. Program quality includes those aspects i.e. programs copied match programs graded, any previously discussed, i.e. choice and use of data programs and grades which do not have matching IDs structures and control structures. Readability are rejected. includes the formatting of the program. Internal documentation refers to the commenting within the This implementation also filters out programs program text. We have also developed a method for which are syntactically incorrect (don't compile). evaluating automatically program format and And it also discriminates between programs which internal documentation at the textual level. This terminate normally and those which halt on some method assumes that students have a set of well-

211 error condition. Copies are kept of programs considered "toy" programs by virtue of their size which halt prematurely; however, part of the (100 - 500 lines, generally speaking). Nevertheless encoded ID contains a record of the premature research must begin somewhere, preferably with halt. large sample sets of programs. Most "real" world samples are available on a one-time only basis as The system does not have the capability to implementations of very large systems. check program correctness automatically since the primary goal of the research has been to automate Applications style grading. Correctness at this point is a relatively easy task for the graders since student Applications for grading systems are not programs are run on standard data sets. simply limited to the evaluation of the end- product of student efforts. However, until Database Description absolute standards of quality are developed, we must be satisfied with this result. Once absolute Programs, program results and grades divided standards are formulated, the evaluation process into categories are entered into the database. can be performed on individual programs, allowing Also included in the database are specifications students to submit programs directly to the for the programming tasks and data sets. Programs automatic grader and obtaining immediate grades. can be accessed by task, section, student, A further refinement would incorporate the semester. Thus far (Fall semester 1982) two assessment into program development systems as projects of approximately 200 students each have part of an interactive set of tor" ;used to write been collected for entry into the database. Future programs. In that setting the measurements could entries include results from automated measures be used to direct students, or even professional" including those for measuring program quality. programmers, to correct poor code, thus enforcing See Figure 2. good programming habits at the outset. A not entirely unexpected side-effect of the automatic evaluation of program quality has been to produce an effective cheating-checker - it is not difficult to see that programs with identical complexity measurements are in fact likely to be the same program.

Conclusions

Summaries In this paper we have described an automated system for grading style in student programs. The system is based on a mathematical model of programs -from- which- measures have been defined to measure the quality of a,program. These measures of qualitylinclude the use and choice of data Programs Grades structures and control structures in a program; they permit the comparison of similar programs to Experimental Output determine which programs are better from a quality Results or style point of view, Preliminary testing of the system indicates that it performs in a manner comparable to human graders. It is expected that, Figure 2. Database Model after further extensive testing, the results will demonstrate that the mode'_ derived from the program complexity model is acceptable for modelling quality or style in a program and is a suitable basis for a system for theautomatic grading'of programming style. It is anticipated-that copies of the database will be made available to outiide experimenters. References We envision that this database can be used-for further studies of the program complexity model, 1. Forsythe, G. and Wirth, N., "Automatic Grading testing and comparison of other. complexity models, Programs", CACM; May, 1965, pp. 275-278. and the establishment of a common set'of programs to allow the reproduction of previous experiments. 2. Harrison, W. et al., "Applying Software Having many copies of prOgrams implementing the Complexity Metrics to Program Maintanance", same task should provide the quantities.of Computer, Sept. 1982, pp. 65-79. ,programs needed for experimentation. It seems obvious that an academic environment is the 3, Hext, J. and Winings,.J., "An Automatic Grading optimal source of these programs since industry Scheme for,Simple Programming Exercises", CACM, could not tolerate the expense or justify the May, 1969, pp. 272-275. existence of so many copies of similar programs. Academic settings also permit samples to be 4. Hollingsworth, J. "Automatic Graders for c, _lected from a variety of applications and Programming Classes", CACM, October, 1960, pri.fgramoning languages. Many of these programs are pp. 528-529.

212 22.1 Oviedo,--t., "Control Flow, Data Flow and gram Complexity", COMPSAC, December, 1980, 146-152.

Oviedo, E., "Control Flow, Data Flow and PrOgram Complexity", Dissertation, SUNY/Buffalo (to appear 1983).

7; Oviedo, E. and Ralston, A., "An Environment to Develop and Validate Trogram COmplexity Measures", NECC, June, 1983.

- 8. Rees, M., "Automatic Assessment Aida for Pascal Programs", SIGPLAN,-October, 1982, pp. 33-42.

9. Robinson, S. and Torsun, I., "The Automatic Measurement sf the Relative Merits of Student Programs", SIGPLAN, April, 1977, pp. 80-93. TEACH TOP-DOWN PROMANiam WHILE YOU TEACH BASIC

by Michael J. Streibel, Ph.D.

Instructional Systems Program The Pennsylvania State University University Park, Pennsylvania

Abstract is a premature commitinent to specific solutions and a program of Basic code that eventually This paper describes a way to teach top-diown becomes unmanageable, unmodifiable, and un- programming principles while teaching the Basic understandable. computer language. Such an approach is especially important since microcomputeraare appearing in The approach described above can at best be large numbers in public schools and in homes and called the bottom-up method of designing programs since Basic is. the first computer language which (i.e., coding the individual statements of Basic mostpeople encounter. Programs, in this approach, before planning the overall structure of the are designed by first spelling out the program program). It is only defensible' when tine programs goal as remarks in the header part of the program are short (i.e., one CRT screen long). In and then developing a complete set of major sUb- response to the bottom-up approach, several rdutine-calls for the main part of the program. authors over the years have proposed a top- The subroutine calls in the main part of the approach to designing and writing programs. program also include remark statements which spell Theirargument goes something like this: the. - out the goal of the subroutine as well as the bottom-up approach does not teach effective parameters (if any) which are "passed to" and problem-solving skills and usually results in "returned hoe the subroutine. Once this phase poorly-written programs. .Some people findthe of the program-design process is finished, the latter an acceptable state of affairs. Why, you actual subroutines are coded and tested to therefore ask, should one learn effective problem- completion. solving skills and why isn't a program that executes correctly enough? The answers to these questions are simple: programming is a subset of Introduction problem-solving (i.e., problem-definition and solution-generation techniques) and programs Many compUter literacy courses have been communicate with humans as well as.with machines. proposed in educational computing magazines which More of this later. include a component on Basic programming. The approach taken in these courses invariable deals The top-down approach to programming attempts ' with teaching the individual statements of Basic to define the problem clearly (i.e., the "top" if before going on to higher-level programming one views the process as a pyramid of levels) concepts. At the same time, other articles in before breaking it into smaller and logically- these magazines encourage their readers to use a distinct components. Giving a program a meaningful top -down approach to designing programs. Why not name, for example, is an excellent first (or combine the two approaches into one method and "top") step. Issues on the most general level help your students learn the basic statements of can then be articulated so that the eventual Basic in the context of top-down-structured program has coherence and conceptual integrity projects? What follows is a-rationale and an (i.e., all the parts work in unison to achieve exam& of this approach. the program goal) and robustness (i.e., ;he program operates_under many conditions).° Top-down vs. Bottom -up Program Design The "top", or first step of the design _Anyone who has ever written a program in Basic process, therefore involves giving a program a has been tempted to turn to the keyboard as soon. name and a description of what it will do. We as possible and start coding the final program. usually do not think of this as programming but The interactive nature of most interpretive Basics this step is as much apart of programming as and the. "friendliness" of microcomputers just writing specific Basic statements. We are, in makes this-temptation irresistable. These same effect, communicating with ourselves and others people have probably also learned the hard way at this point (if we are part of a collaborative that coding first and thinking later has led to team) and thereby programming our minds to innumerable difficulties - not the least of which proceed with the problemrdefinition and problem-

214 23 solution in certain ways. The computing power of A Specific Example microcomputers is an extension of the human mind even though it resides in a separate box and has What happens when you are teaching a computer its own arcane rules of grammar and syntax. By literacy class and have to teach the individual describing what we are going to do in the program statements of Basic? Do you throw out the top - before work gout haw we are going to do it, wl down approach until after your students have avoid locking ourseNES into premature solutions. learned Basic? Definitely notl Programming; habits We will use REMark statements such as those in are developed and reinforced during every the program-example below to carry out the "top" encounter with a computer. You can therefore of the design process. These will be described in structure a project in such a way that it is greater detail later. formulated and partially solved in a top-down manner. The example below will indicate how this Gje proceed in the top-down design process is done (see Listing 1): by asking: what major steps will we have to take in order to accomplish the goal of the problem. Tha objectof the DATA.BASE program is to Notice, we are leaving the how (or the specific teach students about double- subscripted variables

code, whether it be Basic or Fortran or Pascal) . and data-structures (i.e., DB$(I,J) in line 3080 until later. We are also attempting to make the is a mailing list data structure with two major steps represent intelligent sub-processes subscripts I and J): Students are given the of the larger goal rather than arbitrary divisions listing shown (in printed form and anto,their dictated by the Basic code. Bence, the major disks) and then asked to complete the program so steps at this point are usually described as that it will carry out the goal described in the some sub- process like "define the values," header. The main program is a series of GOSUB and "calculate other values," and "printout those REMark statements that has broken the overall goal values." The goal is not to get bogged down in into smaller sub-problems. Each subroutine is made the details of coding Basic statements when we up of a physically-distinct set of REMark state7 are still brainstorming various ways to formulate meets and a RETURN statement. possible. solutions. Otherwise, we would be assembling the "bricks" of the program before The "top" of the design is the program name designing the plan of the entire program. In the because it identifies the intent of the eventual example that follows, a series of GOSUB and program (see line 1020). Use meaningful words here REMark statements will be used to carry out the so that you can refer to the entire program as a second-level of the top-down design process.. meaningful entity. An informative name is Each subroutine in the eventual program is important because it communicates the intent of treated at this stage as a black box that will the program to other human beings. In many cases, dasomething for us. We can satisfy the urge to that human being may be yourself if you develop

type something on the keyboard here by typing the program over an extended period of time. . the GOSUB and REMark statements. If we want to change out minds about a particular solution, The program description elaborates the intent then it will be easier to delete a few lines of of the program (see lines 1060-1110). A clear REMark statements in the Basic program than to description of the program in the early stages of delete a series of Basic'code that we have worked the design process usually helps focus your long and hard to create. attention on what you want to do without getting you bogged down in details. Many people object to The last (?). stage of the top-down design this stage of the design process because they feel process is to fill in the actual subroutines and it limits their creativity and inventiveness later make them do what we want them to do. The large on. They have a point if the description is too problem has been broken down into s.ialler and detailed and prescriptive. You can avoid this more manageable sub-problems that can be solved, pitfall be describing what will be done rather coded, and tested by themselves. The trick is to than how it will be done. For example, a bad treat each subroutine as a black box into which progrE7description would include specifications we pass some variables and values (i.e., input) for the exact messages to be typed out to the and out of which we get some action, variables; user. Iwthe DATA.BASE example in Listing 1, on or values (i.e., output). Since we have specified the'other hand, the description in the header exactly what We want the black box to do, we can does not specify hoW the program will carry out test and debug the Basic code in the black box the intent of the author. The description is until it satisfies our desires. The result is created by the teacher to help focus the student's that, once we are finished with each subroutine, attention on the goal of the program. The we can file it and forget it because we have description is also written to serve as a model solved a small part of our problem. In a larger of haw to formulate a problem. sense, we have formulated the original problem in such a way that the power of computing can A final component of the "top" of the program help us solve the problem. Notice that the last deals with a description of the variables and step in the design process involved coding the files that are used in the program and a listing problem into a particular, dialect of Basic. We of any formulae used in the program (there are could just as easily have coded it into another no formulae in the DATA. BASE program). This language because the design process up to this constitutes a kind of glossary of terms so that point would have been the same. teachers and students have a common "vocabulary"

215 232 1000 REM 1010 REM 1020 REM PROGRAM:DATA.BASE 1030 REM 1040 REM AUTHOR: MICHAEL J. STREIBEE, PH.D. 1050 REM 1060 REM DESCRIPTION: THIS PROGRAM CREATES A DATA BASE 1070 REM IN MEMORY CALLED DB$ AND THEN ASKS 1080 REM THE USER FOR A LAST NAME TO SEARCH DB$. 1090 REM IF FOUND, THE CONTENTS OF THAT PERSON'S -- 1100 REM DATA IS TYPED OUT, ELSE AN ERROR MESSAGE 1110 REM IS PRINTED. 1120 REM 1130 REM VARIABLES: 1140 REM NE = # 07,ENTRIES IN DATA BASE 1150 REM NI = # OF ITEMS/ENTRY 1160 REM MO$ = 011 "Y" FOR MORE SEARCHES? 1170 REM 1180 REM FOR I = 1TO NE 1190 REM DBW,1) ='LAST NAME 1200 REM DB$(,I,2) = FIRST NAME 1210 REM DBW,3) = STREET 1220 REM DB$(,I,4) = CITY 1230 REM DB$(./,5) = STATE 1240,REM DB$(,I,6) = ZIP 1250 REM 2000 REM 2010 REM 2020 REM MAIN PROGRAM 2030 REM. 2040 -GOSUB 3040: REMCREATE DATA BASE DB$(1-NE,1-NI) 2050 GOSUB 4030: REMASK FOR LAST NAME; RETURN NA$ 2060 GOSUB, 5030: REM SEARCH DB$(1-11E,1) FOR NA$; RETURN POINTER PT 2070 GOSUB 6030: REM PRETTY-PRINT TITLES & DB$(PT,1-N1) 20.80 GOSUB 7030: REM ASK IF MORE; RETURN MO$="YES" OR "Y" 2090 REM 2100 IF MO$ = "YES" OR MO$ = "Y" THEN 2050 2110 END 2120 REM 3000 REM 3010 REM CREATE DATA BASE DB$(1-NE,1-NI) 3020 REM NOTE: TYPE IN YOUR OWN DATA STATEMENTS 3030 REM 3040 READ NE,NI: REM NE.4 ENTRIES, NI=# ITEMS /ENTRY 3050 REM 3060 FOR I =1 TO NE 3070 FOR J =1 TO NI 3080 READ DB$(,I,J) 3090 NEXT J 3100 NEXT 1 3110 RETURN 3120 REM 3130 DATA 9,6: REM NE,NI 3140 REM 3150 DATA *%LAST","FIRST","STREET","C/TY","STATE","ZIP" 3160 REM 3170 DATA "LAST","FIRST","STREET","C/ Y","STATE","ZIP" 3180 REM 3190 DATA "LAST","FIRST","STREET","C/TY","STATE","ZIP" 3200 REM 3210 DATA "LAST","FIRST","STREET","C/TY","STATE","ZIP" 3220 REM 3230 DATA "LAST","FIRST","STREET","C/TY","STATE","ZIP" 3240 REM 4000 REM 4010 REM ASK FOR'LAST NAME; RETURN NA$ TO BE SEARCHED INDB$ 4020 REM 4030 RETURN 5000 REM 5010 REM SERACH FOR NA$ IN DB$; RETURN PT=SUBSCIPT INDB$OR 5020 REM 5030 RETURN 6000 REM 6010 REM PRINTOUT DB$(PT,1-NI) OR ERROR MESG IF PT=0 6020 REM 6030 RETURN 7000 REM 7010 REM ASK IF MORE; RETURN MO$="YES" OR "Y" OR "NO"OR"N" 7020 REM 7030 RETURN

LISTING 1.

217 234 4030) for the particular project (see lines1140-1240). box" of code in our'program (i.e., line This mechanism serves a communicationfunction. which will ask the user of the program for a last It also serves as an aid in directing thestudent's name and which will then returnthis name ir a attention to the major data elements in the variable called NA$. The, variable NA$ is not particular problem. Being able to construct a "really" returned in the program because it visual/spatial representation of data-structures constitutes a global variable in Basic (i.e., program); but we helps one to think about data-structures. can be used anywhere in the are treating it as if it werereturned to us by subroutines 'We have now finished the "top" of the design the subroutine. Tfils way of looking at process. We are fairly clear About what wewant fosters a computer-as-tool (i.e., subroutine -as- our Program to accomplish, about 'the typesof tool) attitude which is very healthy. Subroutines data objects (a mailing list in our example), and are treated as autonomous units ofcode that make the specific variables to be used in:our program. no assumptions about the rest of the program We are therefore ready to proceed. to the next (such assumptions are always a source of "bugs") it. The level of the design process. except for what argments are passed to subroutine in the example initially contains only The second level ,of the top -dawn design a RETURN statement (see line4030) which students process is represented in the DATA. BASE program will then expand into a working subroutine. by the 'MAIN PROGRAM"' set of GOSUB and-REMark statements (see lines 2040-2110). In this The "last" levelOf the DATA.BASE program is example, the teacher has already partitioned made up of the individual subroutines (see lines the problem in A. certain way so that students 4030, 5030, and 7030). Since the purpose of our are primarily concerned with theobject of the progrm is to learn about double- subscripted lesson (i.e., double-subscripted variablesand variables, the data-base variable DB$ is pre- data structures). Let's take a look atthe two, defined by the teacher (see lines 3040-3230). main elements of this level of thinking. This provides an example of how a data-structure is created in Basic and yet leaves open several The GOSUB statement is the only mechanism activities for manipulating the data base. In the in most Basics which allows for multiple lines DATA.BASE example, students are asked to make up and then type than to be defined as a single unit. Why isthis a list of names and addresses important? -Well, we have already discussed the into the appropriateDATAstatements. need to break a large problem, into moremanage- able sub-problems and the'value of leaving details The individual subroutines are made up of until later. This helps us avoid premature Mark statements which visually separate the foreclosure (i,e.; premature commitment to code from the rest of the program and which specific solutions). Being able to cluster describe what the subroutine is todo.8 This earlier efforts several lines of code together is also important may seem like a duplication of because it allows us to have a physical counter- but ithelps students focus their attention on part to the conceptual units of the problem- each sub -prellem of the program. The subroutine itself since its solution. Humans can only deal with complexity can then be coded and tested by at a certain level before having to"chunk" input, output, and intent are clearly spelled out. the problem into smaller meaningful units. The subroutine also gives-the student wide Having these chunks In a physical as well as lattitude in working out how the sub - problem conceptual units is a definiteadvantage7 is to be solved.

The REMark statements in the "MAIN PROGFW Stmt' are as important in the top-downdesign process as the GOSUB Statements because theyall us to The top-down project approach to learning articulate our intentions for each Sub-problem Basic which is described above has a number of (see lines 2040-2080). Each subroutine, in effect, distinct advantages over traditional instruction br.somes the "top" of its own pyramid of code.' Time in Basic. First, it gives the teacher wide spent in writing these Mark statementsis not lattitude in formulating both the content and wasted if one is trying to brainstorm solutions the structure of student projects. Bence, the to a complex, problem because the REMark stataments teacher could fill some subroutines completely let us publically inspect our.intentions and for beginning students and yet add extra options meanings. This allows us to "debug" our solutions for advanced students. on the semantic level before debuggingthe actual code. For example, we can ask ourselves Second, the approach described above lets learning whether we should ask for the users last name the teacher address several levels of and immediately search the data base or whether at the same time. For example,the project we should do these things separately(see lines described' above teacher about specific Basic variables), 2050-2060). statements (i.e., double-subscripted computer concepts (i.e., simpledata-structures), The Mark-statements associated with each- systematic program design, and how to think about GOSUB statement also include a brief reference to problems. The assumption throughout the example what variables are passed to the subroutine and is that computer literacy is ultimately a matter what variables are returned. For example,line of the user's ability to think and communicate 2050 indicates that we should jump to a 'black clearly (withhumans and computers) about a

218 235 problem. Computing, in effect, becomes a. welldefined extension of the humen mind. The power of comuting, however, is only engaged after the human being has clearly visualized and then articulated the'purpose and parameters of a-- program.

Finally, the approach described above permits a problem to be solved with style and logic from the-"inside-age (i.e., from an understanding of the problem) .7Programs no longer look like "spaghetticode":but reflect clear, efficient, and effective thinkingAs a teachergradually removes his or her own way of structuring problem-solutions, students learn to develop their Own style of formulating and solving problems with the computer.

References

/Finkel,L. and Brown, J.R. APPLE BASIC: Data File Programming. New:York: John Wiley g Song:M. 2Dijkstra,E.W. A pii.14041.g: of Programming. Englewood Cliff i LUTT-Prantice-Pall, 1976. 3Wirth,N. Systematic Programming.: An Introduction. Englewood Cliffs, LT., Prentici=Hall, 1976. Siagin, P. and Ledgard, H.F. BASIC With Style: ey Puut: Proverbs. Rochelle raiE, 1. J.: :.. Co., 1978. 5Lewis,W.E. Problem-Solving Princ leS for BASIC Programmers. 10Ehelle Park, NJyden E63E CO., 1981. 6Naginand Ledgard, op. cit. 7 Miller, G.A."Themagic number seven, plus or minus two: Some limits on our capacity for processing information." Psychological Review 1956, 63, 81-97. 8Nagin and Ledgard, op. cit. 9 Dreyfuss, Hy Designing for People. New York: Simon & Minster, 19557

219 236 USING' COMPUTER SIMULATED MODELS TO TEACH PROGRAMMING LANGUAGES

BOGDAN CZEJDO

UNIVERSITY.OF HOUSTON WARSAW TECHNICAL UNIVERSITY

ABSTRACT tion with the corresponding program specification._ In this paper computer generated models These models were chosen after a wide variety, to teach programming. languages are defined and of structures had been examined and evaluated for examined. The memory modeling, the program potential use in the introductory courses in pro- generation and the flowchart creation are pre- gramming languages. For each model we investigated sented. Generation and transformation of pic- the problem of how to generate the solution (for torial structures for each model is analyzed. the given task) and evaluate the correctness of The application of these models to computer- .the student's answer. assisted instruction is shown. Finally further 2. THE MEMORY MODEL research in the area is suggested. A good simulation model to start with in the introductory course is a structure which consists I. INTRODUCTION of an input device, memory, and an output device The concept of model-based instruction was (external storage can be added later). This corre- introduced in the paper "Using Model Based In- sponds to the approach chosen in various textbooks struction to Teach Pascal". On the basis of the for programming languages 4. This model enables principles of model-based instruction', CAI lessons the student to visualize the execution of a single have been designed for teaching elements of the instruction as well as the whole program. Two programming languages. The early version of our phases of the operation of the model can be identi- systemi displayed prestored course material, fied: accepted only a.restricted format of the answers 1. Creation of the structure and used built-in comments. This paper describes 2. Transformation of the model the research leading to the new version of the For the sake of generality, standard declara- system. We took advantage of studies in artifi- tions in PASCAL and FORTRANwere chosen to build cial intelligence in the area of knowledge repre- the structure. For example the following declara- sentation, program transformation, algebraic mani- tions in PASCAL: pulation and system simulation5-7. Several simu- I: integer; lation models were defined and implemented in Pas- Z: real; cal on the VAX computer. They constitute the sim- Ch: char; ulation system shown in Figure 1. A: array [1..3] of integer would result in the picture shown in Figure 2. PROBLEM DESCRIPTION

Input I (integer)

Z (real)

Ch (char)

A[1] Figure 1. The Simulation System. A A[2] The program specification is the basis to (in- generate the program or the flowchart. The flow- teger) chart may also be obtained directly from the pro- A[3] Memory model is generated from the program. gram. Output In Figure 1there is no link between problem description and program specification because the Figure 2. Generated Memory Model. current system cannot perform this transformation. The generated picture consists of: We assume that, the instructor will provide that a) An Input Box link by means of associating each problem descrip- b) An Output Box

220

237 c) Memory divided into several boxes. cannot directly use the previous two step method. To make the simulation feasible., we assumed some Several data items can be put into the input or restrictions for the pointers. The following de- output box. Mixing of types is allowed. The mem- claration is assumed: ory consists of several boxes in accordance with the declarations. Every box corresponds to a var- Pointer = " Item; iable. In this way, a simple "v.;sualization" of Item = record variables or constants in programming languages Key: integer; can be introduced. Big arrows indicate the al- Next: pointer lowed transfers. end Program instructions and input values are ne- For example, assuming the declarations 0,Z: Pointer cessary to perform transformations of the model. the following sequence of instructions would re-' The four permitted program instructions are: sult in the picture shown in Figure 4. a) read b) write new (P) ; c) assignment P ".Key: =7; d) new new (Z) ; For example, consider the Pascal instruction read- Z ". Key: = 9; P ". Next: =Z; ln(I) and assume 25,75 is given as input. The result of the transformation is shown in Figure 3.

25 I (integer)

Z (real) Output Figure 4. A Memory Model for Pointers. CH (char) We are convinced that-the above described memory A [1] model enables a down-to-earth interpretation of single instructions (as well as their sequence) as A [2] A a "visible" transfer of some values from some box- (integer) es to other boxes and the creation of linked lists. Output A [3] In the model, the number of variables and the size of the arrays and the number of elements in linked Figure3. Transformed Memory Model. lists are restricted because of the size of the This and any other read statement is executed by screen. To make our model more usable we can re- displaying the transfer of an item from the input lax these restrictions assuming that the screen box to the proper box of the memory. Note that the is the window to display the first elements only. item in the input box is not erased, only its copy 3. THE PROGRAM GENERATION goes tb the memory. At the same time, a small arrow which' indicates the beginning of the input Research in automatic programming has achieved queue is moved to the next element. It helps the some success with experimental systems which pro- student to determine what element is to be read duce programs from some specifications. Various next. Transfer of the item to the memory causes specification methods have been useds: the erasing of the previous value. If the type of 1. formal the item in the input box does not match the type 2. by examples of the memory box, the transformation of the model 3. natural language /- is aborted and the model returns to its previous In this paper we describe a simulation model state. In sucha-case, an error message is dis- to generate programs found in introductory courses played. The latter rules apply also for the of programming languages e:g.4-6. A variety of assignment statement which causes some computations structures of the programs'were analyzed and evalu- and transfers the result to .the specified box in ated for potential generating. To simplify the the memory. The write statement transfers the pro- generating process, we restrict the area to pro - per values from the memory to the output box. Out- grams with one loop or without any repetition. Ex- put is treated like a mini CRT screen using a ceptions to this rule were programs with additional scrolling mode. Every transfer "from" is executed initialization laop for the array. This restric- by making a copy and transferring it to the proper tion is not really very strong because we can al- box. Various ways of animation of the transfer most always break the problem into subtasks in a were considered. Blinking and/or showing charac- reasonable way,-und use the system to generate ev- ters moving on the screen were chosen as the most ery subroutine separately. explicit approach. Our method of specification is based on choos- One of the areas where students have many prob- ing options of the given menu. Some restricted lems and need more assistance and opportunity to formulas are also allowed. External specification practice is in,the concept of pointers. Unfortun- corresponds to the internal representation in the ately in this case, the model of the computer chan- form of a network, which is the basis for the pro- ges'during the execution of the program, so we

221

2:3q gram generation. The method of generation is Program Count (input, output); based on the'observation that most programs can be divided into areas which are independent of each var other. In the first step we need to choose a proper I,X,Count:integer; structure for these areas. The following structures are available: begin 1. No-Loop Initial --2-;--For-Loop Count := 0; Processing 3. Sentinel Loop 4. Eof/FolnLoop 5. Mixed CoiTd_Loop for I.= 1 to 100 do 6. Nested Loop For example, the For-Loop structure is shown in begin Figure 5: read (TRANS) Read Area Program if TRANS >50 then Processing Count:=Count + 1; Area var end; I:integer;

' , Count); Final begin write ('Number of items = Processing Initial end. Processing Figure 6. Generated For_Loop Structure for I:=1 to 100 do 4. THE FLOWCHART GENERATION begin Read The evolution towards-the use of higher level structures has reduced the need for detailed flow- Area charts. Flowcharting might still be useful, however, for the beginners to visualize control flow. Graphics primitives corresponding to three basic Processing control structures, (sequence, condition, loop) Area were designed to implement this model. On this basis, a flowchart can be created for the given: end; 1. program specification 2. program Final Process of flowchart creation' from the program Processing specification is analogous to the three step program generation. The flowchart generated for the example from the section 3 is shown in figure 7. end. Figure 5. Initial For_Loop Structure ( Start ) Once the structure is generated (e.g. for.Loop in Fig.5) we decide whether reading inside the loop Count:=0 is necessary or not. If'our answer As "yes", we have to list all items to be read together with 50 then final program.would look like Figure 6. Count:=Count + 1

write 'Number of items = ,Count) ( Stop ) Figure 7. The flowchart generated from the program specification

222 23 In the current system any decision structure is One can also experiment with the flowchart (the included in the processing box to stress the con- flowchart generator). The testing mode is to trol flow for the loop. The other reasonfor mak- check whether students can perform actions indica- ingthis decision is the fact that the size of the ting their ,ducational achievements. They can be s:reen would impose some restrictions for the num- asked to predict the changes of the content of the ber of explicit decision boxes. memory for a given program and some data (prestored The second option is to generate the flowchart by the instructor) using the memory model. They directly'from the program. In this case, the sys- can attempt to complete the program or the flow- tem displays the flowchart in a more "traditional chart for the given_problem (provided by the in- way" as it is shown in Figure 8. This flowchart .structor). The system should compare the answers is generated directly from the program in Figure 6. with those generated. We plan further investigation of the following extensions to the models and their use in CAI: (Start 1. Implement the testing mode 2. Integrate the simulation system so that I Count= 0 a sequence of simulation models can be ( loop fOr> invoked according to figure 1. 3. Give the instructor or the student the I=1,100 possibility to'define their own language for a memory model. 4. Build a memory model for the lower level t languages. 5. Design educational games (eg adventure TRANS games) connected with the programming process. 6. Use or design the interface to extract program specifications from the problem description given in a natural language. 7. Extend the area covered bythe program generator. COunt = Count +1 SUMMARY

In this paper we described the use of simula "Number of items"Count' tion models in teaching programming languages. We identified and defined the memory modeling the program generation and th: flowchart creation. We ( STOP ) showed how to generate and transform pictorial Figure 8. The flowchart generated from the program structures using these simulation tools. The ar_ plication of these models to computer-assisted In- 5. THE SIMULATION MODELS IN A CAI SYSTEM struction was shown. Finally, further research in the area was suggested. The primary purpose of developing the models described above was to use them in a CAI system to REFERENCES teach programming languages. Three modes of inter- 1. Czejdo, B.,"Using Model-Based Instruction to action with the student for the models were iden- Teach Pascal" Proceedings of NECC/2, Nor- tified: folk, Virginia, 1980. 1. Information 2. Simulation 2. Kolkowski, L.,Nauczanie Problemowe w Szkole 3. Testing Zawodowej, Warsaw: WSIP, 1974. The Information mode is simply presenting the 3. Bork, A., Learning with Computers, Bedford, MA: work of the simulation model. The necessary input Digital Equipment Corporation, 1981. is prestored by the instructor. In the case of 4. Richards, J., Pascal. New York: Academic Press, the memory model, the instructor needs to store .1982. the program and the data. For the program generator, the problem description and the program spe- 5. The Handbook of Artificial Intelligence, Vol.2, cification should be prestored. For flowcharting, Stanford, William Kaufmann, Inc:, 1982. the program should be provided or the problem de- 6. Friedthan, F.; Koffman E., Problem Solving and scription and the program specification should be Structured Programming in FORTRAN. Read- given. The second mode is simple simulation and is ing, Massachusetts: Addison-Wesley, 1981. similar to the "Ten Finger Ekercise". It allows the students to experiment with the language. One 7. Czejdo, B., "Transformation of Universal Alge- can provide programs or data and watch the changes braic Expressions in PASCAL" ACM Computer in the memory and output (the memory model). One Science Conference, Kansas City, Missouri, can supply specifications of the program mainly by 1980. answering the questions and observing the process of generating the program (the program generator).

223 COMPUTERS IN SCIENCE EDUCATION

Raymond E. Bigliani Lewis Dove Stephen Bryant H. Herbert Edwards Kenneth Keudell P. James Nielsen Gerald White 'Robert W. Henkens David Alexander Richard Cornelius correct answer was internally computed. Students ABSTRACT; The Use of an Apple/Corvus Networking were allowed up to three attempts to correctly System in an Elementary Physics Course answer the question before the correct answer was displayed and the student was then directed to the Raymond E: Bigliani, Associate Professor of next problem. If the student answered the problem Physice;. State University Agricultural and correctly in three or less tries, he then was Technical College, Farmingdale, NY 11735 congratulated and automatically cycled to the next problem. All questions in the activity had to be Course Description. A network of eleven. Apple answered before a grade could be assigned. In microcomputers was used in 2 one semester addition, in order to raise their grade, a student elementary physics course (PH095) offered at SUNY could repeat an activity as many times as desired. Farmingdale in the Spring 1981 semester. This However, only the last grade on an.activity was course is taken by students in the Pre-Engineering retained. Technology Curriculum which is designed to help Course Evaluation. The course was evaluated high school graduates satisfy requirements for ,using the results of a student qeustionnaireand admission into a two year Engineering Technology standardized post-tests administered to all program at SUNY Farmingdale.The elementary students enrolled in the. course. Student response physics course described below is designed to to the microcomputer as a testing tool was prepare "pre -tech" students for a one yearphysics overwhelmingly positive, mainly duetotheability-- ourse taken in an Engineering Technology to answer an individual question up to threetimes number curriculum. and the ability to repeat any activity any I taught two of the four course sections using of times. the microcomputer system described below; the At the end of the semester, all studentstook remaining-two-sections were taught in a College Board Entrance Examination testsin. traditional, non-computer mode. Computation and Elementary Algebra. The two Course Objective and Structure. The major ^computer" classes averaged slightly higher onboth objective of the course was to develop an tests than did the "non-computer" classes. A more understanding of the basic concepts and to provide meaningful measure of the efficacy of thetechnique drill in the following areas: 1) mathematical described above will be determined from afuture skills, 2) graphical data analysis, 3) problem correlation study between the ^computer^and solving technique, and 4) physics theory. ^non-computer" groups in PH095 and theirgrades in The course was structured as follows: classes their subsequent physics courses. met for three one-hour classes per week. Two of tka three classes met in the Computer Aided Learning Laboratory (CALL) which consists of eleven ABSTRACT: Program Development by a Biology User's.. Apple microcomputers networked together using a Group for Microcomputer Assisted InstructAon Corvus networking system with a 10 megabyte hard disk drive.The third class met in a regular 1.ewis Dove, Stephen Bryant, H. HerbertEdwards, classroom in a traditional lecture class format. Kenneth Keudell, P. James Nielsen, GeraldWhite, All homework assignments, called activities, Western Illinois UniVersity, Macomb, IL61455 were first disussed in the lecture class; the actual performance of the activity and/or , A biology user's group has beenoperating at presentation of data occurred in CALL. Some of the Western Illinois University for two years. It is activities consisted of drills in the following composed of six faculty, five from the Department areas: calculator operation, algebra, graphical of Biological Sciences and one from theMathematics analysis, and trigonometry. Seven additional Department. The Project Director, a cell physics content activities were used; these biologist, spent two years promoting CA1 and consisted of four to six homework problems taken computer literacy as an associate of thefaculty directly form the text used in the course. development office. Other members of the group Although the wording for each problem was the same include a plant pathologist, a microbiologist, a for every student, any required numerical values human physiologist, a population biologist,and a were randomly generated and the corresponding

224 "241 mathematician who teaches computer programming. the Apple II computer. The Traiher is meant for grant from the Apple Foundation provided two individual use in a course or training program. A microcomputer systems, a faculty development grant unique aspect of the product is the provision for provided funds for the development of an immunology data communications lith the instrument :self, program, and the remaining rpograms were developed allowing the instructor to download a number of under a grant from the National Science Foundation , real data sets of 4-8k each. The concept should be Local Course Improvement (LOCI) program. The group adaptable to a wide range of additional scientific promotes computer literacy'and microcomputer usage instruments. among biology students and, faculty by reviewing The input and output characteristics of microcomputer software in biology, by presenting a_ Scientigio, Instrument Trainer will be like the course on microcomputers in biology, and by writing instrument, using keyboard, display screen, software. Five of these programs are being printer, and plotter. The students can set evaluated by students and others are being written. parameters, such as the accumulation time, pulse These and other biology programs acquired from angle, and decoupler power and frequency, and give external sources are being used'by students to commands through the keyboard, obtain parameter supplement their laboratory work in first-year listings, and observe instrument graphic displays. biology courses, general microbiology, ecology, At the end of the training session, the student can genetics, and cell biology. These include a go over printed parameters and plotted data with program which assists sewage treatment personnel in his instructor and discuss possible ways for evaluating sewage effluent by providing a key to improvement and_return to the trainer to gain indioa'..cr protozoa and a description of their experience and confidence to handle a variety of taxonomy. Another program dramatizes electron flow experimental problems. in light reactions of photosysnthesis and also ABSTRACT: Concentrated Physics Concepts: A includes the dark reactions. The immunology Comprehensive Package of Tutorial Problem Solving tutorial program describes and illustrates the David Alexander, Physics Department, Richard immunoglobulins. A tutorial microorganism Cornelius, Chemistry Department, Wichita State identification program characterizes the bacteria, University, Wichita, KS 67208 fungi, algae, and protozoa.. A,computer-managed Undergraduate students enrolled in,an program written in Pascal incorporates three introductory physics course often experience lessons in cen biology and a student file system. difficulty relating abstract physical principles in Projects under developmnent include energy the concrete problem-solving arena. In many cases, metabolism of mitochondria, adaptations of existing this difficulty arises from a deficiency in general software in chemistry to courses in biology, and problem-solving skills. Unfortun.telY, tutorial/simulations of natural selection and undergraduate physics classes are often too large migration. for the instructor to give students the individual assistance needed to develop the logical thought processes required for successful problem solving. ABSTRACT: Scientific Instrument Trainer Although laboratory experiments are an important element in learning physics, they address a Robert W. Henkens, P. M. Gross Laboratory, Duke different set of goals. University, Durham, NC 27706 Concet,rated Physics Concepts is a package of Apple II+ computer programs designed to foster the Rationale. Access to necessary instruments for development of problem-solving skills for work at the frontiers of knowledge is obviously introductory physics students. The programs are important to the nation's future research contained on about eight diskettes and cover the productivity and for the training of future material normally contained in a two semester generations of scientists and engineers. introductoery physics course. Two types of Most science educators agree that laboratory programs are included: Conceptual programs review training is necessary for sciencestudents. the concepts and terminology of a unit to assure Unfortunately, high costs severly limit this kind that the student understands the material before of training in many areas of instrumental analysis, attempting the problems associated with the unit. including modern spectroscopic analysis, This is in Each problem-solving program presents a graded strong contrast to a decade ago when most sequence of problems to illustrate.a.single universities provided their students with hands-on physical concept. In many cases, the solution of experience with state-of-the-art instrumentation. the problem is approached through several The Scientific Instrument Trainer System is preliminary questions. When requested, the designed to help with this problem by providing programs provide assistance in working out an interactive training for chemistry graduate appropriate strategy for solving the problem. students in modern FT-nuclear magnetic resonance These programs incorporate both user (NMR) instrumental analysis. The system could also friendliness (no prior computer experience is be used in undergraduate upper division chemistry required to operate any part of the package) and laboratories. The basic system architecture is user power (the student controls the pace' and general and might be,adapted to other scientific subject matter at all times). Whenever possible, instrumentation. student interaction with the screen (e.g., properly Expected Product.The main components are an placing vectors on a diagram) is incorpbrated into instrument simulator for the student to use, and an the problem solution. Problem solving techniques instrument communications package for the are emphasized by asking appropriate preliminary instructor. The system will provide an interactive questions and by pr iding assistance in response FT-NMR simulation with extensive CRT graphics for to incorrect answers.

225

C, 4:a

JRSEWARE DEVELOPMENT FROM THE PUBLISHER'S PERSPECTIVE A PANEL DISCUSSION

M. D. Roblyer, Moderator ICON Enterprises

Jack Chapel SRA, Inc.

Harvey Guion Random House Inc.

Jane Isay Harper and Row

Dale LaFrenz Scott, Foresman Co.

Christine Johnston Milliken Publishing Co.

Renzi Sugihara Harcourt Brace Jovanovich, Inc.

As publishers of educational materials continue to expand their development of computer-based instructional products, the role of these organizations inshaping the instructional computing field is becoming increasingly apparent. In this panel presentation, high-level representatives from several major publishing houseswill discuss and answer questions on some of the issues surrounding their involvement in educational technology. A specific focus will be on the methods publishersuse to assure that materials they publish are: (1) responsive to the needs of educat6rs, (2) helpful to the continued evolution of the field, and (3) of high instructional quality.

22§24 j Trends in Interactive Data Analysis In the Classroom

Jon A Christopherson, Chair U.S. Coast Guard Academy New London, CT. 06320

SPONSOR: SIGCUE

ABSTRACT

The course Social Science Methodology at analysis system allows the the U.S. Coast Guard Academy tries to instructor to follow the logic of sensitize the cadets to the nature of class discussion rather than having applied statistical analysis. The tedium it dictated by fixed printed of computing one Pearson's r correlation material. coefficient over 40 cases is usually enough to convince the cadets that they should Michael Smith will discuss an learn to use a computer system. While not elective course in the use of computers course materials for classroom use are in social welfare researchwhich has designed for batch processing, batch been developed for students in the processing is pedagogically an undesirable doctoral program at Hunter College teaching tool. The central problem lies in School of Social Work. The purpose of the loss of continuity in the teaching of a the course is to expose doctoral concept. The best approach is to use an students who are not research majors to interactive data analysis system (IDA). basic data analysis procedures in The desirable characteristics of a good IDA social research, to give them an system include the ability to: understanding of statistical and analytic procedures and to help them 1) create data files either by keying learn a set of skills which can be used in the data orby using a prepared directly in their doctoral projects. data file, An experimental approach to data analysis contains learning that can 2) randomly access any of the never be realized in a survey course variables in the.data base in any with a lecture approach. This paper order, describes three basic issues in the course: 1) the attitude of 3) dress up the output with labels for students; 2) the stress given to the variables and values, role of data ,analysis in the total research process; and 3) the choice of 4) edit online any of the variables in a packaged program: the data base after the variable has been created, The tu...rd presentation will discuss micro versus mainframe conversational 5) automatically handle missing value computing. With the advent of codes specified for any variable, microcomputers, students approach the computer with less trepidation. As 6) create data--- (or conditionally micros become more powerful, colleges create) and perform general must make a choice between offering mathematical manipulations of students computing in the friendly variables and numbers, (only one micro environment or encouraging topic in the Methods course) is conversational mainframe computing. taught in an interactive mode. We find that cadets more easily learn The IDA Interactive Data Analysis the basic underlyinglogic of the and Forecasting System, a teaching and method and the practical results of research tool which runs on both the violation of regression's mainframes and micros is used as a case assumptions. An interactive data study.

227 PARTICIPANTS: Joan Fee SPSS, Inc. Chicago, IL 60611 Michael Smith Hunter College New York, NY 10021

Loren Bullock IBM Corporation Bethesda,MD 20817 William Gattis The Tandy Company Fort Worth, TX 76102

228 245 Science Education and the Growth of the \U.S. Computer Industry: Is This a National Policy Concern? Dorothy Derringer, Chair National Science Foundation

ABSTRACT climate which willencourage this industry ---The U.S. comptiter industry shows a to thrive? If so, how should this be done? strong positive balance of payments and is Panelists will discuss answers to these one of the few growth areas in the economy. questions within the context of the Many feel that a scientifically and economy, the role of education in fostering technologically literate 1:orkforce is key the industry, and government state actions. to increasing the:growth of this industry. Differing views will be presented on Should the government act to create a proposed legislation and possible actions.

PANELISTS Vico E. Henriques CBEMA Fred Weingarten Office of Technology Assessment N. John Castellan Jr. Indiana University William Aldrich National Science Teachers Association

SPONSORS

SIGCAS SIGCUE

267 Teaching Structured Programming in the Secondary School Jean B. Rogers Computei and Information Science University of Oregon Eugene, OR

SPONSOR: ICCE

ABSTRACT This session will consist of various Students learn more, quicker. They are approaches to teaching programming in actually able to solve much more complex secondary schools. Samuel F. Tumolo will problems sooner than when they were using discuss teaching structured programming the old methods, if they were able to solve using Pascal. Pascal was designed the problem using the methods at all. primarily to aid in the teaching of good Pascal is one of the above mentioned programming style. By its' design, it languages and is not without fault It makes learning structured programmin easier solves many of the old problems but in the that many other computer languages do. process creates some new ones. There are Procedures and other constructs of some BASIC and FORTRAN based alternatives Pascal make it easier to develop good that facilitate the development of good programming habits and style. The features problem solving habits as well as force the of pascal that aid in the development of student into the discipline of structured structured programming atmosphere will be program design. discussed. Language independent instruction will be There are characteristics of the Pascal discussed by AnthonyJongejan. Pascal is environment that may detract from its not available to students in most high effectiveness. Their effect on structured schools today formany reasons, including programming will be explored. extra cost, hardware and the inadequate Finally, problems that secondary school preparation of teachers. students encounter in learning Pascal will Currently, the reality in most high be discussed. school computer science programs is that Structured language alternatives will be BASIC will be taught. Thus, every attempt discussed by Michael Ward. With the must be made to convince teachers of hardware and software developments of the computer science to incorporate the problem last ten years, the ways that programming solving model and as many structured is taught and learned have changed. For programming concepts as possilm.e in their the most part, serial batch environments teaching of programming when using BASIC. are no longer the norm. So once initial Concepts that should be incorporated familiarization with the system is gained, when teaching programming using BASIC the methods by which programs are created, include meaningful. variable names, tested and corrected are much less indenting your programming listing-to show cumbersome and awkward. Therefore, there the scope 0-various program constructs, is one less impediment to learning. the orderly use of the GOTO statement and A much more important development has documentation. The modular solution of the been the discipline of structured problem utilizing GOSUB's when programming programmingdesign as a problem solving that solution should be emphasized. device and structured languages as a means Finally, it of implementing those solutions. As the may be desirable to implement the "while - do" languages move from the artificial loop, "repeat - until" loop, "for - next" loo? statements and structures of the machine to and the "if - then - else" statement it BASIC and emphasize the more natural human like native language their use. another major impediment has been removed.

268 . With full knowledge that this will not replace Pascal for implementation in an appropriate structured language, this is a partial solution for those teachrs who do not have the resources to offer Pa-cal

PARTICIPANTS:

Samuel F. Tumolo Cincinnati Country nay School Cincinnati, OH 45243 Michael Ward Willamette University Salem, OR Anthony Jongejan Everett High School Everett, Washington

269 Nationwide Computer Literacy Project Daniel Updegrove EDUCOM Princeton, NJ 08540

Steven Gilbert EDUCOM Princeton, NJ 08540

ABSTRACT

In response to the expressed needs of progress of information technology many of its member college's ,and on their own. universities, EDUCOM has begun a major project on computer literacy in higher This is a project of'great scope and education. Through mai3 surveys,' site potential impact, focused on a complex visits, literature seacones, conferences, need. Consequently, we have designed electronic mail and computer conferencingf. the project with four categories of EDUCOM seeks to determine the current state participation:. of the art in computer literacy programs for students, faculty, and staff. EDUCOM 1. Interested observers will receive plans to publish survey results; provide periodic announcements of project evaluation criteria, exemplary models, and activities, direction, and interim guidelines; and develop consulting teams to results. They may. occasionally assist - colleges and universities. in submit reactions, suggestions, creating or upgrading computer literacy interesting articles, reports of activities. noteworthy local programs, samples of materials, etc., (to project The NECC presentation will (a) provide a staff via phone or mail). progress report on the project, including the initial survey results; and (b) 2. On-line participants will describe how attendees can participate_in participate in -the same way as the project through task forces, electronic interested observers, but will also communication, and computer based project join in one or more forms of- activities. electronic project activities; e.g., exchanging news and ideas .via Several recent studies document a real electronic mail, helping with the and urgent national need for college ongoing development of computer graduates who have had a solid introduction __.___1i_teracy--data--bases--(literat.UreT to information technology, but-----three people, programs, etc.) and/or primary obstacles exist: contributing to a computer conference.- 1. No coherent definition and conceptual framework for computer 3. Implementation Task Force members literacy have gained -wide enough represent their institutions support to provide criteria for (both when attending meetings and

evaluating . present instructional when on home campuses) by being programs. and to direct the available for consultation and

development of new ones. assistance with . key project activities; e.g., developing and 2. Higher education has no effective testing survey instruments, feedback system for identifying organizing site visits to successful computer literacy' noteworthy computer literacy

program models in some institutions programs, commenting . on the and for adapting and disseminating effectiveness of materials or them to others. curricula on nearby campuses, etc.

3. flood of vendor claims about new 4. Leadership TdtA Froce members will hardware,- software, and join otter nationally recognized "courseware" overwhelms the limited leaders (from EDUCOM member. capacity of most 'indNiduals, and institutions, form information even institutions, tomonitor the industries, from other bUsinestes, fromgovetnment agencies, etc.)

270

2s who are ready and able to commit criteria for computer literacy in insight, time, and authority to higher education. The will also developing a compelling new provide overall policy guidance and conceptual framework, set of advice to the EDUCOM staff and the definitions, and program evaluation Implementation Task Force.

SPONSOR: EDUCOM

271 Using the Microcomputer Creatively with Young Children

Marilyn Church June Wright University of Maryland College Park, MD 20742

ABSTRACT kindergarten and primary grades will be This tutorial will address the question, explained. "Do Microcomputers Enrich the Preschool A videotape showing children interacting Environment?" It will report the findings with the microcomputer will highlight the of a two year pilot project which discussion of the kinds of early learning introduced three, four, and five year-olds which theuse of the microcomputer offer. to microcomputers. Thepresentation will Recommendations for fostering creativity focus specifically on the techniques and through the discoveryapproach will be programs developed as a result of that included. A consideration of the role of project. A preschool curriculum based on the parent in computer education and the the Logo philosophy which gives the child a concept of the teacher, the child and the sense of mastery will be demonstrated. The programmer functioning as a team will relationship of this curriculum to the complete the presentation. introduction of Logo Language in

272 2SJ HUNTINGTON III: MICROCOMPUTER COURSEWARE DEVELOPMENT PROJECT

by Thomas T. Liao

Department of Technology and Society, State University of New York at Stony Brook, Stony Brook, New York

Abstract Each computer program is fully documented for its structure (flowchart and pseudocode), meaning of In this paper, an overview of the variables, and equations used. A complete pro Huntington III Project is presented. The gram listing and suggested methods for adapting primary objective of this National Science programs completes the teacher's guide. In Foundationfunded project is the development of addition to the exemplary courseware packages, the a set of interactive courseware modules for use project staff plans to develop a courseware design in grades 8-12 mathematics and science classes. handbook. The project is an outgrowth of the Huntington I and Huntington II computer simulation projects, Dr. Thomas T. Liao, Department of Technology which were also funded by NSF. and Society, State University of New York at Stony Brook, isDirector of theProject; Dr.Ludwig Braun, New York Institute ofTechnology at Old Westbury, is the principal collaborator. Introduction Initially copies of the exemplars will be available for those interested in field testing Huntington III is developing courseware for the materials. theCommodore PETandApple IImicrocomputers because of their widespread use in education and Approach to Design Development their graphical characteristics. The interactive science and math courseware packages are oftwo The Huntington IIIProject uses a team types: oriented systems approach and a welldefined set of design criteria. To help ensure quality Multiple Use Courseware courseware, the project uses a systems approach to instructional design and development and carries These coursewarepackages feature generic out the work with a team of two or three programs that can be adapted for satisfying professionals who have the combination of various instructional needs; for example, an expertise that is required. educational game such as TicTacFlex helps students learn simple arithmetic skills, algebraic The development of quality microcomputer equations, or elements ontheperiodic table. courseware should be similar to the development of Also teachers can decide the version of the game other types of instructional materials. An and questions thatthey wanttheir students to effective approach is one that is modelled after use. engineering systems design and development. The components ofthis systems approach to instruc Applications Courseware tional design and development are:

These courseware packages focuson techno I. Identificationofinstructional need(s) logical topics that will help to provide and characteristics of students. opportunities for students to apply and integrate basic science, math, and programming concepts; for 2. Specification of design criteria and example,an electrical energy inventory program learning objectives. helps students to learn some basic physics concepts and algebraic equations that are used in 3. Identification of constraints suchas analyzing a realworld program. limitations ofthe learning environment and the microcomputer to be used. A comprehensive teacher's 'guide accompanies each computer program, in which the rationale and 4. Brainstorming of alternative designs of performance objectives are clearly stated and the computer program(s) and support operational pattern and sample runs are provided. materials. Also included are master copies of student handouts and worksheets and recommendations for 5. Specification of the content of a instructional strategies and additional resources. courseware package.

273 29u 6. Development of a prototype courseware-\ 4. Is the program highly interactive? The package. user should be consistently involved in providing responses to problemsolving 7. Field testing of the coursewarepackagie. itrl: Learning should be as active as is possible. 8. Revision of the courseware package based on feedback from students and teachers. A teacher's guide accompanies each computer program and contains information that:allows for The development of quality microcomputer effective use in the classroom or learning courseware requires thecreative combination of laboratory; is required for program modification; six areas of expertise: provides additional documentationso that the program can be used as an example of good 1. Instructional design and development programming. experience. An exampleof thetable ofcontents ofa 2. Knowledge of subject matter. teacher's guide follows:

3. Understanding of pedagogical methods. 1. Overview This includes an abstract, speck 4. Knowledge of microcomputer capabilities. fication of grade levels of users, microcomputers to be used, and contents. 5. Knowledge of programming. 2. Performance Objectives 6. Ability to write curriculum materials. Thesestate that studentsshould be able to do after using the courseware. Since it is unlikely that any one person can be expert in all of the above areas, a team of two 3. Rationale or three people works together to develop quality This specifies why teachers should use courseware. Our experience also demonstrates that this package. the giveandtake that occurs among team members leads to the development of more effective 4. Operation of Computer Program and Sample materials. Runs

In the Huntington. III Project, we are guided 5. Student Materials by the preceding suggestions and, in addition, use a) Description a set of more specific design criteria that guide _Description ofhandouts and sample the designand development of eachcourseware responses. package. These criteria are especially useful for b) Handouts/Worksheets finetuning the materials; some are related to the backgroundor guidance teacher's guide, while others deal with the ;01:1=1:111= Computer program. 6. Instructional Strategies When developing a computer program,the team include suggested ways for uses four sets of design criteria, which are using i:OUrseware package. related to the following four questions: 7. of Computer Program 1. Is the program "user friendly"? To sat a) Structure (Flowchart and Pseudocode) isfy this criterion, program users should b) List of Variables have complete control of the pace of the c) Ma...hematical Model(s) presentation. Not only should instruc d) Listing of Program tions and other screen output be easytounderstand, but the programs 8. How to Modify Program should be designed so that they can be easily modified. 9. Additional Resources a) Background Information 2. Is the program "user proof"? To satisfy b) Extension Ideas thiscriterion, the program should be c) References able to continue even if the user strikes the wrong key;the program should only Field Testing of Prototype Courseware Packages accept meaningful input data. During the 1982-83 school year four prototype 3. Does the program take advantage of the packagesarebeingfieldtested--each computer unique characteristics of microcomputers? program and the accompanying teacher's guide will Appropriate use of graphics is one way beused by at least 10 teachersand in 20 of satisfying this criterion. Other classrooms. Abstracts ofthe four packages and techniques include intelligent use of the the evaluation questions follow. timing capability, provision of instant feedback, and useof simulation gaming activities.

274 29 it Besidesobtaining feedbackviathe student andteacher questionnaire, the project staff will Concluding Comments observestudents and teachers actually using the materials. During the summer of 1983, the The objective of this paper, written inearly coursewaredeveloperswill use the information December 1982, is toprovide anoverview and obtained fromfield testing to revise existing a update of the approach and products of materials and to develop additionalexemplary microcomputer courseware developmentproject. In courseware. Jane 1983, the Huntington III Projectwill furnish an additional update for the oralpresentation at Abstracts of Some Huntington III Microcomputer Computing the 5th AnnualNationalEducational Courseware Packages Conference.

Domestic Electrical Energy Use and Cost At that time we plan to demonstratesample (Applications Courseware) courseware materials and to describefeedback from trialof courseware materials. In students the 1982-83 In this applications package, addition, a preliminary edition of theProject's interact with two computer programs to learn how Handbook for Microcomputer Courseware Designwill much electrical energy is used in their homes and be available for inspection. how their electric bill is computed. They also have theopportunity to explore theeffect of various methods of conserving energy to minimize the size of their electric bill; in addition, they carry out data collection and analysis.

Yellow Light Problem (Applications Courseware)

This courseware package provides an opportunity for students to study the factors that affect the motion of a car at a traffic intersection. First, students can use a simula tion to determine their response time, which is then combinedwith four other factors (speed limit, yellow light time, deceleration, and width ofthe intersection) to determine the Go,Stop, and Dilemma zones. By changing each of the five factors, one at atime, students can study the effect of each parameter. Finally, they can graphicially analyze:

a) How the speed limit, yellow light time, and width of the intersection affect the Go zone.

b) How the speed limit, response time, and deceleration affect the Stop zone.

Mass Spring_ (MultipleUse)

When using MassSpring, students interact with a computer program to learn how two usercontrolled variables--spring constant (K) and damping (B)--affect the displacement of an object (M) over time (T). The resulting displacement of the mass is shown in graph format. The students also have the opportunity to decide what values of K and B are appropriate for specific massspring damping systems such as the suspension system of an automobile.

TicTacFlex (MultipleUse)

This courseware package includes two. drillandpractice programs that are modelled after the tictactoe game. Teachers can modify the programs in three ways: (1) choice of game format; (2) questions and answers to be used; (3) instructions to be used. The sampleprograms contain a review of chemical symbols and practice with algebraic relationships.

275 292- Questions from Teacher Evaluation Questionnaire

HUNTINGTON III: MICROCOMPUTER COURSEWARE DEVELOPMENT

Teacher School

Grade level of users Address (Please use separate page for each grade level.)

(Zip) Name of class for which this courseware is being used

Name of courseware package

Please comment on or provide suggestions for each question.

I. Does the program meet the objectives stated in the teacher's guide?

Yes No Partly

2. Are the pre-programming activities effective in preparing the students to use the program?

Yes No Partly

3. Do the worksheets interface well with the computer program?

Yes No Partly

4. Are the instructions in the computer program clear?

Yes No Partly

S. Does the computer program highlight the important concepts?

Yes No Partly

6. Do students find the topic being studied to be related to the real world?

Yes No Partly Does not apply

7. Does the program actively involve students in the learning process?

Yes No Partly

8. Did your students gain knowledge of applications of math/science concepts as a result of using the program?

Yes No Partly Does not apply

(continued)

276 293 Teacher Evaluation Questionnaire (continued)

9. Are the graphics clear, relevant, and pleasing?

Yes No Partly

10. Is the program easy to modify? (answer if appropriate)

Yes No Partly

11. How much time did you spend in preparation for using this courseware package?

12. Do you consider the amount of required preparation time to be reasonable?

Yes No Partly

13. Do you have other suggestions for improving the courseware package?

277 29 Questions from Student Evaluation Questionnaire

HUNTINGTON III: MICROCOMPUTER COURSEWARE DEVELOPMENT

Name of computer courseware

Your grade level

Name of class for which you are doing this program

1. Did the discussion presented before you used the program prepare you adequately?

Yes No Partly

2. Were the worksheets coordinated with the program?

Yes No Partly Does not apply

3. Did the instructions in the program enable you to use it effectively?

Yes No Partly

4. Diethe program help you to understand applications of math/science concepts?

Yes No Partly Does not apply

S. Did you find that parts of the program moved too slowly or too fast?

Yes No Partly

6. Did you choose to repeat parts of the program?

Yes No Partly Which parts?

Why? -:-Fun? More information? Other?

Explain

7. Does the material in the program coordinate with the subject matter you are studying?

Yes No Partly

8. Would you like to use more programs on different subjects?

Yes No Partly

. How does the subject matter in the program relate to out-of-school experiences?

10. Do you have other suggestions for improving the courseware package? A Universal Computer Aided Instruction System *

Henry Gordon Dietz Ronald J Juels

Polytechnic Institute Of New York Route 110, Farmingdale NY 11735

PILE, the Polytechnic Instructional' Language for This paper discusses the design features and the Educators, is a consistent universal language designed embodiment of the PILE system. to provide a simple, yet versatile, language that non-computer-oriented teachers can use to prepare computer aided instruction (CAI) lessons. Developed Ease Of Expertise under a grant from the Initial Teaching, Alphabet Foundation, PILE supports the Initial Teaching Many CAI lessons are written as little more than Alphabet (and othercharacter sets), graphics, and the equivalent of a series of flashcards, hence it is audio cues; further, PILE is designed to operate easy to write such lessons in BASIC, Pilot, or nearly efficiently in microcomputer systems currently any other computer language. Unfortunately, more available and to be portable across vastly different complex (hence more interesting and more effective) hardware and software systems. However, the most CAI lessons are difficult or impossible to write in important LsiAct of PILE is that it encourages more most CAI languages unless the programmer is an sophisticated use by hiding details of implementation, expert. In defining PILE, the primary goal is to unlike Pilot which is simple by excluding abilities create a system which is almost effortless to master that do not have obvious implementations. This paper so that higher quality and reusable CAI lessons can be discusses the techniques used in to achieve PILE produced by educators who do not have an extensive without simplicity sacrificing ability or background in computer science. portability.

Introduction There are several major factors in design of a The current dearth of quality CAI software is well computer language system that is easy to master. known. This unhappy situation parallels general These factors are: coherence of program structure with software availability prior to the existence of modern conceptual program structure (5], ability to build software practices. Computer Scientists now recognize generalized software tools (modularity) EU, use of the importance of language definition, modularity, data objects (rather than types and rigidly specified structures) (1], and good reuseability, and portability. PILE, a universal readability inherent in the modern system for CAI software development, has been language (self-documenting code). designed to address these important attributes. PILE was developed by researchers at the Polytechnic Institute of New York under a grant from the Initial Teaching Alphabet Foundation (ITAF) and will soon be Program Structure: There is a strong tendency to distributed worldwide by the ITAF. adopt a trivial, inflexible, program structure for CAI lessons. Pilot, for example, is such a language; The system as implemented is portable to a wide designed on the theory that trivial program structure variety of microcomputers, minicomputers, and is easier to Learn. While, on the surface, it appears mainframes, and encourages reuse of program modules. that such an approach would succeed, forcing a program , Educators with minimal experience in computer with a different natural structure into a structure technology can easily learn to use the facilities made the language accepts shows the futility of this available by PILE and can incorporate modules as strategy. In Pilot, it is often impossible; in BASIC, needed. While the cost of hardware is dropping it is possible only with great effort and an excellent dramatically, the requirement for rewriting programs understanding of the particular computer's already in place can greatly inhibit the widespread implementation of BASIC. and cost effective use of CAI. The PILE system, designed for portability, can effectively operate across the spectrum of current and future hardware. There are two separate aspects of the structure of a program: the control structures used for looping and conditional execution of parts of a program, and the * This research project has been supported by a overall structure of the program. Neither aspect is grant from the Initial Teaching Alphabet Foundation. unique to CAI.

279 29G In the earlydays of computer programming, there function "max" begin were many different theories concerning program biggest = argl; while (argent 11 "0") do begin structure. There remain a variety of approaches to overall structure, but Pascal-like control structures if (greater(Warg" !argent), have become somewhat standard. biggest)) then biggest n Ware !argent); The advantage of a Pascal-like set ofcontrol argent = argent - "1"; structures is simply that people tend to think in end; terms of them naturally. Structures like return(biggest); if...then...else, while...do, and repeat...until are end; all common ways of expressing, in natural language, the concepts they represent. Since the semantics of the control structure mirror the natural language The other keyfeature of PILE in building modular conception, fewer mistakes are made and more complex software tools is dynamic linking at run -time. Since programs can be written. PILE is conmonly run on microcomputer systemswith modest resources, avoiding the overhead of having In PILE, we have chosen to adopt a set of control multiple copies of compiled functions is important. structures as similar as possible to those in Pascal (although Pascal programmers will notice that PILE is However, dynamic linking in PILE does much more more, tolerant of semicolon and missing keyword than save disk space. In PILE, the compiled code for errors). However obvious this choice may seem, these each function is merely the value associated with the control structures are unusual in a CAI language. name of the function; functions can be manipulated as data. The overall effect is that functions exist in a The overall structure of aPILE program is also huge virtual name-space (rather than in small main unusual for a CAI language, but consistent with the memory), updates of a function inherently update all state of the art in computer language design. Each references to that function. Modular use of functions program consists of a main program and, optionally, does not requirespecial libraries nor linking, and functions. function names can be derived at run-time.A function need exist only if it is actually called (useful in The most difficult concept for the programmer using testing programs before all the functions have been Pascal-likefunctions is nested scoping, so PILE has written). only two scopes: local to this invocation of a and function or global (actually local to the main (Self-modifying code and programs thatcreate program). Functions may be defined anywhereexcept then execute functions within themselves are possible, and inside another function and may be invoked from however, sucil techniques are rarelyemployed, anywhere (as in the C language). Recursion is there are safeguards to insure against accidentally supported. executing data.)

No distinction is made between functions and subroutines (the returned value is ignored when called Data Objects: Until computers and humans begin as a subroutine) and each invocation of a function can naturally communicating in the sane language, there pass a different number of arguments, using call by will always be a trade-off between native language for value and referencing by position of arguments in the computer and programmer. Making the computer more function call. efficient is easier than making the human more efficient, therefore, PILE opts for a form most natural to humans. Data types, and data structuring, are both at the center of this trade-off. Software Tools: The easiest way to write a program Traditional CAI languages standardize on a single is to reuse a program that is already written to data type, but no data structures, which limits the perform that function. If parts of a program are ability of the system to an unacceptable level. designed as separate modules, or functions, then only functions and control structure that have not Traditional computer science imposes a wide variety previously been written must be implemented; the of data types and structures. Pascal, C, and PL/1 majority of functions withinmost programs can simply have approximately half their complexity (measured by be reused. There is nothing special about CAI that the BNF dedicated to data typing and would make this general rule of computerscience productions) declarations. There is no need for this complexity invalid. other than to increase the executionefficiency. programs are rarely limited by In PILE, the implementation of functions makes However, CAI computational speed, even on microcomputers. reuse exceptionally uncomplicated.

Part of the simplicity of modular design in PILE is The AI (Artificial Intelligence) community has due to the generality of the arguments to a function. developed several languages that effectively use one one universal data structure. For example, in most languages a function which would universal data type and return the maximum of a group of numberscould not be Lisp uses atoms and lists, Logo uses objects and uses strings and the written to accept a variable number of arguments, so lists, and Snobal tables; max(a,b) and max(a,b,c) would require two different appelations differ far more than the implementations max functions to be defined; a single maxfunction can EU. In much the same way, PILE uses values and easily be written in PILE as: environments.

280 29 7 A value is really nothing more than a Graphics: Graphics used in CAI systems tend to be variable-length character string. Variables in other simple drawings with little or no animation. PILE languages ace names in PILE. An environment is merely graphics are designed to make drawings, at that level, a list of name-value pairs considered as a whole. easy to create and to transport to different 4ithin a non-local environment, a name isreferenced displays. as "environment"Vname"), even if the environment is stored in a file on disk. There are no data To accomplish this, the PILE system includes an declarations. image editor. The image editor writes a program which results in the drawing when executed by the PILE Operations on data are also similar to those found Interpreter (PILEI). Since the image is specified as a in Lisp and Snobol, except that PILE uses conventional mathematically-precise entity, although it is drawn in algebraic notation for all expressions. a conventional manner, PILE programs can adjust the image to the characteristics of different displays.

(Turtle graphics are also supported, but are not Powerful, Flexible, Features For CAI embedded in the language)

Thus far, the features discussed have been applicable to general-purpose computer languages as Audio Cues: As many ether CAI systems, PILE well as to CAI systems. To this extent, PILE is a supports control of a cassette recorder. However, general-purpose language. However, there are special PILE also supports phonetic voice synthesis. Each abilities that are so commonly used in CAIthat, for font can have different pronunciation rules for reasons of efficiency, Cley ought to be directly phonetic translation, and intelligible speech can be embedded in the language. These features involve produced from text. Text in the ITA font is database-like referencing, character fonts, graphics, pronounced particularly well due to the phonetic and audio cues. nature of the ITA. In addition, natural languages which have no written form can be supported. Database-like Referencing: PILE environments hake Regardless of font, pronouncing a phrase is as simple relational database operations simple (although as: perhaps wasteful of disk space). For example, the answer given by a student can be entered in a database (on disk) by: sayiphrase);

"question"Cstudent_namel = student_answer;

Although PILE also supports conventional disk I/O, Portability when using the array-like referencing of environments, Hence, for purposes the simple assignment in the above example is all that There is no standard computer- is needed... there is no open file, close file, nor of portability, PILE defines its own standard which is statements to scan for the student's name; it is all easily emulatld on currently available systems. transparent to the programmer. Reading is equally simple. The only disadvantage is that there must be one environment for each relation expressed in the Pseudocode: PILE is implemented as a compiler database. (PILEC) that generates an internal pseudocode, an interpreter (PILEI) that executes this pseudocode, and a series of utilities to aid in building lessons Character Fonts; The Initial Teaching Alphabet written in PILE. Unlike the pseudocode implementations (ITA) uses a set of phonetic characters, in addition of Pascal, the speed of execution of PILE programs is to the traditional lower-case alphabet, to aid in not significantly degraded by the overhead of teaching written English. Since PILE was developed for interpreting the pseudocode; like Snobol, thetypical the Initial Teaching Alphabet Foundation, the system PILE pseudocode instruction takes much longer to must efficiently deal with non-standard (relative to execute than to decoOe. computers) fonts. The entire PILE system, including the pseudocode In order to provideflexibility, PILE supports the interpreter (PILEI), is written in C. This, permits ITA, and user-defined character sets, using porting to oost microcomputers and minicomputers. stroke-fonts. A stroke-font consists of a list of While an assembly language version of PILE would be character definitions. For eachcharacter, the faster, most of the spend increase is easily achieved proportional spacing width, the sequence of pen by re-writing only a few C functions in assembly strokes to draw it, and the 8-bit code to represent it language. On a 4Hz Z80a, approximate relative speeds internally are specified. are: 1 for PILEI in pure assembly language, 1.5 for PILEI in -C with 5 simple functions re-written in Using this implementation, PILE programs can assembler, and 3 for PILEI in pure C. control the way in which text is written to the display by choice of stroke-font, character size In additionto the efficiency of generated code, C scaling, and stroke rate (the speed with which the is language of choice because of its popularity pen-strokes of each character are drawn).Still, text and the quality of the standard definition of C. While is written to the screen by a simple Pascal-like write BASIC and Pascal have many mutant versions, C remains or writeln. relptively pure.

281 29s Internal Data Formats: Just as the PILE system and References compiled PILE programs must be portable, all other forms of data used by the system mustbe hardware and (operating system) software independent. To simplify this, there is only one form usedfor disk storage: Harold Abelson, A Beginner's Guide to the file format for PILE environments. Compiled PILE Logo, BYTE, Volume 7, Number 8 (August programs, data, images, configuration tiles; all tiles 1982), pages 88-112. used by the system are in this same format. Barry U. Boehm, Software Engineering Wherever possible, data is formatted in such a way Economics, Copyright 1981, Prentice-Hall, as to be crudelyhuman-readable. Since these files pages 114-144. must be able to be transferre from one machine to another, using printable ASCII for data files usually B. U. Kernighan and P. J. Plauger, is more convenient. Software Tools, Copyright 1976, Addison-Uesley Publishing Company. Graphics Meta-forms; The largest obstacle to portable CAI lessons appears to be portable graphics. T. U. Pratt, Programming Languages: No two displays are even remotely similar. The only Design and Implementation, Copyright solution is to abstract a graphic meta-form which can 1975, Prentice-Hall. be applied to nearly all displays. N. Wirth, Foreword to A Primer On Pascal, In PILE, this graphic meta-form consists of points, Copyright 1976, Winthrop Publishers, vectors (lines), ellipses, andtilled areas (seed pages XI-XII fill). These constructs apply to any device (although filled areas are not implementedon some displays), and are used for character stroke-fonts as well as graphic images. PILE graphics have been successfully ported to Apple lIs, NorthStar Advantages, Tektronix 4006 terminals, HiPlot DMP4 plotters, and Zenith Z19 terminals.

Conclusion

In this paper, a system for developement of sophisticated CAI software has beenpresented. This system is very complex; however, this complexity is not experienced by the typical user of the system. By burying most of the complexity of CAI within the system, the CAI author is freed from most programming concerns.

The key to writing an effective program is the ability to make the program do what is desired. Although writing a trivial first program in PILE might seem awkward, writing a complexprogram is much simpler inPILE because the system is compatible with modern programming practices and provides many tools. According to the intermediate COCOMO model ID, a significant software project requires only 44% of the effort requiredfor the same project without these techniques. Ue expect that these claims will be supported by a large user community shortly after PILE's first general release later this year.

282 A STUDY OF STUDENTCOMPUTER INTERACTIVITY

David Trowbridge Educational Technology Center University of California, Irvine, CA

Robin Durnin Claremont Graduate School, Claremont, CA

Abstract role of a scientist or investigator. The simulations are embedded in Socratic A research project being conducted at the dialogs. Students play the role of Educational Technology Center is experimenters, manipulating objects on the addressing questions about interaction screen, and engage in a dialog with the among individuals and small groups as they program about the experiments. Both the use computer based learning materials. A simulated experiment and the dialog videotaping system that utilizes an require frequent keyboard activity. interface device between a microcomputer and a videotape recorder is described. An The learning materials developed at observational instrument is presented for the Center have been field tested in analyzing the interactions among group junior high schools, science museums and members and the computer program. Results libraries and in various social settings. of a pilot study and future directions of We have found a high level of a formal study are summarized. interactivity among students using these dialogs, whether one student is working Introduction alone at the keyboard, or several are working together. The primary advantage of computers in education may well be the high level of Frequently, when there are more than interactivity that properly designed one student at the computer, conversation learning materials can provide to is lively; group members talk with each students. Educational research has other often, offering assistance, indicated that active involvement of the encouragement, opinion and argument. The learner is an essential condition for the nature of the interaction in this development of reasoning skills, the environment is diverse. A given formation of concepts and the acquisition individual interacts both with the of problem solving skills (Hilgard, 1975). computer program and with the other group We would expect that learning environments members. Primary interaction with the which provide greater opportunity for program is via the screen and the active engagement would be more effective keyboard. Interaction with other than those which provide less. Certainly, individuals has both cognitive and social the computer has the potential for functions. Communication is both verbal providing a highly interactive and non-verbal. Social interaction is environment. sometimes cooperative, and sometimes competitive. This project uses a collection of highly interactive computer dialogs that Pilot Study were developed under two previously funded projects at the Center: Science Literacy In the initial stages of this in Informal Learning Environments, and investigation, we have attempted to Formal Reasoning Skills for Young identify variables of interest, choose a Adolescents. The learning modules we have research strategy, develop a system for chosen for this project have been data collection; and construct and described elsewhere (Bork, 1981; validate an observational instrument. Trowbridge, 1981). They are Batteries and These aims of the pilot study have been Bulbs, Speed, Optics, Tribble Families and met and are reported in this paper. In Sherlock Holmes. They all use extensive the formal study, we will seek to test graphics to simulate experimental certain hypotheses arising from the situations and place the student in the earlier work. There, we will investigate

283 300 the effects of group size on selected handling input and output during recording educational outcomes: interactivity, and playback (Figures 1, 2). frequency of success and achievement. Later in this paper, we will outline the Verbalizations are recorded on one of directions of the formal study. the audio tracks of the cassette tape, and keystrokes are recorded on the other. Thirty-five students in grades 6-8 During playback, the keystroke codes are took part in the pilot study. They worked used to drive the computer in at the computer as individuals or in synchronization with the video and audio groups of two or three. In consultation components. The computer programs have with their teacher, we selected groups been modified so that during the recording representing a wide cross-section of session, all keystrokes are transmitted to students, of high, medium And low ability. the videotape recorder. During playback, Some groups had worked together before, the keyboard is disabled and all others had not. Some groups were chosen keystrokes are received from the videotape to be heterogeneous with respect to player. This underlying software is a ability levels, others homogeneous. Some modification of the TextPort and GraphPort groups were composed of individuals of systems developed under UCSD Pascal in mixed sex, others of the same sex.We earlier projects at the Center. also used a variety of computer based learning materials which varied in several Taping Sessions respects, such as in how much graphics they used. Thus, the results of the pilot Students were brought from their study are drawn from a fairly diverse school to the university at the end of the collection of groups and computer school day. They worked alone or in materials. groups of 2, 3, or 4, spending up to an hour at the computer. The television Research Strategy camera was in full view, and students were told that they were being filmed. We have chosen to use a method of Usually, they became quickly engrossed in interactional analysis, coupled with a the computer dialogs and paid little videotaping system for investigating attention to the fact that they were being interaction in this special environment of recorded. No adults were present in the groups and computer based learning room during the learning session; materials. We have attempted to consequently, students were generally operationalize the idea of interactivity expressive and uninhibited. by defining a quantitative measure of the level of interactivity that can provide an Observational Instrument indicator of the degree to which a particular student is actively involved in Our observational instrument has been the learning session. We use this to derived from the methods of interactional examine the effect of group size on the analysis developed by Bales (1950) and level of interactivity. Flanders (1970). Bales' system for observing small groups engaged in problem Videotaping System solving situations has twelve categories (e.g., Gives suggestion, Asks for As students use the computer based suggestion, etc.) grouped into two social- learning materials, the sessions are emotional areas (Positive and Negative) recorded on videotape. Three components and one task area (Neutral). Flanders' of the group activity are recorded: system for analyzing teacher behavior also (1) video of the students working has multiple categories, most of which together,(2) audio of their conversation, relate to teacher talk in the classroom. and (3) all key pushes on the computer keyboard. The videotape data collection Our own instrument for interactional system consists of a microcomputer that analysis is suitable for recording both runs the interactive learning materials, a verbal and non-verbal behavior. Use of video cassette recorder with two separate videotapes allows us to categorize audio channels, an interface device for gestures, actions and facial expressions connecting the computer to the recorder, which do not involve vocalizations. The and associated utility software for system of categories we have developed in the pilot study is shown in Table 1.

30i 284 Table 1. Interactive Behavior Codes session. They use the digital timer on an auto search controller to keep track of Keyboard Interaction the frame numbers on the video. In each ten-second period, each observer makes one types at keyboard or more single-letter entries (typically, no more than 4 observations during each Verbal-Cognitive Interaction 10- second interval) in a box.

T tells, directs others 1. Interaction with Computer Q queries, as!-s for suggestions While it can be argued that reading from the computer screen is a form of A accepts, responds to suggestions of interaction with the program, we have others chosen not to include this behavior in our measure of interactivity. We have found L looks away to ponder or discuss with that for group sizes of 1, 2 or 3, each others member typically watches the computer screen more than 95% of the time. It is a I interprets in one's own words relatively passive activity, akin to reading from the page of a textbook or X explains, formulates reasons listening to a lecture. It usually indicates attentiveness to the learning M formulates question or answer activity, but it is not active in the sense that typing at the keyboard is. P formulates prediction Our category, K, for keyboard E evaluates using criteria activity was used to record instances of keyboard input excluding simply pressing D disagrees with program or objects to the space bar to continue (a standard message convention used in these programs to ensure that messages are not cleared from Cooperative-Social Interaction the screen before the user is ready to go on). We have included all cases of using n neutral conversation, opinions the arrow keys to move the graphics cursor, pressing keys for single character a approval, agrees with another input and typing in words and sentences. With the elimination of simple "page- d disapproval, disagrees turning" input, the category, K, is counted as active participation on the s shares keyboard with others part of the student.

t takes turns 2. Verbal-Cognitive Interaction h gives help, assists another (aid to Ten categories of behavior represent action) instances in which the subject is apparently engaged in some kind of polls others, solicits, votes cognitive activity related to the content of the learning material. Most of these y delegates task to another categories are verbalizations concerning the program. Verbalizations that were not e encourages another related to the program in any way were categorized as Off-Task. Other verbalizations that were not cognitive could be identified as Cooperative-Social.

3. Cooperative-Social Interaction

Coding Procedures Nine categories of behavior represent socially supportive and cooperative forms Observers view two screens during of interaction. While we did record a few playback of the videotapes: the computer categories of anti-social behavior (e.g., screen as it appeared to students using wresting the keyboard forcibly from the learning material and a television another individual, verbally abusing screen showing the students themselves. another person, or consciously ignoring During the playback session, observers the talk or other behavior of another have a coding form before them with a person), this happened relatively sequence of boxes corresponding to each infrequently. Competitive behavior was ten-second time interval during the also relatively rare. The behaviors labelled Cooperative-Social are all

285 considered to be generally beneficial to Table 2 the learning process. Any social behaviors that were distracting to the Interactivity Rate learning activity were categorized as Off- by Group Size Task. Elec 1-4 .41 (2) .57(2) .43(2) Measures of Interactivity Speed 1-3 .45 (2) .82 (3) To quantify the observations of interactive behaviors, we define Speed 4-5 .24(1) .45(2) .49(2) Interactivity, Interactivity Rate, and three components of Fractions of 1 2 3 Interactivity Rate. Group Size A numerical value for Interactivity of a particular individual participating in a given session may be defined as the total number of interactive behaviors observed for that session according to our table of behavior codes. By dividing the 1. Interactivity Rate as a function of Interactivity by the corresponding number group size of 10-second coding intervals over which the observations were made, we obtain an We would expect that an individual Interactivity Rate. The Interactivity working alone would have a lower rate of Rate represents the degree of active interactivity than when working as a involvement of an individual participant member of a pair, simply because when in the learning activity. working alone, modes of verbal-cognitive and cooperative-social behavior are not We also define three components of available. Furthermore, we would expect Interactivity Rate corresponding to the that if the group size were to grow very categories Keyboard, Verbal-Cognitive and large, then the Interactivity Rate for any Cooperative-Social. To determine the particular individual would fall. Fractions of Interactivity Rate in a Presumably, a small group with some number particular learning session, we count only of members greater than one would result the number of observations falling within in the greatest interactivity. a particular category and divide by the Observations on this small sample is number of 10-second intervals. consistent with this expectation (Table 3).

Results from Pilot Study On several occasions, we have observed that in groups of three, one In this section we present results person tended to withdraw from the group's from a sample of sixteen sessions. Six activities. In order to test whether this involved the first four activities of the is a likely occurrence, we have Batteries and Bulbs program (Bork, 1981). categorized the members of each 3-person Five sessions involved the first three group on the basis of the Interactivity activities of the Speed program and five Rate of each into the High, Middle and Low sessions involved the final two activities member of the group. Then we have plotted (Trowbridge, 1981). Four sessions the averages for the High person's Rate consisted of 3 students working together from each group, the Middle person's Rate at the computer, seven sessions had 2 and the Low person's Rate (Figure 3).A students, and five sessions had 1 student conclusive generalization will have to working alone (Table 2). Time to complete wait until we have obtained observations the four activities ranged from 25 to 45 from a larger number of groups. However, minutes. Students typically remained if the graph retains the shape suggested attentive to the lesson for at least 30 by these few data, we shall conclude that minutes. Overall, interactive behaviors for groups of three, there is a tendency occurred an average of once every 20 for one member to be left out. seconds.

31)3 286 Table 3 Interactivity Rates

Group Size 1 2 3 low high low med high I' k .21 .17 .20 .13 .13 .14

I' .15 .21 .36 .08 .21 .28 c I' .00 .14 .17 .07 .14 .14 s

.36 (5) .52 (8) .73(8) .28 (3).48 (4) .59 (4)

Note: (1) Combined results from Elec 1-4, Speed 1-3, and Speed 4-5

(2) (#) indicate the number of sessions analyzed

2. Mode of Interactivity as a function We are examining the effects of both of group size group size and prior group experience on interactivity and achievement. Three modes of interactivity may be Interactivity is measured using the separated from the overall Interactivity observational instrument developed in the Rates described above: Keyboard: pilot study. Achievement is measured Cognitive and Social. As expected, the using a pencil and paper test which is fraction of Keyboard Interactivity administered immediately after the session decreases as the size of the group with the computer. The post-test consists increases (Figure 4). In addition, both of eight questions, each showing an Verbal-Cognitive and Cooperative-Social arrangement of batteries, bulbs and wires Fractions are highest for groups of two and asking whether the bulb would light in (Figures 5, 6). The two individuals who each case. worked alone were instructed to express their thoughts verbally as they used the We have videotaped fifty-six 7th and 8th computer materials; thus we see a grade students in 26 sessions with the significant portion of their interaction computer. They were generally college- as verbal-cognitive. In the formal study, bound, of middle ability students, from we hope to ascertain whether the fairly affluent middle class backgrounds. observation of maximum Interactivity Rate Taping sessions took place in the morning, for groups of two is statistically with an adult present in the room in a significant. non - supervisory capacity.

Directions for the Formal Study Table 4. Sample for theformal study As a result of the pilot study, we decided Group Size that the variable of group size was worth Numbers of examining in greater detail, and this Groups Ind. 2 3 became the focus of subsequent research. In addition, prior group history seemed to Cooperative Yes 5 5 4 2 have an effect on interaction, so this Experience? No 3 3 2 2 became a secondary variable of interest. The mixture of ability levels also seemed Numbers of to have an important role, but due to Students 8 16 18 16 limited resources, we chose to control for this variable by selecting all of our The hypotheses we hope totestin the subjects from comparable, middle ability formal study are: students for the formal study. We have not found sex to have a strong effect on 1. Students working in pairshave higher interaction, so have chosen to combine measures of interactivity than students single and mixed sex groups. For the working in other grouping arrangements. formal study we are using the same instructional materials with all students 2. Interactivity measures for acquainted (Batteries and Bulbs). student groupings are higher than for unacquainted groupings.

287 304 3. Post-session achievement measures for Acknowledgements students who have worked in pairs are higher than for students working in other The interface device was designee and grouping arrangements. built by Michael Potter. Associated software was developed by John McNelly and Steven Bartlett. Elinor Coleman and Cynthia Powell have helped in the coding Conclusions of the videotapes. The authors wish to thank each of these peole for their The videotaping system described contributions to this project. here, which couples the computer to a video recorder, provides a rich source of information about the use of computer based learning materials. We have only References begun to explore the possible applications of a system that can reproduce human- 1. Bales, R., 1950. Interaction Process computer interactions at a detailed level. Analysis: A Method for the Study of Small Groups, University of Chicago. The behaviors coding scheme we have described enables us to quantify the 2. Bork, A., Kurtz, B., et al., 1981. overall level and quality of interactivity "Science Literacy in the Public in this learning environment.Preliminary Library - Batteries and Bulbs," evidence suggests that students working in Proceedings of the National pairs have the greatest opportunity for Educational Computing Conference. high interactivity. Soon, we expect to be able to make some generalizations about 3. Flanders, N., 1970. Analyzing how various kinds of computer based Teaching Behaviors, Addison-Wesley learning materials affect the Publishing Co.: Menlo Park. interactivity of the learner as well. 4. Hilgard, E. R. and Bower, G. H., Results of investigations such as 1975. Theories of Learning, 4th Ed., this should provide guidance to developers Prentice-Hall: Englewood Cliffs, New of computer based learning materials as Jersey. well as to classroom teachers who must manage the logistics of computer usage in 5. Trowbridge, D. and Bork, A., 1981. their classrooms "Computer Based Learning Modules for Early Adolescence," Computers in Education, R. Lewis and D. Tagg (editors), North-Holland Publishing Company.

288 3 tiiJ computer video monitor video cassette of recorder interface ( %I1 001 mic hone camera

Figure 1. Setup for recording sessions Figure 2. Setup for playback sessions

Keyboard Interactivity Interactivity vs. Group Size vs. Group Size >1'k

.,11 0.8 .1 0.8 > 0.6 0 0 High co 0.6 rt )-1 *Middle 0 0 00.4 9 g 0 4 H b,0 0.2 0 > 0.0 0.0

1 2 3 1 2 3 Group Size Group Size Figure 3 Figure 4

VerbalCognitive Social Interactivity I' s vs. Group Size c Interactivity vs. Group Size 0.8 .> 4.) > > 0.6 g 0.6 0

)-1 0 0)-1 0.4 g 0.4 H 0 co 0.2 0+ 0.2

)-1 0 )-1 > 0 0.0 fg 0.0

1 2 3 1 2 3

Group Size Group Size

Figure 5 Figure 6

289 306 THE IMPLEMENTATION OF TECHNOLOGY AND

THE CONCERNS-BASED ADOPTION MODEL

By Dr. Cheryl A. Anderson

Department of Curriculum and Instruction University of Texas at Austin

Abstract there is a body of literature on the change and adoption process which may provide some guidance. This paper presents a brief synopsis of There is also a change model called the Concerns- the literature on the change and adoption pro- Based Adoption Model (CBAM) which relates speci- cess as it relates to the task of infusing fically to the teacher and the adoption process. computer technology into the school system. The purpose of this paper is to present a brief The Concerns-Based Adoption Model (CBAM) is synopsis of the relevant change literature and to also discussed. The model is based on ten provide an introduction to this model. First, years of research on teacher behavior with however, it is necessary to look at why the exist- educational innovations. According to the ing school system is so difficult to change. author, CBAM can provide a theoretical framework for planning a computer implementation effort. The model provides the change agent Current School Model with a set of tools for measuring the success of the adoption; for measuring teacher atti- Part of the problem in trying to implement a tudes and feelings toward the innovation; and new technology, such as computers, is that the for measuring teacher behaviors with the inno- current school system is designed around the tra- vation. By using these tools, a change agent ditional teacher-learning model (Pitts & Schneider, is assured of using a more systematic approach 1981). The model is teacher- and textbook- to the implementation effort. dominated. Computer technology challenges this traditional focus because, instead of information that comes from the teacher or the text, the student's work revolves around the computer. With computer technology, teachers' roles will also It seems that schools are anxious to jump change (Sheingold, 1981). Unfortunately, schools Many on the computer technology bandwagon. are very stable institutions that are resistant to people are concerned about the potential for any kind of change (Goodlad, 1976). Not only is failure, in part because similar attempts to the system as a whole resistant, but the individ- change established school practices have failed. uals within the system are resistant as well. The A long line of innovations has preceded the change process is very dependent upon the individ- computer, including instructional television, aul, particularly in the educational system. This programmed instruction, teaching machines, team is because the individual teacher assumes a degree teaching, individualized learning, and modern of ownership over an innovation during the process Few of these made their way permanently math. of assessing its benefits and testing its value into the classroom. Why should we expect the (Podemski, 1980). This concept of ownership is computer to be accepted with enthusiasm when so critical in education because the individual Indeed, Oettinger many innovations have failed? teacher is allowed a certain amount of autonomy (1969) suggests that computer technology will not with regard to the curriculum materials and proce- be effectively used because its very nature works dures he or she uses in the classroom (Podemski, against the rigidity of the school system. Karen 1980). It is often the case that the individual Sheingold (1981), in a study of the implementa- teacher's attitude toward an innovation determines tion issues relating to computers in the schools, not only how it is used, but whether it is used "While the new technology does differ states: (Hall, Wallace & Dossett, 1973). in many important respects from the old, expectations about educational impact must be viewed cautiously. There are many steps between putting The Individual and Change a machine, albeit a powerful, engaging machine, into a classroom and making a difference for That the individual is the primary focus of children and teachers" (p. 4). efforts at technological change is evident in For those agents of change who are respon- social science literature. Wolcott (1981) notes sible for seeing that computer technology is that social scientists have much to offer the successfully infused into the school system, change agent who is trying to implement a

290

307 technology. He summarizes some of the problems 5. that an innovation is being implemented; which have been uncovered through studies of human it can be a curriculum material, a tech- behavior and change. These problems include the nology, or a teaching method that is new following: to the school; 6. that the innovation is appropriate to the 1. human will resist any change that threatens their basic security; system and has the potential to be effective within the school system; and 2. humans have a tendency to resist any technology that they do not understand; 7. that there is an informal or formal leader who represents the change agent for the 3. humans will not vary their behavior unless they can use the new technology process. The function of this person is to satisfy a need that is not being to design the effort in an adaptive and thoroughly met by existing technology; systematic way so that the state of change is assessed and reassessed (Hall, 1979). 4. humans cannot be forced to accept technological change; This allows the change agent to select certain interventions based on the latest 5. in order to be accepted, an innovation- - particularly one that originates from diagnostic data. Thus, effective train- extensive knowledge of science and tech- ing in the use of an innovation occurs nology--must be made intelligible and be when the trainer uses a diagnostic and given a place of value by the adopting prescriptive model so that training can culture; and be targeted to individual needs. 6. resistance to change is often centered Within the model, several diagnostic tools around the way in which the innovation have been developed to assess where an individual is administered rather than the innova- is in relationship to the training effort. Three tion itself. diagnostic tools have been developed: Stages of Wolcott (1981) reminds developers and dis- Concern, Levels of Use, and Innovation Configura- tions. What follows is a brief description of seminators of technologies ". . . that lessons about change go unheeded only at considerable these concepts. risk" (p. 25). He also points out that the typical teacher rarely values innovations because most do not address the teacher's real problems or Stages of Concern provide real help in the classroom. Thus, if the Stages of Concern about an Innovation are the computer is to be implemented successfully, it is feelings, motivations, and perceptions that indi- the individual teacher who must be won over. The viduals have as they progress through the adoption plan of action must present the computer in a non- process. The CBAM researchers have identified threatening, understandable manner in addition to seven stages: providing uses which will be valued by and of interest to the teacher. 1. awareness: at this stage there is little involvement with or concern about the innovation; Concerns-Based Adoption Model 2. informational: here the individual is only interested in receiving general As mentioned previously, CBAM could provide information about the innovation and he a useful framework for those who are respor.sible or she is not thinking about how the for implementing computer technology in the innovation will effect him or her per- schools. This model is based on ten years of sonally; research by the Research and Development Center 3. personal: at this stage the individual for Teacher Education at the University of Texas. is concerned about the demands that the The model is based on the following assumptions: innovation is making on his/her time; consideration is given to the personal 1. that change is a process, not an event; benefits of using the innovation as well neither a principal's edict nor a one- as the potential perils; day workshop can assure change, because 4. the individual focuses on change takes time; management: the processes and tasks of using the 2. that individuals within the institution innovation; concerns deal with the best must change before the institution can use of resources, management of time, change--thus the individual is the pri- efficient use of the innovation, and mary focus of the CBAM research; organizing use; 3. that the individual's personal feelings, 5. at this stage the individ- perceptions, and motivations concerning consequences: ual begins to consider the impact that the an innovation play an important role in innovation will have on the students; the determining the success or failure of an concerns focus on the relevance of the innovation; innovation to students and on the out- 4. that change is developmental; that is, comes based on student performance; an individual will move through identi- 6. collaboration: the individual has the fiable stages of concern about the desire to cooperate with other people who innovation and levels of skill in using are also using the innovation; and the innovation during the change process;

291 303 7. refocusing: this is when the individual 7. integration: the individual initiates begins to express concerns about replac- efforts to combine his or her use with the ing or modifying his or her use of the related activities of other teachers to innovation (Hall, i979). obtain a collective impact on the students; and Knowing a teacher's Stages of Concern can be 8. renewal: the individual starts to re- of use to the change agent in planning in-service evaluate the quality of the innovation and training. The CBAM research indicates that, for begins to make modifications or seeks training to be effective, there must be a match other alternatives to the innovation between the personal concerns, the expertise of (Hall, Louchs, Rutherford & Newlove, the trainee, and the method of training used to 1975). facilitate the use of an innovation (Hall, 1978). For example, nonusers of computer technology would Researchers at the R & D Center have discov- be most concerned with obtaining some general ered that, regardless of the type of innovation, information about the innovation and exploring the the majority of the teachers stay at the level of implications that it has for them on a personal routine use. In the first year of use 60-70% of level. A workshop dealing with the consequences the individuals will only reach the mechanical-use of using computers would be inappropriate at this level. They have also discovered that the pro- time; however, a simple hands-on experience would gression is not always a lock-step one; an meet their needs. CBAM research indicates that in individual can move out of sequence. However, the the beginning of the implementation process, CBAM researchers have found that individuals do teachers are not concerned with the implications not collaborate until they have personally mas- for students and will not be until their own per- tered the innovation. sonal concerns are overcome. Three different procedures have been developed for assessing the stages: an interview Innovation Configuration technique, an open-ended question technique, and a 35-item questionnaire. By analyzing the stages The third component of CBAM is Innovation through these methods, a profile can be obtained. Configurations (Hall & Louchs, 1978). It is based This profile will typically show that some con- on research which indicates that innovations are cerns are more intense. This intensity should often adapted or modified by individuals to suit progress to different stages. This enables the their particular needs or situation. Often an trainer to tailor the different interventions, innovation is so drastically modified that it is whether they be workshops or individual guidance, no longer used in the manner originally intended to the profile. The Stages of Concern focus by the developer. Innovation Configurations on the attitudes or feelings that an individual reflect the different patterns that result from has toward an innovation. The next facet of CBAM the selection and use of key elements or compo- focuses on the behaviors that the individual nents of an innovation. These components could exhibits when attempting to use an innovation. include: instructional objectives, materials, equipment, grouping patterns, and tests or some means of indicating that the innovation has been Levels of Use of an Innovation implemented. The purpose of identifying various patterns is to determine whether or not the imple- The Levels of Use measurement focuses on mentation has been successful. The procedure .eight behaviors which can be identified through suggested by the CBAM researchers provides the observation and interview techniques. These change agent with a checklist that can be used for behaviors include: this evaluation. The procedure is as follows:

1. nonuse: the individual has no knowledge 1. conduct interviews with either the devel- or involvement and is not attempting to oper or the person who is responsible change; for facilitating the adoption of an 2. orientation: the individual makes a innovation--the interviewer tries to decision to seek information about the identify key components by determining innovation; What the innovation will look like in 3. preparation: the individual makes the relationship to the behaviors of the decision to devote time to learn to use students and the teachers; the innovation; 2. conduct interviews and observations with 4. mechanical use: the individual is using a small number of users to determine the the innovation in an uncoordinated manner possible variations of the components; and his or her behavior focuses on mas- 3. develop interview questions and interview tering the task of use of the innovation; a larger number of users--questions are 5. routine: the individual stabilizes his asked about teaching practices with the or her use, little time is required to innovation; prepare for use, no changes are made in 4. construct a checklist which contains the how the innovation is used, and there is key components and the possible varia- little thought about improving use; tions within each component; and 6. refinement: the individual initiates 5. fill out a checklist on all users and efforts to increase student outcomes, and determine what dominant patterns exist thought is given to both the short-term (Hall & Louchs, 1978). and long-term consequences for students;

292 30j The benefit of developing a checklist is that Hall, G. E. & Louchs, S. F. Innovation configura- it can help clarify how the innovation is supposed tions: analyzing the adaptions of an to operate when it has been fully implemented. innovation. Research and Development Center This model is important in evaluating the success for Teacher Education, University of Texas at of the implementation effort. Facilitators of Austin, November, 1978. change and in-service trainers can use this concept to identify the components that need to be Hall, G. E., Louchs, S. F., Rutherford, W. L., & targeted for further effort (Hall, 1978). Newlove, B. W. "Levels of use of the innovation: a framework for analyzing innovation adoption." Journal for Teacher Education Conclusion 26(1), 1975.

This has been a very brief introduction to Hall, G. E., Wallace, R., & Dossett, W.A Devel- the Concerns-Based Adoption Model and the change opmental Conceptualization of the Adoption literature which supports its use. The model Process Within Educational Institutions. gives those of us interested in implementing com- Research and Development Center for Teacher puter technology a theoretical framework and a set Education, University of Texas at Austin, of tools for measuring our success in our imple- 1973. mentation efforts. It provides us with a method for systematically analyzing our efforts to faci- Oettinger, A. G. Run, Computer, Run. Cambridge, litate change. By using the Stages of Concern, Massachusetts, Harvard University Press, we can determine how teachers feel about using 1969. the computer in the classroom.By using the Levels of Use, we can determine how skilled they Phillipp, C., Muntner, J. & Cutlip, P. Computer are at using the computer. And by using the literacy for K-6 teachers. Paper presented Innovation Configuration process, we can first at the 20th Annual Association for Educa- identify the key components needed for full imple- tional Data Systems Conference, Orlando, mentation of computer technology in the classroom, Florida, May, 1982. and then evaluate Our success in relationship to those key components. CBAM has been used by Pitts, M. & Schneider, J. Educational Technology: the Montgomery County Public Schools in Rockville, Bright Promise or Dim Future. CEDAR Proceed- Maryland, in their efforts to implement computer- ings of a Cooperative School Improvement literacy training for K-6 teachers. School Seminar, Washington, D.C., June, 1981. representatives state "Rooted in both theory and practice, CBAM includes strategies that may be Podemski, R. "Computer technology and teacher used to guide teacher education as well as to education." Journal for Teacher Education provide a basis for determining when elements of 32(1), 1981. curriculum have been implemented" (Phillipp, Muntner, & Cutlip, 1982, p. 321). There are valu- Podemski, R. "Educational technology and the able lessons to be learned by looking at this body development-adoption process." Educational of literature. If we choose to ignore these les- Technology 20(5), 1980. sons, then it is likely that computers will join other technologies which have failed to impact Sheingold, K. Issues related to the implementa- teaching practices and which are relegated to the tion of computer technology in schools: a storage closet. cross- sectional study. Paper presented at NIE Conference in Issues Related to the Implementation of Computer Technology in Schools, February, 1981.

REFERENCES Wolcott, H. "Is there life after technology?" Educational Technology, 21(5), 1981.

Goodlad, J. I. "Schooling and education." In R. Hutchins (Ed.), The Great Ideas Today. New York: Encyclopedia Britannica, Inc., 1976.

Hall, G. E. Concerns-based inservice training: an overview of concepts, research and practice. Paper presented at Conference on School-Focused Inservice Training, March, 1978.

Hall, G.E. Using the individual and the innovation as a frame of reference for research on change. Paper presented at the Annual Meet- ing of Australian Association for Research in Education, Melbourne, November, 1979.

293 310 ELEMENTARY TEACHER EDUCATION: INCLUDING LOGO IN TEACHING INFORMAL GEOMETRY

by M. Moore and W. Burger

Department of Science and Mathematics Education and Department of Mathematics, Oregon State University, Corvallis, Oregon

Abstract classroom. As such, the course provides a computer literacy component. The course does provide practical application of strategies in Elementary education majors have little back- solving problems using the computer. Yet, this ground knowledge in geometry. Also, there are few course does not intend to teach mathematical opportunities for courses with computer activities concepts other than strategies in problem solving. integrated in the teaching of the content. Yet, PUrthermore, because of the popularity of the it is recognized that teachers tend to teach using course for all education majors, most students are the methods by which they learn. Although unable to enroll until their senior year. This students are exposedto one computer literacy aspect unfortunately restricts the possibility of course during their senior year, their mathmetics integrating the strategies developed in this courses do not utilize computer methods. course with other mathematics programs for pre- :,nition of these problems directed the service elementary teachers which would be more in vi :on of the informal geometry course for -,:ieme:,tary teachers at Oregon State University. line with recommendation three. The mathematics content courses have analysis of Logo exposed strengths and weaknesses addressed a broad range of topics: problem solving; in the language for demonstrating geometric an informal development of the real numbers and concepts. The revised course includes a Logo lab- arithmetic; mental, written, and electronic oratory in addition to a variety of additional computation; number theory; decimals and percent; environments for concept development.Evaluation ratio and proportion; probability and statistics; of the course available in June 1983. and a substantial amount of informal geometry. Topics in informal geometry are included that reflect the spirit of the second NCTM recommendation. These topics enable prospective teachers to develop skills in geometrical awareness and perception, analysis of geometrical In An Agenda for Action: Recommendations for Shapes, patterns and transformations, and the use School Mathematics for the 1980s, the National of informal deduction ( not based on axiomatics)to Council of Teachers of Mathematics (NCTM) has Show relationships among properties of shapes and addressed the broad issues facing the mathematics figures. The specific topics include: education community of the near future. Their 1. types of polygons and their properties first three recommendations are that: 2. abstract measurement 1. problem-solving be the focus of school 3. geometric constructions mathematics in the 1980s; 4. patterns of polygons and tessellations 2. basic skills in mathematics be defined 5. motion to encompass more than computational 6. isometries, congruence and symmetry facility; and 7. magnification and similarity 3. mathematics programs take full advantage 8. polyhedra. of the power ofcalculators and Approximately 13 weeks are spent on these topics. computers at all grade levels. Our experience has been that many of our Incorpating these objectives into our elementary elementary education majors have little background teacher education program at Oregon State knowledge in geometry, and benefit greatly from University has been accomplished through a working in a variety of. geometrical environments: cooperative effort involving the School of graph paper; geoboards, and dot paper; blocks, and Education, the Department of Mathematics, and the cutouts; drawings, and constructions; paper Department of Science and Mathematics Education. folding; or Logo. This educational method is See (3). The abstract Interpreting the third recommendation as a consistent with learning theory. notion of symmetry via isometries, for example, is call for at least minimal computer literacy, a first understood in concrete embodiments, or required three-quarter hour Instructional models, of the concept. Varying the irrelevant Strategies/Computer Education course was attributes, including those that determine the incorporated in the 18-hour mathematics/methods/ particular environment, and stressing the relevant This course has had computer science component. attributes enables the student to formulate a as its primary goals an introduction to Basic precise idea of the concept at hand. This programming and applications of computers in the

294

311 description of learning is advocated by Dienes, combine his own body motions in a familiar space to Bruner and van Hiele, among others. develop formal geometry. Each of the environments mentioned has Adapting ideas expressed by Weir and Watt certain advantages. Graph paper, for example, (9), learning to program in Logo provides an environment for: allows one to study such topics as symmetry in terms of numerical relationships among coordinates 1. developing logical thinking and problem solving; of points in the plane. The transformations T1 2. using strategies and variables; (x,y)=(-x,y), T2 (x,y)=(xy-y), and T3 (x,y)-(y,x) 3. exploring symmetry, design, angles, produce reflectional symmetries, while T4(x,y)= geometric forms; (-x,-y) produces another type. Interesting 4. developing and testing student's own questions arise quite naturally about which types theories; of isometries are "easy" to represent with such 5. understanding computers. transformations. All can be done in the complex An understanding of these characteristics of plane, but not all are so easy in Cartesian Logo and turtle geometry encouraged the inclusion coordinates, the "graph paper" environment. See of Logo activities in the geometry course for pre- (2) for other graphpaper applications. A possible service elementary teachers. The informal disadvantage of the graph paper environment is the geometry course content had previously been care- loss of "hands on" manipulation of objects, and the fully defined and described as mathematically necessity for a certain amount of prerequisite necessary in the preparation of elementary technical knowledge about coordinate systems, teachers. It is important -.to recognize the intent positive and negative numbers, and so on. This was not to change the content-but to enhance the serves to point out that each environment teaching of the content in another environment. facilitates concept formation and the development Thus, it was necessary to evaluate the Logo of problem-solving abilities in certain ways, but environment in comparison with other possible that no environment is complete by itself. environments for each geometric concept. The Recognition of the strengths offered by using process exposed strengths and weaknesses in both more than one environment led to the interest in Logo and other laboratory embodiments. In some including Logo and turtle geometry as one of the cases the capabilities of Logo facilitated concept environments in the informal geometry course. formation more clearly thee: other environments. Additional intereet in incorporating the use of the However, in some cases operations in Logo inhibited computer in learning mathematics was generated with concentration on the concept. the recognition that teachers tend to teach using As an example, consider the construction of the methods by which they learn. Although students the set of regular polygons. In Logo the are exposed to a computer literacy program during construction is trivial when compared to the their senior year, their mathematics courses do not environment offered by use of straightedge and utilize computer methods. The decision was, there- compass. More importantly, the constructions of fore, to experiment with incorporating turtle the regular polygons through Logo procedures geometry and Logo in the informal geometry course emphasizes the properties of the regular polygons because of the special pedagogical advantages of in a non-abstract manner. that environment and the usefulness of displaying This Logo procedure (which students create) computer methods of teaching mathematics. draws the set of all regular polygons of less than Turtle geometry is a computational style 12 sides (see Figure 1) when called with the geometry in which the fundamental building block is command polygon 3. an entity called a turtle. The turtle, similar to TO POLYGON :N.SIDES the point in coordinate geometry, is described as IF :N.SIDES<3 THEN MAKE "N.SIDES 3 having two independent characteristics, a position IF :N.SIDES>12 THEN STOP (as with the point in coordinate geometry) anda PENUP heading (which is not a characteristic of the SETXY -100 0 point). The turtle responds to commands which PENDOWN form the language called Turtle Talk. The commands REPEAT :N.SIDES [FORWARD 50 RIGHT make it possible for turtle transformations in (360 /: N.SIDES)) local space rather than in relationship to a fixed POLYGON (:N.SIDES +1) global referent. A set of primitive commands END provides the language building blocks. All other commands are created using these commands. There- fore, turtle geometry is a mathematical system based on turtle movements. It is a geometry with an undefined term (the turtle), a set of axioms which when implemented in a computer language, Logo, has aspects of coordinate geometry and vector geometry. Logo is a computer language used for a representation of turtle geometry. The turtle is typically represented by an isosceles triangle. In Logo the turtle becomes what Papert calls "an object to think with." See (8). Use of the computer as a programming tool and the turtle as a geometric drawing device aids the student to Figure 1

295 312 The compass-straightedge environment for the geometry, namely that the student can del: lop the construction of the same ten regular polygons is idea that a 360-sided polygon is a circle. significantly more abstract although it does Another example of the contrast in environ- present the mathematically interesting Theorem of ments ic the problem of triangle construction. Gauss. According to this theorem, three polygons Compass and straightedge constructions of triangles in the set of regular polygons with less than 12 allows the discussion of the attributes which sides, the 7-gon, 9-gon, and 11-gon, cannot be completely define a unique triangle. This inform- constructed using a compass and straightedge. ation is then utilized in showing congruence of This discovery can provide a nice mathematical triL.1gles (i.e., the three conditions sufficient to generalization but may result in an abstraction determine congruence of two triangles: side-side- for which the students are not prepared when first side, side-angle-side, and angle-side-angle). learning to use the straightedge and compass. Given three sides (meeting the necessary condition The straightedge and compass procedure for the that the sum of the lengths of two of the sides is construction of the regular n-gons begins with the greater than that of the third), construction of construction of a circle, easily done with the the given triangle is trivial with compass and compass. At this point the student must locate the straightedge. Without trigonometry, construction n vertices of the regular n-gon on the circle. A of this triangle in the Logo environment is a procedure for finding the vertices of the regular significant problem. Elementary teachers typically pentagon for example, is as follows: do not know trigonometry. Therefore, it is not 1. Label any point on the circle V1 and reasonable to change the content of the course to draw -013 perpendicular to 1071. include the discussion needed in order to construct 2. Join V1 to C, the midpoint of mr. "general ", non-equilateral, triangles in Logo. However, the Logo environment provides an excellent 3. Bisect angle OCV1 to obtain the point N, on OV1. environment for the construction of medians, altitudes, and perpendicular bisectors of the sides 4. Construct the perpendicular to 71 at N and obtain the point V2. for any triangle. Halving angles is easily done The segment V1 V2 is one side of a regular computationally and then incorporated in a Logo pentagon and from it points V3, V4, and V5 can be procedure to construct the angle bisectors of a found. These points are then connected to complete triangle, given the length of each side and the Similarly procedures can be the construction. See (6). Figure 2 demonstrates measure of each angle. this construction. easily defined for the construction of medians, altitudes and perpendicular bisectors. As a result, the relationshp of the centroid, the orthocenter and the circumcenter can be studied in a less complex manner than through paperfolding, protractor and straightedge, or compass and straightedge environments. And, the concentration remains on the concept rather than difficulties presented in theli'COnStruction. Logo is p4ti6iarly well-suited to illustrate concepts in simUnrity and magnification. Nhgnification and reduction are easily described in Logo procedures using inputs. For example, this procedure draws squares, (See Figure 3) magnifying them by a factor of two each time. Figure 2 TO SQUARE :SIZE In contrast, the Logo environment can create IF :SIZE>100 THEN STOP an inaccurate mathematical understanding when IF :SIZE<1 THEN STOP discussing the procedure to construct a circle. REPEAT 4[FORWARD :SIZE RIGHT 90] Use of the compass to construct a circle embodies MAKE "SIZE :SIZE*2 the definition of a circle, the set of points SQUARE :SIZE equidistant from a specified point. The END construction is trivial with a compass. However, in Logo, drawing circles begins with the use of the human body to walk a circle and progress to the procedure, CIRCLE. TO CIRCLE REPEAT 360[FORWARD 1 RIGHT 1] END Although this procedure "appears" to produce a circle on the output, it is, in fact, a 360-sided The discussion can be extended to regular polygon. Figure 3 include the idea of the circle as the curve By simply changing the size factor in this proced- approazhed as the number of sides increases (with ure to the proportional decrease in length). Without some MAKE "SIZE :SIZE*.25 recognition that the "look-alikeness" does not the squares will reduce in size by a scale factor indicate congruence, this example indicates one of one-fourth. risk in using only Logo environments for teaching

296 313 During spring quarter, 1983, students enrolled in the informal geometry course for elementary teachers will register concurrently for a one credit course, Logo In Informal Geometry. This course will follow the content of the informal geometry course but will provide Logo environments to describe the content. The goals of the course will be as follows: 1. Learning to program in Logo; 2. Having students create Logo procedures which; a. construct regular polygons; b. construct components of triangles: angle bisectors, medians, perpendicular bisectors of sides, altitudes and the related "centers" of a triangle, incenter, centroid, circumcenter, and orthocenter; c. create tessellations of the plane using recursion; d. illustrate the isometries of translation and :votation; e. construct magnification images of plane figures, both enlargements and reductions using inputs; f. combine magnifications with isometries to produce similarity transformations. The addition of the Logo course will be evaluated in three ways: (1) student success in learning Logo and developing the procedures described above, (2) student success with informal geometry, and (3) student evaluation of the usefulness of the course for prospective elementary teachers. Results of the evaluation will be available from the authors in June 1983.

REFERENCES 1. Agenda for Action: Recommendations for School Mathematics for the 1980s. Reston, Virginia: National Council of Teachers of Mathematics, 1980. 2. Burger, William F. Graph Paper Geometry. Mathematics for the Middle Grades (5-9). Reston, Virginia: National Council of Teachers of Mathematics, 1982. 3. Burger, W. F., Jenkins, L., Moore, M. L., Musser, G., and Smith, K. Teacher Education: A Coordinated Approach. The Arithmetic Teacher, March 1979. 4. van Hiele, P. M. Begrip en Insicht (Under- standing and Insight). Dordrecht, The Netuerlands: Muusses'Fnamerend, 1973. 5. Hoffer, Alan. Geometry is More Than Proof. The Oregon Mathematics Teacher, March 1979. 6. OiDaffer, }hares and Clemens, Stanley. Geometry: An Investigative Approach. Menlo Park, Califormta: Addison Wesley, 1976. 7. Harper, D. 0. and Stewart, J. H. (eds). Run: Computer Education. Monterey: Brooks/ Cole, 1983. 8. Papert, Seymour. Mindstorms:Children Computers and PaiWiTirrilii415FR:Basic Books, 1980. 9. Steen, L. A. and Albers, D. J. (eds). Teachin: Teachers, Teaching Students: eflect ons on Mathematical Education.Boston: tirkhauser, 1981.

297 314 A COMPUTER LITERACY CURRICULUM FOR PRESERVICE TEACHER EDUCATIONCANDIDATES

Brent E. Wholeben, Ph.D., E.M.T., Associate Professor

Department of Educational Administration and Supervision The University of Texas at El Paso (El Paso, Texas 79968)

With the advent of microcomputers in the classroom tion to the processes of the instructional to facilitate the ongoing instructional process, classroom. In addition, this same teacher must have the ability to apply the use of the computer animmediate need has simultaneously arisen for the training of the classroom teacher regarding for both CAIas wellas the management of that the valid and reliable use of this newest instruc- instruction (CMI). As school districts continue tional technology. For teacher education can- to increase their expenditures for the purchase of teachers must didates who are involved in their semester microcomputers for classroom use, internship,a unique opportunity exists for pro- fully unaerstand the potential application of this viding the necessary preservice computer literacy new instructional technology. training coincidentally with their practice teaching experience. As a formal course, computer The student has basic needs also related to literacy curricula under the auspices ofthe the employment of computerized instructional tech- College of Education at the university-level must nology. Although many needs address the issues be reviewed for approval by the appropriate state surrounding increased effectiveness, efficiency and satisfaction regarding study and learning of coordinating agency. Such review procedures are often of a year or more in length, yet the need appropriate curricula, other needs require atten- for such training continues to grow. To meet the tion apart from theschool environment itself. need for formal instruction in the use of computer Students exist in a technologically-oriented technology within the classroom, alternate methods world, and therefore must be helped to cope ade- must be developed for beginning teacher training. quately with such demands as will be placed upon them by a technological society. Such objectives as understanding the potential of computers(both positive andnegative),how computerizedtech- The Emergent Learning Technologist nology is currently utilized to benefit mankind, and where computers could be utilized to invade privacy rights of the individual -- can only come With the advent of themicrocomputer for computer-assisted instruction (CAI), the beginning from understanding the computer itself. classroom teacher has available the most formida- ble ally for the conduct of teaching since the printing press. The computer has been shown to be Training at the Undergraduate Level capable of selectively facilitating the instructional efforts of the teacher in such a way as to Itis the mission of this paper, to examine provide greatly enhanced learning on the part of the potential for training the beginning classroom the student. But the use of CAIin producing teacher during the'classroom internship cycle at increased learning can only be made possible the undergraduate teacher education level of uni- through the careful training of the teacher-user. versity training. While the,role of an inservice paradigm willalways be necessary for continued The teacher today must therefore be viewed as on-the-job training of the classroom teacher; the a learning technologist -- a professionalwhose cost-effective rationale for including-coMputer. role must assume the structured intervention of and CAI/CMI literacy training at the preservice computerassisted technology within thelearning teacher education level exists in three areas. process of the child. To accomplish a successful intervention however, a careful process of First,preservice training at the undergra- training must ensue for instructing the teacher duate, teacher education level provides systematic how to best utilize CAI, for whom, when, and why. quality controlrelated to the information pre- To accomplish these several missions, the needs of sented regarding instructional technology the teacher and student alike must be addressed literacy. Secondly, the opportunity for providing and understood. sufficient time for the structured implementation of required learning activities exists during the The beginning classroom teacher must have a semester of the student-teacher internship. Lastly, this same time will provide the necessary basic understanding (or general literacy knowledge) of the microcomputer, and its applica- opportunities for the preservice teacher to prac-

298 315 tice various CAI approaches whileina regular withthe employment of instructionaltechnology classroom (under the supervision of the regular within the classroom and learning process); and, teacher) andto report on thevarying success experienced -- promoting optimal information 3.0 The CAI Incorporation Component (a three-day profeTiTORT workshop of 12-sessions of 1k-hours transfer to the student teacher. each, during which participants are afforded opportunities to explore the use of CAI/CMI software in their classroomsas first-year teachers (the Rationale for a Variable Instructional Sequence following semester or year), to learn techniques in the design of specialized lesson plans for CAI As will be presented in the next section, the applications, to understand the basic rudiments of proposed presentation of the literacy objectives computer programming for specialized applications and activities occurs in a variable time sequence; in the classroom, to interact with colleagues of that is,activities and meeting times will vary like grade-leveland discipline areas concerning depending upon the material to be presented, and the incorporation of CAI within the classroom pro- when the actual presentation will take place. cess, and to researchwith local vendorsthe Four issues of rationale are apparent in requiring potential of microcomputers for facilitating the such a variable instructional format. instructional process.

First, the literacy (instructional) objectives and activities arethemselves of variable-time Component A: Awareness format, requiringsignificantly differentdura- tions to effectively present the material. The first component in the series of computer Secondly, these same objectives and activities are literacy instruction for thepreservice teacher linearly sequential. Therefore, rearrangement of education candidate is intended asa first-stage activity sequence is precluded. Thirdly, the suc- introduction to microcomputers and their potential cessful assimilation of the material requires for usein the classroom, thereby providing the appropriate time lapses between activity initialfcundation for awareness development on presentation(s). Lastly, these same time lapses the part of thefuture beginning teacher. A permit ample opportunity for student teacher prac- sample scheduling of activities for this four-hour tice and curricular development intervention(s) block of instruction exists as follows: during their classroom teaching activities. Activity 1 (0:30) Introduction to Microcomputers, and Their Potential Utilization Components of the Semester Curriculum within the Schools;

The implementation of the computer literacy Activity 2 (0:45) Introduction to the Components curricular program for teacher education candi- of Microcomputer Hardware and dates will exist over a single semester, prefera- Software, and Their Termin- bly the semester within which the candidate is ology; performing the required internship (practice teaching) experience. Activity 3 (0:30) < audio-visual presentation >

The components of the semester-length program (0:15) ** break ** exist in three parts, each representing a serial stage in the development of awareness, skills and Activity 4 (1:30) Demonstration(s) of Appropriate personal techniques concerning the utilization of Microcomputer Hardware and computers for instruction. These three parts are Software (round-robin format of defined as: 15-minutes pereachstation), and including such topics as: 1.0 The Awareness Component (a single session, programming, simulation, CAI four-hour sequence of activities designed to in- packages, CMI programs, and troduce the participant to the idea of microcom- color graphics capabilities; puting,andits potential for classroom instruc- tionutilization; and toprovide a structured, Activity 5(0:30) Question/Answer Period (small first-stage hands-on experience for each partici- or large group), and Post- pant); Session(s) Evaluation and Feed- back. 2.0 The Developmental Skills Component (a four- session sequence of two-hours each, during which the participant'is provided an in-depth coverage Component B: Skill Development of those topics initially introduced within the preceding 'awareness' session, including such This second component in the series of com- topics as: CAI/CMI intervention techniques, pro- puter literacy topics for preservice teacher edu- cedures in theassessmentand selection of in- cation candidates exists as a direct extension of structionalsoftware and hardware, demonstration those topics covered during the first (awareness) of various curricular oriented CAI packages from component. available vendors, and responsibilities associated

299 316 each demonstration During the skill development component, four structured at the end of individual sessions of two-hours each are employed sequence. to providegreater :J-depthtreatment of those Conducted as single The second day is concerned mainly with the topics introduced previously. implementing a seminar format over the specifics of planning, designing and eveningsessions in These sampletopics for these CAI strategies for classroom application. course of four-weeks, how to delineate four-on-two survey sessions will include: sessions include such topics as: curriculum for CAI-intervention structuring, how to develop the CAI-oriented lessonplan, and how 1.0 INTRODUCTION TO MICROCOMPUTER APPLICATIONS populations and/or FOR THE CLASSROOM to apply CAI to special interest groups. 1.1 Data-Processing via the Microcomputer The third day primarily focuses upon the content and process associated with the evaluation of 1.2 Computing Needs Assessment for Classroom microcomputer software and hardware for matching Activities the intended goals of instructional use. Special interest groups allow the further sharing of spe- 2.0 INTRODUCTION TO MICROCOMPUTER HARDWARE provides a basis for DESIGNS AND CONFIGURATIONS cialized information, and evaluating the perceived benefits of the three-day Finally, vendor demonstrations and pre- 2.1 Componentsof Microcomputer Hardware sequence. Design sentations at the conclusion of the incorporation phaseprovide the studentswith the basisfor designed their equipment need requests for the 2.2 Operation of the Microcomputer Hardware System upcoming school year. sampleoutline ofthis three-dayincor- 3.0INTRODUCTION TO MICROCOMPUTER SOFTWARE A APPLICATIONS FOR TEACHING poration component follows:

3.1Components of Microcomputer Software for Classroom Use Day-1/Ses-1 Presentation: Design of CAI-Oriented Instruction 3.2 Demonstration of Instructional Software as a Classroom Teaching Strategy; Applications Suggested Presentation Topics: 4.0 INTRODUCTION TO SCHOOL MICROCOMPUTER six-stagesof computer-assisted SELECTION AND ACQUISITION A. instructional intervention three-levels of computer-managed 4.1 Preparation of the Decision Matrix for B. Hardware and Software instructional monitoring Microcomputer cross-reference matrix for CAI Selection C. and CMI development 4.2 Solicitation and Validation Strategies Individual Simulations: for Area Vendors Ses-2 Students interact with suitable software and participate in Component C: CAI Incorporation packages; small-group, follow-up discussion; The third and final component of the computer literacy program for preservice teacher education Ses-3 Presentation: candidatesinvolvesa three-day sequenceof12 Design ofCAI Programs-- What to sessions of 11/2-hours each. This phase of computer Expect and What to Require; literacy training is designed to acquaintthe future beginning teacher with the skillsand procedures necessary to actually utilize computer- Suggested Presentation Topics: student-user orientation assisted instruction within their particular A. menu-driven routines discipline at the classroom level. B. C. compatibility, ofinstructional The incorporation stage of computerliteracy objectives training commences with a full day of intensive D. testing and progress monitoring lectures on the use and design of CAI/CMI-oriented Ses-4 Individual Simulations: instruction. In addition, two sessions allow suitable opportunities for students to acquaint continue their exploration themselves with instructional software specificto Students of suitable software packages related their teaching curriculum. Group activities for to their specialized needs; and par- the sharing of information (reaction to thesoft- ticipate in small-group discussion; ware, suitability of thepackages for instruction at the particular grade level, and thepotential for CAI intervention in specificinstances) are

300 31 ? Day-2/Ses-1 Presentation: C. foursteps in approaching the software/hardware compatibility Techniquesin Applying CAIPackages assessment for Instructional Facilitation; Ses-2 Special Interest Group Sessions: Suggested Presentation Topics: A. individual student versus small- Students meet by grade-level and group orientation curricular discipline areas to discuss B. tutorial versus remedial versus common interests and needs related to enrichment modes CAI application; and a formal evaluationof thethree-day sequence is Ses-2 Presentation: conducted;

How to Delineate Curriculum for Ses-3Vendor Demonstration of Software: Identifying CAI-Oriented Potential; !round-robin scheduling); Suggested Presentation Topics: A. clarificationofneeds versus Ses-4Vendor Demonstration of Hardware: desires B. establishment of concept mission, (round-robin scheduling). instructional goals, classroom activities, and student tasks 'Scheduling of the Semester Program Practicum Laboratory: Although differences exist in terms of time- Studentsare afforded theimmediate availability at various institutions regarding opportunity to simulate curricular their instructional sequence (semester,quarter, delineation relative to their trimester),little modification to the computer specialized areas; literacy program presented would be required.

Ses-3 Individual Simulations: It is suggested however, that the program be implemented as early as possible during the Studentscontinue their exploration 'internship semester'-- preferably prior to the of instructional software, research- mid-term period. Such early intervention provides ing software suitability based upon maximum opportunities for the student-teacher to their recently delineated curricular employ the full range of skills learned during the objectives; preservice training sessions.

Ses-4 Presentation: It isalso suggested that skill laboratories be scheduled during the second part of the How to Design and Implement the CAI- 'semester period'. Suchsessions willprovide Applications Lesson Plan; opportunities for preservice teachers to address and rectify problems which have developed during Suggested Presentation Topics: their initial CAI incorporation efforts within the A. considerations of equipment and classroom. material availability B. considerations of time, facilities, and varying student needs

Practicum Laboratory:

Students are afforded the opportunity to design a simulated lesson plan based upon theirknowledgeof CAI programs previously explored during the individual simulation sessions;

Day-3/Ses-1 Presentation:

Special Techniques for the Evaluation of Instructional Software and Compat- ible Machine Hardware;

Suggested Presentation Topics: A. content and process criterion references for software quality B. process andtooling criterion references for hardware quality

301 318 DYNAMICS OF LEARNING AND MISLEARNING IN A SIMULATED MICRO-WORLD

Andrea L. Petitto & James A. Levin

Graduate School of Education and Human Developement University of Rochester Rochester, New York 14627

Center for Human Information Processing Unvisity of California, San Diego La Jolla, CA 92093

The main idea behind the use of such simulations Abstract is that they embody general principles and relationships in the material to be learned and This paper presents an overview of research on permit a kind of exploratory behavior that is not children's learning processes in a computer usually possible when the same material is implemented micro-environment. A class of fourth presented in standard, expository writtei or and fifth grade children played a set of "shark spoken formats. Students explored the properties shooting" games as part of their regular school of this Newtonian object by manipulating it and activities for a three month period. The games developing strategies to manage it. Exploration required the estimation of numerical values on is not simply a matter of trial and error, as number lines, and the coordination of vertical might be encountered in drill and practice. In a and horizontal dimensions. Observations are made computer implemented simulation feedback from an concerning variations in the development of game error or an exploratory attempt is informative. skills, and transfer of number concepts to That is, it not only provides information about non-computer activities. Discussion focuses on accuracy, but gives additional information transcript analyses revealing specifics of about relationships embedded in the micro- learning processes during play. Interactions environment. among cognitive skills, game features, and the role of goals in structuring conceptual The type of simulation we will be concerned development are discussed. with here takes the form of an educational game. Games such as these usually simulate micro- environments which incorporate a goal. In the dynaturtle game, for example, the Newtonian turtle is to be guided to a specific "port", or in another version, around a circular track Introduction without crashing. The point of introducing game versions of simulated environments is that they At the Laboratory for Comparative Human require players to develop and sharpen cognitive Cognition at the University of California in San Diego, several of us have been investigating skills in the service of attaining the game goal. Because of the inclusion of goals internal to the cognitive implications of the use of micro- system, games are more able to stand alone than computers in elementary classrooms. Someof this work has concentrated on the use of computers as are pure simulations which students manipulate to achieve externally derived academic goals. The learning environments. In order to understand how inclusion of internally defined goals also learning occurs in these environments, we have affords the opportunity for an internal tutor found it necessary to look closely at the details function. Since the program can assume that the of the interactions among the children, the goal of the player is the internally defined goal various helpful adults in the room, and the of the game, it can monitor the effectiveness of computer itself. In this paper we describe a the player's performance.This allows the study of learning in a simulated micro- program to recognize and respond to errors by environment in which we take a close look at the offering hints, and can alter game parameters to qualitative differences in performance between adjust automatically to different levels of the most and least successful players. skill.

A well known example of a simulated micro- The Shark Games environment is diSessa's dynaturtle which behaves strictly according to Newton's laws of motion. Working on the notion of exploratory activity, several of us have developed a family of three This research was supported by the National estimation games called the "Shark Games". The Science Foundation, Research in Science Education games are intended to strengthen children's Grant SED-8112645. Many thanks go to Robert Rowe, knowledge of numerical relationships. Earlier Marcia Boruta, Karen Johnson, Jose Vasconcellos work by us and by other researchers had found and Dan Rieswig for their help and support. that fundamental concepts of numerical rela-

302

31,0 tionships are often weak among poor achievers in levels involve two coordinates. Difficulty on the arithmeticl. We hoped that these games could two axis games is varied by reducing the size of strengthen these fundamental concepts directly, the shark with respect to the numerical span of and that this would in turn affect classroom the axis, thus requiring progressively more arithmetic skills. accurate "thows". Movement to higher or lower levels on successive games is automatic, All the Shark Games utilize the same game contingent upon rate of success. world - sailing on the high seas, hunting down and harpooning sharks. All three games require The Study estimation along two dimensions marked by numerical scales (see Figure 1). All three of the Our original purpose in carrying out the study games were used in our research, though for the was to examine the possibilty of transfer of sake of brevity only one, called "Sonar", will be game-related skills in numerical estimation to described here. In Sonar, the player is sees a other kinds of numerical manipulations, pair of coordinate lines - one horizontal and the particularly paper and pencil tests of number other vertical - on the screen display. The line skills and classroom arithmetic. It is location of an invisible (underwater) shark is impossible to overlook the variation in academic indicated by a pair of numerical coordinates achievement that is typically found in most written at the top of the screen. Using these elementary school classrooms. For this reasons, coordinates, the player must estimate the shark's we were also interested in finding out how position visually. Each dimension is dealt with learning to play the simulation-games themselves in turn, first the horizontal estimate (labeled interacts with these differences in academic "aim") then after the aim is set the vertical skill. Thus we were looking for a two-way effect, dimension (labeled "distance") is entered. In game skills transfering to classroom performance, effect, each of the coordinate lines serves a and academic skills affecting the ability to dual purpose, as an axis on one dimension and as become skillful in the game. an indicator marking a numerical selection on the other dimension. The player uses game-paddles Two structural aspects of these games were (though keyboard entry is a possible option) to particularly interesting to us from a theoretical move indicator lines indicating the position he perspective. One is the inherent requirement for thinks is specified by the numerical information. successive approximations2, and the other is the The paddle button (or RETURN) is pressed to "set" introduction of cartesian coordinates requiring each numerical estimate once it is made. To the coordination of two dimensions. Strategies avoid the confusion that can arise from the dual which use successive approximation are encouraged role of each axis line, while each dimension is by the game setup which allows for multiple tries being estimated, the endpoint numbers and label (or throws) with informative feedback on the other axis line (now acting as an accompanying each miss. The coordination of indicator) disappear. estimates with respect to the two axis system was accomplished through a social coordination The "throw" of the Harpoon, "hits" and between two players. Children usually played "misses" all are represented in an interesting these games in pairs, one child of each pair graphic display with sound accompaniment. Once playing one of the dimensions. We were interested both dimensions are entered, a "harpoon" moves in the way that the social coordination of action from the bottom of the screen to the point might serve as a basis for spatial coordination specified by the intersection of the horizontal on a two-dimensional plane. and vertical estimates. Feedback is in the form of a "splash", visually specifying the actual With all this in mind, we placed the Shark location of each throw. The splash is labeled games in a combined fourth-fifth grade classroom with its actual numerical coordinates and remains where they were used in a computer-based "center" visible on the screen as an additional point of scheduled to be visited by specific pairs of reference through the next several tries. children each day. In this way, all the children in the class were able to work with the Shark The games also perform a tutor function. When games on a regular basis, amounting to about one a player's shot misses the shark, verbal and half-hour per child per week over a period of directional hints are written at the top of the threemonths. The Sonar game was first introduced screen. When a player's estimate is too high, the into the classroom with a range of 0 to 100 on word "smaller" with an arrow pointing in the each axis. direction of lower numbers (<=.) for that dimension. If the estimate is too low, the screen At the first session, and at several important displayes "bigger" with an arrow pointing in the transitions throughout the study, an direction of higher numbers (=>) for that observer/helper attended the children's shark dimension. game sessions. As the term implies, the helper/observer had a dual role. One role was to To adjust for variations in skill, the games record events as they occured during game play. all included multiple levels of difficulty such The other role was to help the children if they that the computer took over some functions at the encountered any major difficulties preventing easier levels. This was accomplished by creating them from effectively playing the games. The "beginner" levels in which only one axis specifications of the helping role were to keep (horizontal or vertical) is used, while higher the game going while interfering as little as

303 32u possible. A graded sequence of hinting procedures systematic in their initial approach to the was to be used when intervention was necessary. games. First, the observer/helper was to simply point out relevant information on the monitor screen. Transcript Analysis Then, if that were not enough, point out relevant relationships between game elements. And finally, In order to account for these data, we looked if all else failed make suggestions about what to at the interactions among the players, the observer/helper, and computer as recorded by our do. audio-tapes and field notes. We found that there Several kinds of data were collected on game were at least two major ways that game strategies play itself. The games had been set up to record differed between players ranked as good or poor all sequences, of keypresses, all game parameter in overall game achievement: the emphasis on values, and time information from an internal numerical judgements in making game decisions, clock. This data was taken on all games played and the ability to learn and manage the mechanics throughout the study. There were also the audio- of the game. Differences in the use of numerical tapes and field notes taken at several transi- strategies were striking. The better players used tional points during the study. The audio-tapes numerical specifications frequently as a way of recorded conversation and other sounds and could explaining thier own actions or specifying some But later be coordinated with the written field notes game related information to another player. and computer collected records. numerical judgements did not substantially enter into the remarks or game decisions of the poorer players. Results It is not surprising that players with poorer General Trends number concepts should use numerical information less than those with better understanding of Most children increased their skill at playing numerical relationships. But the mechanics of the the shark games themselves throughout this time shark games should be just as unfamiliar to all period, though there was considerable variability the children, regardless of numerical skills. among children in how well they played. Our Nevertheless, the poorer players also appeared to findings showed the classic pattern - those who have more difficulty adjusting to the mechanics played relatively well in the beginning also of playing the game itself - how to set up and both improved more than those who played more poorly execute the shots, coordinate the actions of on the first few attempts. Nevertheless, even players on the two-dimensional levels, and so on. with the skilled players, we could find no Could the lack of certain conceptual knowledge overall transfer of game skill to paper and interfere with the ability to learn the mechanics pencil tests which assessed number line and of a game in which that knowledge must be written arithmetic skills. employed? Or is there some other reason, perhaps some fundamental problem underlying both poor These findings were quite disappointing since mathematical ability and game learning ability? we had developed these games specifically tohelp the poorer students, and had hoped that game The data from this one study ran not provide derived skills would show transfer to some paper definitive answers to these questions. But an and pencil tests. We decided to look at the analysis of the details of the children's actions children's game playing in some detail to find in the games reveals some interesting relation- out what had actually happened. ships among cognitive skills, understanding of specific relationships among game elements, and A preliminary analysis has been done on the recognizing the goal of the game. keypress data for the initial session and on overall achievement within the game itself All the children in this study quite easily throughout the study. Overall game achievement recognized the main goal of the game - to shoot a was assessed as the percent of sessionsin which shark with the harpoon. There appeared to be no the players attained level 6 or better (out of a difficulty accepting the moving arrow as a possible 9 levels). This method of scoring harpoon or the idea of a hidden shark which, when overall game achievement resulted in scores for hit, appears momentarily as a small triangular 19 individual children ranging from 10 to 90%. We dorsal fin. The unanimous acceptance of this game then designated the 7 players who scored over 70% goal, however, masks a host of ambiguities which as "high", the 6 players with scoresbetween 40 only become apparent from the children's remarks and 69 as "intermediate", and the 6 below 30% as and questions addressed to each other and to the "low" in overall game achievement. (There were no observer/helper. Because the misconceptions scares between 30 and 40 percent.) This arising from these ambiguities prove to be presentation concentrates on the contrast between functionally related to the development of the high and low game players. We found that in cognitive skills, two representative examples of the introductory session low achievers overall these misconceptions are discussed here at some had spent more time on the single-dimension lower length. levels; used more "throws" per game to hit the Some Failures: Several children showed shark; and showed a much higher average deviation of throw values within each game. In general, the considerable difficulty even on the low level overall poorer game players appeared to be less single-dimension games. After completing seven

304 single-dimension games and in the middle of least on case, this became a rather pernicious multiple attempts to hit the shark during the bug in the child-computer interactions. The eighth game, one such child asked if the shark problem first showed up when one child in a pair were "roaming around in there". In this context, of players read aloud the screen instruction: the remark suggested that the player thought the "Set the aim, then push RETURN." The second child shark was a moving target, invisibly swimming then repeated the instruction in the following across the screen. Such a notion could derive form: "Then push RETURN? Push RETURN? Then I from repeated frustrated attempts to estimate the start the aim?" Though the first child and the shark's position, or It might be a misconception observer/helper manipulated the situation so that from the start. After all, sharks an not the aim was at least sometimes set before RETURN characteristically stationary, especially when was pressed, the "RETURN starts things" idea under fire. However it arises, the idea of a persisted into the higher, two-dimension levels moving target calls into question the player's of the game. This player would then press RETURN understanding of the relevance of the numerical to "start" her partner's turn, thereby firing the specification of the shark's position. If the harpoon before he could make any move to set his shark is moving, then the numerical information own aim. This short-circuited the possibility of is no longer relevant, and Sonar becomes a cooperative, coordinated activity and the guessing game in which the "smaller", "larger" children in this case (as in at least one other) hints provide the only clues. This pattern persisted in thinking they were playing essentially characterized the game playing of competitively against each other. The goal of several of the poorer players. Instead of using coordinating two dimensions was never numerical information to progressively narrow established. down the search for the shark (the process of successive approximations we were interested in), These examples are typical of the difficulties these players simply used the directional the poorer players had learning the mechanics of information from the hints without calibrating the Shark games. In some cases, specific miscon- the distances or coordinating successive tries in ceptions were quickly corrected by adult inter- any way at all. vention. In other cases, misconceptions persisted through several sessions and some were never Notice that this erroneous conceptualization resolved. These technical difficulties, though makes fundamental changes in the underlying they are not directly related to the concepts we goal of the game. Specifically, it makes wanted the children to learn, reorganized the irrelevant the major goal that the game is dynamics of the games so that they no longer intended to support: get as close as possible to addressed the relationships the programs were the position on the screen which corresponds to intended to embody. the numerical coordinates printed in the upper part of the display. When a player does not This line of reasoning suggests that technical recognize the goal of a game, no amount of misconceptions on the part of some players practice will bring him closer to achieving it. reduced the effectiveness of these games in The child has missed the point and is simply promoting the development of numerical concepts. playing a different game, one involving But evidence from the better players indicates different skills, strategies and goals than those that the causality implicit in the above intended by the game's designer. statement might also run the other way. Weak numerical concepts might have left the poorer There were many other technical pitfalls that players open to technical misconceptions by not lead to major misconceptions about the game and providing a coherent framework to guide learning its goals. One which proved to be a destructive of the game. factor as the game progressed roncerned the function of the RETURN key. Programmers and users Some Successes: In the introductory of software tend to have difering views of the sessions, as in subsequent ones, the better function of the RETURN (oi ENTER) key. From a players overtly and verbally refered to numerical user's point of view, r'ie RETURN key appears to information in all aspects of play: in aiming the initiate an action, tc, start something. Where harpoon themselves or in guiding each other's the programmer inte,-As the RETURN keypress to actions, "Seventy-eight is about here"; in taking tell the program to read a line of characters roles in two- dimensional play "You got just typed in, the user sees it as a keypress sixty-four", and so on. This often lead to that directly starts an action. The tacit successive approximation strategies. After two assumption on the part of a naive user is that misses, for example, one player remarked, the computer is reading along as information is refering explicitly to the value of the second entered at the keyboard or other peripheral input miss and tacitly to the first: "Twenty-one. We device. The RETURN key simply starts the next have to be right between here." machine function. The numerically specified position information In the Shark games, the introductory text and and the familiar (though tacit) numerical instructions end with the instruction: "Push relationships on the number line provided a RETURN to start a game", further reinforcing the coherent structure within which the technical idea of RETURN as "begin". Not surprisingly, aspects of the game could be worked out. That is, more than one of the children in this study possible ambiguities such as the function of the treated the RETURN key as an initiator. In at RETURN key and the nature of the shark's

305 322 movements are worked out in the service of It can be argued that these difficulties arose maintaining coherence within the numerical because of flaws in software design or that context. The process of working out the better tutor functions would more effectively technical details of the games usually went so monitor and guide a player's actions. Both of smoothly among the better players that it was these arguments are probably valid. However, we often difficult to detect transitions to serve as cannot assume that software - however well examples. A few observations can be cited here to designed - will ever be completely free of illustrate the point. either of these problems.

In one case, a player had begun by randomly When unfamiliar goals are introduced requiring aiming and shooting across the screen, apparently a new application of undeveloped or under- without regard for numerical information. After developed skills, it can not be assumed that a two low-level games in this mode, the novice player will apprehend all the relevant observer/helper simply pointed out that the relationships among game elements and correctly shark's position was numerically specified. The infer the goal or goals intended by the game's player's next aim was again wild, but this time designers, however clearly these goals appear to she verbally predicted it's numerical value, be presented. If a player infers a wrong goal, saying "That's probably gonna be two." This then feedback from an autos stic tutor is not prediction signaled her first attempt to likely to be effective. This feedback is only coordinate position and numerical information, corrective with respect to those goals the game and her accuracy increased steadily with designers intended. But players would interpret subsequent tries. such feedback in terms of their own intentions. If a goal is misconstrued, then automatic tutor In another case, we discovered that a player functions are rarely adequate to correct it. had not clearly understood that the hidden shark Continued practice in the game rarely corrects was not a moving but a stable target until his this misunderstanding and usually causes it to be ninth game. Though he had participated effect- more deeply entrenched. At the same time, since ively in quite accurate and strategic play in all changes in game goals alter the dynamics of the these games, this boy had been setting his aim game itself, the skills and strategies that are according to the specified numerical informaton, being developed through practice may not be those not recognizing its relationship to the position which the designers of the game or the teacher of the hidden shark. During the ninth game, he intended. expressed surprise that the shark so often turned up where he had shot: "So wherever you hit the These principles might at least partly explain thing, it goes?" This player had also assumed the lack of transfer from game playing to paper that the shark was swimming invisibly around the and pencil analogs of number-line estimation screen, but in contrast to the previous case tasks. With inadequate supervision at entry into where the notion of a moving shark was so the game environment, children whose number line debilitating, this boy's playing had been skills were too poor to provide a way of recog- consistently strategic and effective. This was nizing the salient structures of the game did possible because he was using numerical not play the game as it was intended to be information to structure his activity before he played. They effectively rearranged the dynamics completely understood how all the game elements of it to avoid practicing those skills we wanted were interrelated. Ultimately, he discovered the them to develop. Children whose number concepts "stable shark" feature because his actions were were sufficiently strong to recognize the salient consistently guided by appropriate goals which relationship expressed numerically generally did were supported by his understanding of the very learn the game as it was intended, but did not numerical concepts that the game was designed to need to improve those skills very much to play. teach. At least not enough to be detected by pre- and post-tests. Discussion Conclusions Though this very preliminary analysis of our recently gathered data cannot give definitive The function of education is to transfer to evidence about learning processes in interactive children socially organized and formalized media, several important points can be made. knowledge in the case of this study, the number First, stand-alone simulation-games cannot be system and the arithmetic and coordinate systems considered a panacea for the problem of remedia- based on it. Thus far, our analysis is able to tion of low achieving children. The specific show that when these knowledge systems form the game-goal of shooting sharks was not by itself basis for a simulation ur simulation-game, enough to organize behavior in ways that would players must rely either upon prior familiarity lead to development of the intended skill. In with that system, or external guidance to order to insure that the players interpreted the discover the relevant parameters for manipula- game the way the game designers had intended, tion. We expect to show from further analysis and some adult intervention was necessary even for in future research efforts that once this basic the best players. Initial misconceptions are understanding is established exploratory activity difficult to alter once they are established3, can be productive. and can be perniciously detrimental. Beyond this, our observations have implica-

306 3 2 tions for educational practice. They suggest that it is important to consider the role of the Sonar : AIM = -22 Right on! teacher when investigating the dynamics of child- Smaller .1 computer interactions. The use of computer-based Reading : DISTANCE = -57 media might provide a new role for teachers. With Set the DISTANCE, then push RETURN computers, parameters and rules for manipulation 100 are internal to the machine. They are not unam- biguous, and novices - either children or older students - need help to learn to manipulate them effectively. Because of this, the teacher can become an ally and helper to the student, a role which contrasts to the usual one of task master and judge. Future research should investigate the potential of this teacher-student-computer interaction system in which the teacher's competence 4:Ai -17 serves as a resource for students ia problem solving situations.

References

1. Petitto, A. L. "Developmental study of arithmetic competence among children with I-100 DISTANCE school related learning difficulties", unpublished working paper, Laboratory of Comparative Human Cognition, Univeristy of California, San Diego, 1982. igur.

2. Petitto, A. L. "Long division of labor: in This is a typical screen disp'ay in Sonar. The support of an interactive learning theory", numbers at the center top of the screen indicate unpublished manuscript, Graduate School of that the shark is hidden at -22 on the AIM Education and Human Development, University of (horizontal) line and -57 on the DISTANCE Rochester, Rochester, NY, 1983. (vertical) line. The player has already made one shot which has left a "splash" (ragged concentric 3. Miyake, Constructive Interaction, CHIP ovals) at -17 on the AIM dimension and -15 on report #113, ONR report #8206, Center for DISTANCE. Verbal hints appear to the right side Human Information Processing, University of of the top of the screen. "Right on!" indicates California, San Diego, 1982. that the AIM estimate (-17 showing in the "splash") was close enough. "Smaller" indicates that the DISTANCE estimate (-15 showing in the splash) needs to be revised downward. The downward pointing arrow indicates the direction that the indicator line must be moved.

Note that the AIM line is not labeled here. This is because the player is in the process of selecting an estimate on the DISTANCE dimension and the AIM line is for the moment acting as an indicator rather than a numberline.

307 324 Observation and Inference - A Computer Based Learning Module

by Alfred Bork and David Trowbridge Educational Technology Center, University of California, Irvine

Arnold Arons Department of Physics, University of Washington

Abstract present paper. This program has been tested with a very wide range of students, in several This paper reviews a computer dialog to different environments, as will be discussed. teach the distinction between observation and inference. It is a self-contained program Environment designed to work with a wide range of students. This module is about 15 to 20 minutes long An important distinction in undertaking the for the average student. It is embedded in a, nature of scientific knowledge is that between longer program called "Spacelab." The longer what is observed, seen directly, and that which program presents a fantasy about space travel in is inferred from the observed evidence. While which the student is the captain of a starship this distinction is, like all human distinctions, which comes across another, derelict, starship not absolute, it is nevertheless extremely useful carrying some supposed "energy crystals." in understanding the nature of scientific Measurements of mass, volume and other properties information. It might also be considered an are made on these energy crystals to try to imottant intellectual tool, one we want to bring decide which ones'might be a possible source of to students at as young an age as possible in energy (Fig. 1). Students are introduced to the order to enhance intellectual development. concept of density and use this idea to search for crystals consisting of the same material. 3xperience shows, however, that many From the standpoint of Piagetian tasks, the students at all levels, have difficulties making the distinction between what is seen, and what is reasoned. Even at the college level many students do poorly on examples of this type. Some of the cremtionist literature furnishes striking examples of fuzzy distinctions, Elo cziA 0 AM IF ND INN exhibiting a failure to distinguish observation NM Ni 111, It 21 II 13 21 from inference.

The same kind of distinction arises in other disciplines. In studying history, for example, it is necessary to distinguish between primary information or evidence on the one hand and interpretations or inferences drawn from such material by the historian on the other.

The program to be described here is a computer based learning module, intended for tress11" ter eNuM er 1" ter reersug a blest. students from about 12 yeaers of age or over, Press I[11101 do tie seta istitMime. concerning the distinction between observation and inference. It involves a variety of situations to illustrate and establish the Figure 1 distinction. This project was funded by the National Science roundation, through the A weighing experiment in the Spacelab dialog. Development in Science Education Program. It is primarily concerned with aiding students in early exercise is principally concerned with ratio adolescence, about the age of 12, to develop reasoning, involving the ratio mass/volume, but intellectual skills which are important for later an opportunity arises to lead the student into life. Although initially the primary focus was making the distinction between what is observed on various standard Piagetian tasks, we have also and what is reasoned. considered other exercises in abstract logical reasoning such as the one presented in the The observation-inference module comes as

308

323 the last sequence in "Spacelab," but it can also When the examples in the passage are be used independently of the larger program in exhausted, we finally introduce the words which it is embedded. In fact, we have two "observation and inference" (Fig. 3). In much of versions of the module, one in which there are our materials at Irvine we have taken taken care reference to the "parent" Spacelab program and to avoid using a technical term until the idea it one in which there are no such references, so denotes has been .-.,stablished operationally that the module can be used alone. through specific examples and shared experience.

Program Outline

The program begins with a sequence from a >4[f tlTdlMi Sherlock Holmes novel,The Greek Interpreter," la is nano, blest lest mew deer by Arthur Conan Doyle. There are many passages miles pastilles Megeve sae Footage is a rattle ltes hisetwinfe in the Sherlock Holmes novel in which Holmes, or use pastas is a metre beet Ne MAW wog Two in this case, his brother Mycroft, make a series Oft IteI Ia shtleirth of startling deductions about a person or has asother And situation based on what appear to be relatively few, apparently insignificant, observations.We have picked one of these passages to begin the III the thisap Ia the first alum III the dams la tie NNW seise an site dinette. we nivel beGL' current program (Fig. 2). The entire passage is Meg ere armee_ at g rump. first presented to the student in an attractive free the 'Mutations. Is all thee kallthee 11151311TIN6. AtNOM Its eisoplete elmeing Mess that be boo loot seam opt ewe. Tbe fast Not heIsbetas his we Mesas late as thoughItwe his elles Please wen spat Is Inbees bows, this. for WIPP. bars ...bort is a rattles shish shoos that en it them is wry pose. Figure 3 Tbe gift probable IWO laMile Tbe fast that be Inplaten bestsirIts ire Introduction of terms, Observation and Inference. shunthatthere is astir ohne to be the* of. The next sequence in the program concerns counting the elapsed time interval between lightning and thunder and calculating the distance between the observer and the lightning strike, the student having to answer questions about observation and inference in this Flame piss sautebars connection. This sequence, however, is not needed by all users. If someone shows no difficulty going through the first "Sherlock Figure 2 Holmes" sequence, then the lightning and thunder activities are bypassed. Excerpt from a Sherlock Holmes story.

way and with some associated visual information. Then the program analyzes the passage, classifying for the student several examples of what is seen directly and what is reasoned out. In each case, this is done by using a blinking box surrounding a phrase or sentence of text, and then classifying what is in the blinking box. At this point the computer is being used in an expository manner, but this is only a short episode. Note that we are not using the "technical" terms, observation and inference, as yet.

The second activity in the Sherlock Holmes sequence requires the student to make his own classification of items still remaining in the passage. The blinking box is still used to set Please tress stage bars off a phrase or sentence, but now the student must decide, on the basis of the earlier example, whether each item refers to something that Mycroft sees or to something which Mycroft Figure 4 reasons out. If mistakes are made, they are corrected. Graphics for activity on counting grasshoppers.

309 323 1 weer

1 kter

thatstatt gas It rant t

lee Is a elese q glee of Ile rater spare. Sat& It for a 'bile, aal til la attests bee easel risibenes there artta taw fetid asks get beuse pasOieWstare art ta 1k dale MIL It is ugessile t mat nenpask, bat pa eat to title wetMod pt a mg% tin about In sag thanere.

Press epee bar eke pa are reed'.

Figure 5 Figure 6

Opening question for grasshopper problem. Exercise on counting a fluctuating number of grasshoppers. The next component of the program involves statistical inference. The student is presented to stay at the display. with a large field containing many jumping grasshoppers (Figs. 4-5). The student is told This program has gone through a number of that it is impossible to count all the revision stages after such testing. It now works grasshoppers directly, and he is invited to well with a wide group of individuals.We are suggest alternate strategies for getting some also using it as one of the programs in another idea of how many grasshoppers there are in the research program involving the behaviour of field. Thus approximation and sampling come into groups in using computer based learning material. this activity. The student is led, though a In this case we videotape groups employing the series of questions, to the idea of counting the material, capturing their conversation, emotions, grasshoppers in a one meter square. The students and key pushes at the computer. This, too, has then count directly, watching the one meter been useful in understanding how the program square with the jumping grasshoppers (Fig. 6). works and in leading to additional revisions. The number of grasshoppers in the square fluctuates slightly. The program checks Plans their value and gets them to count again if they are far off. Students then determine the total The program currently runs on its area of the field, and obtain an estimate of the developmental hardware, the Terak 8510/a. The number of grasshoppers in the entire field. choice of delivery hardware will be made by the Again, questions are asked about what is an Educational Technology Center and the observation and what is an inference in this distributor. We are currently conducting oquenens discussions with several companies about the possibility of having these materials available There is a final sequence providing further commercially. exercises in discriminating between observation and inference.We also show the student that not At the moment, users with Terak 8510/a's everything we do can be classified as either an can obtain copies of the program at our cost. An observation or inference.For example, defining order form is available from the authors. a concept or inventing a name do not fall in Currently the program is not availaW.e on other either category. personal computers, although segments run on the IBM Personal Computer in connection with another Observations of Usage program.

The modules developed under this grant, and under a FIPSE grant concerned with public understanding of science, have all been tested in a variety of environments.The principal testing environment has been public libraries.We have found this to be a very useful environment, not only to identify conceptual weaknesses in the programs, but also to be sure the programs are motivationally strong, holding viewers even in a library environment where there is no pressure

310 327 DOES USE OF MICROCOMPUTERS IN JUNIOR HIGH SCHOOL INCREASE PROBLEM SOLVING SKILLS

Barbara Kurshan Joyce Williams Nancy Healy

Hollins College Hollins College, VA 24020

inevitable impact on society. However, if this study and future research can give some evidence ABSTRACT to the increased problem solving ability that students gain from "doing" then the issues con- This Is a study to determine if the use of the cerning the microcomputer in the classroom will microcomputer increases problem solving ability perhaps have a central focal point. This study of seventh grade students.Two seventh grade is an initial attempt to begin to form that base classes from similar schools in Roanoke, Vir- of data to answer the question "Does the use of ginia were selected for the study.The first group the microcomputer for learning increase problem- was exposed to introductory computer literacy/ solving ability?" computer programming activities for an entire Background year in a microcomputer lab. The second group did not have a computer lab in the school. The The emphasis of the research concerning the preliminary results indicate that students effect of classroom computer use on student per- exposed to computers show increased problem formance has focused on comparing CAI (Computer- solving ability. Hopefully, this study will assisted instruction) to traditional instruction encourage others to explore the benefits of the and the effects of computer programming.by computer as a tool for increasing problem students on problem solving skills. The major solving skills. emphasis of research concerning general computer use has investigated the effect of CAI use on achievement when comparedo traditional instruction (see for review Dence E dwards, Norton, Weiss and Dusseldorp3Forman' 4; Jamisoni,Suppes and Wellsl° and Kulik, Kulik and Cohen"). Re- search that has investigated the effect of CAI and traditional instruction has generally shown Introduction that the combination of CAI and traditional instruction is the most effective and requires "Children learn by doing and by thinking about less instructional time than traditional instruc- what they do" (Papert, 1980, p.161). This process tion. of doing and thinking should ultimately provide children with the ability to solve problems. The The reviewers of the CAI vs traditional instruc- introduction of the microcomputer into the learning tion studies generally support the effectiveness of classroom computer use. process enhances the "doing" area of learning. It Edwards, Norton, gives children the "power" to view and solve Taylor, and Dusseldorp3, for example concluded as exciting problems. The skills used to solve prob- a result of their study in elementary schools, lems are perhaps enhanced by the use of the com- traditional instruction supplemented by computer puter. Students appear to be able to grasp the based instruction was more effective than tradi- true significance of broader problems.They are tional instruction along. Jamison's et al.10 able to apply algorithms that in the formal rather survey of the effect of CAI studies found that dis- than experimental setting are lost to the learner. advantaged elementary school students appeared to show the most achievement gain when using CAI. The assessment of the impact of microcomputer Further, Jamison's et al.1' concluded that CAI use on the problem solving abilities of students was most effective when used as a supplement to has diverse benefits.Educational decision makers regular instruction and that the instruction time can use the data for substantiating the need for was less. Edwards' et al.3 and Jamison's et al.i° zomputers in the school. Curriculum designers findings indicate that using CAI in conjunction could redesign portions of courses that teach prob- with regular classroom instruction improves lem solving and include a greater emphasis on achievement, that students take less time to learn computer learning. The classroom teacher will the material and students that are less academic- hopefully be more inclined to use the computer for ally prepared and from lower socioeconomic appear project design, creation and implementation. In to benefit more from computer use. general, the educational community certainly realizes and enthusiastically embraces the need Dence2, Forman4, and Kulik's et al.12 reviews for the computer in the classroom because of its present a current perspective of the state of

311 328 computer use on student achievement and identify effect of computer programming on students'problem outcomes of computer use that are very similar. solving behaviors indicate that general computer In accordance with the aforementioned researchers, use enhances students' achievement and problem Dence2 reported that CAI students that receive CAI solving performance. The present study while and traditional instruction obtain higher scores continuing in the same tradition as those pre- than those students who receive only CAI or only viously cited, differs on several specific features. traditional instruction. Dence also reported First, the treatment consists of only one group of that CAI students appear to have a greater re- students being exposed to the computer in math tention of material than students that were taught class. However, the students were not trained in only by the traditional method. Students that had a specific computer skill nor were they given spe- prior familiarity with CAI or subject matter cific CAI material to use. They were given a benefitted more when using CA/ and students with variety of computer experiences including games, initially low levels of achievement tend to show programming, CAI and simulations. The control greater test gains (pretest and posttest)2. group did not use computers in their math class. Similar conclusions were made by Formang. It The second feature of this study is data appears that regardless of the age of the student obtained from school records of eighth grade stu- or type of hardware used students tend to show dents in both schools. Third, problem solving be- improvement in achievement scores. Further, it haviors were assessed by students' performance on seems that students that have prior experience the problem solving subtest of the SRA Achievement with the computer or subject tend to benefit even Series test and the Hartman Test of Causal Reason- more from computer use. ing. Finally, prior familiarity was manipulated The general conclusion drawn from the litera- by assessing students' general use of computer through their ratings on the Prior Exposure to ture related to the issue of CAI vs traditional It should be instruction is that CAI, when used as a supple- Computers Index (Anderson, et al.1). noted that for the purposes of this study, general ment to traditional instruction, does produce computer exposure is defined as using CAI, gaming greater achievement. Further, there appear to be factors other than the delivery method that in- and simulation, computer programming, and experi- fluence achievement when CAI is used in schools. ence with arcade games. For example, the socioeconomic factor introduced The Prior Exposure to Computers Index was used by Jamison et al.1(); the prior familiaripffactor for the reasons of ecological validity; it was introduced by Dence2 and the ability factor, reasoned that while general computer use has been addressed by Jamison et al.1002 andFormed,: used to establish the power of a variable, valida- seem to influence the amount of achievement gain tion in a natural setting with students of similar and instruction time. SES and intellectual ranking, and the Prior Expos- ure to Computers Index would strengthen the con- Investigations of the effects of computer programming by students on problem solving skills clusion of the difference in performance being due to general computer use in the math classroom. (Johnson and Harding6, Milner13,Ronan14, Thus, the predictions based on the research are Wilkinson17, and Holoien9) have also been con- that: (1) students that are exposed to computers ducted. Research has shown that studenti learn in their math classroom, regardless of their prior the content better when they write and run their exposure to computers, will show significantly computer programs (Foster52 Johnson and Harding6, higher problem solvinc; scores on the problem Milner13, Holoien9, RonanI4, and Wilkinson17.) solving measures;(2) students that have had little Foster5,for example, investigated how students' or no prior exposure to computers and are exposed use of computers and flow charts effect their to computers in their math classroom will show problem solving ability. Sixty-eight eighth significantly higher problem solving scores on the graders were placed in four treatment conditions: (1) use neither computer nor flow chart; (2) use problem solving measures than no prior exposure students that did not use the computers. flow charts only; (3) use computers only and (4) use computers and flow chartss. Over a Procedure period of twelve weeks each student was provided The experimenters interviewed school officials with 24 tasks that required both computer and non- in the spring of 1982 to receive permission to pro- Foster found that the third computer solutions. ceed with the study and to receive advice on condition (i.e. only used computer) had significant choosing a control school. The experimental school mean differences on processing hypothesis, identi- was chosen by the fact that it was the only junior fying a pattern, and selecting relevant data.The high school in the city where all students within data also showed tha: group 2 and 4 performed a grade would use computers. The control school better than the grown that used neither the was chosen because it was more similar to the Milner13 and Ronan' s14 computer nor flow chart. experimental school than any other junior high findings are very similar to Foster's5. That is, school in the city. In the fall of 1982, the students that were taught programming and given principals, teachers, and guidance personnel at the problem solving tasks to perform showed greater two schools were involved in the study. achievement gains than those students.that did not use the computer. Standardized test scores and other relevant data were gathered from school records by the Description of Study experimenters and recorded on the Check-Off Sheet The research comparing CAI to traditional infor Computer Study. They were assisted struction and the research investigating the by two Hollins College student assistants.Data

312

3 2 includes scores for reading, math, problem-solving Math Portion of the SRA Achievement Series included and EAS - Educational Ability Series Tests. a subtest specifically aimed at the area of problem- The curriculum at Woodrow Wilson does not, at solving. the present time, include instruction using com- The Prior Exposure to Computers Index puter instruction if funds become available. (Anderson, et al.1) was used as the measure of Students at Breckinridge were offered the oppor- students previous exposure to computers. It was tunity to enroll in microcomputer/math classes taken from a study of affective and cognitive during the 1982-1983 school year. A follow-up effects of microcomputer based instruction. study of standardized test scores for the spring of 1983 is planned for both the experimental and The Hartman Test of Causal Reasoning control gro'Ips. was chosen as an additional measure of problem-solving ability. It is not a mathemat- Selection of Students ' ical test and requires more verbal reasoning skills Students selected for this study come from the than mathematical reasoning skills. It requires present eighth-grade classes at two of Roanoke 55 minutes of class time to administer. The City's six junior high schools, Breckinridge and instrument is a multiple-choice Test of Multivari- Woodrow Wilson. These two schools have similar able Causal-Logical Competence, constructed by socioeconomic profiles. Cheryl Hartman 7 to assess the four forms of reasoning (i.e. Form I: the ability to validly The population at both schools can be describ- induce an equivalence causal relationship, Form II: ed as middle-income but with some lower-income the ability to validly induce a multiple sufficiency and some higher-income families. Single family relationship, Form III: the ability to validly in- homes predominate the neighborhoods from which duce a multiple necessity relationship and Form VI: these schools draw students, but both school pop- the ability to validly induce that a data contra- ulations include students who live in federally- diction can only be resolved by invoking a hypo- funded housing projects. thetical variable; Forms IV and V of the model were The city's school population is 20% black and not studied due to pragmatic limitations.) almost 80% white. At Breckinridge, almost 23% of Method of Testing the students are black and 77% are white. At Woodrow Wilson, almost 26% of the students are The SRA tests were administered by homeroom black and almost 75% of the students are white. teachers within each school. Students took these Less than 1% of the students from either school tests in the spring of 1981 and 1982 as part of the belong to other ethnic or racial groups. regular school program. Teachers giving these tests received standardized instructions for administer- Both schools consider their student population ing them. very stable, with transient students not a problem. The ability levels and achievement levels of the The Hartman Test of Causal Reasoning will be total student population within the two schools administered to 30 students Irom each school popu- are similar to within several points on stan- lation. Students will be chosen randomly within dardized tests. male and female groups. These students will be tested in groups of five with the experimenters The experimental group consists of the students serving as the tesbas. This setting should approx- at Breckinridge Junior High School where computers imate the individual testing done in Hartman's were used in math classes.The control group con- research. sists of students at Woodrow Wilson Junior High School were there were no computers in classrooms. Analysis Students who transferred into or out of either T-tests were done on the groups (school, prior school were excluded from this study. exposure, and sex) the means of GainEAS scores, EASproblem solving scores and the Hartman Problem Selection of Tests Solving test scores were compared by group. The Several testing instruments were used in this probability level of (p<.10) was set. Even though study. To test for problem-solving ability, the this was an expofacto experiment, a 90% chance of problem-solving subtest of the Mathematics portion computer use in the classroom affecting students' of the SRA Achievement Series were used (SRA,1980). problem solving ability is important. Sixth-grade scores were used as a pre-test; seven- EAS Gain Score Analysis (GainEAS) th-grade scores were used as a post-test. An additional measure of problem-solving ability was The T-tests comparing the "GainEAS" by school the Hartman Test of Causal Reasoning, (Hartman7). and sex were not found to be significant at the To determine the level of prior exposure to com- .10 level. There was a notable difference in the puter use, questions were formulated from the "GainEAS" means scores in favor of Breckinridge. Anderson, Klassen, Hansen, and Johnsonl study. However, the small sample sizes and large variability between the scores may possibly have in- The SRA Achievement Series was used for several fluenced the findings of no significance (Table I). reasons. It has previously established levels of reliability and validity. Since students take There was a significant (p<.074) between the these tests every spring as part of the school's mean "GainEAS" score of low exposure Breckinridge regular testing program, they cost students no out- students (X=5.8983, 0=9.234, N=59) and the mean of-class time. They provided pre-test information "GainEAS" score of low exposure Woodrow Wilson which could not have been obtained otherwise. The students. These results indicate that the Breck-

313

3 3 inridge low prior exposure to computers, students Table II benefitted more, in terms of "GainEAS", than any GAINPS of the other students in the study. Group Mean SD T. Prob. Table I Schools GAINEAS Breckinridge 2.33 4.224 1.22 .225 Group Mean SD Prob, (N=90) Schools Woodrow 1.6111 3.714 Breckinridge 4.8280 8.872 1.26 .209 (N=90) (N=93) Sex Woodrow 3.3269 7.708 (N=104) Boys 1.875 3.890 0.86 .390 (N=80) Sex* Girls at Breck. 5.7736 8.617 1.21 .230 Girls 2.200 4.060 (N=53) (N=100) Girls at Woodrow 3.9153 7.544 Boys Breck. 2.5750 4.150 2.08 .041* (N -59) (N=40) Boys at Breck. 3.5750 9.156 .55 .587 Boys Woodrow .8000 3.436 (N -40) (N=40) Boys at Woodrow 2.556 7.936 Girls Breck. 2.140 4.314 -0.15 .883 (N -40) (N=50) Prior Exposure Girls Woodrow 2.2600 3.832 (N=50) High-Breck. 2.9706 7.998 -0.43 .672 (N=34) Prior Exposure High-Woodrow 3.8750 4.518 Low-Breck. 2.3519 4.274 .78 .423 (N=8) (N=54) Low-Breck. 5.8983 9.234 1.81 .074* Low-Woodrow 1.7976 3.757 (N=59) (N=84) Low-Woodrow 3.2813 7.930 High-Breck. 2.3056 4.208 3.50 .002* (N.96) (N=36) * Sex of some subjects wereunknown. High-Woodrow -1.000 1.549 (N=6) EAS Problem Solving Gain Score (GainPSZ Analysis Significant (p<.10) differences were not found Hartman Problem Solving Test Analysis (PSScore) in the mean "GainPS" scores of groups (school and prior exposure) (Table II).There was, however, a The data from the Hartman test analysis is not - significant difference (p<.002) shown between the complete. However, preliminary analysis show that "GainPS" in high exposure students at Breckinridge there is no significant difference between the per- (X=2.3056, SD -.208, N=36) and high exposure students formance of girls and boys on this test. The at Woodrow Wilson(X=-1.000, SD=1.549, N=6). The analysis does show an unexpected significant dif- high-exposure students at Breckinridge apparently feience between the mean test score of the two gained in problem solving skills over the year. schools. That is, it appears that the control This gain may possibly be attributed to these stu- school (Woodrow Wilson) students performed better dents use of the computer in the classroom. There on the Hartman test than the,Breckinridge students was also a significant (p<.04) difference in the who were exposed to computers in the classroom. problem solving gain scores between the boys at However, it is again noted that these data are in- Breckinridge (X=2.5950, SD=4.150, N=40) and the boys complete and inferences made from it are not reli- at Woodrow Wilson (X=.8000, SD=3.436, N=40). It is able. This test is being repeated for Spring, 1983 apparent that more boys at Breckinridge increased data. their EAS scores than boys at Woodrow Wilson. As Discussion mentioned earlier, the level of significance may have been influenced by the amount of variability The data support to a degree the beginning between the scores and the small sample sizes. assumption that students that have been exposed to Nonetheless, it appears that students that were computers in the classroom will show high problem exposed to computers in the classroom did show im- solving scores. The data even though incomplete provement in their problem solving scores on a indicates that high exposure students that have standardized measure. been exposed to computers in the classroom tend to increase their problem solving scores. There is no significant evidence that students that have had low exposure to computers and are exposed to computers in the classroom perform better on problem

314 331 solving measures. However, there is evidence that 10. Jamison, D., Suppes, P., & Wells, S.The students that have little prior exposure to com- effectiveness of alternative instructional puters and are exposed to computers in the class- media: A survey. Review of Educational room do improve their over all intelligence scores. Research, 1974, 44, 1-61. Perhaps upon completion of the Hartman analysis the picture of students problem solving performance 11. Johnson, D.C. Programmed learning: A com- will be clearer. parison of the school mathematic study group programmed and conventional textbooks Conclusion in elementary algebra (Doctoral dissertation, The use of computers in education will continue University of Minnesota, 1965). Dissertation to grow at an even more rapid rate than today. Abstracts Internationa, 1966, 26, 5294. Educators need to know that the computer is more 12. Kulik, J.A., Kulik, & Cohen, P.A. than a "fun" way to learn. If computer use does Effectiveness of Computer Based College increase problem solving ability then the designers Teaching. Educational Technology, 1981, of curriculae should incorporate this factor into 307-318. learning programs. However, it is difficult to satisfactorily design a program to teach problem 13. Milner, S.D. The effects of teaching computer solving skills. Therefore, if the computer can be programming on performance in mathematics. used and one of the inherent benefits, whether (Doctoral dissertation, University of Pitts- formally identified or imbedded in the "nature of burgh, 1972). Dissertation Abstracts Inter- the beast", is an increased problem solving ability, national, 1973, 33, 4183A. then learning with micros is certainly ideal. 14. Ronan, F.D. Study of the effectiveness of a computer when used as a teaching and learn- References ing tool in high school mathematics. (Doctoral dissertation, University of Michi- 1. Anderson, R.E., Klassen, D.L., Hansen, T.P., gan, 1970). Dissertation Abstracts Inter- & Johnson, D.C. The affective and cognitive national, 1971, 32, 1264A-1265A. effects of microcomputer based science 15. SRA Achievement Series, Forms 1 & 2. Scienre instruction. Educational Technology Systems Research Associates, Inc., 1980. 1981, 9, 329-35. 16. Taylor, Robert P., editor, The Computer in the 2. Dence, M. Toward defining a role for CAI: A School; Tutor, Tool, Tutee, Teachers College Review. Educational Technology, 1980, 20, Press, New York, 1980. 50-54. 17. Wilkinson, A. An analysis of the effect of 3. Edwards, J., Norton, S., Taylor, S., Weiss, M., instruction in electronic computer program- & Dusseldorp, R. How effective is CAI? A ming logic on mathematical reasoning ability review of the research. Educational Leader- (Doctoral dissertation, Lehigh University, ship, November 1975, 33, 147-153. 1972). Dissertation Abstracts International, 4. Forman, Denyse. Search of the literature. The 1973, 33, 4204A. Computing Teacher, 1982, 37-49.

5. Foster, T.E. The effect of computer programming on student problem solving behaviors in eighth-grade mathematics (Doctoral dissertation, University of Wisconsin, 1972). Dissertation Abstracts International, 1973, 33, 4239A.

6. Harding, R.D. Computer-aided teaching of applied mathematics. International Journal of Mathematical Education in Science and Technology. 1974, 5, 447-455.

7. Hartman, Cheryl W., The construction and empirical investigation of a model of multi- variable causal logical competence, Appendix, Roanoke, Va., 1982. 8. Hatfield, L.L., & Kieren, T.E. Computer-assisted problem solving in school mathematics. Journal for Research in Mathematics Edu- cation.1972, 3, 99-112.

9. Holoien, M.O. Calculus and computing: A comparative study of the effectiveness of computer programming as an aid in learning selected concepts in firnt-year calculus. (Doctoral dissertation, University of Minnesota, 1970) Dissertation Abstracts International, 1971, 31, 4490.

315 332. DIVERGENT ANSWERS TO THE QUESTION, "WHERE SHOULD COMPUTER EDUCATION DOLLARS BE SPENT?"

Arthur Luehrmann Computer Literacy

Eric F. B'irtis President, Centurion Industries, Inc.

Beverly Hunter Human Resources Research Organization

ANSWER 1:

Not on General Purpose Computers at the Elementary Level. All this emphasis on general purpose computers is taking elementary school dollars away from where they belong: teaching the basic skills of reading, writing, and arithmetic. What good are computer skills and knowledge to a kid who can't even spell or add? Save the money and invest it in effective systems for teaching basic skills. -- Eric F. Burtis.

ANSWER 2: On the Social and Ethical Issues of Computer Use. Before committing vast sums to training a generation of programmers, we should be certain that all students know how computer use, proper and improper, can affect the individual and the society. Mere technical expertise is not the answer. -- Beverly Hunter.

ANSWER 3: On Teaching a New Basic Skill: Computing. Learning to use a computer is learning a new way of writing and thinking. People who have this skill cansolve more complex problems than others, manage information better, and get better jobs. Schools should provide these new skills for the same reasons they teach other basic skills. -- Arthur Luehrmann.

316 333 An Evolving Model for Providing Computer Education for Gifted Children

Mary Crist, Chair The Ames Hill Center for Gifted Children Wilbraham and Monson Academy Wilbraham, MA 01095

NBSTRACT The Ames Hill Center for Gifted computing--problem definition and solution 2hildren, a part of the Wilbraham and through algorithm definition. The pros and Monson Academy, serves 340 students (ages cons of the programming languages we have 3-16) from the Greater Springfield experienced will be presented. metropolitan area in western Massachusetts, Observed characteristics, including and northern Connecticut through the individual learning styles of gifted Afterschool, The Saturday School, and two children in computing courses, will be summer sessions. described. Factors, such as the ability to Although there often appears to be work with a partner, tolerance for little agreement as to the true nature of frustration, and the needs for exploration giftedness, this presentation will discuss and limited structure will be examined. five specific characteristics of the gifted The presentation of case studies of child that appear to be critical to students in Logo will provide additional learning computer programming. information regardingprocesses used in Ccasideration will also be given to the problem solving and the relationship influences of I.Q. and chronological age. between (1) age and level of achievement, A challenge to educators is issued. and (2) selection of goals. Thecomputer education program at Ames The cost of microchips continues to Hill consists of four programming decline, and as it 'does, the economics languages. TheLogo language serves to dictate that a whole new set of tools will introduce students with no previous become available. We will examine these computer experience into the tools, including speech, speech non-threatening world of computers. BASIC, recognition, robotics, and others, and PASCAL, assembly language, and FORTH guide their anticipated impact upon the education students into creative areas of of gifted children.

Pat Semmes Department of Computer Science Trinity University San Antonio, TX 78284 Elaine Henshon The Ames Hill Center for Gifted Children Wilbraham and Monson Academy Wilbraham, MA 01095

317 334 Training University Faculty in the Use of Computer Graphics

Richard G. McGinnis Bucknell University

ABSTRACT In the first phase, the project director Interactive computer graphics is a met with the departments on campus to powerful means of presenting and explain the characteristics of computer complex data in visual form. manipulating graphics and 1,;(3 identify potential The technology of this field has grown applications of computer graphics within rapidly in thelast decade; and computer the various disciplines. Next, interested graphics is now widely used in the areas of faculty members submitted proposal engineering and science that require indicating how theywould like to use design, analysis, and the presentation of computer graphics within various digital data. Other, less well known disciplines. Next, interested faculty applications have been developed in members submittedproposals indicating how non-engineering disciplines: digital data they would like to use computer graphics in processing in linguistic studies, art and courses, and from these proposals software dance applications, computer-assisted needs and equipment needs were identified. musical score generation, demographic and The third phase, faculty training and geographic mapping, medical x-r:y image course development, occurred during the processing, and social science statistical summer when the twenty-two participants displays. were given intensive instruction, including Recent efforts to incorporate this an overview of computer graphics concepts, technology into the curricula of higher the use of computergraphics devices, education have been largely restricted to "hands on" experience with available the engineering disciplines. There is, software packages, and discussions of however, enormous potential for the appropriate ways to improve teaching application of computer graphics into other effectiveness through the use of computer curricula at both the undergraduate and graphic technology. Following the graduate levels. For example, in the instruction period, the faculty social sciences, computer mapping routines participants developed the teaching could be used to showthe distribution of materials necessary to integrate their' various demographic characteristics such as proposed graphics applications into their income levels, ethnic populations, contours courses. Implementationoccurred during of percents of political party the 1982-83 academic year. registrations, contours of the incidents of Case studies of two of the faculty various diseases, etc. In art, students participants are presented. The first is could use computer graphics packages to that of an English/Theatreprofessor who experiment with various shapes and shadings had no prior computer experience and who to produce differenttypes of drawings; wanted to use computer graphics as an aid while chemistry students could use graphics to teaching scene design. One of the to investigate themolecular structures of difficulties of teaching scenic design is various compounds. that students with a keen interest and In 1982, Bucknell University received a talent for design often have poor rendering grant from the Exxon Education Foundation skills and little or no understanding of for the purpose of expanding Bucknell's perspective drawing techniques; and thus, interactive computer 'graphics capability they cannot accurately visualize their and its curricular impact to appropriate ideas on paper. However, computer graphics disciplines in the sciences, social can lift theatredesign classes from the sciences, and humanities. Since computer realm of drawing and composition into what graphics is a relatively new teaching tool they are intended to be: courses in and many facultymembers have little, if scenography. any, knowledge about its potential applications inundergraduate instruction, the project emphasized faculty training and curricular revision.

318 Tfte secona case study presented is that of an economics professor who had previous computer experience but none with computer graphics. Her project involved the development of a program to produce "bulging" pie charts that could be uses for econometric analyses. Theproject was very successful in stimulating faculty and student interest in computer graphics. During Fall, 1982, there were over 14,000 programes run (total student enrollment at Bucknell is 3,500) using computer graphics, and the level of activity should increase as faculty members implement additional uses of graphics in their courses.

PANELISTS:

Daniel C. Hyde F. Elaine Williams Jean A Shackelford Bucknell University Lewisburg, PA 17837

319 336 Recommendations for Programs in Computing at Small Colleges John Beidler, Chair University of Scranton Scranton, PA 18510

ABSTRACT An ad hoc committee of the ACM Education Board has been formed to update and revise the curriculum recommendations for small colleges thatwas published in 1983. Since a number of fine curriculum recommendations have been published by the ACM and other organizations, this committeeplans to develop its report as a consulting/planning document that takes into consideration the special environmental and resource problems that many small colleges must face. Thecommittee will complete itsreport before the end of 1983. At this time the committee will present a preliminary draft of its report. The audience will be encouraged to respond to the report and provide their input to the committee.

PARTICIPANTS Richard Austing University of Maryland College Park, MD 20742

Lillian Cassell Goldey Beacon College Wilmington, DE 19899 COMPUTERS AND QUANTITATIVE METHODS:HEALTHY FOR THE HUMANITIES?

by Rudy S. Spraycar

Data Processing Department United States Fidelity and Guaranty Company Baltimore, Maryland 21203

Abstract between 'humanistic psychology" and "experimental "6 psychology. Giles Gunn would add the extremes The advent of the computer has enhanced the among historians represented by "the computer ability of the humanist to apply the quantitative programming of the cliometricians" on the one hand, and statistical research methods that have been a and the studies of cultural historians of the mainstay of the natural and social sciences. Controversial in nature, these techniques have at traditional school, on the other; Gunn also finds times been misused. Further, a solid theoretical European structuralism "scientistic to the core.' foundation for quantitative approaches to research Far less polemical is Morton W. Bloomfield's in the humanities is still to be sought. distinction between the "part of the humanistic process /That? is obviously analytic and scientific," Introduction concerned with knowledge "from the outside," and, on the other hand, the concern_ with "inward

Since C. P. Snow first pointed out "how very . 8 experience," as he recasts the commonplace little of twentieth-century science /had/ been distinction between "objective" and "subjective" assimilated into twentieth-centuryartjor into the 9 elements of humanistic study. It is the former "Literary Culture" in general, and W. H. Auden wrote with which the present paper is concerned, drawing "2 "Thou shalt not sit / With statisticians..., many examples chiefly from literary research. After all, humanists have made great strides in the learning of Monroe C. Beardsley has remarked that "in so far as science and technology for which Samuel M. Hines, the student of literature is interested in... 3 describing or interpreting what he finds, he relies Jr., has called. Indeed, C. E. Kaylor, Jr., went upon plain (though by no means simple) empirical so far as to argue that "much of the research in the humanities has now become at least quasi- methods."-10 "4 scientific, and if this is so it can be laid in Eric Weil puts the interdisciplinary borrowing great measure at the door of the computer. Perhaps of empirical tools in perspective when he reminds us no technological development has had so great an that "psychoanalysis, statistics, and structural impact upon research in the humanities as the analysis provide the humanist with new tools he must computer has had. Apart from facilitating the employ simply because they are there. preparation of such traditional research tools as These tools g.ve him the opportunity to look at his material bibliographies, indices, and concordances, the computer made possible the proliferation of from new perspectives.... What we have...are new Auxiliary in relation to the quantitative and statistical analyses of the matter auxiliary sciences.... humanities.... But does it not then follow that the of the humanities. humanities should have their awn kind of This is a salutary counterweight to the efforts 'scientificity,' besides and above that of these 11 to fill such perceived vacuums in disciplines sciences...?" Or as Clarke puts it, "What is outside the humanities as the need for ethics and needed is a new philosophic stance /that/ must see 5 scientific methodology for what it is, merely a values, as Hines has stressed; it is a hopeful and sharpening of man's powersof accurate observation healthy sign that humanists are now methodological and logical deduction.... On this depends the borrowers from as well as lenders to their 12 scientific colleagues. future of the humanities."

The Problem But what of the peculiar problem raised when a humanist borrows only the most rudimentary of the Nevertheless, this trend has exacted a quantitative tools and methodologies developed by considerable price. Bowman L. Clarke recently the natural sciences and in some measure observed that humanists' welcoming new methods and appropriated by the social sciences?Here, the technologies in spite of their colleagues' rejection presence of a computer printout only seems to of such "Faustian" behavior has polarized some enhance the objectivity of the results; the adage disciplines; he offers the example of the gulf "Garbage in, garbage out" seems apt. In other

321 333 15 words, the dangers of the American version of the misplaced faith in crude arithmetic. Here one two cultures really lie less with the stolidly might wish to be reassured that a five percent resistant humanist whose knee-jerk.reaction against difference in "percent formularity" cannot be computers and .quantitative methods may well be attributed to the margin of error. fundamen1-1,,Ay-anti-intellectual than in the ignormmca_ :anwever tolerant or enthusiastic, of the Among studies in this area, there are few blftta_b.mm=m1=7.-who applies with apparently elegant instances of collaboration between humanists and sir ,.city...aarathodology borrowed from our statisticians. To be sure, some recent sr------r4, 71y-trained colleagues to an age-old dissertations have been written by scholars with humaniszic =xux. or conundrum. serious training in both the humanities and in, for 16 Though such scholars are Pseudo-Science example, statistics. rare, and their works make unprecedented and often nearly impossible-darands on their humanistic The study of Classical and medieval epic offers readers, they nevertheless represent a healthy trend an example that illustrates some generalities about that will in time contribute greatly to a solution hasty and literally "undisciplined" inter- of a part of the problem. disciplinary borrowings. It is to a would-be scientific impulse that one may attribute the In the meantime, however, while the ranks of attractiveness of an "acid test" to establish simply such interdisciplinary scholars are thin, other and conclusively the provenance of a troublesome problems are arising as less well prepared I refer to the relatively well-known "oral- text. humanists attempt to develop new methods on their formulaic" theory of Milman Parry and Albert Bates own. One of the worst consequences of this activity 13 Lord, originally a rather subtly conceived and is its effect on the humanities' credibility in the carefully qualified hypothesis that the study of intellectual community at large. Of much greater Homeric style may be illuminated by the living oral destructive potential, however, are the consequent epic tradition in Yugoslavia. In latter days, resentments and divisiveness among humanists. however, the attractiveness of an easy, quantitative mode of analysis for troublesome Classical and Methodological Caveats medieval texts, inspired less by scientific method than by a credulous misunderstanding of what a What deserves our careful consideration, then, "living laboratory" of poetic composition might is less our attitude toward new research tools in signify to a "man in a white coat," does little general than our profound need to carry over along justice to Protessor Milman Parry's original genius. with these tools the methodological caveats and To be sure, it is a radically attractive notion that restrictions that responsible scientists place on one might, simply by counting the number of poetic their use. The real problem in need of resolution phrases repeated elsewhere in a "literary" text, and requiring the concerted efforts of us all is adding them up, and dividing by the number of not a conflict between technological or quantitative phrases in the whole sample, thus calculating a approaches and the matter of humanistic study; it "percent formularity" as a measure of the is the disregard for the very intellectual repetitiveness of the text, demonstrate apparatuses without which the scientists' tools are incontrovertably that a given poem--say Beowulf or little more than toys, that can be inimical to the Chanson de Roland--was composed by illiterate traditional humanistic concerns. means, despite the common-sense testimony of the text's survival in written form. For example, a bulwark of scientific method is the formation and testing of hypotheses. But John But some of the followers of Professors Parry Reichert has pointed out that "one of the tests of and Lord have been plagued by such methodological a scientific hypothesis is its predictive capacity. difficulties as the labelling as "statistical To the extent that it predicts well other events analysis" of what should be called "raw data": beyond the event it was designed to explain, its numbers generated by elementary arithmetical validity is established. This aspect of hypothesis operations like counting or adding up and dividing. testing points up nicely one of the fundamental Indeed, few studies of this literary problem have features of criticism.... Our critic in the role mentioned even so fundamental a qualification as of interpreter or teacher is not concerned with the the "margin of error" resulting from the sampling discovery of general laws that will apply beyond technique that we have come to expect from the the work. His target is an understanding of the purveyors of public opinion polls. Similarly, the individual literary event in its uniqueness. In discounting of a wide disparity in the relative this respect his task is somewhat like the lengths of two samples of poetry analyzed for historian's as it was conceived by philosophers repetitive diction (without apparent regard for the like R. G. Collingwood and Michael Oakeshott, both probabilistic principle that the longer anyone, poet of whom stressed the uniqueness of historical events or not, speaks, the greater is the likelihood of 17 and the need to "get inside" them." repetition), flies full in the face not only of 14 mathematics but of everyday experience as well. Or, in Frank Kermode's more succinct terms, Further, to find, for example, in a disparity of "Certainly science is about reality experienced as five percentage points between the degrees of repetitive, and art about reality experienced as repetitiousness of two poems of unequal length "18 constituted of unique events.... Of course as grounds for making a determination whether a text Eric Weil has observed, when the humanities, or was composed orally or in writing betrays a

322

3 3 more usually the social sciences, examine non-unique measurement is often of paramount concern. But phenomena, their methods approach those of the statistical analysis evidently requires the natural sciences: "The social sciences are situated assumption at some level of random behavior, on the same terrain as the humanities--that of whether that of the agitated vibrations of history. They are not interested in individual moleculeS or of atomic particles or that of the acts as such, but they deal with such acts globally collective human expression of political opinions. considered. They do not go into the private Northrup Frye ties the predictive power of the convictions of John Doe; they look at the group of natural sciences and of their adjunct quantitative which he is a member or at him as representative of methods to the unconscious character of the this group. Their method is statistical; their phenomena observed: "Where the phenomena are purpose is to discover necessary relations within a unconscious or where the units involved are small "19 and numerous, like atoms, molecules, or cells, so certain series.... that there is no practical difference between the highly probable and the certain, the language of But because the oral-formulaicists have science is primarily mathematical. From the attempted to use statistics and are interested natural sciences we move toward the social sciences, precisely in the relation of a poem's style to that where the phenomena are relatively large, few, and of_others in a "series," the methodologies of the complicated, like human beings. Here prediction on natural sciences and consequently their inherent a statistical basis is as important as ever, but, strictures would seem to obtain here; Northrup Frye except for some aspects of psychology, the "The physical sciences sums up the latter well: repeatable experiment is no longer at the centre of at least are not simply descriptive, but are based the study....We then move into what are generally It is not the experiment on prediction as well.... 24 but the repeatable experiment that is the key to rcgarded as the humanities." the understanding of nature in the physical sciences, and the repeatable experiment is what makes Thus the student of, say, literature, faces a prediction possible, and gives to science a problem that does not trouble the chemist or the "20 physicist or even the sociologist or the political prophetic quality. The predictive value of a scientist: while natural phenomena or perhaps even hypothesis is at once tested and established by its human behavior in the aggregate can be measured as Bruce A. Beatie insists, "only replicability: against the constant yardstick of random when an hypothesis has been confirmed by repeated distribution, the examination of an artifact of the experiments is it...provisionally accepted as a human will poses different perils. What does it valid interpretation of observed data.... The mean to speaL of a theoretical "random distribution" original Parry-Lord hypothesis, however, remains of a poet's words or phrases, in terms of which the "21 unconfirmed by repeated experiment.... As I significance of his choices or of his repetitions have argued elsewhere, not only have the followers may be assessed? Tests of statistical validity of Parry and Lord repeatedly modified their generally require a random standard as a background, against which the "significance" of data becomes 22in studying poetry in various methodologies apparent by virtue of its difference from the languages, undermining the hypothesis' claim to random pattern. Any artifact of the human will, 23 universal validity, but my own attempts to test, such as poetry, must therefore pose some in the very Serbo-Croatian tradition upon which the philosophical difficulties for statistical analysis. hypothesis is based, its prediction that no literate poet can write as formulaically as an oral Conclusion poet can compose, has cast further doubt on the usefulness and reliability of the methodology: in I will not attempt to offer even a rudimentary fact I found no significant quantitative 25 solution here, but this problem serves as an distinction between oral and written Serbo-Croatian example of the central problem created by the verse based on the number of repeated phrases that impingement upon the humanities of the scientist's they exhibit. world. Very likely, we must also steer clear of rigid "scientism," which "assumes a principle of That the replicability or predictive power of determinism. i.e., from the quantitative so explicitly quantitative an hypothesis was left description of an existing situation and from the untested for decades is certainly cause for concern. known laws governing the changes of the magnitudes But I am troubled far more by the question whether involved in the situation, a universal prediction scientific and quantitative methodologies, even can be made regarding the changes occurring in the when borrowed by humanists so as to preserve fully situation in question.... Apart from measuring, their explanatory powers, are finally suited to the uul:ervation and statistics, only the deductive analysis of poetry. "26 method is considered to be valid. Problems such Random Behavior as this one will not be resolved without recourse to the fundamental philosophy of interdisciplinary Of special difficulty here is the dilemma methodology. We have only begun to raise the right faced by those tackling the essentially quantitative questions about how the quantitative analysis of problem of the degree of repetition as a factor of literature that the computer makes more practicable poetic style. "Degree" must be measured, and it is is to be carried out. no mistake to borrow here from our colleagues in the natural and social sciences, for whom

323 340 References the Study of Oral Literature, Harvard University, 1973). 1. The Two Cultures and the Scientific Revolution (New York: Cambridge University Press, 1959), 14. Creditably, Joseph J. Duggan is unusually p. 17. See also Snow, "The Two Cultures," New keenly aware of the statistical problems Statesman, October 6, 1956; F. R. Leavis and involved; but see his The Song of Roland: Michael Yudkin, Two Cultures? The Significance Formulaic Style and Poetic Craft (Berkeley: of C. P. Snow and an Essay on Sir Charles University of California Press, 1973), pp. 18ff. Snow's Rede Lecture (New York: Pantheon, 1963); Snow, The Two Cultures: and a Second Look 15. Donald C. Green, "Formulas and Syntax in Old (Cambridge: Cambridge University Press, 1964). English Poetry:A Computer Study," Computers and the Humanities, 6 (1971), 92. 2. W. H. Auden, "Under Which Lyre:A Reactionary Tract for the Time (Phi Beta Kappa Poem, 16. See, for example, Agnes M. Bruno, Towards a Harvard, 1946)," in Collected Poems, ed. Edward Quantitative Methodology for Stylistic Analysis Mendelson (New York: Random House, 1976), p. (Berkeley: University of California Press, 262. 1974).

3. "Biological Science and the Humanities: Some 17. Making Sense of Literature (Chicago: University Considerations of Human Values and the New of Chicago Press, 1977), pp. 26-27. Biomedical Technologies,"in Images and Innovations: Update '70's, ed. Melinda R. 18. "The University and the Literary f.!.blic," in Maxfield (Spartanburg, S. C.: Center for the Stroup, ed., p. 69. Humanities, Converse College, 1979), p. 123. 19. "Humanistic Studies," pp. 245-46. 4. "The Humanities and Medical Science: An Epistemological Diagnosis," in Maxfield, ed., 20. "Speculation and Concern," in Stroup, ed., p. p. 137. 36.

5. Hines, p. 127. 21. "Measurement," p. 186.

6. "The Future of the Humanities," National Forum: 22. "The Aesthetic Implications of Formulaic The Phi Kappa Phi Journal, 69, No..3 (Summer Diction: How Are We to Read Beowulf?" 1979), 21. unpublished paper read at the Conference on Language and Style, City University of New York, 7. "Reflections on the Humanities," National Forum: April 1977; see also Beatie, p. 189. The Phi Kappa Phi Journal, 69, No. 3 (Summer 1979), 15. 23. "La Chanson de Roland:An Oral Poem?" Olifant: A Publication of the Socigtg Rencesvals, 8. "The Two Cognitive Dimensions of the Humanities," American-Canadian Branch, 4, No. 1 (October Daedalus, 99, No. 2 (Spring 1970), 256-57. 1976), 63-74 (Beatie seems to have been unaware of this article); "Automatic 9. See for example Bruce A. Beatie, "Measurement Lemmatization in Serbo-Croatian," ALLC and the Study of Literature," Computers and the Journal (published by the Association for Humanities, 13 (1979), 185; Beatie's article Literary and Linguistic Computing), 1, No. 2 addresses Limner ptoblems from a perspective (Autumn 1980), 55-59; "Statistics and the comparable to mine. Computer in Formula Analysis of Serbo-Croatian Heroic Verse," Proceedings of the International 10. "The Humanities and Human Understanding," in Congress on Applied Systems Research and The Humanities and the Understanding of Reality, Cybernetics, ed. G. E. Lasker, 5, 2576-80 ed. Thomas B. Stroup (Lexington: University of (Oxford: Pergamon Press, 1981); "Formulaic Kentucky Press, 1966), p. 13. Style in Oral and Literate Opic Poetry" (with Lee Dunlap), Perspectives in. Computing 11. "Humanistic Studies:Their Object, Methods, (published by IBM), 2, No. 4 (December 1982), and Meaning,'" Daedalus, 99, No. 2 (Spring, 24-33. This research was aided by grants from 1970), 247. the American Council of Learned Societies in 1979 and from the American Philosophical 12. "Future," p. 23. Society in 1980.

13. See The Making of Homeric Verse: The Collected 24. "Speculation," p. 37. Papers of Milman Parry, ed. Adam Parry (Oxford: Clarendon Press, 1971), and Lord, The Singer of 25. Though a comprehensive review of promising Tales, (Harvard Studies in Comparative studies relevant to the problem at hand lies Literature,' No. 24, Harvard University Press, well beyond the scope of the present paper, the 1960). For a general but dated account of reader may find useful the bibliographies subsequent work, see Edward R. Haymes, A recently offered by Susan Hockey on pages Bibliography of Studies Relating to Parry's and 141-43 of A Guide to Computer Applications in Lord's Oral Theory, Publications of the Milman the Humanities (Baltimore:Johns Hopkins, Parry Collection (Cambridge, Mass.: Center for 1980) and by Robert Oakman on pages 170-71 and

324

3 4 217-227 of Computer Methods for Literary Research (Columbia:University of South Carolina Press, 1980).

26. The definition is that of Joseph Kockelmans, Phenomenology and Physical Science: An Introduction to the Philosophy of Physical Science, Duquesne Studies, Philosophical Series, 21, p. 75 (Pittsburgh: Duquesne University Press, 1966); for his account of the history of the trend of "scientism" in scientific thought, see pp. 72 ff.

325

3 4 ° A PERSONAL COMPUTER FOR EVERY COLLEGESTUDENT

David W. Bray

Educational Computing System Clarkson College of Technology Potsdam, NY 13676

Abstract all of the computer services that they require. The requirements for servicehavebeen _steadily Clarkson College of Technology will provide growing, along with the waiting lines at the pub- for sometime each entering freshman studentwith a personal lic terming s. It has been clear computer beginning in the fall of 1983. This pap- that simpl r addingmore and more terminals to puters is a losing battle. A solu- er describes the history behind the establishment mainframe c of this program, how the prog7sm will operateand tion for increasing computer resources is to put computers be administered, and some expectations as to how the new generation of powerful personal who require it.At it may change the educational process at thecol- at the disposal of those student bodyand lege level. Clarkson this is the entire faculty. Introduction The introduction of thepersonal computer com- ClarksonCollege of Technology is one of the will do much more than provide the required educational largest engineering colleges inthenortheastern putingpower, it will change the Before we consider the 'ways in whichwe United States. It has concentrated in under- process. graduate education in engineering, science, and believe that education will change because of the this program came management for many years. Next fall, in August computer, we next discuss how computer was 1983, every freshmanstudent enteringClarkson into being, and how theZ-100 Collegewill be provided witha powerful pro- selected. fessional-quality Zenith Data Systems Z-100 desktop computer. By 1986 every undergraduate Computer Selection student, and most faculty andgraduate students, will have his or her own personal computer.These The introductionof the personal computer computers will be interconnected via a campus wide into the educational process at Clarkson had been year network which will include the College mainframe under consideration for several years. Last during the Spring 1982 semester a group nf ;- computers as well as thepersonal computers. They met on a reg- Clarkson is, to thebest of our knowledge, the and students formed a "class". first college to provide every student with a per- ular basis twice weekly to consider the recom- sonal computer and to integrate its use into the quirements and specifications of apersonal an associated campus wide network that educational program. puter and would meet the College needs. The mainrec- The Zenith Data Systems Z-100 dualprocessor ommendations of this group with regard to the per- desktop computer, has 128K of program memory, high sonal computer were that: resolution bit-mapped graphics with 640 by 225 16 bit proces- pixel resolution, one five and one-quarter inch 1. The computer should contain a floppydisk drive, and three communications sor. It was clear that the 16 bit processors thus ports. Each computer will be supplied with CP/M were coming into the market in force, and for use on the eight bit 8085 processor, Z-DOS (a it would not be wise to begin a new program on an eight bit derivative of MSDOS) for use on the sixteen bit with a computer operating processor, even though there was a large amount 0 Mod processor, along with Fortran, Pascal, ZBasic (IBM PC compatible), Multiplan, and a word proces- of aoftware available for the eight bit processor, as well as Cobol for those who need it. The sor computers. It was further felt that the IBM computers will be used in all facets of the under- computer should be compatible with the graduate educational programs, and will also be PersonalComputer (PC). The reason for this used by the faculty andgraduate students for was that there are a large number of--companies the IBM PC. There is their research. producing software for also a fair amount of interest on thepart of If this interest At the present time the College .uframe universities in the IBM PC. and compatible computers, an IBM4341/Mod II and a Digital is strong enough, the IBM PC Equipment Corporation VAX-11/780, support over two computers, might be the foundation for a benefits of a hundred terminals. Even this amount of computa- standard in education. The tion capability is not sufficient for the needs of "standard" computer in education would be enor- could exchange software and theCollege. It has been a long standing policy mous, 1olleges to provide all students and faculty members with courseware freely without the current problems

326 3 i3 of computer incompatabilty. Implementation Plan

2. The computer shouldhave a high quality At Clarkson the plan for the introduction of cathoderaytube display that is capable of at the personal computer is to start with the least 80 characters per display line, and res freshman class entering in the fall of 1983, and olution high enough for engineering and to continuewith each entering freshman class scientific graphic applications. The 80 thereafter. Ownership of thepersonal computers character displaywould allow the computer to will reside with the College. However, when the provide professional word processing student graduates the ownership will be trans capabilities. ferred to the student. Shouldthe student withdraw from the College without completing de 3. The computer should have a professional greerequirements, the computer will remain with keyboard to complement the high quality display the College. These computers will bereassigned for word processing. to transfer students. Beginning with the entering freshman class, there will be a tuition surcharge 4. The computershould have bitmappped graphic of $200 persemester to cover, in part, the capabilities that would be suitable for en additional cost of the program.In addition, each gineering and scientific applications. If student will be required to make a $200 possible the graphics should be upgradeable to maintenance deposit to be used in theevent that color, for those special applications that re servicing of the computer is necessary. It is the quire it. students responsibi_ity to keep the computer in good condition. Afterthe ninety day guarantee 5. The computer should have communications ports period has expired the student will be financially that would be general enough to support network responsible for keeping it maintained. The interconnection, and to support localprinters College will insure the computers against fire and should a student desire to attach one to his or theft. her computer. This implementationplan has a number of 6. The computer should have softwaresupport that benefits. First, having the personal computer in included at least: Fortran, Pascal, Basic, troduced over a four year period with each new en Cobol, word processing, and spread sheet tering class provides lead time for the faculty of accounting. the upper class courses to modify and develop their courses for the most effective use of the As a result of this study a set of guidelines computer. Second, the graduating students will were prepared. These guidelines were sent to the take with them a computer and muchselfdeveloped major personalcomputermanufactures which were software thatwill help bridge the gap from col believed tohave, or be developing, computers lege to business and industry. Third, the comput which mightmeet our needs. These guidelines er can be a potential means of communications and suggested that the program might be implemented as contact with alumni. It might even be a vehicle early as the fall of 1984.A wide variety of re for continuingeducation. Fourth, the College sponses were obtained. The first manufacturer to will be free to select new model computers as respond indicated that they had a computer "on the technologyadvances since each year new computers drawing board" that exceeded our specifications, will be needed for the enteringfreshman class. and suggested that our timetable could be pushed Thus the computing facilities will not grow old in up to the fall of 1983. Being encouraged by this time as do thecurrent facilities involving response we decided to attempt implemention of the mainframe computers. plan by the fall of 1983. Following this plan, this fall each freshman Early in the summer of 1982 we prepared a will be provided with a computer.For many years formal set of specifications for the computer, and every freshman student has been required'to take a sent it to those whoreceived our original computer programming course. With the introduc guidelines, and others who had heard of our pro tion of the personal computer the subject matter ject. From this effort we received bids from five of this courseswill not change substantially. companies. Fromthese theZenithData System What will change is the availability of a computer Z-100 was selected for its price/performance com to the student. The course has beenmodifiedto bination. The Z-100 computerexceeded our take advantage of ready access to a computer and, specifications in almost all ways. It has two of course, to teach the students how to use the features that were considered most important. The particular features of the Z-100 computer to full first of these is a second processor, an 8085. advantage. Most students will not havehadmuch This will allow the current eight bit software to corputer programming training prior to entering be used on the computer if desired. This Clarkson, and almost none will be proticientwith additional processor removes one of the concerns theZ-100, its operating systems, and language that a 16 bit computer would limit use of avail processors. Therefore, the other first term able software. The second feature is the in freshman courses must restrict their use of the clusion of four available S-100 industry standard computer to those problems that do not require ex bus slots. This allows expansion of the computer tensive programming knowledge, or to demonstration for use in the laboratory. The fact that the programs that are supplied to the student in a student computer would be software compatible with preprogrammed form. In the second semester the laboratory computers would be very beneficial. freshman year, the students willbe able to do

327 moreprogramming ontheir own. Many will have Repair of the computers, after the ninety day become very proficient with theircomputers. As warranty period, will be conducted by college per they progressfrom semester to semester the pro sonnel. Being a technical college there is no fessors will be able to require moreprogramming lack of qualified technicians and students already knowledge. on campus.

To allow the faculty time to discover the Educational Benefits many potential uses of the personal computer in their courses, computers have already been pro The computerhas become a necessary tool in vided to a number of faculty members. Those who thebusiness world. Aftergraduation current first obtained the computershave contact with studentswill be confronted withcomputers in freshman and sophomore students (or other students their employment, no matter what the field. A who will be using the Z-100 in the laboratories), tool that has become so much a part of business and wish to introduce the use of the computer into and industry, surelyhas an equally important their courses. The initial delivery of computers place in the educational process. At this point for the faculty was completed before theend of in time it is very difficult to evenimagine the theFall 1982 semester. This gave a minimum of effect that the personal computer will have on the nine months lead time for coursepreparation educational process. However, there are a few planning for the freshman courses, and more than clues as towhat may occur. Just as the that on the average,Those faculty memberswho calculator replaced the sliderule inengineering are teaching upper divisor courses will be receiv and science courses, the computer will replace the ing computers in the near future. calculator. Its potential in education, however, is much more than simply being a better tool to do Administration of the Program the same job.

To administer the introduction of the per Some of the important advantages of the com sonal computer into the educational process at puter in education are evident now. These are: Clarkson, a newposition has been established. This position is the Dean of the Educational 1. When a student is asked to perform a homework Computing System.Theresponsibilities of this assignment that involves a large amount of position are, in general terms: calculation the process of performing the calculations becomes the focus of the lesson. 1. To provide coordination for academic activities That ia, the student studies the theory, or concerned with the program. ganizes the problemto be solved, and then begins the calculation. At thatpoint the 2. To develop a personal computernetworkwhich process of operating the calculatorreceives will interconnect with the current mainframe the full attention of the student. The student network that now exists on campus. often loses sight of the lesson, concentrating only on pushing the calculator buttons. If the 3. To administer the purchase, distribution, and student is asked to perform the same problem on repair of the personal computers. a computer with which he is familiar, the solu tion takes a different form. The student or A major responsibility will be in coordinat ganizes the theory in terms of language of the ing the academic activities. Thiswill include computer. A language much like the equations training of the faculty in the use of the Z-100 of the theory. Therefore the student's efforts operating systems and associated software, and toward obtaining the calculation actually coordinatingcourse development in the sense of becomes an effort of rewriting and rethinking disseminating information about the activities of the theory of the problem. The computer can the faculty and thesoftware that is being de then focus its attention on obtaining the veloped. It will also include conducting actual solution. A student who can instruct a workshops at the local college level so that computer how to solve a problem has a good faculty membersmay learn from fellow faculty grasp of the problem. memberswhohave introduced the Z-100 in their courses, and holding workshops ledby professors 2. Computer simulations can be usefuland from othercolleges who have used personal com effective teaching tools. The computer can aid puters in various academic settings. the student in obtainingmore insight into scientific principles that are being studied. At the present time, the collegemainframe Once a student has obtained a solution to a computers and several groups of remote terminals problem that has been calculated by a computer, are tied into a X.25 packet switchingnetwork. there is very little effort involved in obtain The present thinking is that the personal com ingmore solutions to the same problem with puters will be tied into this network by forming a different input parameters. The ability to ob series of local areanetworks consisting of tain many solutions to a problem in a matter of several hundredpersonal computers each. Each a few minutes can provide the student with in local networkswill be tied into a X.25 packet tuition about the subject matter that is switcher and thereby connected to themain generally not possible without a computer. network. Thisnetwork is under study at this Particularly, if a solution to a problem is time. It is expected that it canbe operational presented in a form that is easilyunderstood, with the personal computers within two years. such as graphically, a student can learn much

328 about a subject matter by changingparameters and seeing the effect upon the result.

3. The computer can in many cases be a substitute for laboratory experiments at theintroductory level. For economic reasons many colleges have had toeliminate introductory chemistry and physics laboratories. The personal computer can provide, in simulated form, much of the in formation that is gained in the laboratories. It, of course, cannot substitute for the ex perience gained in working with the physical objects of an actual laboratory.

In those rajectswhich dealwith complex systems such as the stock market, management systems, weather patterns, etc. the computer can, through simulation, give the student valu able albeit limited experience that could only otherwise be obtained in the real world.

4. The computer has the potential for demonstrat ing principles that are difficult to present in lecture form. Lectures can be supplemented by the student executing programs prepared by the course instructors which woulddemonstrate, possibly in graphical form, the fundamentals of the lecture subject matter. In d sense, this is another form whereby the student can gai7a more intuition into the subject matter at hand.

Without question the personal computer will have a dramatic effect on the educational process. Only after a few years of experience at the college level will we be able to accessthe full impact that the computercan have in the educational process. Clarkson .College is com mitted to provide leadership in the use of the personal computer in college education.

329 COMPUTER-ASSISTED SIMULATION IN THE POLITICS OF REAPPORTIONMENT/REDISTRICTING (CASPOR)

Jerry E. Bolick and James 0. Icenhour

Lenoir-Rhyne College, Hickory, North Carolina 28601

ABSTRACT and minimize the use of available computer memory, developed the program, and tested its various Written in Basic and implemented cn a components. PDP-11/34 system, CASPOR is designed to enhance the learning process in both computer science and The application of this simulation in the political science classrooms. It provides a political science classroom involves the use of an realistic project for computer science students outline map of North Carolina with county bound- and renders manageable the classroom simulation aries detailed, student role assignments, general of the politics of reapportionment and redis- rules and limitations, and the use of CASPOR for tricting by political science students. CASPOR analyzing and verifying proposed redistrictinc is flexible, can be adapted to various classroom plans. situations, and can be modified to reflect the political realities of vai.'-1.is states. DEVELOPING THE PROGRAM

The program CASPOR was developed with the aid INTRODUCTION of a beginning class in data structures. After the class had analyzed the problems of reappor- The program presented and described herein tionment/redistricting and understood exactly is called Computer-Assisted Simulation in the what was necessary to produce a useful program, Politics of Reapportionment/Redistricting (CASPOR). the class of 12 students was divided into 4 groups This program grew out of the desire of one co- of 3 each. Under the professor's direction and author for improved instructional methods in the coordination, the groups completed the following teaching of the political dimensions of the re- assignments: apportionment/redistricting phenomenon at the state level and the need of the other co-author 1. Collected voting, registration and population for more realistic projects for his students in data for each county and stored it in a data a data structures course. CASPOR has been tested file. and used in our classrooms with excellent results. 2. Wrote a subroutine to determine if the coun- We believe this program is sufficiently flexible ties in a proposed district are contiguous. to permit its adaptation to the political reali- 3. Wrote a subroutine that permits modification ties of various state political systems and that of districts. it can be readily modified to fit the classroom 4. Wrote a subroutine to print the information needs of a wide range of classes as well. about each district and calculate cumulative statistics for each district. CASPOR is written in the programming language Basic and is Implemented on a PDP-11/34 The simulation makes extensive use of data computer system. Data files, all of which are that must be available each time the program is stored in virtual arrays, are used extensively. used. One of the essential elements of the The structure of these files is an essential simulation is the 23 items of data which describe ingredient in this simulation. the population, registration and voting statistics for each of the 100 counties in North Carolina. The data for this simulation were derived from the public records of North Carolina. Vari- The structure selected for this data is a ables include population, voter registration by 100 x 23 array - one row for each of the 100 political party preference, and presidential, counties and one column for each of the 23 data congressional and gubernatorial electI ?\ 1:0sults, items for each county. Since this data set must all by county. be available to theill aX times and is quite large for a ,RW41-:-i4 ?rCiMter, the students This paper describes the step by L,i.up pro- chose to store the,data in array. cess by which computer science students, worki,-; in teams, amassed the data base from original When the political science students begin sources, selected appropriate structures for the their reapportionment/redistricting simulation, data to maximize the efficiency of the program they can use the computer as their data source.

330 347 To collect the needed data the student runs virtual array. The following portion of the matrix CASPOR and selects the data collecting option. describes the adjacency of North Carolina counties.

When the political science students collect the data needed for their decisions, they are also 9 78 26 82 71 24 0 0 0 0 0 supplied with a list of North Carolina's 100 counties which are listed in alphabetical order and 24 78 9 71 10 0 0 0 0 0 0 indexed with numbers 1 - 100. The counties are stored in an array that contains 100 elements so 26 67 47 63 43 82 9 0 0 0 0 that each county can be referenced by its index number.Again, the data structure class chose a 71 65 10 24 9 82 31 67 0 0 0 virtual array so that the list of counties can be stored separately from the main program and yet can 78 83 47 26 9 24 0 0 0 0 0 be accessed by the program when proposed districts are being developed. 82 71 9 26 43 51 96 31 0 0 0

With the statistical data related to popula- Using this matrix, it is relatively easy to tion, voter registration and election results write a subroutine that searches for a ,bath accessible, the next problem that must be solved through all of the counties that &re proposed for is that of contiguousness. Since all counties in a a district. If a path exits then the counties are congressional district must be contiguous, the com- contiguous. The followis.., program segment searches puter must be able to decide if the counties for a path through a proposed district. selected for a district are in fact contiguous. 4105 S1(L1)=0 FOR L1=1 TO 90 The data structures class found that the 4120 OPC-. °Are.iNCY-DAT° AS FILE 48 problem of county contiguousness is really an 4200 S=D1-.1) application of graph theory. If one lets the 4210 FOR Cr-1 TO 20 nodes of a graph represent counties and agrees 4220 IF D1(C)<>0 THEN 4240 that two nodes are joined by an edge if two coun- 4225 CLOSE 08 ties are adjacent, a graph that describes the 4230 F$=YES° adjacency of counties in North Carolina results. 4235 RETURN Figure 1 shows a segment of this graph. 4240 T=D1(C) 4250 IF T=S THEN 4450 4260 L=1 4263 S1(L)=S 4270 I1=2 4280 12=81(0 4290 IF S1(L)<> 0 THEN 4310 ELSE F$moNtr 4295 CLOSE 08 4300 RETURN 4310 J5=0 4320 J5=J5-1-1 4330 IF A4(I2,J5)=T THEN 4450 4335 IF A4(I2,J5)<>0 THEN 4340 4337 Lz-L+1\8OTO 4280 4340 FOR K=1 TO 20 4350 IF A4(I2,J5)<>D1(K) THEN 4400 4360 FOR M=1 TO I1-1 4370 IF A4(I2,J5)=S1(M) THEN 4400 4375 NEXT M Figure 1 4380 S1(I1)=A4(I2,J5) 4385 I1=I1+1 An obvious way to represent this graph is to 4390 0070 .4320 use a 100 x 100 adjacency matrix 4400 NEXT K 4410 GOTO 4320 [a..] 1; if counties i and j are adjacent A = 13 4450 NEXT C 0; otherwise If a group of counties constitutes a valid But this representation produces a sparce ma- district, it is then necessary to determine the trix and is an inefficient use of memory. Hence, cumulative population, election results and voter the class chose a matrix representation in which registration statistics for the distri't. In the entries in row i are the indices of the particular, it is essential that the population of counties that are adjacent to county i. For each district be recorded since the population of example, Bladen County has index number 9 and each district must be compared to the average popu- counties with indices 24, 26, 71, 78, and 82 are lation per district in terms Of deviation and adjacent. Thus Row 9 of the adjacency matrix has average deviation. These deviations must fall 24, 26, 71, 78, and 82 asits non-zero entries within predetermined bounds. (a zero indicates no more counties in the adjacency list). This matrix is also stored as a If these bounds are not satisfied then the

331 348 district must be adjusted. Therefore, the data All of this information is public knowledge structures class wrote a subrout4nP to permit and is shared with all other participants counties to be moved from one district to anther. (Appendix 3). In addition to this public role, each student is provided with a statement of confi- North Carolina actually has only 11 Congres- dential political considerations which is not sional Districts, but for this simulation 12 shared. Each participant is urged to maintain this districts are used with a maximum of 20 counties assigned role in utmost confidence (Appendix 4). per district.The indices of the counties that are selected for district i are stored in row i of The overall rules of the simulation are then a 12 x 20 matrix, C. After composition of the 12 explained to the participants. Each is urged to districts has been determined, the print subroutine maintain high fidelity to the assigned role in all uses matrix C to sort the population, registration, negotiations among the group. Specific limitations and voting data for the counties in a district, include the following: calculates the totals for the district, and prints a complete description of the district. The 1. Districts must be as compact as possible. - district totals are saved and the final printout is 2. Districts must include whole counties. a set of cumulative statistics for each of the 3. Exact mathematical equality of district popu- districts. Appendix 1 contains a complete descrip- lation is desired. tion of the proposed 5th District. 4. Overt gerrymandering will render proposed districting plans suspect and should therefore be The summary analysis in Figure 2 shows total consciously avoided or minimized. population, deviation in population, and relative

deviation in percent for each proposed district. - An outline map of North Carolina is provided each participant and is to be used to outline pro- TOTAL STATE POPULATION 5,084,360 posed district boundaries (Appendix 5). The RATIO OF HIGHEST TO LOWEST 1.338 participants are now dire,:ted to the computer and NUMBER OF CONGRESSMEN 12 given access to all the aata stored in the CASPOR AVERAGE POP PER CONGRESSMAN 423,696 program.

RELATIVE After considerable individual efforts to DIST POP DEVIATIONDEVIATION develop acceptable district plans, the participants convene as a group and begin the process of nego- 1 354,040 -69,656 -16.4402 tiations necessary to reach workable and acceptable 2 445,257 21,561 5.0887 solutions to the redistricting problem. When a 3 473,794 50,098 11.8240 consensus is reached it is tested and verifiedby 4 419,394 -4,302 -1.0154 the application of the CASPOR program. Remodifi- 5 415,252 -8,444 -1.9930 cation and revalidation are pursued until either 6 457,354 33,658 7.9438 an acceptable redistricting plan is devised or 7 405,817 -17,879 -4.2199 until stalemate is reached. 8 428,202 4,506 1.0634 9 429,285 5,589 1.3190 When the simulation is terminated the 10 410,409 -13,287 -3.1361 participants then share their confidential roles 11 420,830 -2,866 -0.6765 with each other and the instructor summarizes the 12 424,722 1,026 0.2421 learning experience mutually shared by the participants in the simulation. AVERAGE RELATIVE DEVIATION PER CONGRESSMAN 4.5802 AVERAGE DEVIATION SUMMARY PER CtieGRESSMAN 19,406 While CASPOR was developed for use in both Figure 2 computer science and political science courses, it can be used independently in either. Furthermore, any number of students can participate. Small IMPLEMENTING THE PROGRAM classes can participate as a whole while large classes can be sub-divided into smaller teams The implementation of the simulation (CASPOR) with each team developing or using the complete in the political science classroom begins with simulation. Roles can be assigned to reflect the assignment of roles to the various partici- any constellation of political variables. And pants. Collectively the class constitutes a finally, the concept can be applied to the politi- simulated state legislative committee charged cal systems of other states. with the task of devising a congressional redistricting plan to accomodate North Carolina's CASPOR has enhanced the learning process in presumed increased apportionment of congressional our classrooms. We hope that others may find it seats. Each student is assigned a role that useful in their classrooms as well. includes specification of the following variables: county of residence, political party affiliation, age, sex, race, religion, social ntotus, economic status, and legislative seniority and assignment (Appendix 2).

332

3 4 j Appendix 1

DISTRICT 5 POPULATION STATISTICS COUNTY TOT POP TOT WHITE X WHITE TOT BLACKX BLACK TOT OTHER X OTHER BLADEN 26,477 16,151 61..00 10,326 39.00 0 0.00 COLUMBUS 46,937 32,997 70.30 13,940 29.70 0 0.00 CUMBERLAND 212,042 161,364 76.10 50,678 23.90 0 0.00 ROBESON 84,842 62,953 74.20 21,889 25.80 0 0.00 SAMPSON 44,954 29,445 65.50 15,509 34.50 0 0.00

TOTALS .:15,252 '302,909 72.95 112/343 27.05 0 0.00

REGISTRATIONSTATISTICS COUNTY TOT REG TOT DEM. X DEM. TOT GOP. X GOP. TOT OTHERX OTHER BLADEN 13,638 12,629 92.60 857 6.28 152 1.11 COLUMBUS 24,831 22,377 90.12 2,147 8.65 307 1.24 CUMBERLAND 57,936 44,536 76.87 8,938 15.43 4,462 7.70 ROBESON 48,340 45,300 93.71 2,357 4.88 683 1.41 SAMPSON 23,734 14,571 61.39 8,621 36.32 542 2.28

TOTALS 168,479 139,413 82.75 22,920 13.60 6,146 3.65 PRESIDENTIAL ELECTION STATISTICS COUNTY TOT VOTE TOT DEM. % DEM. TOT GOP. X GOP. TOT OTHER %OTHER BLADEN 7,589 6,009 79.18 1,546. 20.37 34 0.45 COLUMBUS 14,401 11,148 77.41 3,184 22.11 69 0.48 CUMBERLAND 38,683 24,297 62.81 14,226 36.78 160 0.41 ROBESON 25,699 20,695 80.53 4,907 19.09 97 0.38 SAMPSON 15,902 8,869 55.77 6,968 43.82 65 0.41

TOTALS 102,274 71,018 69.44 30,831 30.15 425 0.42 SENATE ELECTION STATISTICS COUNTY TOT VOTE TOT DEM. % DEM. TOT GOP. X GOP. TOT OTHERX OTHER BLADEN 5,140 3,093 60.18 2,047 39.82 0 0.00 COLUMBUS 9,630 5,610 58.26 4,020 41.74 0 0,00 CUMBERLAND 25,345 12,358 48.76 12,987 51.24 0 0.00 ROBESON 12,156 7,296 60.02 4,860 39.98 0 0.00 SAMPSON 14,609 6,423 43.97 8,186 56.03 0 0.00

TOTALS 66,880 34,780 52.00 32,100 48.00 0 0.00 HOUSE ELECTION STATISTICS COUNTY TOT VOTE TOT DEM. % DEM. TOT GOP. % GOP. TOT OTHER %OTHER BLADEN 6,665 5,853 87.82 812 12.18 0 0.00 COLUMBUS 14,466 12,904 89.20 1,562 10.80 0 0.00

. CUMBERLAND 37,591 30,125 80.14 7,466 19.86 0 0.00 ROBESON 25,613 23,103 90.20 2,510 9.80 0 0.00 SAMPSON 15,966 8,654 54.20 7,312 45.80 0 0.00

TOTALS 100,301 80,639 80.40 19,662 19.60 0 0.00 GOVERNOR'S ELECTION STATISTICS COUNTY TOT VOTE TOT DEM. % DEM. TOT GOP. % GOP. TOT OTHER %OTHER BLADEN 7,363 6,432 87.36 861 11.69 70 0.95 COLUMBUS 14,216 11,994 84.37 2,152 15.14 70 0.49 CUMBERLAND 39,020 28,646 73.41 9,654 24.74 720 1.85 ROBESON 24,908 22,212 89.18 2,539 10.19 157 0.63 SAMPSON 15,810 9,718 61.47 5,980 37.82 112 0.71

TOTALS 101,317 79,002 77.98 21,186 20,91 1,129 1.11

333 35u Appendix 2

CASPOR ROLE ASSIGNMENT

STUDENT'S NAME: AGE:

REPRESENTATIVE FROM: (County) POLITICAL PARTY AFFILIATION:

SEX: FEMALE MALE RACE: WHITE BLACK OTHER (SPECIFY)

RELIGION: MAINSTREAM PROTESTANT FUNDAMENTAL PROTESTANT CATHOLIC JEWISH

ECONOMIC STATUS: UPPER UPPER MIDDLE MIDDLE LOWER MIDDLE

SOCIAL STATUS: UPPER UPPER MIDDLE MIDDLE LOWER MIDDLE

LEGISLATIVE SERVICE (SENIORITY):

CONFIDENTIAL POLITICAL CONSIDERATIONS (Must Not Be Shared With Other Participants)

Appendix 3

SUMMARY OF CASPOR ROLE ASSIGNMENTS

z PNP O V O Q.p EnoEn 0 z H H STUDENT'S ti 0 4 g giE" 0 WW P NAME COUNTY 4 wm m0N N COMMITTEE ASSIGNMENTS

Able Mecklenburg R 63 F W J U U 6 V Ch House Comm on Humanities and Fine Arts Baker Madison D 55 M W Fp UM M 16 Ch House Comm on Appropriations

Charlie Durham D 45 M B MP UM UM 8 Ch House Comm on Finance (Taxes)

Delta Guilford D 40 F B MP M M 4 House Comm on Social & Economic Affairs Easy New Hanover D 52MWCUUM 14 Ch House Comm on Natural Resources

Foxtrot Stanly R 48 M W MP M M 4 House Comm on Commerce

George Forsyth D 40 F W C U U 6 Ch House Comm on Labor

How Cumberland D 35 M B HP M M 4 V Ch House Comm on Transportation

India Surry R 30 M N FP LM M . - 1st term

James Wilson D 60 M W iiP U UM 12 Ch House Comm on Agriculture

King Buncombe R 48 M W C UM UM 8 Leader House Minority

Love Wake D 56 F W MP UM UM 6 Ch House Comm on Justice

334 351 Appendix 4

SUMMARY OF CASPOR PARTICIPANT'S CONFIDENTIAL POLITICAL CONSIDERATIONS

Able. Needs to retain seat for enhancement of per- George. Passionately anti-organized labor. Wants sonal social status. Political party is of secon- Forsyth County to be included with rural anti- dary importance. Passionately desires to see per- labor counties in a congressional district so as sonal friend (also Republican) capture seat in U. S. to elect to U. S. Congress a senior vice president Congress (from Mecklenburg County) next election. of one of the Winston-Salem tobacco firms. Is loyal to Democratic Party and rather liberal. Baker. Wants to become first black U. S. Congress- man from North Carolina in modern history. To do How. Owns local independent trucking firm. this, Durham County must be part of a congres- Served in U. S. Army as officer for 4 years.Re- sional district that encompasses northeastern lates well to military personnel at Ft. Bragg and counties with large black populations. Everything Pope AFB. Wants to be a king-maker by providing else is of secondary importance. But does have winning margin in next election for new congress- intense loyalty to own race and political party. man from Fayetteville. Hopes for appointment as U. S. Assistant Secretary of Transportation. Charlie. Wants to become Speaker of the House, so must keep own seat and curry favor with all other India. Wants to be Lt. Governor, or U. S. Con- legislative Democrats as well as Democratic party gressman by age 40. Politically ambitious but leaders statewide. Seeks good publicity and adheres to a very strict personal code. Will party harmony. sacrifice principle only for personal political advantage or gain. Delta. Wants to see U. S. congressional districts drawn in a way that will make the major metro- James.Tobacco farmer and warehouseman. Dedicated politan centers of North Carolina the focal points to his own economic self-interest. Above all else, of congressional districts, thus enhancing the wants to keep high party power position and over-all political impact of black voting ability to horse-trade in the legislative process. concentrations. King. Hopes to win seat in U. S. Congress from a Easy. Over-riding desire is to see the entire western North Carolina congressional district. outer-banks area of the state encompassed in a Must structure a congressional district that in- single U. S. congressional district. Or at least cludes Asheville in a way that will enhance Repub- as much of it as can be achieved. Former N. C. lican chances of winning next election. Strong State Democratic Party Chairman with vast real party supporter and one who thrives on publicity. estate holdings in the outer banks area. Would probably run for Congress from such a district. Love. Wants to run for Office of N. C. Attorney General or N. C. Supreme Court.Lawyer of some Foxtrot. Views Stanley County as part of the note, with thriving practice in Raleigh. Seeks eastern section of the state. Wants to make sure greatest exposure statewide to create positive Stanly County is not placed in a congressional personal image. Seeks party unity and would district with any counties to the west of it. like to see all N. C. congressional districts Would also like to see district lines drawn to elect Democrats. enhance Republican voting power in congressional elections.

Appendix NORTH CAROLINA

335 352 INTEGRATING COMPUTING PACKAGES AND STATISTICS INSTRUCTION

William D. Schafer and C. Mitchell Dayton

Department of Measurement, Statistics, and EValuation College of Education, University of,Maryland College Park, Maryland20742

Abstract Characteristics of the Course A computer-based, intermediate-level applied statistics course is described. All course Materials. It was expected that students would assignments were carried out using BMDP and SPSS have had experience with some computing package routines and the results were integrated into in a first course in applied statistics. Materi- als used in a first course administered locally the course instruction. An evaluation of the course is discussed and revisions based on the were used when students did not have this back- evaluation are presented. A description of the ground. data base and the assignments are included. 1 The textbook (Afifi and Azen) was supplemented with various handouts and descriptions of BMDP and SPSS packages (those used in the course) Introduction developed locally. Use of the full manuals was encouraged but not required. Single copies of Many statistics courses have as one of their these materials and the syllabus may be requested goals providing a computational basis with which from the first author. statistical topics can be applied in practical situations. This goal is often met using devices such Topics and Assignments. Following an initial as calculators which have undergone recent changes introduction to BMDP and SPSS, the order of topics in power and portability, but have retained limita- was: bivariate regression and correlation; part tions in terms of storage capacity and. output. and partial correlation; multiple regression and With good reason, field applications of statistics correlation; dummy coding and product variables; primarily use computers and there exists a large ordered regression; one-way analysis of variance; body of software to support this use. It is natu- comparisons and contrasts; simultaneous inference; ral, therefore, that applied statistics curricula two-way analysis of variance; interpretation of should increasingly reflect involvement with com- interaction; non-orthogonal analysis of variance; . puters. fixed, random, and mixed models; analysis of covariance; generalized ordered regression; Computing packages in institutions which are Hotelling tests; discriminant analysis; multivari- users of statistics have become commonplace. While ate analysis of variance; and principal components they are somewhat restrictive in that one is limit- analysis. The topics were illustrated throughout ed to the analyses they provide, they do employ a the semester using a total of 36 final computer rather broad coverage of popular techniques and are runs by each student, equally split between BMDP relatively easy to use. and SPSS programs.

A course is described here which was developed The data set for the majority of assignments in order to provide package programming skills to was taken from the 1979 Information Please Almanac students in support of the topics often found in a and consisted of sixteen variables relating to the second-semester applied statistics curriculum. In "quality of life" of each of 113 nations, along this effort we had three guides: the traditional with their continents. Arbitrary codes were used content of such a. course, the capabilities of the, for missing data.A description of the data set local computing center, and an article by Thisted' and the assignments appear in the appendix. describing similar efforts. Students were assigned corresponding computer The course was implemented first during summer runs for virtually every topic which was presented. of 1979. This paper describes an evaluation of the By-hand computations were included in most cases, course during the fall of 1979, when a non-computer- as well. In three instances data from the text oriented section was available as a "control". were used to give students experience in preparing Finally, revisions to the course based on these ex- data for processing by computers. periences are suggested and an updated version of the course is discussed.

336 Students. Forty-two students completed the Results. Scores on the common exam items were course. Of these, thirty-eight (90%) were pursuing paired with the correspomiing item on the pretest. a doctorate. Their programs were in the areas of The observed difference between the sections was human development (12, 29%), counseling and per- not significant (see table 1) at conventional sonnel services (11, 26%), secondary education levels. (6, 14%), health education (3, 7%), industrial education (3, 7%), special education (1, 2%), and The ten items on the affective posttest were measurement and statistics (6, 14%), the home each paired with the corresponding item on the department of the course. Twelve students dropped pretest. The observed differences on these items during the semester, representing 22% of the ori- were significant at conventional levels in favor ginal registration, which is not particularly of the "computer" group (see table 1): unusual for this course. I have come to understand many useful statis- Format. The course was taught on a twice-perweek tical techniques in the course. basis over one semester (16 weeks) by a single instructor with no graduate assistant support. I learned many new skills which will benefit Computer facilities were located some distance me in my research in the future. from the classroom. Most students operated in a batch mode because of inaccessibility of termin- I had difficulty understanding the material als and file space on the campus. Batch turnaround" presented in class (reversed scale). varied._ between five minutes and six hours during the semester. This course was particularly well-organized.

Topics were treated for the most part with an This course increased my interest in statis- introductory presentation of the material followed tics. by a discussion of package output in a later class petiod, thus allowing students to refer to I enjoyed my experience in this course. their own completed assignments (and annotate them) during the session. Assignments were collect- The observed difference on this item was sig- ed and returned at the next class. nificant at conventional levels in favor of the "control" group: Student Evaluation I had difficulty understanding the material Procedures. Two sections of the course were offer- presented in the textbook (reversed scale). ed during the fall semester of 1979, one on two mornings (9:30 - 10:45) and one on two afternoons The observed differences on these items were (4:15 - 5:30) each week. The section given on the not significant at conventional levels: afternoons was instructed using the computer- oriented approach described here; the other section I spent too much time and effort on this was instructed by another faculty member of the course (reversed scale). same rank using a non-computer-oriented approach traditionally used in the department. Aside from I had adequate background preparation to take incidental information learned during the semester, this course the students were unaware of the difference in approaches. With minor exceptions, identical con- The material in this course was covered too tent was covered. The syllabus for the "control" superficially (reversed scale). section may be requested from the first author. Observations Students were asked to complete an attitudinal inventory covering their expectations in the course Thisted3 presented three organizational during the initial class period. An anonymous approaches for such a course: by package, by method was used to match these "pretests" with topics, and by software systems components.We "posttests" covering identical content at the end chose to organize by topics. This method has two of the course. also, students completed a ten- major advantages: it maintains a logical flow of item pretest covering prerequisite material during content, building later topics upon earlier mat- the first class session. These were matched with erial, and it allows comparisons to be made their performance on forty-eight common items between the computer packages on features relevant embedded in the two examinations in each of the to potential statistical analyses. sections. The data set has some advantages but we feel Subjects. Students registered for the two sections it could be improved upon. One clear advantage is normally. The "control" section originally regis- its real, public nature (it can be located and tered 35 students of whom 19 (54%) completed the understood easily). Also, it contains some cate- course. The "computer" section originally regis- gorical variables, exhibits some interesting rela- tered 54 students of whom 42 (78%) completed the tionships, and presents some common difficulties course. The difference between the proportions is such as missing data and outliers, some which significant (corrected x2 = 4.40). arise through obvious error in the data set. However, the data set is not educationally

337 354 oriented. It would be valuable to have a set B. For the variables POP and LITER, generate which is more meaningful while maintaining the histograms, ogives, and normal probability plots desirable elements noted. using BMDP5D. Similarly, generate a histogram for the variable FREEDOM using SPSS FREQUENCIES. We have found that it is very important to C. Use SPSS CONDESCRIPTIVE to obtain summary include some "by-hand" work based on each output. statistics tor all 17 variables; also, use BMDP1D If this is not done, there is a tendency for aome for the same purpose. students to regard their output as a final product with no investigation of its meaningfulness until D. From summary statistics 311 the printouts class time. Also, since it is important that all in C., above, use a one-sample t test to test the students have the same output for discussion pur- hypothesis that the true value for mean energy poses, we have found it helpful to give students consumption is 2000.0 kg of coal equivalent.Also some significant numerical value to look for in test the hypothesis that the true variance for the output in order to check their work prior to this variable is 5 million square units. These class. computations must be done by "hand."

. E. Using two-sample t tests, compare Asia As noted in the evaluation, textbook support and Africa on all of the remaining 16 variables. proved to be a problem. we have had success re- Run the analyses with both SPSS T-TEST and BMDP3D. cently with Pedhazur2. Show in detail (by "hand") how the t test for the variable GNP is computed. Our experience with this course has been positive. Students seem to develop facility with BMDP F. Prepare a bivariate plot for the varia- and SPSS in this context and seem to appreciate the bles BIRTHS (vertical axis) and LITER (horizontal tools they have learned how to use in support of axis). Use SPSS SCATTERGRAM and BMDP6D. For the what they have learned about statistics. While the BMDP6D run, have the 6 continents identified on design of the course evaluation is admittedly the bivariate plot by unique symbols (i.e., the modest, the results do not dissuade us from this letters A,B,C,D,E,F). view. G. Using a chi-square test, test the inde- pendence of the FREEDOM variable with respect to References the continents. Use both SPSS CROSSTABS and BMDP1F. 1. Afifi, A.A. and Azen, S.P. Statistical Analysis: A Computer Oriented Approach. Computer Assignment No. 2 Academic Press, 1979, 2nd ed. Due on class meeting number 7

2. Pedhazur, E.J. Multiple Regression in For the Quality of Life data set, we want to Behavioral Research. Holt, Rinehart, and predict the percent of the population in'higher Winston, 1982, 2nd ed. education (HIGHER) from selected variables. Using FREEDOM, GNP, and AGLAB as predictors, obtain 3. Thisted, R.A, "Teaching Statistical Computing solutions from SPSS REGRESSION and from BMDP2R. Using Computer Packages." The American Obtain all residual listings and plots which are Statistician, 1979, 33(1), 27-30. available from each program. By hand, use the BMDP output to compute (1) the squared part correlation (HIGHER,GNP.AGLAB. FREEDOM); and use the SPSS output to write the final, raw-score regres- Computer Assignment No. 1 sion equation. Due on Class Meeting Number 4 Computer Assignment No. 3 A. The element EDMS *STAT.QUALITY contains Due on class meeting number 8 data for 113 nations which relate to the "Quality of Life" in those nations.A description of the A. Use the adult literacy variable, LITER, variables, the data format, and a source for the in the Quality of Life data set to predict BIRTHS. Obtain solutions, including all available resi- data are presented in the element dual listings and plots, with both SPSS REGRESSION EDMS*STAT.QUALFORMAT. A listing of that element is to be obtained by setting up a run including and BMDP2R. the following statement: B. Create a variable equal to the square of LITER (in SPSS, use a COMPUTE statement; in BMDP, @PRT,S EDMS*STAT.QUALFORMAT use a TRANSF statement) and redo the prediction This data set will be used in thisand several of BIRTHS with a quadratic model.Set up a sum- subsequent assignments. To include thedata in an mary table showing sources due to linear and SPSS or BMDP run, use quadratic regression; compute the appropriate statistical tests, by hand, and interpret the @ADD EDMS*STAT.QUALITY results. when the data are needed (e.g., after a READ INPUT DATA card in an SPSS run or after the /END card in a BMDP run).

338 35 Computer Assignment No. 4 Computer Assignment No. 7 Due on class meeting number 11 Due on class meeting number 18

A. The six continents in the Quality of Life The Quality of Life data set contains two data set represent a nominal variable. Create categorical variables, Continent and Freedom dummy variables for this factor; use SPSS REGRES- Status. Treating these as factors in analysis of SION and BMDP2R to undertake comparisons among variance, use both SPSS ANOVA and BMDP2V to obtain the continents on adult literacy rates (LITER). results for the following criterion variables: By hand, use the SPSS output to obtain the mean on HIGHER and LITER. Group the continents into a LITER for each continent using the regression co- dichotomy: West (N. Amer. and S. Amer.) vs East efficients and constant. (the others). For each dependent variable, produce a two-way table of cell means. Then, gener- B. Create appropriate product variables and ate row and column means(a) by the simple average test for parallelness of slopes for the regression of the cell means, and (b) by weighted averages, of LITER on BIRTHS using both SPSS REGRESSION and weighting by the cell. sizes. If the interaction BMDP2R. By hand, use the BMDP output to obtain the effect is significant for either criterion varia- regression equation for each continent using the bles, present an interaction plot of the cell regression coefficients and constant; and prepare means. using the SPSS output, produce the two an ordered regression summary table using the SPSS ordered regression tables for each dependent vari- output. able which resulted in the sums of squares given. NOTE: On SPSS runs, expand the workspace Note that the analyses will be "unbalanced". Whe._ available for transformations; for example, start effect does this have on the interpretation of the your run with analyses of variance. @SPSS*SPSS.SPSS,F 5000 Computer Assignment No. 8 Also, for the BMIP runs, in the REGRESSION para- Due on class meeting number 20 graph, include the sentences ENTER=0.0. Afifi and Azen present a four-group analysis REMOVE=0.0 and TOLERANCE=.0001. to ensure that all of covariance design on page 267, with the analy- variables enter the prediction equation. sis occurring on the next several pages of their text. Use both SPSS ANOVA and BMDP1V to reproduce Computer Assignment No. 5 their analysis. For the BMDP run, include con- Due on class meeting number 14 trast cards so that each of the 3 treatment groups The nations of the 6 continents differ in is compared to the Cr,ntrol group. In addition, adult literacy rates. Confirm this by rvaning an run the scattergrams for the groups with BMDP6D analysis of variance using SPSS ONEWAY. in addi- and sketch in the group-specific regression lines tion, do the following: (by hand). Note that the BMDP program provides the homogeneity of regression test, but that SPSS a. Enter VALUE LABELS and take OPTION 6 so lacks this feature. Produce the two ordered re- that the continents are properly identified on the gression summary tables used by BMDP2V for the printout. tests of means and slope (you will need to refer b. Using ./.05 level of significance, run to the SPSS output in order to do this). Also, Tukey and Newman-Kuels pairwise comparisons. produce an ordered regression summary table con- Explain '..by the results differ. forming to the SPSS order, but incl-.:Sing the test c. Prior considerations suggest the follow- for equality of slopes; in this table, use a ing contrasts: single denominator for all tests. NA versus SA EUROPE and OCEANIA versus NA and SA Computer Assignment No. 9 AFRICA versus average of all others Due on class meeting number 22 Test these contrasts with Bonferroni control of Type 1 error (i.e., use a significance level of The purpose of this analysis is to determine .05/3 per test). Explain the basis upon which you the contributions to predicting GNP of groups of reached each of these decisions. variables from the Quality of Life data set. The predictor sets are: Computer Assignment No. 6 AEXPECT, MORT, PHYS Due on class meeting number 16 B PRIM, HIGHER, LITER A. Afifi and Azen present data and analysis CFREE/NOTFREE(d.v. with NOTFREE including for a 2 x 4 design on pages 221-222. Confirm their partly) results using both SPSS ANOVA and BMDP2V. DA by C Product Terms E B by C Product Terms B. A three-way factorial design is presented by Afifi and Azen on pages 240-241. Confirm their Using SPSS REGRESSION, enter the sets of pre- analysis using both SPSS ANOVA and BMDP2V. The AB dictors in the above order (use EVEN inclusion interaction is significant at conventional levels; levels to force each set in on one step). Then, present a relevant plot and interpret this inter- by hand, set up an ANOVA summary table with appro- action effect. priate statistical tests. Finally, re-do the assignment using this order: B, A, C, E, D. In all, you will have two regression orders and two summary tables.

339

3 5 G Computer Assignment No. 10 Due on class meeting number 24 The problem is to compare, simultaneously, the Free and Non-Free nations on the basis of the 15 Quality of Life variables. The Non-Free 'group includes those nations which are Partly Free. Also, in this analysis, the continents are not used. Since the nations are currently categorized into 3 freedom groups, it is necessary, when using BMDP, to include a GROUP paragraph with a CUTPOINTS sentence: e.g., use 1.5 as a cutpoint to create two groups: those below 1.5 (i.e., the l's) and those above 1.5 (i.e., the 2's and 3's). For the analysis, use BMDP3D and include the HOTELLING sentence in the TEST paragraph.

Computer Assignment No. 11 Due on class meeting number 25 The problem is to determine how well the 15 Quality of Life variables can discriminate among the nations in terms of the Freedom variable. Use BMDP7M to conduct a Discriminant Analysis with the grouping variable being the 3 Freedom groups and the criterion variables being the remaining 15 variables (excluding Continent). When setting up the DISCRIM paragraph, use the following sentences only: ENTER=0.0. REMOVE=0.0.JACKKNIFE.

Computer Assignment No. 12 Due on class meeting number 26 Once again using the Quality of Life data set, perform a Principal Component Analysis using the following variables: GNP, ENERGY, BIRTHS, DEATHS, EXPECT, MORT, PHYS, PRIM, HIGHER, and LITER.Con- 'duct the analysis using BMDP4M. When setting up

. the FACTOR paragraph, use only the following sentences: METHOD=PCA. ROTATE=NONE. Also, include a PLOT paragraph with the sentence INITIAL =2. Repeat the analysis using SVGS. FACTOR. With the FACTOR procedure; include the statements TYPE = PA1 and ROTATE = NOROTATE.

Note 1: Codes for continents are: 1= N. Amer. 2= S. Amer. 3= Europe 4= Asia 5= Oceania 6= Africa

Note 2: Freedom Status codes are: 1= Free 2= Partly Free 3= Not Free

Variables are right-justified in their fields without decimal punched with the exception of population (cols. 6-10) and Annual Inflation Rate (cols. 26-30) which have the decimal punched. Thus, all variables may be read using an F5.0 format, except CONT and FREEDOM which must be read as.F1.0. Missing data are represented by -99., except for Freedom Status (column 5) which has one missing case which is coded 9. The data appear on pages 134-136 of the Almanac cited above.

340 Table 1.Adjusted Means, Standard Deviations, and ANCOVA Results

Computer Section Data Control Section N = 29 Source N = 17 Mean S.D. Mean S.D. F Prob. 48 Common Examination Items 31.16 6.80 29.43 8.72 .77 .38

I have come to understand many useful statistical techniques in this course* 4.47 .51 3.73 1.10 11.38 .00

I learned many new skills which will benefit me in the future* 4.31 .55 3.76 1.16 6.85 .01

I had difficulty understanding the material presented in class** 2.90 1.25 2.05 .94 6.02 .02

This course was particularly well organized* 4.48 .74 3.59 1.23 9.64 .00

This course increased my interest in statistics* 3.98 .96 2.80 1.32 15.00 .00

I enjoyed my experience in this course* 3.68 1.00 2.55 1.42 11.81 .00

I had difficulty understanding the material presented in the textbook** 1.45 .73 2.30 .93 12.25 .00

I spent too much time and effort on this course** 2.45 1.30 2.23 1.19 .36 .60

I had adequate background preparation to take this coursa* 3.42 1.24 2.81 1.19 2.89 .10

The material in this course was covered too superficially** 3.81 .83 3.79 .85 .00 .95

*Scale is 1-5; 5 = agreement. **Scale is 1-5; 1 = agreement.

Appendix Description of the Data Set and Computer Assignments The data set EDMS*STAT.QUALITY contains a number of variables whichare related to the general quality of life within nations. The source of the data is the 1979 "Information Please Almanac." The names of the variables and the units of measurement are listed below.

Column Code Variable Fnme Unit Unit 1-3 Code $ for Nation 4 CONT Continent See Note 1, below 5 FREEDOM Freedom status See Note 2, below 6-10 POP Population, 1976 Millions 11-15 AREA Area Thousands of sq. km 16-20 GNP GNP per Capita, 1976 U.S. Dollars 21-25 ENERGY Energy Consump. per Capita, 1975 kg of coal equiv. 26-30 INFL -AnrWal Inflation Rate, 1970-76 Percent 31-35 AGLA3 Labor in Agriculture, 1970 Percent 36-40 BIRTHS Crude Birth Rate, 1975 per 1000 Population 41-45 DEATHS Crude Death Rate, 1975 per 1000 Population 46-50 POP2 Proj. Population in 2000 Millions 51-55 EXPECT Life Expectancy at Birth, 1975 Years 56-60 MORT Infant Mortality, 1975 per 1000 61-65 PHYS Population per Physician, 1974 66-70 PRIM Persons in Elem. School, 1975 Percent of Age Group 71-75 HIGHER Persons in Higher Ed., 1975 Percent of 20-24 Population 76-80 LITER Adult Literacy Rate, 1974 Percent

341 358 A COMPUTER-BASED TUTORIAL ON MATHEMATICAL INDUCTION*

by J. MACK ADAMS AND MARVIN LANDIS

Department of Computer Science New Mexico State University Las Cruces, New Mexico 88003

Abstract application-oriented and seem not to have acquired a working knowledge of induction in their mathe- The development of a tutorial on mathematical matical prerequisites. induction is described. The tutorial is based on the notion of informal verification of correctness The specific problem stated above led to an of programs with one simple loop. investigation of the following approach: reteach- ing induction using the students informal notions of program correctness as a relatively concrete I. Introduction basis and only then using induction as a tool in more formal proofs of correctness. Some success This paper contains a description of a with this approach led to the Following hypothesis: computer-based tutorial on mathematical induction and a discussion of the motivation for such a Hypothesis tutorial. It also describes the development of the tutorial, especially the problems encountered in Concrete examples of informal justification the development. of program correctness of programs involving one simple loop can be used as an effective basis for The tutorial was authored in the latter stages teaching mathematical induction to students with a of a project to develop computer-based learning knowledge of computer programming. material in computer science and mathematics. It represents an attempt to apply the techniques and To test this hypothesis we decided to develop expertise gained in preparing tutorials on lower a computer-based tutorial that embodied the pro- level material, beginning programming and trigo- gramming approach. nometry [Mac81], to a topic that seems more difficult to teach. III. Authoring the Tutorial

II. Motivation and Approach The authoring followed the style developed at the Educational Technology Center at the University Motivation for developing the tutorial comes of California, which had proved effettive in the from a general difficulty encountered by the development of previous tutorials for our project. authors and their colleagues in teaching induction Although we encountered difficulties that were and, more specifically, from the following two much more challenging than those of previous tuto- problems: rials, they did not have to do with deficiencies of the authoring style and thus no changes were The General Problem necessary.

The abstract nature of the usual approach tr Before actually beginning the authoring we teaching induction does not seem effective for studied various approaches to teaching induction, application-oriented students. More concrete ex- particularly geometric approaches [Spe69, Wis70] amples than the usual formulas proved by induction since our material is designed for microcomputers seem necessary. with graphics capabilities. We also studied the origin of mathematical induction [Bus17, Caj18]. The Specific Problem These studies provided good general background but no specific techniques or examples amenable to our Using induction to teach proofs of program approach. correctness [A1g78, Wir73] to students in upper division computer science courses is quite diffi- We finally decided to use a relatively simple cult, since these students are generally very programming example in flowchart form. The example, given in Figure 1, was first used to introduce the notation and the idea of assertions at *This paper is based upon w,:,rk supported in part key points in the program. Then it'was used for by the National Science Foundation under Grant No. the informal verification of the output assertion. SER-8005317.

342 { N > 0 } COM CHART

RESP

TABLE

Figure 2

Port Layout True Essentially, we decided on an induction on the in- { S = put parameter, rather than the usual induction on an "iteration variable ". Figure 1 After authoring the first version of the Simple Pr.-. with Input and Output Assertions tutorial using the approach described above, the tutorial was presented to students and colleagues in a colloquium. One of the authors simulated execution of the tutorial using the blackboard as It thus provides a rela,:ively concrete example of the screen, while the other author recorded an induction argument. audience response to questions posed within the tutorial and also suggestions for changes. This The first major change in our preconceived had not been done for our previous tutorials and notions about the approach came when we thought we highly recommend it, particularly for tutorials seriously about using the usual technique of a rn difficult subject matter. A second version was loop invariant, or inductive assertion, in the prepared, based on response from the Colloquium, informal verification of correctness.We realized and a brief description of this version is given the degree of abstractness was only slightly less in the next section. than that of traditional approaches to induction, and we recalled that students did not usually IV. Brief Description of the Tutorial react well to loop invariants. In short, loop invariants do not seem strongly related to the, After displaying the title and credits, the perhafs unconscious, pr:Dcess by which students simple program given in Figure 1, without the justify the correctness of programs. assertions, is displayed in the port named Cliputi (see Figure 2). The terminology of assertions is We eventually reached the conclusion that then introduced as the assertions are added to the most students justify the correctness of a simple display. loop by: The student is then invited to verify, by 1. tracing the simple cases of 1, 2, or 3 computation, the output assertion for N= 1, 2, iterations, and then and 3. These results are recorded in a table in port TABLE. 2. tracing a general case by checking the final iteration, assuming that all has At this point the student is asked to verify gone well in previous iterations. the output assertion for a value at the upper end of the table, and subsequently asked if he would On this basis we decided to: like' to verify the result when N is 209, a large value chosen arbitrarily by the authors. If the a. motivate the basis step of induction by answer is "no", we heartily endorse the student's checking simple cases of a few iterations, reluctance, and otherwise we question the advisa- and bility of taking on this task.This is used to motivate the need for :.4ome sort of general cor- b. motivate the induction step by checking rectness argument. the final iteration for an arCtr..lry value of an input parameter, N in Figure 1, with The studentthen participatips in the develop- the explicit assumption that the algorithm ment of a correctness argument for an N of k+1, has worked correctly to that point.

343 360 assuming the algorithm worked when N was k and V. Summary tracing another iteration. This is done in the following stages: We have described the development of a tutorial on mathematical induction based on informal 1. the output assertion when N is k is verification of correctness of a program with one "backed-up" tr. an intermediate assertion, simple loop. The resulting tutorial has also been preceding the test, which we hope the summarized. We hope that relating the difficulties student can correctly identify as being associated with the development of a tutorial over the same as the output assertion. challenging subject matter may be helpful to others contemplating such developments. 2. The intermediate assertion is then taken through the false branch of the test and We would like to recognize the valuable the body of the loop, resulting in a new suggestions of colleagues and students, particular- intermediate assertion and hence an output ly those of Pifessor Warren Krueger and Mr. Jack assertion for an N of k+1. Medd.

3. The student participates in the algebraic manipulation necessary to obtain the form References of the output assertion upon which correctness has been based. [A1g78] Alagic, Saud and Arbib, Michael A. The Design of Well-Structured and Correct The resulting "induction step" is displayed for Programs, Springer-Verlag, New York future use in port COM. This marks the beginning (1978). of severe screen space problems since we have only port RESP avi)able for use. [Bus17] Bussey, W. H. The origin of mathematical induction. The American Mathematical At this point the student is asked to apply Monthly, XXIV, 5 (May 1917). the induction .tep to the specific case when N is 3, which has previously been verified by computa- [Caj18] Cajori, Florian. Origin of the name tion. Thus he obtains correctness when N is 4 "mathematical induction". The American without computation of S. He is then asked to Mathematical Monthly, XXV,5 (May 1918). apply the induction step twice, starting again when N 3, to obtain correctness when 11 is 5. [Mac81] MacKichan, Barry, Adams, J. Mack, and Hunter, Roger. Starting a computer We then observe that the induction step is based learning project. Proceedings of easier to apply starting from the value of 1 for N. NECC 1981, National Education Comput- This is a somewhat ungainly transition, but we felt IRT6Wnrence, Denton, Texas (June it was desirable to first apply the induction step 1981). to obtain results for 4 and 5, which had not been previously verified by computation. [Spe69] Speck, Royce A. The number of squares on a goeboard. School Science and Mathe- The student is then asked how many applica- matics (February 1969). tions of the induction step are needed to verify correctness for a value of 209 for N, and subse- [WiriJ] Wirth, Niklaus. Systematic Programming: luontly for an arbitrary value, m, of N. An Introduction. Prentice-Hall, Englewood Cliffs, N.J. (1973). The process is then summarized in the form of a basis step, induction step and application of [Wis70] Wiscamb, Margaret. A geometric introduc- the induction step, and given the name "mathemati- tion to mathematical induction. The cal induction". Mathematics Teacher (May 1970).

The length of the specifications for the version of the tutorial described above is approximately the same as those of our previously developed tutorials, so we expect the running time of the implementation to be about the same, 20-30 minutes. Implementation is now underway on the IBM PC using UCSD Pascal Version IV and our version of the packages Textport and Graphport obtained from the University of California at Irvine. We anticipate completion in time for trial usage of the tutorial in the latter half of the semester beginning in. January of 1983. There seem to be no particular implementation problems except t11,1 one of screen space mentioned previously.

344

;1. IMPLICIT FUNCTIONS AND COMPUTER GRAPHICS

Sheldon P. Gordon

Suffolk Community College Selden, NY 11784

Abstract In the present paper, we will focus The present paper deals with the use of on an algorithm which will lead to the the computer to generate the graphs of graph of the majority of implicit func- implicit functions in the form F(x,y) = O. tions. We will also discuss some of the In particular, several algorithms are limitations and problems with such a discussed which can be used as the basis program. Finally, we will demonstrate for a computer graphics program which will the use of one such program in producing produce the'graph of most such implicit the actual graphs of several implicit functions. The value of having such pro- functions. grams available for use in both introductory and intermediate calculus classes Suppose we start with an implicit is also discussed. Finally, a variety function in the form of examples are displayed to illustrate F(x,y) = O. the results of such a program. Since this represents a functional relation with y as a function of x, then there should correspond (at least) one One of the least satisfying topics value of y for each value of x. In for students in calculus has to be that order to start the graph of such a func- of implicit functions. Most students tion, it is necessary to determine one feel that there is nothing tangible for point on the curve. To accomplish this, them to grab hold of with the topic. The we arbitrarily select a value for x, say old standby of performing algebraic mani- x and seek to locate a corresponding pulations is usually useless on all but 0, the simplest expressions; for example, value yo, by applying the Bisection Method consider 4 3 to the equation F(x0,y) = 0. We first xy + xy = 1. There is nothing to visualize in the way locate the desired root by considering of a graph and no elementary method to the sign for F(xo,y) on a sequence of solve for one of the variables in terms intervals Cy, y+k3 with sufficiently of the other. The fact that an Implicit smallk with y between -1000 and 1000, Function Theorem guarantees the existence say. Once such an interval is found, of such function or its derivatives the Bisection Method zeros in on the subject to appropriate sets of conditions desired root fairly fast. is small consolation. Now that a particular point (x0,Y0) Fortunately, the power of sophisticated computer graphics now presents us on the graph has b.:en found (presumably), with a tool which can draw the graph of we seek to continua the graph by deter- most implicit functions. This capability mining additional points and connecting allows us to add a valauble new dimension them. Since the graph can continue both to the subject at both the elementary and to the left and the right, we will have intermediate calculus levels. At the to consider both directions. Suppose introductory level, the availability of we continue it to the 'eight initially. such a program provides the student with We consider the sequence of points x. = xo + ih (i=1,1,) and attempt to a better grasp of the concept of implicit 1 function because he or she can actually determine the corresponding values yi. "see" it. At the more advanced level, the underlying algorithm behind such a We could continue to apply the Bisection program can be as valuable as the program Method to generate the new points, but itself in teaching the students to analyze it converges far too slowly to use repea- and anticipate the possible shapes of tedly. In addition, it also requires curves which can arise when dealing with finding two values of y bracketing each implicit functions. successive solution yi. Instead, we will apply Newton's Method in the form

345 362 fairly easily using a simple modification of the - F(xi,Yn)/Fy(xi,Yn) above procedure. If the Method does not converge where F is the partial derivative of F with res- for a particular value of x, say xN = xo + Nh, Y pect to y. (as determined by either a count on the number of iterations being performed or on the relative With this approach, we note that each previ- changes in the values of the successive iterates), ously determined value yi at (xi,yi) is used as the then it makes sense to skip on to the next point using y as the initial esti- initial approximation to the next root at xN +l while still N-1 mate. If Newton's Method converges at this point Since h is taken very small and the (xi+h, yi+l) within a reasonable number of iterations (cer- curve is presumably continuous and fairly smooth, tainly fewer than 10), then yN can simply be set yi+1 should be close to yi and the convergence will to the average of yN_1 and yNil for an acceptable usually be extremely rapid, if a solution indeed graph. exists. As an alternative or even subsequent "fix" for However, the "if" can be a big one. A great the above problem, we can also resort to a secon- many oddities can occur when one deals with impli- dary method. One such might be to call on the cit functions and provisions must be made to ac- Bisection Method again. Another might be to use count for them, if possible. First, Newton's an extension of Newton's Method such as the one Method breaks down when the denominator, Fy(x,y), i n described by the present author in C13 which avoids is (nearly) equal to zero. In the present instance, the problem of Fy = 0 while developing a cubically this corresponds to a vertical tangent which can convergent process based on approximating a curve take any of the forms illustrated in Figure 1.The with parabolas instead of tangent lines. cases shown in Figure la and lb can be handled If all of these techniques fail, it is probable that the curve actually bends back on itself, as shown in Figure lc. To attack this case, we have to locate a point on the second (or later) branch of the curve. This can probably best be done by recourse to the Bisection Method once again where we would have to take into account the likely orientation of the anticipated branch. Thus, if the y's are decreasing before the turn, then the search for the next point, also corresponding to x = xN, should occur from y = -1000, say, up to y = yN. Once such a point has been located, we simply set the step h = -h and continue on to the left.

If this latter attempt also fails, we should probably just accept the inevitable and not try (a) to extend the curve any further at that end. Instead, we should return to the original starting point (x,y) and attempt to extend the graph to o o the left using the same algorithm as above.

We note that the above efforts should essentially cover the case of a vertical asymptote as well. If the graph has the appearance shown in Figure 2a, then the continuation might be picked up with a "jump" across the singularity. If it looks like Figure 2b, then it is unlikely that the algorithm will converge quickly enough to have points on either side of the singularity connected. Alternatively, we can provide for such a singularity by keeping track of the relative sizes of the values for y and if they exceed some preset limit, then the graphing routine should allow for the "pen" to be lifted and reset across the jump.

There are several other possibilities which (c) can also occur when one deals with implicit functions. Primarily, these involve the case of a

FIGURE 1.

346 363 Another major problem, which appears to be insurmountable in terms of any reasonable modification of the present method, involves the situation where the curve consists of two or more distinct and non-intersecting segments. For example, consider the case of the simple hyper- bola. The algorithm will trace out one portion of such a curve, but will miss entirely the other portions. Further, any effort to take such an eventuality into account would seem to make the resulting program incredibly unwieldy.Conse- quently, we will ignore this case and hope that it does not arise or simply be content with tracing out a single portion of the curve.

A flow chart for the above algorithm is presented in Figure 4. From this, it should not be too difficult to write the corresponding computer graphics program to implement the method on most of the available high resolution computers. The author has already done so for a PDP 11/34 system supporting Tektronix graphics terminals (model 4006) as well as for the TRS 80 Color Computer. However, it should be emphasized that (a) the actual procedure is an extremely time-consuming one, particular if a reasonably large FIGURE 2. number of points are to be located. In fact, the amount of number crunching is probably comparable to that needed to graph three dimensional horizontal cusp, as shown in Figure 3.Both of surfaces. As such, it is not well suited for these cases should he handled by the procedure microcomputers on a real-time basis unless one discussed above whet the curve turns back onto a is using a preprocessor to handle the calcula- new branch. However, if the curve bifurcates at tions. Alternatively, it is possible to generate a particular point, then the algorithm will con- such graphs in advance and 'record" the images tinue along just one of the branches and it does for later demonstrations. Unfortunately, this not seem possible to design any reasonable modi- certainly involves a tremendous loss in spontan- fication of the algorithm which can pick up such eity and a consequent reduction in impact on a a situation. class. 7,-

It is worth noting that there is an alternate algorithm which also works quite well in producing the graphs of most implicit functions. In fact, the author has found that it is usually more effective than the one presented above. In particular, the idea is to transform the given function F(x,y) = 0 into polar coordinates in the form F(r cos9, r sing) = 0 and graph it with r as the implicit function of 9. A natural choice for a starting point might be 9 = 0and a full loop will often occur for 9 between 0 and 21r. One further advantage is that 4. most of the shapes shown in Figures 1, 2 and 3 could be traversed smoothly in polar coordinates while they present: major difficulties in rectan- l% gular coordinates. One potential problem does arise with a vertical cusp and a procedure for handling this would be analogous to the method used earlier when a rectangular curve bends back on itself. Another advantage of the polar representation is that it would be oftrq easier to check for a complete loop of a closed curve (simply compare r for 9 = 0 and for 9 = 217, say.) (a) There are several further complications that occasionally arise, particularly with the polar representation. For one, if we utilize a function involving the trigonometric functions especially, FIGURE 3. then there are inherent difficulties built into

347 364 the use of Newton's Method in terms of instability, Reference: If we have a point on such a curve anywhere near Gordon, S. P. and E. Von Escfio, On a cubic (and that can be quite far, relatively) a root of extension of Newton's Mettoii. (lubmittW. the derivative, then the corresponding tangent line will shoot off to a point far distant. As a osult, m^wfon's Method can easily converge tl a Acknowledgement:

uccesion of distant pufat , di ,, branr.hes Tfie author gratefully acknowledges the of the curve and the resulting graph, when these support provided in part for the developments points are connected, will be highly unreliable, discussed in the present paper by the State Uni-

albeit quite interesting in shape. versity of New York under a grant from its pr , 'm for Improvements in Undergraduate Instruction. A second difficulty is a computational one. In order to pick up the initial point, a search method has been suggested to bracket the value, involving a large range of values for y or for r. However, if the given function involves exponen- tial terms such as EXP(Y), then the capacity of the computer can easily be exceeded causing an overflow error. Similarly, an all-purpose program applied to a general expression F(x,y) = 0 cannot anticipate all possible instances where a function is not defined - say if LOG(Y) occurs or if frac- tional exponents are involved.

rFIDvo US:NG In view of these comments, it should be clear EtISECTION TETHOD that while this type of program can be an exciting and instructive one for students, it has to be Fl%3 Y(L1) USING used with considerable advanced planning or a METHLIO willingness to encounter errors with good humor.

Incidentally, it might appear strange that the major emphasis above was devoted to the rectangular case despite the fact that the polar form is usually more effective and easier to apply. The author feels, though, that from a purely pedagogical point of view, the rectangular approach appears far more natural to the students. Almost all problems in calculus dealing with implicit functions are in rectangular form and, in fact, in introductory calculus, the student encounters implicit functions long before he or she sees polar coordinates. Thus, if the student is to see the program rather than simply the output, then the use of the rectangular form is to be preferred. This is probably even more important in intermediate calculus where the emphasis is likely to be placed more on the possible behavior of such functions than on the actual appearance of some of them.

Finally, we illustrate the use of such a f4lr__=H2 program to produce the graphs of a variety of RESET TU implicit functions. In Figure 5, the graph of XD,Y0 { 3 3 4 4 (x+y) + (x-y) = x + y is shown. In Figure 6, the implicit function displayed is given by 5 3 5 x + 4xy - 3y = 2, while in Figure 7, we show 7 32 6 the graph of x - 4xy + 4y = 1. Lastly, in Figure 8, we show the graph of the implicit func- Fig.a...! 4. 4 tion xy - sin(xy) + x = 1. Once such graphs are available, an immediate followup would be to investigate the results of modifying the values for the constants on the right to examine the corresponding families of curves so generated. However, we will not do any of this here.

348 36 Y RANGES FROM-4.15882 TO 4.13787 Y RANGES IN STEPS OF FROM-1.23899 - .831589 TO 1.15388 IN STEPS OF .239207

X RANGES FROM .924778E-6 TO 4.16212 IN STEPS OF .416212 X RANGES FROM-.306732 TO 1.34491 IN STEPS OF .165165

3 4 4 3 5 FIGURE J: (WI) (X-Yr; Y Y 5 FIGURE 6: X 4XY - 3Y 2

Y RANGES Y RANGES FROM-1.03143 FROM-4.13839 TO 1.03693 TO 4.1336 IN STEPS CF IN STEP7S OF .286236 .832373

X RANGES FROM-.993599 TO 1 28314 IN STEPS OF .227674 X RANGES FROM .330507E-2 TO 1.63301 IN STEPS OF.16297

7 32 6 FIGURE 7: X - 4XY 4Y FIGURE G: XY4 - 5IN(XY) + X . I

349 INTERRUPT DRIVEN I/O PROJECTS IN AN ACM '78 CS4 COURSE*

by Greg Starling

Department of Mathematics and Computer Science Western Carolina University Cullowhee, North Carolina28723

Abstract (d) Computer Architecture (e) Example (an actual microcomputer system) In the ACM '78 curriculum recommendations for At Western Carolina University we are now in CS4, Introduction to Computer Organization, the our fourth year of teaching this course. In the three objectives of the course involve some aspect 1979-80 academic year we used a Z-80 based micro- of simple input/output devices at both the software computer (Exidy Sorcerer and TRS-80) to support the and hardware level. These recommendations also hardware and programming topics of the course. In call for the study of a simple minicomputer or 1980-81 we established a microcomputer laboratory microcomputer system. with 15 6502 based APPLE II plus microcomputers and a CORVUS Systems hard disk network storage facility In a computer science program which is ori- under NSF Grant number SER-80-04761. In both years ented toward a time-sharing facility some objectives we found that some of the topics dealing with input, of the course are impossible to meet. Even with output and logic design were rather artificial with- microcomputers some of the aspects of I/O program- out some means of non-keyboard, non-printer or CRT ming do not come alive without a facility for I/O. In fact, the concept of interrupt driven binary input and output ports with interrupt systems remained a mystery to most students, even capabilities. We describe some hardware and soft- some of the best ones. In 1981-82 it was decided ware I/O projects which we use in our version of during the logic design and memory organization CS4. Included in the paper are details to allow phases of the course to design a binary input /output others to implement these projects. We use APPLE II port with interrupt capability using 6520 Pas plus microcomputers, but the principles could be (Peripheral Interface Adaptors). After studying adapted for other microcomputers. logic design, memory mapped I/O and the organization of the 6502 and 6520 chips the class, under the Introduction guidance of the instructor, established the design criteria and designed the memory address decoder to An abbreviated description of the ACM Curric- place the I/O port in a peripheral expansion slot ulum '78 CS4, Introduction to Computer Organization, of an APPLE II computer. The instructor then did course is: the mechanical design and fabrication of Coe I/O Course Objectives. device. The students then wrote program exercises using the port. (a) to introduce the organization and structuring of the major hardware components of Hardware. computers; The design criteria decided upon were: (b) to understand the mechanics of infor- (a) Two bytes for input and two bytes. for out- mation transfer and control within a digital put to facilitate easy transfer of signed numbers computer system; and as large as 32,767 in magnitude and to provide both (c) to provide the fundamentals of logic a status register and a data register for programmed design. I/O experiments. This requires two PIAs. Topics. (b) Toggle switches for input and LEDs (light- emitting diodes) for output so that the relation- (a) Basic Logic Design ship between characters on the video screen or key- (b) Coding (BCD, ASCII, etc.) board can be easily observed by way of the I/O port. (c) Number Representation and Arithmetic (c) Momentary pushbutton switches and LEDs for interrupt requests and interrupt acknowledge signals fhis material is based upon work supported by using the peripheral input lines and peripheral the National Science Foundation under Grant No. control lines of the Pas. It was agreed that each SPE-82-63111. Any opinions, fiLuings, and conclu- byte of the input port and each byte of the output sions or recommendations expressed in this publica- port would' have an interrupt request button and tion are those of the author and do not reflect the acknowledge LED. One button would have normallyopen views of the National Science Foundation. contacts and one would have normally closed contacts to provide the possibility of an

350 3 , interrupt on falling edge or on a rising edge through seven has a 256 byte page of memory allo- signal. cated beginning at $ C 1 0 0 and ending at (d) A slope-front, Aesktop console to mount $ C 7 F 0 (refer to chapter 5 of the APPLE II the switches and LEDs for ease of use while con- Reference Manual), There is also a 2K byte block trolling the computer from its keyboard. of ROM space reserved for the use of all of the (e) An edge-card contacts circuit board peripheral slots so that service routines longer which fits an APPLE computer peripheral expansion than 256 bytes can be accommodated.Although we slot to hold the PIAs and address decoding logic. did not implement either a 256 bytes primary page (Soldering should not be done by a novice with a or 2K byte expansion block in the class, we did soldering iron. Wire-wrapping is probably the study their use and one student implemented both best choice, but use one level wire-wrap sozkets in a one hour summer school special topics course. so as not to interfere with other peripheral cards The EPROMs are visible in figure 4. inserted in adjacent slots.) (f) Use a 50 conductor ribbon cable (2' - 3' The details of the hardware design are con- long) and IDC (Insulation Displacement Connectors) tained in the following tables, figures and to interconnect the console and the card. diagrams. Table 1 and Figure 2 give the pertinent architectural and organizational details of the Each of the eight peripheral slots on the 6520 PIA. Figure 3 contains information on inter- APPLE Computer's expansion bus is allotted sixteen facing the PIA s to the APPLE bus.For more memory mapped I/O locations beginning at location details than these, the Motorola data sheets for $ C 0 8 0 ending at location $ C 0 F F. For slot the 6520 (same as 6820) PIA and the APPLE II k the address range is$C0n0-$COn F, Reference Manual 1 should be consulted. where n = $ k + 8. In addition, each of slots one

Control Register Bit Register Address Lines X = Don't Care Selected

RS1 RSO CRA-2 CRB-2

0 0 1 X Peripheral Register A Data Direction Register A 0 0 0 X

0 1 X X Control Register A

1 0 X. 1 Peripheral Register B

1 X 0 Data Direction Register B

1 1 X X Control Register B

Table 1. PIA configuration information

7 4 l 3 .2 1 J 0

CRA IRQA1 IRQA2 CA2 Control DDRA CA1 Control Access

CRB IRQB1 IROB7 CB2 Control ,, DDRB CB1 Control Access

Notes. Data Direction Access Control Bit (CRA-2, InterrutPlasoRA-7,CRARB-6. CRB-2). Bit 2 in each control register (CRA, These flags are set by active transitions on CA1, CRB) allows selection of either a Peripheral CA2, CB1, CB2 respectively. They are reset by Interface Register or the Data Direction peripheral read data operations. Register when the proper register select signals are applied to RSO and RS1.

Figure 2. PIA Control Registers (CA1,CA2,CB1,032 control)

351 +5 v

1K 330 1K

PBS/NO _a_ -4)3. 330 P xTyiS,ical

tad CB-2 1'BQI +5v CI, VCC FBI ou C2> VSS IRQ PB2

PB3

PB4

PB5

PB6 . PB7 PIAO (6520) /W 1K n Inn Typical"AM PAO DO <:: DO PA1 D1 <7.) D1 PA2-0Ar0---111. D2 <::) D2 -14-°-111. D3 D3 PA3 --O D4 .CD. D4 PA4--0 D5 D5 -14r-0-111. D6 D6 PA5 D7 D7 PA6 AO RSO Al [D., EREICSOCS1 +5v E> A3 E2> A4 ED, 13 Y1 To CSO and CS1 of PIA1 LS139 NC

CND E> NC

DEVICE SELECT E>

Notes.

1. All resisters are in 10 pin, 9 resist.= 4. Push button switches are NC (normally closed) SIP packs. on PIA'. 2. Inverters are 74LS04. 5. Toggle switches are SPDT, long handled, 3. LEDs are Tl 3/4 red miniature. 6. LEDs and inverters on outputs are powered from an external +5 v supply.

Figure 3. Typical connections for one of the PIAs Figure 4. The finished product

Software Projects. (1) Configure the PIAs in the I/O port. (See Many of the programming exercises in an chapters 11 and 12 of 6502 Assembly Language Pro- gramming3.) assembly language programming course at the college This gives quite a dramatic effect, sophomore level are repetitions of the exercises because the LEDs are all lit on power up, and all go out upon successful configuration. done in the beginning programming course in the (2) Load the APPLE's accumulator from the freshman year. Usually, only those exercises involving code conversions are significantly input switches and transfer the contents to the output port. This may be done for both an eight different from those done in the first course. bit transfer or a sixteen bit transfer. That is because traditional assembly language programming courses have been taught on mainframe, (3) Do exercise (2) in an unending loop so central, time-shared computers.This has prohi- that the output changes continuously with the input (flip a switch on and the corresponding LED bited the undertaking of most basic I/O and lights). interrupt programming. In fact, in most systems the students must be taught how to interface with (4) Take a character from the keyboard and the system's I/O routines just to get input to and transfer it to tha output lights. The COUT sub- output from their assembly language programs. routine in the APPLE's Monitor or Autostart ROM is useful for this exercise. There is no such thing as raw binary input and (See page 61 of the APPLE II Reference Manuall.) output. And as far as interrupt programming is concerned there is none. These are two deprivaticns (5) Take a code from the input switches and that insulate computer students from some of the transfer it to the APPLE's video screen either as most exciting capabilities of computers. Fortu- an ASCII character or as hexadecimal digits.Tne nately, the growth of microcomputer usage in the subroutines PRBYTE, PRHEX, PRNTAX, and RDKEY (see introductory computer organization course gives us pages 61-62 of the APPLE II Reference Manuall) are used in these exercises. the possibility to rectify these shortcomings. With the I/O port described here attached to an (6) Take a binary number from input switches, APPLE microcomputer both I/O programming and rotate or shift, display on the LEDs and repeat interrupt programming exercises are easily under- after an appropriate time delay so that the bits taken. Indeed, basin data transmission programming, move slowly enough to observe. The APPLE's WAIT subroutine (see page 63 of the APPLE II Reference an'increasingly more important partof programming, 1 is more easily understood. Manual ) is useful here. (7) Take two 8-15 bit numbers from the Some of the programming assignments we use switches, add them and display the result on the with the binary I/O port are as follows. LEDs and the screen. The sixteenth switch is used

353 as a strobe bit to signal the program that an Listing 1 configures the r/ As for APPLE operand is set on Ale switches. This gives peripheral clotnumber 2 and tvInsfers the input experience on programmed or polled input status switches valuesto the output LEDS. checking as opposed to interrupt driven input Listing 1. acceptance. (8) Configure the control registers of the $9600 LDA #0 PIA s for interrupts (see figure 3). The control $9502 STA $COA1 ; SELECT registers contain interrupt enable bits (CRA-0, $9605 STA $COA5 ; DATA CRB-0, CRA-3, CRB-3) interrupt status bits (CRA-7, $9608 STA $COA3 ; DIRECTION CRB-7, CRA-6, CRB-6) direction select. bits (CRA-4, $960B STA $C0A7 ; REGISTERS CRB-4) and active transition bits (CRA-4, CRB-4). $960E STA $COAO ; DATA REGISTERS For more details see pages 11-15 through 11-20 of $9611 STA $COA4 ; A WILL BE INPUT 6502 Assembly Language Programming3. $9614 LDA #$FF (9) Configure the two PIA s so that the binary $961A STA $COA6 B WILL BE OUTPUT. I/O port is two 8 it input/output ports with CA1 $961D LDA #04 interrupts enabled on both PIA s.Modify exercise $9611 STA $COA1 ; SELECT (3) so that input/output port 0 and input/output $9622 STA $COA5 ; DATA port 1 can interrupt each other. Input and $9625 STA $COA3 ; REGISTERS corresponding output comes from first one and then $9628 STA $C0A7 the other depending Upon which one had its IRQ $962B LDA $COAO ; READ LOWBYTE SWITCHES button pressed last. An added feature to this $962D STA $COA2 ; TRANSFERTOLOWBYTELEDS exercise would be to light the ACK LED of the $9631 LDA $COA4 active port. This is controlled by making CA2 an $9633 STA $COA6 ; TRANSFER TO HIGH BYTE output signal on both PIAs. We have not tried this ; LEDS particular feature, but we believe it should work. STA $COA6 ; TRANSFER TO HIGH BYTE

The program might be written to poll the status ; LEDS bits (CRA-7) on the two ports to determine which JMP$962B ; REPEAT FOREVER (UNTIL issued the IRQ. ; RESET) There are several cautions about interrupts Summary. on the 6502 MPU and the 6520 PIA. Upon responding The binary I/O device provides the means for to an interrupt the 6502 disables interrupts, so very basic input/output and interrupt programming to make it possible for the other port to interrupt exercises which are not traditionally available an interrupt service routine bit 2 of the P (flags) to computer science students. With it we believe register should be reset. Also the service routine that students get a better understanding of code should read the data register of the port sending conversions, the I/O interface and interrupt driven the IRQ to reset the interrupt status bit, CRA-7 systems. It prepares them better for the required and thus deactivate the IRQ for that port. upper division courses and also makes it possible When the interrupt service routine is loaded for them to start at a higher level in some of the into the APPLE's RAM one must also load the address hardware oriented elective courses. of its first executable statement into locations Plans for the future are to continue the $3FE and $3FF. An IRQ causes an eventual jump to projects but to change from the PIAs to a VIA the location stored there. This. is because the (Versatile Interface Adaptor) because it has all locations $FFFE and $FFFF which hold the true the features of the PIA, two interval timers and a interrupt vector, $FA40, are in ROM space on the serial output in addition. It will allow us to APPLE II. On interrupt request execution is have timed interrupts exercises which will be useful vectored to that location which eventually leads in studying time-sharing in a systems programming to an indirect jump to the location whose address course. was stored at $03FE $03PF. Refer to page 143 of the APPLE II Reference Manual.' If one were to implement the page of memory allocated to an APPLE REFERENCES. expansion slot, then the service routine could be 1. APPLE II Reference Manual, Copyrighted 1979, located in there. If the routine is too large for 1981 by APPLE COMPUTER INC. this 256 byte page, then one would have to include a 2K expansion ROM (or RAM) so that the vF.rvice "Add a Peripheral Interface Adapter to Your routine could jump out of the single page Llto that 2. Apple II," by Kenneth J. Ciszewski, BYTE 2K of memory. This 2K of address space is reserved Magazine, January 1982. so that any board plugged into an expansion slot can switch its own 2K of memory into it. The proper 6j9/j±nE±Etnhlnlivaiell,2rogra14mmin by Lance A. protocol for this substitution can be found on 3. Leventhal, Osborne/McGraw-Hill, 1979. pages 84-85 of the APPLE II Reference Manual.' We provided a unique decoding of the special location address SCRIPT. One student who took the CS4 course has subsequently done that.He put configuration routines for the PIAs and some of the exercise routines listed above into 2716 EPROMS. We actually used 2716 EPROMS both for the single page and for the 2K expansion ROM space because our EPROM programmer doesn't program the 256 byte 1702 EPROMS. ASSEMBLY LANGUAGE ON THE APPLE -- A THOROUGH INTRODUCTION

W. D. Maurer, Professor The George Washington University (SEAS) Washington, D. C. 20052

Abstract ented call and return instructions, push and pull pop) instructions, and status An increasing number of colleges ana flags (an improvement on condition codes). universities are setting up APPLE computer Even some more traditional assembly lan- laboratories. These are being used for tike guage concepts, such as immediate addres- teaching of BAS/C, as well as for experi- sing for comparing and subtraction, indi- mentel projects in, wide variety of cour- rect addressing, and direct memory incre- ses. Assembly language, however, does not ment, decrement, and shift, are available seem at this time to be widely taught on on the APPLE and not on the 4341. Finally, the APPLE. One reason forthis may be that there is the sheer challenge of working on until recently the APPLE Corporation did a micro -- learning to do without multiply not produce a good assembler of their own. and divide, without floating point, and, One was forced to go to other, organirat- in the case of the 6502, without 16-bit tiona, such as Lazar Systems, which act as operations or add-without-carry. distributors for assemblers written. independently. The present paper outlines the Existing Literature resultm of a project involving the preparation of a thorough syllabus for the tea- The existing literature on the as- ching of assembly language on the APPLE. sembly language of the 6502 (the particu- As a measure of the success of this ef- lar microprocessor used in the APPLE) was fort, the department in which this author quickly found wanting. There are several teaches has just instituted a new course short books on APPLE assembly language, in assembly language programming of micro- many of which have no exercises, and none computers, to be offered as an alternative of which was felt to be thorough enough. to the study of IBM 370 assembly language. (Our syllabus, which runs to over 500 manuscript pages, has, however, been accep- Introduction ted for publication [1] by Computer Sci- ence Press, Rockville, Md.) There is also There are many arguments for teach- a very thrbuugh book on the assembly lan- ing assembl" language on a microcomputer guage of the 6502 [2]. This book, however, such as the KPPLE. Of course, there is the is intended for the teaching of logic de- basic argument concerning progress in com- sign replacement. As a consequence of puter science; microcomputers provide bet- this, the student, in using it, does not ter throughput (instructions per dollar) require a knowledge of BAS/C, but does re- than mainframes, and therefore constitute quire an elementary knowledge of computer an advance in computer science that should hardware (equivalent to Volume I of [3],' be reflected in curricula. But there are for example), whereas the reverse is true other arguments. Assembly language pro- in a typical assembly language programming gramming on computers ouch as the IBM 4341 course as it is given in colleges and uni- has been strongly discouraged for some versities (and in our syllabus) . time now. Assembly language programming on microcomputers, on the other hand, is of- In teaching a course on APPLE assem- ten necessary because of the small memory bly language, it is essential to specify size of the microcomputer system being an assembler. We use the Lazer Systems In- constructed. teractive Symbolic Assembler (LISA), for several reasons. It is one of the few pri- It may be argued that microcomputer vately produced assemblers for the 6502 to assembly language is of particular impor- be systematically maintained, down through tanc in a compiler-writing course, since the years; new versions continue to be most of the new compilers being written produced, involving new features as well these days are being written for microcom- as correction of problems discovered in puters. Many new assembly language con- older features. It is very fast, being cepts are available in microcomputer as- what used to be called an "in-core" assembly languages which are not available sembler; that is, assembly is directly on the IBM 4341; these include stack-ori- into memory rather than to a diskette

355 372 file. It has consistently been rated as machine, and this is followed by a number the best APPLE assembler, overall, in the of chapters on specific programming tech- evaluations that we have seen. Finally, it niques. In our syllabus, we have taken a is reasonably priced. different approach. The introduction of the carious instructions, status flags, regis- Our APPLE laboratory cons'ists of 12 ters, and addressing modes is spread out APPLE II+ systems and eight APPLE III sys- over the wheLe syllabus. Thus the PSW (the tems. At the moment LISA runs only on the status register), for example, is not in- APPLE II+, and PASCAL runs only on the troduced until section 67; the interrupt APPLE III (and on one of the APPLE II+ status flag and the last three instructions systems, which has the Language Card). (CLI, SEI, and RTI) are introduced in sec- Hence PASCAL classes are scheduled for the tion 68; and the last three addressing Ills, LISA classes for the II+ systems, modes (zero-page and the two kinds of indi- and BASIC classes for either machine. In rect indexed addressing) are saved for sec- the fall of 1982, six sections of assembly tions 74 through 76. Our motivation here is language were run, of which four were ori- to provide a thorough introduction to every ented towards the 6502. separate instruction, register, status flag, and addressing mode. Thus there is a Description of the Syllabus separate section on the logical exclusive OR, for example, giving several applica- A heavy emphasis is placed on writ- tions, and introducing the truth table and ten exercises. In the teaching of assembly the two alternate mnemonics (EOR and XOR). language, one finds far more concepts that Similarly, each status flag is introduced have to be learned than is the case with in a section of its own, together witn a FORTRAN, BASIC, or even PASCAL. Some of discussion of its uses. (The uses of the these can be learned by hands-on experi- carry flag are spread over at least a dozen ence in assembly language programming, but sections.) there is a limit.to the amount of material that can be learned in one semester in Detailed Outline of the Syllabus this way. Accordingly, we have arranged the syllabus in 100 small sectionu (a max- (1) Coces (Section 1). The basic idea imum of three manuscript pages per sec- of representing all data in a computer as tion), with three written exercises per codes is introduced by analogy to the se- section, each of which has no more than cret codes which children pass back and three parts. These have been assigned to forth. the students as homework in each of three (2) Number systems (Sections 2-5) .An semesters (spring, summer, and fall 1982), introduction to binary and hexadecimal, and checked very thoroughly, both as to conversions between one number system and possible incorrect answers supplied by the another, and addition and subtraction in instructor (who makes more mistakes than binary and hexadecimal. he cares to admit) and as to difficulty (a (3) Registers (Sections 6-7). The number of the exer ises were revised .and basic ideas of registers and memory cells; made easier). Several students submitted the A, X, and Y registers (the others are acceptable alternative answers, which have introduced later); the basic fact that n been included in the section on answers to bits can hold 2n codes; multibyte quanti- selected exercises. ties and twos' complement arithmetic. (4) Load, store, increment, and de- The student is introduced to the crement (Sections 8-9) .The first twelve computer at a much later time, during the instructions: LDA, LDX, LDY, STA, STX, STY, semester, than is customary in programming INC, INX, INY, DEC, DEX, and DEY. (We do _courses. Assembly language differs from not introduce ADC or SBC until later, be- other programming languages in that a tre- cause of the need to discuss carrying and mendous amount of material must be under- borrowing.) stood before the student is capable of (5) Machine and assembly language, writing reasonable programs. It is possi- and pseudo - operations (Sections 10-11). At ble to write a program which adds two num- this early stage, the student is shown, in bers, for example, in the first week, but a simple form, the difference between as- this, in our opinion, does not facilitate sembly language and machine language, and learning how to write and debug more so- is introduced to three pseudo-operations: phisticated programs. In the outline of ORG, END, and DFS. These will be enough the syllabus, given below, it is made for the writing of simple programs. (We clear how much material we cover (about may note that many syllabi start out with two-fifths of the total) before the first machine language, and only later introduce actual hands-on programming assigndent. assembly language, hoping to make the point that the tediousness of machine language In quite a number of texts on assem- suggests the need for assemblers. We take a bly language, the student is presented, at different tack here because we see no need the start Of the course, with a complete to make this subject any more tedious -- list of all the instructions of the given there are enough f4.ne points to learn as it is.) (12) Offsets (Section 23) . This is (6) Two-byte quantities (Section an often sadly neglected topic. By offsets 12). We are going to be doing 16-bit ope- we mean the use of LDX J followed by LDA rations throughout the syllabus (as is T-1,X (rather than LDA T,X), for example, clear from what follows). At this point to load T(J) when the first element of the our purpose is merely to introduce the no- array T is T(1) (that is, men there is no tation K+1 (as in LDA K+1) for the upper element T(0)); also LDX J followed by LDA byte of the two-byte quantity K, and to T+8,X (for example) to load T(J+8) (or start clearing away the common misimpres- T(J+9) if T starts with T(1), thus combi- sions about K+1 (distinguishing LDX K+1 ning this with the preceding technique). from LDX K followed by INX, for example). Here the offsets are respectively -1 and (7) Indexing. (Section 13). An un- 8. Offsets are indispensable; th s in mo- usual note here: this early_ in the sylla- ving an array U to an array V, with a loop bus, we in roduce LDX J followed by LDA ending in DEX and BNE, it is necessary to T,X for the loading of T(J) into the A use LDA U-1,X and STA V -1,X rather than register. It is our strong opinion that LDA U,X and STA V,X (since the final value subscripted variables are usually intro- of X is 1, in such a loop). duced far tco late in the semester, re- (13) Character codes (Section 24). gardless of the programming language be- Pretty soon the student will want to put ing taught. One must face the fact that messages (like ENTER THE FIRST NUMBER) in over 95% of all programs contain arrays programs, and this section, introducing and indexing. all the basic character code concepts on (8) Adding and subtracting (Sec- the APPLE (normal, inverse and blinking tions 14-17.) This includes 8-bit and 16- mode, double quotes and single quotes, bit addition and subtraction, the carry control characters, and so on) is put in flag, and the relation between carry and to anticipate the student's wishes. borrow. Two-byte numbers are considered as (14) Input- output, subroutines, and two-digit numbers in a number system with EQU (Section 25). Again: pretty soon the base 256, so that their addition and sub- student will want to do IjO, which, on traction can be seen to follow the same the APPLE, is done by means of monitor rules as with other number systems. We may subroutines. Just the basics are given also note that a careful and precise here: JSR (but not RTS, and nothing about treatment of the carry flag as used in stacks yet); descriptions of each of three subtraction, specifically, is markedly ab- common monitor subroutines (RDKEY, COUT, sent from many treatments of microcomputer and GETLNZ); and the use of EQU (which is machine language programming. necessary in specifying where these sub- (9) Transfer instructions and com- routines are in memory). This is a good ments (Section 18). Two topics here: way to sneak in an introduction to EQU, first of all, four more instructions, TAX, which causes far more student confusion TXA, TAY, and TYA, and their use in the than, one might assume. evaluation of expressions such as T(J+K) (15) BYT and ASC (Section 26). A (where T is an array); then, comments very careful introduction to BYT is need- (which follow a semicolon, if the LISA as- ed, because students frequently tend to sembler is used). The reason for introdu- confuse BYT with EQU. cing comments at this point is that our (16) The program counter and rela- programs are just now starting to get big tive addressing (Section 27). We introduce enough that the student can see the need the program counter relatively late in the for some memory aid to remember what has syllabus; our feeling is that only now been done. does the student have enough familiarity (10) Branching and labcls (Section with assembly language concepts to under- 19). Since we have introduced the carry scani it properly. Relative addressing, flag, we can introduce BCC and BCS, and which takes some getting used to, is also this leads naturally to a discussion of introduced here (because its definition the syntax of labels. Here the student who involves the program'counter). It is ne- has had BASIC or FORTRAN needs to get used cessary to understand this in ca-der to be to the idea of an alphanumeric label. An able to hand-translate assembly language application of BCC and BCS (adding or sub- to machine language, which is coming up tracting two quantities, one eight bits quite soon. long and the other 16) is also given. (17) Sign status (section 28). This (11) Loops, and zero status (Sections section is mainly devoted to a ca::eful ex- 20-22). The student who has had only FOR planat-f.on of why one cannot use the se- statements in BASIC, or DO statements in quence LDA P / CMP Q / BMI ALPHA to test FORTRAN, needs to understand how to simu- for P < Q, whether P and Q are signed or late these in machine language. Two basic unsigned. This then leads to an explana- types of loop are presented: the loop tion of the use of BCC and BCS for this starting with INX and ending with CPR and purpose. BNE, and the loop ending with DEX and BNE. (18) Two-byte operations and shift- (Other types will be presented later.) ing (Sections 29-35). Covered here are

357

37_4' two-byte unsigned comparisons, increment, than they might seem. We start this dis- decrement, complement, shifting left and cussion by covering a few topird that our right by one bit, and arrays of two-byte discussion of stacks will illuTzinate: saved quantities, as well as ordinary 8-bit and restored variables, the concept of a shifting, multiplication and division by return address, and indirect jumps (no powers of two, and multiplication by ten. other indirect addressing yet). We then in- Also, the use of shifts and the carry flag troduce stacks, first in the abstract and for processing of the bits in a byte is then with specific reference to the 6502, discussed. including PHA, PLA, TSX, TXS, and the ac- (19) Table lookup, space_ -time trade- tual operations of JSR and RTS, as well as off (Sections 36-37). Here we have a phi- the stack pointer and the page-one stack losophical digression: is it more impor- area. This is followed by a discussion of tant to save space, or to save time? As stack techniques, first in general and then usual in such digressions, we aim to show with specific reference to the input-output that there is no simple answer, but rather conversion programs which we mentioned ear- several complex answers. In order to dis- lier; these are two programs which illus- cuss the subject intelligently, we need to, trate, between them, several advanced uses know about table lookup, and how to do ti- of stacks. Finally there is a discussion of ming calculations; both of these are ex- why we use stacks (there are good reasons plained in detail. that have nothing to do with recursion, (20) Multiplication, division, in- which is important because most programs do put-output conversion (Sections 38-41). not call themselves, either directly or in- Beginning with a discussion of the multi- directly) . plication and division of general binary (25) Decimal mode (Sections 65-66). numbers, we proceed to the consideration This includes the decimal mode flag, CLD of two very tightly optimized subroutines, and SED, and routines to pack strings, un- one for multiplication and one (already pack strings, and add two packed strings. reported in [4]) for division. We then (26) The status register (Section proceed to the use of two conversion rou- 67). This section appears here because we tines, one for decimal input and one for can now treat several of its applications, decimal output; these will be discussed such as saving decimal mode status in a further later on, since they involve con- subroutine (using PHP and PLP) so that the cepts we have not had yet. RTS is also in- subroutine can use decimal mode, whether troduced, but only as a return instruc- its calling program does or not. tion; nothing is said yet about the stack (27) Interrupts and input-output or return addresses. (Sections 68-72). The section on the status (21) Running _programs on the compu- register naturally leads us into a discus- ter (Sections 42-51). Only now is the stu- sion of interrupt subrdutines and why they dent deemed sufficiently prepared to be must save the status register (and the fact able to write and run assembly language that this is done automatically on the programs. This complete discussion of the 6502). This, in turn, naturally leads us subject includes hand assembly, desk into a discussion of input, output, simul- checking, walkthrougha, commands in LISA, taneous input-output using queues and pol- stepping, tracing, breakpoint debugging, ling, and the speaker on the APPLE (al- patching (at the assembly level), and com- though we treat only the simplest applica- munication between LISA and BASIC. Along tion here, namely the generation of a mu- the way, there is a thorough discussion of sical tone). intermixing errors (STA Q followed by Q (28) Further string declarations in DFS 1, for example) and overwriting errors LISA (Section 73), namely INV, BLK, DCI, (instruction codes being destroyed in the STR, and HEX. course of running a program). (29) Page zero and indexed indirect (22) Logical operations (Sections addressing (Sections 74-77). A major deci- 52-55). There is one section each on AND, sion in this syllabus was to postpone the ORA, XOR, and BIT. Several applications of treatment of page zero almost to the end. each of these are presented. It is seldom used on the APPLE by user (23) Overflow status (Section 56). programs (because using it would overwrite This is introduced quite late. The reason locations used by the monitor, LISA, and is that, compared to the other features of BASIC). In turn, we cannot introduce in- the machine, it is not very useful; CMP dexed indirect addressing until the stu- does not affect overflow, and signed com- dent knows about page zero. An important parisons, which use overflow, are trickier application of post-indexed indirect ad- than they seem (a complete discussion of dressing, namely the processing of arrays this point is given, with a sample pro- having more than 256 elements, is given in gram) . a section of its own. (24) Stacks (Sections 57-64). We may (30) Modification of instruction note that data structures, in general, are words (Sections 78-81). Most people shy discussed in a later course; the student, away from discussing this subject at all. at this stage, has no conception of them, 'peir feeling is that it is bad program- and stacks are thus harder to understand ming practice anyway, and hence better

358 3 "f left out of the curriculum. The result, type of a variable in memory along with however, is that progra,=Dr4 inevitably its value, and interrogating the type be- discover it for themselves, like it, think fore performing any operations). they've made a new discovery, and use it in bad ways. We introduce it and cover se- The Appendix To The Syllabus veral applications of it, being careful, along the way, to show the difficulties. The syllabus is accompanied by an Our hope is that programmers either give appendix, including the following tables: up on it (which is what many people wanted (1) an introduction to BASIC for those who in the first place), or else learn to use might know only FORTRAN or PL/I or PASCAL; it responsibly. Modification of ordinary (2) number base conversion;(3) the 6502 addresses, of immediate data, and of rela- instructions in alphabetical order, toge- tive addresses is covered. ther with their meanings and the status (31) Arrays and sorting (Sections flags they set;(4) the 6502 instructions 82-87). This includes arrays of-strings, in alphabetical order, together with arrays of hexadecimal digits, two-dimen- their addressing modes and machine lan- sional arrays, and the process of sorting guage forms;(5) the 6502 instructions in and of searching a sorted array. The con- the numerical order of their machine lan- nection between sorting alphabetic and guage forms;(6) the LISA pseudo- opera- numeric data is shown (that is, the inter- tions and extended (7) all spe- pretation of a sorted alphanumeric array cial characters used in LISA, :ogether as being "alphabetized" in the usual with their meanings; (8) all addressing sense) . modes used in 6502 instructions, together (32) Two-byte signed numbers (Sec- with their meanings; (9) character codes tion 88). So far we have considered signed or letters and digits in all modes;(10) and unsigned 8-bit data, and unsigned 16- character codes for all characters other bit data; here we complete the picture. than letters and digits, in all modes; This also gives us an opportunity to prove (11) all the APPLE monitor subroutines rigorously something we have assumed as used in this syllabus, together with their given -- namely that the same add and sub- actions (we actually use only a small num- tract instructions work on both signed and ber of the available monitor subroutines); unsigned data. (12) all registers and flags and their ca- (33) Loops ending in INX and ENE pabilities (that is, which instructions (Section 89). Another type of loop combi- use them directly); (13) all the LISA com- ning advantages of speed and forward pro- mands used in this syllabus, and their cessing, but a little more difficult to meanings;(14) all the APPLE monitor com- work with, and with more restrictions. mands used in this syllabus, and their (34) Tapes and disks (Sections 90- meanings; (15) a table of the various 92). This includes a discussion of tape subcodes of the 8-bit operation code of parity checking, the APPLE disk operating the 6502 and of the "families" of 6502 in- system, and a number of further LISA structions (depending on the rightmost two pseudo-operations (DCM, LST, NLS, PAG) and bits of the operation code). commands (W, control-D), as well as two more monitor subroutines (KEYIN and COUT1). We have found that one semester is (35) Simulators, interpreters, as- about right for the coverage of this ma- semblers, compilers (Sections 93-95). A terial. If advanced topics (Sections 49- general philosophical discussion of these 51,65-66, 68-73, and 78-100) are omitted, is often given in a first course on BASIC the material can be covered in a quarter, or FORTRAN, but this discussion should be for those schools on the quarter system. expanded now that the student knows assembly language and machine language. References (36) Structured programming (Section 96). This section is mainly devoted to the 1. Maurer, W. D., APPLE Assembly connections between structures programming Language, Computer Science Press, Rock- and assembly language, such as how a typi- ville, Md., September 1983. cal structured programming statement (DO- 2. Leventhal, L. A., 6502 Assembly WHILE, CASE) would be implemented in as- Language Programming, Osborne/McGraw- sembly language. Hill, Berkeley, Calif., 1979. (37) Real numbers (Sections 97-100). 3. Osborne, A., An Introduction To There are no floating point instructions Microcomputers, Osborne/McGraw-Hill, Ber- on most microcomputers,(even the 16 -bit keley, Calif., 1976. variety), but many students have used 4. Maurer, W. D., An improvement floatingpoint BASIC on the APPLE and are upon division program la Leventhal, Dr. curious about how this is done. We start Dobb's s Journal 7, 3 (March 1982), pp. 20- by discussing binary and hexadecimal frac- 21. tions, then floating point formats, then floating point operations (including nor- malization), and finally a discussion of * * * * * * * * * typeless processing (that is, keeping the

359 376 STUDENTDOWN SYSTEM DESIGN

Robert Geist

Department of 06?vuter Scienoe Duke University, Durham, North Carolina

"paper measures" seen by system adminis Abstract trators. The issue is this: can we lend quantification to the aeasures students gather, and can we use such measures, A new method for the design of com rather than the classical objective func puter systems is put forth, which recog tions, in designing systems? It seems we nizes that an individual student's percep can. tion of system performance may differ rad ically from that of the system administra 1. ILIA IlcumdaiA MIER tors. Recent analytical results from application of the method to the design of dual processor systems aro surveyed, and a In [7], Harvard's David Hemenway call is issued to the educational comput offered an "alternative mean" as the ing community to undertake the empirical answer to an oftenposed student question, research in psychopbyslcs that is neces "Why are my classes larger than the 'aver sary to complete a comprehensive methodol age class size' printed in the school ogy. catalogue?" Speoifioally, if M students populate N classes of sizes 2 bl. the alternative mean class size is 0. Introduation N X 2 (3i /M)Eil One of the most painful problems in 1=1 educational computing today is the alloca tion of the scarce (nonexistent?) school that is,the expected size of a glass con dollar among the over increasing multitude taining a randomly selected student. of available computer system configura Hemenway provides convincing evidenoo that tions. Although it is easy to observe this measure more accurately reflects the that the explosive growth in alternative information actually available to the stu systems has been accompanied by an equally dents, who are unable, to view all classes explosive growth, within the computer sci from "on high." ence community, of analytic design tools intended to solve the configurationprob In [3], we showed that the natural lem (e.g. [1,5,8,11]), it is our conten extension of this alternative mean to the tion that such tools are inappropriate for userperceived mean number of customers in our use in educational computing. a queueing system turned ont to be

After all, the classical approach to NsBali analytic system design (including, until E[N] recently, our own (5,6]) has been to tacitly assume a goal of minimizing mean system response time, or mean reciprocal where N is a random -,,ariable denoting throughput, or system cost, subject to the number of customers in the system and some constraints. Yet, though each of E is the ordinary expected value opera these measures may be held to champion the tor. Now in classical analytic system cause of a particular group (the students, design, we regard the components of a com the administration, Ia bursar), each must putin3 system (CPU's, drums, disks, etc.) be recognized asa surely external meas as servers and jobs as customers, who ure, not directly available to, and hence queue when the desired component is busy. not directly measured by, the internally Thus we can represent an entire computing situated users of the system. system as a network of queues, and, as night be anticipated, "studentdown" Most of us would agree that or stu design (wherbin the student moves to the dents gather system performance measures "top" position) is then that design metho which differ substantially from thong dology obtiined by replacing E[N] with N

360 37-7 wherever the former appears in the classi cal queueing networkapproach. Perceived E(R1]< E(R2] iff pi>12+(7/2)(1Xi(2p2)) mean response time, R , is then given, in accordance with Little's formula (see In particular, if pl = p2 + X/2 then

(67), by R = N /X, where X denotes the E(R1]( E(R2],] so we should choose system mean arrival rate of jobs to the system. 1. But from (3],if we should also have

X < The question at hand is now this: if (1 + \II7)p2/4, then R: > 1121 we were to design a system to minimize Thus there setups to be a window of perceived mean response time, , R rather parameter than ordinary mean response time, values in which customers would prefer the queue E(R] = EINI/X, might we reach a combined configuration substan over tially different conclusion? Indeed. the separate queue configuration, even though Even in elementary design problems, the the ordinary mean response effects of this change are dramatic. time would not be as good.

2. A SummaLy of Preliminary Results B. Should we unplug a weak procpssor?

We now present the results from thz,e Consider again a combined queue, elementary problems on the design of dualprocessor configuration, suchas sys tem 2 dual processor systems. The analytic of example A, but now suppose the service derivations of these results will eventu rates of the two processors are ally appear in (3,4], and will not concern not necessarily identical. Let them be given by us here. Our aim is merely to give the pi, and p2 where µi = ap2 for reader the flavor of the new methodology. some a 1 1. A classical result from system design A. SLould we have separate gueues or a (12] states that if wa want to minimize combined gueue? mean system response time, then there is some disparity ratio a beyond which wo Ir. figure 1 we show two possible con should simply unplug the weaker processor. figurations for a dualprocessor system. For example,if X =.2 and p2 = .25, then The processors service jobs at a mean rate for pi> 2.82512, we should unplug the of pi jobs/second, i=1,2, and jobs arrive weaker processor and run jobs solely on to either system at a mean rate of X the faster one. jobs/second, where necessarily 02

E(R1] = 2/(2p1 X) and that for the combined queue configura so that Rc ( Rs for all relevant tion of system 2is given by X, P2, and a, and we should always leave the weaker processor plugged in. E(R2] = 4u2/(44 1`2) C. How should .nompr be allocated If pi = p2 then clearly E(R2]( E(R1], but between 212R/112r4? we usually incur additional management costs associated with system 2, so that Consider the system of example S, but point Pi > The at which such addi now suppose that we are free to choose tional costs would cause us to choose sys p and p subject to a budgetary l 2 con tem 1 can be determined quickly from the straint of the form pl + expressions above: 112 = IC,a con stant. Classical analysis shows (4] that mean system response time is minimized if

361 WO choose nor that of R as a function of N is necessary. If we merely wish to compare alternative systems on a basis of per P2 + \114-(1l/k) ceived mean response time, it suffices to assume that each of 0and R (or, more (and, of course, pi = Y a2). On the generally, their composite) is monotone other hand, when we incorporate student nondecreasing, for system coparisons on perception we find (as in example B) a basis of 0(R (N )) can then clearly be made on a basis of N alone. ge , 1 [Lill A c On the other hand, ve contend that which is independent of the allocation. the most vital system design imperatives e.re of the form, "Minimize system cost So allocation becomes, in effect, a sable's! to an upper bound constraint on nonproblem. Perhaps a more important 0(R (N ))." For such, a precise specifica observation is that this last formulation, tion of 0 is mandatory. together with thst for Rsin example B, suggests thst students perceive anydual To our knowledge, no attempt has been processor system as equivalent to a single made to establish a power law for time processor system having service rate equal duration-in the context of computer system The results of any such attempt to the sum, pi n response. 2 , would be invaluable to our own research, snd we call upon the educational computing 1. A Can Fkt BPIRIX111.1 LitlAikk community to undertake this important study. Only in this way might we be able to exchsnge our scarce dollars for the A major factor in the student down maximum student satisfaction. , methodology is still missing: R no psychophysics. We must digress. 4. ggnAlusions

If f(a)is a physical scale and g(a) a psychological scale, where "sumac," is We have proposed a new "studentdown" used in the measurement theoretics sense methodology for the configuration design (see Roberts [9]) then their relationship of computer systems. It is our basic 0: thesis that student evaluation of computer system performance differs markedly from g(a) s 0(f(a)) administrative evaluation, and that it depends heavily upon the student's loca is called the psychophysical function. In tion, which is necessarily internal to the Our case, f and g represent system system itself. We contend that the per response time measurements, and, up to ceived mean, originally introduced by this point, we have assumed 0 to be the Hemenway [7], captures the effects of this identity, that is to say, we have assumel location. perfect perception of the information available. The issue raised in the Using the new methodology, we often preceding sections is that this available find that the optimal system design' information, and hence any conclusion differs radically from that obtained derived therefrom, depends heavily upon through classical analysis. Studentdown the location of the perceiver, even if his designs are also usually easier to imple Akility to perceive is unimpaired. ment. To incorporate the ability of the The final (and yet unspecified) com perceiver, we propose to follow the great ponent of the new methodology, the psycho body of literature from the psychophysics physical function, is the subject of our of prothetic continua (of which time dura call to the educational computing commun tion is one example) and assume the power ity for extensive experimental research., law, that is, P(x) 1. a x°,a )0. From Ekman and Sjoberg [2] on prothetic con tinua: "As an experimental fact,the power law is established beyond any reasonable This work was supported in part under NASA doubt, possibly more firmly established Langley Research Center Grant #NAG1-70. than anything else in psychology." Stevens [10] provides evidence that for estimation of time duration of white noise stimuli. 5. References P 7, 1.1.

Now, for a certaincollection of 1. Chandy, Y., Hogarth, I., and Sauer, problems, neither the precise form of C., "Selecting Capacities, in Computer

3623, Communication Systems." mg Wilt. Soft. Elm.. SE-3 (1977).

2. Ekman. O., and Sjoberg, L., "Scal ing," Ana. Ray. Psvehol., 16 (1965).

3. Geist, R., "PerceptionBased Confi guration'Design of Computer Systems," submitted to jnform lion Processing Letters.

4. Geist, R., and Trivedi, K., "The Integration of User Perception in the Heterogeneous M/M/2 Queue," 11E21. of PERFORMANCE '83, May, 1983.

5. Geist, R., and Trivedi, K., "Optimal Design of Multilevel Storage Hierar System 1. Separate Queues chies." IEEE Trans. on C2E2., C-31 (1982).

6. Geist. R., and Trivedi, K., "Queueing Network Models in Computer System Design,"Mathematio"Magazine, 55 (1982).

7. Hemenway, D., "Why Your Classes are Larger Than 'Average'," Mathematics magazine. 55 (1982).

8. Ramamoorthy, C., and Chandy, R., "Optimization of Memory Hierarchies in Multiprogrammed Systems," /ACM 17 (1970).

9. Roberts, F., Measurement Theory, Vol. 7, Encyclopedia of Mathematics and System 2, Combined Queue Ids Application;, AddisonWesley, 1979. FIGURE 1 10. Stevens, S., "On the Psychophystcal Law," Pswe49.1. RAiiew, 64 (1957).

11. Trivedi, K., Wagner. R., and Sigmon, T., "Optimal Selection of CPU Speed, Device Capacities, and File Assign ments," JACM, 27 (1980).

12. Trivedi, K., Probability and Statis tics wiih Rejiability, Queueing, and gAN2119A 3gl4A9A Appliations, PrenticeHall, 1982.

363 3 8 COURSEWARE DEVELOPMENT AND EVALUATION

L. Carl Leinbach Barbara C. Garris Ann Lathrop John C. Miller

The remainder of the Computer Studies ABSTRACT: Computing Literacy and the Liberal Arts curriculum consists of: Introduction to L. Carl Leinbach, Chairman, Computing Studies, Algorithms; Data Structures; Design and Analysis Gettysburg College, Gettysburg, PA 17325 of Algorithms; and Computer Organization and Assembly Language Programming. These courses are What is the role of Computer'Science in a taught using the Burroughs 6700 computer and the liberal arts college?This question and Gettysburg language of choice is ALGOL. The mathematics College's definition of an answer to the question department supports this program with courses in are the subjects of this presentation. Discrete Mathematics and Numerical Analysis. Gettysburg College is a liberal arts college This presentation will focus on the with an enrollment of 1850 undergraduate students. development of the curriculum in view of the It is extremely proud of its liberal arts heritage Acacemic Purposes of the College.A special and offers a broad, diversified curriculum emphasis will be given to the development of the consistent with its heritage. In 1976 the faculty Computing Literacy course, the recruitment of published an academic purposes document which faculty for the program from within the College, clearly defines the role of the curriculum within and the decisions which led to the establishment of the microcomputer laboratory. the liberal arts. That purposes document states that the curriculum must emphasize the following elements: 1. Logical, precise thinking and clear use of ABSTRACT: Courseware Evaluation Techniques language. 2. Broad, diverse subject matter. Barbara C. Garris, Teachers College, Columbia 3. A rigorous introduction to the assumptions University, Box 27, New York, NY 10027 and methods of a representative variety of The Educational Product Information Exchange academic disciplines. (EPIE), established in 1966, has been a pioneer in In 1979 the Academic Policy and Program Committee of the faculty conducted a study of the developing evaluation techniques for textbooks and audiovisual materials and equipment which it introduction of Computer Science into the publishes through a series of subscription curriculum. Previously, computing courses had been publications. taught in the mathematics and business With the advent of a microcomputer hardware administration departments. The study resulted in the establishment of an interdisciplinary Computer and courseware community, EPIE has adapted its Studies Group which was given two charges': evaluation system to the rapidly evolving 1. To promote computing literacy on the technology of microcomputers. In partnership with Gettysburg College campus. Consumer Union, EPIE hopes to raise the 2. To establish a curriculum consisting of not consciousness of today's educators into an informal more than four courses for those students consumeroriented Alliance for Quality in- who desire to learn more about Computer Educational Computing, pressuring courseware Science. producers to higher standards than are currently The Computer Studies Group developed its the norm in the courseware mr,ket, and to make curriculum during the 1981-82 academic year and possible 30day review copies of software similar to those available for textbooks. instituted its program in the fall of 1982. The computing literacy course is taught in a new In the short presentation, EPIE will outline (1) how we train groups of teachers to analyze microcomputer laboratory which has 18 Apple II+ courseware using our own evaluation techniques, and microcomputers and 3 Epson MX-100 dot matrix (2) how we evaluate courseware for publication as printers. Each microcomputer is equipped with 64K PRO/FILES Courseware. The Evaluation Coordinator of memory and two disk drives. The course itself emphasizes the algorithmic approach and topdown will use slides and transparencies to illustrate the detailed evaluation protocol and follow through problem solving. The BASIC programming language is the workshop training process which any school used during the introductory portion of the course, system can adapt to its own evaluation and but the main portion of the course is taught using selection of microcomputer curriculum materials. Pascal. The session will be concluded with a questionandanswcr period. Handout materials will be provided.

364 33' ABSTRACT: The California Courseware Clearinghouse ABSTRACT: Let's Write Usable Courseware: Project The City College Algebra Project Ann Lathrop, San Mateo County Office of Education, Jon C. Miller, 110 Riverside Drive, Apt 14C, New 333 Main Street, Redwood City, CA 94063 York, NY 10024 The Teacher Education & Computers (TEC) Center The City College of New York has an extensive program divides California into 15 regions, each remedial mathematics program, enrolling almost 1000 with a TEC Center responsible for providing students per term in one of two courses which inservice to teachers in the fields of cover science, basic algebra. In an attempt to provide mathematics, and computer literacy. The individualized instruction and student invL'vement Microcomputer Center in the San Mateo County Office and interaction not attainable in a conventional of Education has been designated the software classroom, the Mathematics Department is turning to library and evaluation center toprovide support computer based instruction. services to the 15 TEC Centers for the periodfrom November 1, Existing programs seem not to exploit the full 1982, through June 30, 1983. potential of microcomputers in algebra instruction, The Clearinghouse will have three major responsiblities: and the college is therefore developing a set of materials more suited to the students' needs. (1) the training of a cadre of software Many existing programs are limited to practice evaluation specialists who will then problems on a limited range of topics, and are train teachers in their regions as usable only as a supplemenv to classroom software evaluators; instruction. The City College system covers all the construction of a threedimensional (2) the topics needed for a comprehensive course of subject/computer system/grade level basic algebra instruction. matrix of highly recommended software Most existing programs do not fully exploit that will be suggested for preview microcomputer graphics capability. The City throughout the state; and College system uses graphics in a variety of ways. (3) the develoment of subjectoriented The Apple II's high resolution graphics capability collections, each with 10 to 20 promising is used to represent all algebraic expressions in new software packages, that will be standard algebraic notation. The system represents circulated to the TEC renters on a signed numbers as vectors in order to explicate the rotating basis for preview and rules for signed number operations. It allows evaluation. exploration of algebraic expressions by a Successful completion cf these three combination of evaluation and graphing. responsibilities will widely expand the base of Most existing programs restrict user input so trained evaluators throughout California, help them that, for example, if a fraction is "expected", to identify the best currently available software then only a fraction can be entered. The City for. previewing, and also provide interesting new College system features a completely general software for evaluation. expression entry and display capability, with Plans are currently underway for a Software simple control codes to produce exponents, Evaluation Institute to be held in San Mateo County fractiols, radicals, absolute values, and in January 1983, for the purpose of developing the transcendental functions. Virtually any matrix of highly recommended software. syntactL'ally correct expression can be entered at Representatives from successful software evaluation the keyboard and displayed in standard notation, projects at stae and regional levels will b: but a syntactically incorrect character produces an invited to participate. It is our hope that this immediate error message. group of educators, all experienced in software The City College algebra system allows evaluation at their individual sites, can agree individualized flow from topic to topic. Excessive upon the software to be included and help to errors result in immediate transfer to an establish its place in the matrix. appropriate prerequisite topic. After Other activities of the Clearinghouse include demonstrating mastery of a topic, the student i3 the investigation of electronic dissemination of given o choice of topics to try next, and a choice the matrix and of software evaluations, contacts of whether to start from the beginning or to start with software publishers, publication and with a diagnostic quiz for possible exemption. dissemination of software reviews received from the Assistance is always available by typing a question 15 TEC Centers, and the development and mark. distribution of SOFTSWAP public domain programs. Small scale preliminary testing of the earlier There will be a second, smaller Institute held in portions of this system is in progress, and further May 1983 to evaluate the work of the Clearinghouse portions are being written. The project and to make recommendations for 1983-84. presentation will include a demonstration of the major features of the City College algebra instructional system and a summary of the preliminary results of students using the system.

365 382 Request. C'or Equipment Proposals

Joseph Wolfsheimer Division of Automated Services District of Columbia Public School District Washington, D.C. 20004

SPONSOR: SIGCAS

ABSTRACT Numerous technical and procedural issues exist to determining the content of confront education organizations seeking to technical specifications. All encompassing procure equipment. The nominal procurement specifications are advocated. At the same environment becomes more intricate as the time, it is held that a generic approach to quantity of equipment increases. identifying hardware and software Education's computer related efforts differ components provides advantages over in nature from earlier, pilot efforts. manufacture-based specifications in cases They now tend to address large populations not involving a large installed base. of students throughout the educational An approach to proposal evaluation is organization. The need to succeed in suggested. Empirical evaluation is providing them computer, related experiences advocated. It is suggested that evaluation places further emphasis for success on the techniques be a part of information procurement process. released to potential vendors in the This one and a half hour session education organization's request for presents a case study of several proposals. A case study is mentioned with educational organizationswhich recently a breakdown of advantages and shortcomings. procured large quantities of equipment to Finally, approaches to phased address multipleeducational requirements procurement are discussed. These include on a universal basis. Prime emphasis is rent, buy, or lease decisions, discount placed on formulating technical arrangements and open-ended contracts. specifications for hardware as an intrinsic Budgetary issues are discussed in brief and portion of instructional planning for it is advocated that computational computer related efforts. equipment move from capital budgets to some It is noted that several approaches extent.

366 3S3 Courseware on Social Issues of Computers

Ronald E. Anderson, Chair University of Minnesota

ABSTRACT impacts. Software and accompanying text New courseware from the Minnesota materials for both college and precollege Educational Computing Consortium and instruction will be demonstrated and Control Data Corporationoffer assistance discussed. Special attention will be given to those instructors desiring to include to techniques for integrating these computer literacymaterial dealingwith materials intovariowi types of courses. social issues. Such issues include Consideration will be given to additional privacy, computer crime, copyright needs and future developments. violation, employment, and economic

DISCUSSANTS

Hans Lee Michigan State University

Beverly Hunter HumPRO

PANELISTS

Thomas Heaney Control Data Corporation Catherine Dunnagan Control Data Corporation Richard Pollak Nancy Kozen Minnesota Educational Computing Consortium

SPONSOR SIGCUE

. 367 384 Word Processor in the Composition Classroom

Mary Dee Harris Fosberg Department of Mathematical Sciences Loyola University, New Orleans, LA

Donald Foss Composition Programming English Department University of Minnesota, Minneapolis, MN

SPONSOR: ACH

PAPERS: "Writer's Workbench: Teaching Aid and Learning Aid" Kathleen Kieffer and Charles Smith Colorado State Fort Collins, CO

"Aids to Organization"

Helen Schwartz Oakland University Rochester, MI "Studying the Composing Process in the Computer Age"

Lillian Bridwell and Parker Johnson University of Minnesota Minneapolis, MN Interactive Computer Graphics and Computer Animated Films in Education

Maria Mezzina, Chair Teachers College, Columbia University New York City, NY

ABSTRACT The most important issues related to the use of computer graphics in education and the production of computer graphics instructional material at different levels will be presented. The speakerswill discuss various applications of computer graphics to education according to their major area of expertise and interest. Current learning approaches, both with and without the computer, place a heavy emphasis on verbal skills. But it has long been recognized by teachers that not all students have such verbal skills. On the contrary, many students need visual information to aid them in learning. Alfred Bork will illustrate the use of such visual information in computer based learning using examples developed at the Educational Technology Center. Patricia Harrison will discuss an example of an easy language to be used to design computer graphics software: ZGRASS. Characteristics of the language will be explained. Easy ways to generate graphics and animations for educational purposes will be illustrated visually. During the period from 1970 through 1977,the Topology Films Project, supported by the National Science Foundation, produced a series of educational films explaining by visual examples various concepts in topology. Nelson Max will show excerpts from the films, and discuss how they were designed, programmed, filmed and edited. Marial Mezzina will discuss the state of computer graphics in education, giving an overview of systems, languages and experiences. The problem of portabilitywill be addressed. The role of computer graphicsin the development of instructional software including conditions for effective use and conditions for easy programming will be addressed by illustrations and examples.

SPEAKERS:

Alfred Bork University of California Irvine, CA Patricia Harrison Electronic Visualization Laboratory University of Illinois Chicago, IL

Nelson Max Lawrence Livermore National Laboratory Livermore, CA

369 386 Teaching Ada With Computers

George Poonen Computer * Thought Corporation Plano, TX 75075

ABSTRACT Ada is the new programming language that developments in software methodology of the has been adopted as a standard by the last decade. Many of these features are Department of Defense. It is expected that somewhat foreign even tomanyof todays Ada will be used for both systems experienced programmers. As a result there programming as well as large scale real has been a great deal of concern expressed time applications. Consequently there has about education in Ada. been a burgeoning interest in the language This panel will cover a wide variety of both in industry as well as academia. In topics ranging from generic issues in fact, recently, Ada was approved as an ANSI transferring complex technology to' specific standard. approaches using computers to aid in the Ada incorporates manyof the important teaching of Ada. PARTICIPANTS:

Lee Blaine Computer * Thought Corporation Plano, TX 75075 Peter Wegner Brown University Providence, RI 02912

Lt. Col. Vance Mall AJPO Arlington, VA 22203

Kenneth Bowles TeleSoft San Diego, CA 92121

370 Electronic Mail and Computer Conferencing Paul Heller, Chair EDUCOM/EDUNET

ABSTRACT introduction to the concept and operation This session will describe the of Mailnet, a service which links local electronic mail services and systems that (campus based) electronicmail systems to are in use on numerous campuses. A variety each other. This service permits persons of applications will be described. using their ownmail systems to exchange Fundamentally important differences between messages and documents with colleagues on electronic mail and computer based other campuses. The mailnet service has conferencing will be explained in the been designed to allow virtually any local context of system features and their mail system to be connected with modest effects on patterns of communication within investment of moneyand technical talent. and between groups of participants. To design does not require acquisition of Numerous examples will be presented based new hardware nor any changes to local on use of systems including EIES, operating systems. Typical costs for COM/PORTACOM, Telemaail, Stanford's message exchange using the ordinary CONTACT/EMS, DREAMS, and VAXmail. telephone network are 25 cents per typed A special feature of the session is an page.

PRESENTERS:

Paul Heller Daniel Oberst EDUCOM/EDUNET

SPONSOR: EDUCOM EDUNET

371 388 SEX DIFFERENCES IN MICROCOMPUTER LITERACv

Marlal,ue Lockheed Antonia Nielsen Meredith Stone

Educational feting Service Princeton High School Educational Testing Service Princeton, New Jersey Princeton, New Jersey Princeton, New Jersey

Abstract relatively minor role in the remainder of the movie. There are no other females in the picture. In this paper, sex differences in computer literacy, use of computers, attitudes towards In the real world of computers and program- computers ard motivations toward using computers ming, life mimics art: only 11.2% of doctorates were examined. Sex differences in observed gain in in computer science were awarded to women in 1980- computer literacy were correlated with computer 81 (Chronicle of Higher Education, October 6, use, liking and motivation. Statistically signif- 1982). Computers are a man's world. But why is icant sex differences were found for both use and this the case? Are females less computer literate liking of computers, but not in expectations for than males? Don't females like computers? Don't future utility. Liking and motivation were females recognize that computers will be used in unrelated to gain in computer literacy. Home virtually every occupation by the year 2000?What access to a computer was related to gain for girls accounts for the numerous anecdotal reports, such only, and after school Computer Center use was as Sheingold's cited above, that girls do not use related to gain for boys only. computers when they are made available in schools? Since computer literacy is emerging as a major educational concern of the present decade (Seidel, Anderson, & Hunter, 1982) the purpose of the study was to answer some of these questions.

Method

During the academic year 1981-82, all students at a suburban, upper middle-class high school in Central New Jersey who were enrolled either in first or second year general mathematics, or in Introduction first, second or third year college preparatory mathematics were concurrently enrolled in a re- "In all of our sites we observed quired 28-hour computer literacy course that sub- differential use (of microcomputers) stituted for every sixth mathematics class through- according to sex, particularly at out the year. Students were administered a 15-item the secondary level. This is not an test of computer literacy at the first session of issue of access per se, since girls the computer literacy course and at the final are not systematically excluded from session; at the penultimate session they were ad- using computers. At the elementary ministered a 37-item questionnaire regarding their level, each sex could and did use experience and attitudes towards computers. the micro more or less equally. But, starting in seventh grade, when the 1. Measures micros moved out of classrooms and hallways into math and business de- Computer literacy pretest and posttest. The partments, there was an overwhel- same computer literacy instrument was used for mingly male representation among both the pretest and posttest. It contained 15 students who used the micros" items, evenly divided between general computer (Sheingold, 1981). knowledge, vocabulary, and programming algorithms. Several items were adapted from the computer lit- In Tron, the 1982 Walt Disney computer eracy test developed by the Minnesota Educational graphics fantasy about computers and computer Computing Consortium (Klassen, Anderson, Hansen, programming, the hero, his sidekick, and all other & Johnson, 1980). major characters, including anthropomorphized programs, the master control program and the Tron Computer survey. The survey instrument con- program, are male. All characters, that is, except tained 37 questions regarding student access to a heroine-programmer whose procedure for laser- and use of computers, student attitudes towards beaming matter into electronic impulse is respon- computers, and descriptive information such as sible for starting the action, but who plays a grade, sex and mathematics course.

372 30r, 2. Computer Literacy Course On the posttest a statistically significant (t = 2.56; p <.01) difference between male (M = The goals of the computer literacy course were 7.68; S.D. = 3.04) and female (M = 6.92; S.D. = to acquaint students with the potentials and appli- 2.74) scores was found. The highest score on the cations of microcomputers, to introduce students to posttest was 14, which was achieved by eight boys the BASIC Programming language and to give students and one girl. One girl scored zero on the posttest. practice in elementary scientific method. The course was taught in the high school Computer To verify that this difference was not due to Center by the Computer Center director; the Center changes in the student sample between fall and itself contained 11 Apple microcomputers with disk spring, a multiple regression was conducted to drives which students used during class. When assess sex differences in computer literacy gain. classes were not in session--before school for In this analysis, which was conducted on data from approximately 25 minutes, during the day for two those respondents having both pretest and posttest hours and fifteen minutes and after school for scores, the posttest score was the dependent vari- forty five minutes--students were also permitted to able and the pretest score was the control variable; use the microcomputers. Students were permitted to gender was considered the independent variable. The reserve a computer for programming, but not for results of this analysis are presented in Table 1. games. MOst of the students who availed themselves For the 318 students having both pretest and post- of the opportunity to use the Computer Center dur- test data, females gained approximately .75 points ing these times played games, although there were (one-fourth of a standard deviation) less than some who used out of class time to go beyond the males, a difference that was statistically classwork and to complete programs. significant.

The general pattern of the classes consisted 2. Sex Differences in Student Computer Practice, of four steps. In each six-day cycle, the students Attitudes, and Motivation were introduced to new concepts by watching a videotape made by the Computer Center Coordinator The second question to be answered addresses which superimposed relevant Apple programs on her the issue of sex differences in student use of or video image, by a lecture or by a handout. The practice with computers, student attitudes toward introduction to new concepts was followed by a dis- computers and student motivation regarding cussion in which student's questions were answered computers. and the classwork was explained. Using the new concepts, tho students entered demonstration pro- Practice. Nine questions regarding student grams, ran them and made notes on their results. use of computers were included on the survey. The students were then asked to write programs Statistically significant sex differences were applying the concepts which had been presented. found in the responses to every question but one, regarding prior use of a computer (Table 2). Males Results reported greater access to computers outside of school than females, greater extracurricular use of Usable data were obtained from 413 (87.1%) of the Computer Center, more frequent computer progr- the returned surveys; pretest scores were available amming and more frequent computer game playing than for 345 (83.4%) of these respondents and posttest girls. The largest sex difference was found for scores were available for 383 (92.5%). Complete reported game playing: 82% of the boys, compared pretest, posttest and survey data were available with 48% of the girls, reported having played for 114 females and 116 males. Of students with either a computer game, video games or arcade games complete records, 74% were ninth or tenth grade at least three times. students and 97% were enrolled in college preparatory mathematics courses. Liking. Four questions related to student liking of computers were included on the survey. 1. Sex Differences in Computer Literacy Statistically significant sex differences were found in the responses to all four questions, with The first question to be answered addressed boys reporting more favorable attitudes than girls the issue of sex differences in computer literacy. toward computers and programming. Fewer than half From the survey, we found that 62% of the males the boys and a third of the girls reported that and 56% of the females had used a computer prior to they liked working with computers or programming, the computer literacy course, a nonstatistically however. significant difference. Moreover, no statistically significant difference was found between the mean Motivation. Three questions relating to stu- pretest scores on the computer literacy test of dent expectations for future computer use were in- males (M = 3.49; S.D. = 2.21) and that of females cluded on the survey. Sex differences were found (M = 3.12; S.D. = 1.69). Even though high propor- on only one of the three questions. Two thirds of tions of both male and female students had been the boys compared with less than half the girls re- exposed to computers before entering the course, ported that they expected to use computers the most students were quite computer illiterate, as we following year, a difference that was statistically measured computer literacy. The highest score on significant. On the other hand, 80% of both boys the computer literacy pretest was 11; it was achi- and girls thought that knowing how to use a compu- eved by one boy (one girl scored 10). The lowest ter would be important for them in the future. score was zero, which was achieved by seven boys While few boys or girls repotted that they planned and three girls. to take any other computer courses, there were sex

373 39 differences in what types of courses boys and girls 4. Sex Differences in Factors Related to Computer planned to take (Table 3). Twice as many boys as Literacy Gain girls reported future plans to study programming languages, while more girls than boys reported The four practice factors that we found were future plans to study computer applicati2ns to related to adjusted gain in computer literacy were business, research or word processing (X (4) = reexamined for evidence of sex differences (Table 21.37, p< .001). 4). In all cases, males reported greater practice with computers than females reported. One third of 3. Factors Related to Computer Literacy Achieve- the males compared to about one-fifth of the ment Cain females reported having access to computers outside In general, achievement gain in any subject of school. Nearly 40% of the males reported coming can be viewed as a function of practice, attitude, to the Computer Center at times other than class, and motivation factors, such as those discussed in compared to fewer than eight percent of the females, the previous section. To assess their effects on and over 20% of the males reported that they stayed achievement empirically, we related measures of to use the Computer Center after school, compared practice, attitude and motivation to gain in compu- to fewer than three percent of the girls. Finally, ter literacy. the majority of males (61.3%) played computer games once a week or more, while the majority of females Practice. The nine questions relating to (52.2%) had not played computer games more than student use of computers were analyzed separately twice ever. using analysis of variance with adjusted gain* on the computer literacy test as the dependent vari- 5. Sex Differences in Determinants of Computer able and the question as the independent variable. Literacy Gain Four of the nine practice variables were related to adjusted gain: (1) access to a computer out- Having identified four practice-related vari- side of school (F = 7.314; p< .01); (2) coming to ables that were related to gain in computer liter- the Computer Center at times other than class acy and for which sex differences were observed, we (F = 5.042; p< .05); (3) coming to the Computer examined the relationship between these variables Center after school (F = p<.05) and (4) and gain in computer literacy, separately by gender. playing computer games, video games or arcade games (F = 2.778; p<.05). In all cases, students In these analyses of variance, adjusted gain reporting more practice achieved greater gain. was the dependent variable and the four practice variables were the independent variables. Access Liking. The four questions relating to stu- to a computer outside of school was modestly re- dent liking of computers were analyzed similarly; lated to adjusted gain in computer literacy for none were related to adjusted gain in computer females (F = 5.178; p<.05) but not for males. literacy. While use of the Computer Center after school was modestly related to achievement gain for males Motivation. The three questions relating to F = 4.311; p < .05), it was not related to achieve- student expectation for future computer use were ment gain for females. Use of the Computer Center also analyzed through analysis of variance; none at times other than class (that is, before school were related to adjusted gain in computer literacy. or during free periods) and game playing were unrelated to computer literacy gains for either males *Since simple gain scores (the difference be- or females. tween pretest and posttest scores) are relatively unstable, we computed adjusted gain scores (the Summary difference between predicted posttest scores and actual posttest scores) to estimate the signifi- This paper has examined sex differences in cance of these apparent differences in learning. computer literacy among secondary school students. The predicted posttest scores was obtained from an Our findings may be summarized as follows: ordinary least squares regression of pretest on posttest, which yielded the following prediction 1. No sex differences in initial levels of equation: computer literacy were found, but boys gained more Y' = 0.4610X + 5.926 than girls on computer literacy from pretest to where Y' is the predicted posttest score and X is posttest. the pretest score. This equation was applied to each student's pretest score to yield the predicted 2. Males reported more frequent use of com- posttest score. The student's predicted posttest puters than females and more positive attitudes score was then subtracted from his or her actual towards computers. posttest score to yield an adjusted gain score: G = Y - Y' 3. Although no sex differences in perceived where G is the adjusted gain, Y is the students future utility of computers were found, more males actual posttest score and Y' is his or her than females planned to take a computer course in estimated posttest score. the future.

4. Access to computers outside of school was significantly related to computer literacy gain for girls, but not for boys.

374 391 5. After school use of the Computer Center REFERENCES was significantly related to computer literacy gain for boys but not for girls. Klassen, D. L., Anderson, R. E., Hansen, T. P., & Johnson, D. C. Study of computer use and 6. Computer game-playing was unrelated to literacy in science education. St. Paul, MN: computer literacy gain for both boys and girls. Minnesota Educational Computing Consortium, 1980.

Seidel, R. J., Anderson, R. E., & Hunter, B. Computer literacy. New York, NY:Academic Press, 1982.

Scheingold, K. Issues related to the implementation ofcormlpi.1auterteclischools:A cross sectional study. Paper presented at the National Institute of Education Conference on Issues Related to the Implementation of Computer Technology in Schools, Washington, DC, 1981.

Table 1

Pretest and Gender Determinants of Posttest Computer Literacy

Score for 318 9th, 10th, 11th and 12th Grade Studentsa

(1) (2)

Pretest .460*** .444*** (5.778) (5.605)

Gender -0.755** (2.441)

Constant 5.926 6.346 2 R .096 .112

.092 .107

Note. aThe numbers in the table are the unstandardized re-

gression coefficients, with their associated t-statistic

below in parentheses.

375 392 Table 2

Sex Differences in Use of Computers. Liking of computers and Expectations for Future Computer Use

Response Male Female Chi- Category (H206) (N*207) square

Use of Computers

Had you aver used any kind of computer before taking this course? 8 Yes 61.7 36.0 n.s Do. you have amens to a computer outside of school? 8 Yes 33.3 22.2 6.14* Did you case to the Computer Center this year only for class? 8 Yes 61.3 92.6 54.14*** Did you come to the Computer Center this year before school? 8 Yes 18.0 1.9 27.92*** Did you coma to the Computer Center this year after school? 8 Yes 21.5 2.4 33.67*** Did you come to the Computer Center this year during your free period(s)? 8 Yes 40.8 8.3 56.76*** Have you reserved n computer for programming after school? 8 Yes 23.8 15.1 11.01* Row often have you programmed or used a computer for schoolwork outside of class? X Once or more 28.4 14.1 20.17*** How off:en have you played computer games, video games or arcade games? 8 3 times or mere 81.9 47.8 72.68***

Liking_of Comitors

Do you like playing computer games? 8 Yes 77.7 63.6 11.94** Do you like working with computers? 8 Yes 49.5 35.0 13.73*** Do you I.Liii learning to Program? Z Yes 41.5 28.2 19.61*** Do you like writing computer program to solve problems? 8 Yes 27.2 17.6 8.43*

Expectations for Future Computer Um

Do you plan to take any other computer /programming courses? 8 Yes 20.2 14.8 n.s. Do you think you will use computers mmt year? 8 Yee 68.4 46.3 16.64*** Do you think that knowing how to use a computer will be impertant foryouin the future? 8 Yes 81.6 80.2 n.s.

* p < .05 *5 p < .0/ *5* p < .001

Table 4 Table 3 Sex Differences in factors Related to Sex Differences in Future Plane for Computer Literacy Cain Computer/Programming Courses

Frequency of Response Frequency of Response Hale Female 22 Type of Course Male Female Access to a computer Other programming language 29.3 11.2 outside school 33.72 22.22 6.14 Yes Word processing 9.0 19.2 66.3 No 77.8 Business/research applications 22.6 32.8 Use of Computer Center only during class Intermediate or advanced basic '0.1 21.6 61.3 92.6 Yes 38.7 No 7.4 Other s.1 15.2 Use of Computer Center after school 21.5 2.4 33.6755 Yes 78.5 97.6 No

Frequency of playing games 3.4 72.68664 Never 5.9 14.7 Once or twice 46.3 About once a month 20.6 23.9 About once a week 31.4 16.6 Several times a week 29.9 7.3

p < .05 5p < .01

oulut p< .001

376 393 COMPUTERS: LESS APPREHENSION, MORE ENTHUSIASM

Janet Parker University of Louisville

Constance Widmer Northern Kentucky University

Abstract of the general public surveyed by Ahll (1976). He found that, "in general, educators seemed less Although the number of microcomputers be- enthusiastic about the computer's role in society ing purchased and placed in schools is rapidly than did the general public" (p. 48). The teach- increasing, many teachers arc hesitant, even ers were found to be less sure than the public frightened, by the thought of having to use (64% vs 87%) that computers will improve educa- one. Careful consideration must be given by tion; more were convinced that computers dehuman- inservice coordinators about the type of ize society (55% vs 37%), and that computers iso- initial computer experiences that should be late people from social interaction (30% vs lg%). provided for these teachers, since these ex- About one third of the teachers reported feeling periences can make the difference between computers are beyond the understanding of the acceptance or rejection of computers. This typical person and more than one fifth did not feel paper will identify possible causes of appre- that if one was in their classroom it would help hensions and fears about computers and present them be a better teacher. Though most (67%) did specific suggestions for conducting effective not feel computers will take their jobs, a dis- inservices to overcome such feelings. turbing number (16%) did feel this would happen.

Introduction Smeltzer13 (1981) also used Ahl's survey as a basis instrument, but added additional state- The message is clear. It is being broadly ments appropriate to media specialists. Perhaps stated by public media and professional jour- reflecting the AV training of the sample, his find- nals: This is the Computer Age. These wonder- ings are more encouraging. Only one-fourth of ful machines have amazing capabilities that those surveyed felt that computers dehumanize can dramatically enhance the way we do our jobs, society; most (86%) did not feel that computers are run our homes, even educate our children. The beyond the understanding of the typical person; implication is strong: teachers "should" and most (83%) feel that computers can improve understand and use computers. education. In interpreting these encouraging findings, however, it must be kept in mind that Yet despite all the laudatory publicity the sample size was small (N=29) and select given microcomputers, indications are that not (active members of an educational teU'nology all teachers are welcoming them with enthusiasm. association). Although most teachers acknowledge the importance of "computer literacy" for all students, some In a much larger, less restrictive study, teachers outrightly reject using computers; Stevens12 (1980) compared the knowledge about and many other teachers are apprehensive, anxious, attitudes towards computers of teachers, teacher even panicked at the idea of being expected to eudcators and student teachers in Nebraska. Parker use computers. Why do these teachers feel this and Hazuga10 (1982) replicated her study with way?Are their attitudes and fears real? What Kentucky educators. Both studies found that al- causes them?What can and should be done to though all the educator groups strongly supported meet the needs of these teachers and, ultimately, the concept that students should be taught about their students?Based on the research and ex- computers, they themselves did not feel qualified perience of the authors, this paper will document to do this teaching. They also found that many the prevalence of computer anxiety among teachers, educators are uncertain or doubtful that computers identify the main fears and misconceptions caus- can be useful in all instructional areas, and can ing the anxiety, and present specific suggestions provide more advantages than disadvantages in the for conducting teacher inservices which overcome classroom. On items related to computer anxiety, and alleviate these fears. both studies found that as many as one-fourth of the educators were not comfortable around compu- Prevalence of Computer Anxiety Among Teachers ters, Stevens finding teachers the more anxious subgroup while Parker and Hazuga found it to be the In 1976, Lichtman8 (1979) compared the student teachers subgroup, despite the fact that attitudes towards computers of educators to those this subgroup had received more training on

377 394 computers. As in the Lichtman and Smeltzer computer companies, rather than teachers, will be studies, both these studies found that as many handed down to them to be used without question. for others as 25-30% of the educators areconcerned with the The computer is seen as a powerful tool dehumanizing aspect of computers, expressing to exercise control over classroompractice, by- agreement with the statement that, "Computers in passing the teacher entirely. education almost always result in less personal treatment of students" (Parker and Hazuga, p. 11). Many teachers have become alienated by personal experience with unfriendly programs and sys- In addition to such "hard" data, most tems. They feel computers are sitting there to schools illustrate the phenomena of teacher appre- embarrass them, and they worry about making mis- hension towards computers. Some teachers will be takes and appearing foolishin front of peers and "users." They will make use of the computers any- students. time they can get them into their classrooms. They will come early and stay late to provide Some teachers associate computers with other extra computer time for their students- They technological developments of the past which will spend weekends reading computer journals or failed to live up to promises of revolutionizing writing courseware. They will get together with education. Educational television, teaching other teachers to share new findings and materials. machines for programmed learning, fully equiped They will go to computer shows and workshops eager language labs -- the list goes on. A lot of money to learn about tie latest technological develop- was spent, but now the machines collectdust be- ments. These are the "users". In the same cause they never fully succeeded to meet expecta- building there will be the "non-users."These tions. Teachers see computers as just another teachers will not make use of computers even of these fads that, too, will pass, costing a lot when available. Why? and producing little learning.

As suggested by the surveys cited above, Another fear might be ealled the "1984 Syn- many of them have fears, apprehensions, and/or drome", namely that computers will dehumanize misconceptions about computers.These feelings education, isolate students, cause them to lose are very real to them and effectively act as interest in human interaction, and lead to asocial barriers to their using computers in their class- behavior. In the extreme, the vision is of a rooms. A first step in overcoming these nega- situation in which people are treated like inter- tive feelings is to identify them and perhaps changeable parts of a huge machine. gain some insight into their causes. Effective Inservices In their work in schools beginning to implement computers, the authors have noticed the Considering all of these fears which exist in following attitudes and feelings among educators. different degrees among a large number of teachers, That these feelings are fairly common is suggest- what can be done? Positive attitudes towards using ed by articles expressing similar observations by computers are essential to successful implementa- other inservice coordinators (e.g. Rosen, 1982). tion of any school computer-based project.How can What specifically are these feelings? negative attitudes and fears be overcome, and er thusiasm toward computer usage be instilled? Fears and Misconceptions Listed below are suggestions and caveats for inservice coordinators and teacher-trainers in For some teachers, the technology itself planning those critical "first exposure" computer of computers is intimidating. They perceive workshops for teachers. These suggestions focus computers as overwhelmingly complex, beyond the on approaches as much as content, for experience understanding of ordinary people like themselves, has shown that the key to success is to alarge understandable only to geniuses who use the extent attitudes: lessen apprehension and in- mysterious jargon and master the intricate cir- crease enthusiasm. cuitry. Tbey are understandably reluctant to use equipment they are convinced they cannot 1. Begin the inservice with a guaranteed understand. For some, this mind set is antago- successful experience. As quickly as possible, nized by associating computers with mathematics, have the teachers sit at the computers and run a toward which they have deep-seated negative "friendly" intriguing program with valid educa- feelings. tional content. Using one of the Minnesota Educa- tional Computing Consortium* disks with a simula- Also creating anxiety is sensationalist tion such as Odell Lake or Oregon Trail has over- reporting about computers which simulate intelli- come a dozen or more cases of computer pbobia. An gent behavior, such as the use of anthropomorphic experience such as this establishes a relaxed imagery in some recent films. Some envision com- attitude. In the process the teachers find they puters as intelligent machines destined to replace them as teachers. They feel threatened, and fear a loss, or at least a drastic change, intheir * Minnesota Educational Computing Consortium jobs. They feel a loss of authority and control (MECC), 2520 Broadway Drive, Saint Paul, MN over their classrooms, over their comfortable, 55113. "Odell Lake" is on the Elementary tried-and-true methods, and ultimately over their Volume 4 disk, "Oregon Trail" on Elementary role in the educational process. They fear that Volume 6.

378 3Jj themselves can easily control the machine, and being replaced by computers. Do not avoid this that nothing is ruined by pressing the wrong key. issue; address it directly.After all, the compu- They begin to feel in control, finding that they ter was invented by man to relieve man of some of can control the machine and are not being con- the things he was having to do. In the same way, trolled by it. This hands-on experience does it can do some of the things teachers are now ex- much to dispell the mystique a computer holds for pected to do. If it could not, there would be many teachers. There will be time later, after justification for having it in the school. confidence is built, to discuss RAM's, ROMTT,ind Holmes° (1982) identifies the heart of the issue CPU's. Avoid technical jargon during this first with two questions: "First, does the teacher need experience. to perform all the teaching functions?" and,

". . . Can machines perform all, or even most, of 2. Have the teachers work in twos or the teacher's functions?" (p. 9) To address the threes at the machines. They are comfortable with first, show the teachers a well done gradebook or asking each other questions, whereas they may be other recordkeeping program, and emphasize how hesitant to ask an unfamiliar instructor. Work- the computer can be used to do recordkeeping and ing in small teams is very effective. Besides other repetitive paper work, freeing them to do helping to catch each other's errors, the team the more creative parts of teaching. Then have a approach lessens the tension which may be ex- frank discussion among the inservice participants perienced when first sitting at a computer. Sy listing the kinds of things that teachers do that himself at a computer, an adult may be so tense computers will never be able to do. This will do that he cannot assimilate the simplest directions, much to alleviate their fears. and will sit motionless, staring at a screen direction such as, "Press return to continue." 6. In choosing the content for computer inservices, remember that not all teachers need all 3. If computers are to actually be imple- levels of computer expertise. Do not overwhelm. mented in classrooms, the teachers must be con- For example, although most teachers should ex- vinced of their relative advantage over the perience programming in order to understand what methods and materials they are durrently comfor- a program is and how it controls the computer, table using. The key is to spark their interest they do not need to become proficient programmers. by getting them to see the potential of a compu- The myth that to use a computer a teacher must be ter in their classroom. That is, the computer a computer programmer has turned off more than must beTierceived as being relevant to high one teacher. The vast majority of teachers will priority items that they wish to achieve with use computers with pre-programmed materials, For their students, and as being better than what it them, computer literacy is being able to use the replaces, not just a nice aid that can be re- computer with appropriate materials, not being placed or done without. It must be compatible able to write computer programs. This is analo- with the existing curriculum, fitting the estab- gous to most of us with language literacy: we lished goals, values and needs. One way to show can use books, but most of us do not write them. this relevance is to identify a familiar topic the teachers feel is difficult to teach, and 7. Also in planning inservices, keep in mind illustrate a computer program that can help them that different people react in different ways to teach it. Then actually provide them with the computers. The inservice should provide a variety lesson plan and the materials. Ready to go, the of activities and approaches. If the session fo- computer-based lesson is more likely to be tried, cuses on only one approach, such as programming, and making that first try is an important step. only some of the teachers will begin to feel comfortable with computers. At the end of the ses- 4, Also necessary to getting teachers to sion, have the participants discuss which of the use and teach with computers is helping them with variety of activities they liked best and why. the classroom management aspects of computer use. This will help them understand their own reactions The computer will present an atmosphere of dis- to computers in comparison to others, and will covery and innovation that is unfamiliar to some help them when dealing with the various reactions teachers who use a traditional textbook/lecture that will surface among their students. As we approach. Computers will require a certain know, some students will become computer addicts, flexibility, including perhaps a role change in others will not. Teachers need to be made aware which the student knows more about computers than of this. the teacher. To help teachers with this, provide models of different classroom management These are only some of the possible sugges- systems during the inservice, including total tions for implementing effective computer in- group instruction using one computer with a large services that focus on providing not just the con- monitor, and small group work in a lab mode. Also, tent needed by teachers but also the attitude during the presentation, do not hesitate to admit needed to implement computers in the schools. Yes, it if something happens you, the presentor, cannot we need more funding; yes, we need more quality explain. It is a rare inservice using computers software; yes, we need more research data on the in which this opportunity does not occur. Admit- effects of computers in the teaching/learning ting we do not know everything, but use computers process. But these will come.Right now, in the anyway, does much to build confidence. schools, we need less apprehension and more computer enthusiasts. 5. As documented above, many teachers fear

379 396 References

1. Ahl, D. H. Survey of Public Attitudes Toward Computers in Society, in David H. Ahl(Ed.), The Best of Creative Computing, Volume 1. Morris- town, New Jersey: Creative Computing Press, 1976, pp. 77-79.

2. Bork, A. Computer Literacy for Teachers. In R. J. Seidel, R. E. Anderson and B. Hunter (Eds.), Computer Literacy: Issues, and Directions for 1985. New York: Academic Press, 1982, 91-98.

3. Bruwelheide, J.H. Teacher Competencies for Microcomputer Use in the Classroom: A Litera- ture Review. Educational Technology, 1982, 22(10), 29-31.

4. Diem, R. A. Developing Computer Educa- tion Skill: An Inservice Training Program. Edu- cational Technology, February, 1981, 21, 30-32.

5. Forman, D. Search of the Literature. The Computing Teacher, 1982, 9(5), 37-51.

6. Holmes, G. Computer-Assisted Instruc- tion: A Discussion of Some of the Issues for Would-Be Implementors. Educational Technology, 1982, 22, 7-13.

7. Lazarus, M. So Complicated, They're Simple: A reassuring Word on Classroom Compu- ters. Principal, 1982, 62(2), 39-41.

8. Lichtman, D. Survey of Educator's Attitudes Toward Computers. Creative Computing, 1979, 5, 48-50.

9. Martellaro, H. C. Why Don't They Adopt Us? Creative Computing, 1980, 6(9), 104-105.

10. Parker, J. and Hazuga, M. Educators' Attitudes and Knowledge of Computers in Jefferson County. Unpublished paper, University of Louisville, 1982.

11. Rose, N.R. Barriers to the Use of Educational Technologies and Recommendations to Promote and Increase Their Use. Educational Technology, 1982, 12(12), 12-15.

12. Stevens, D. J. How Educators Perceive Computers in the Classroom. AEDS Journal, Spring, 1980, 221-232.

13. Smeltzer, D. K. The Media Specialist and the Computer: An Analysis of a Profession's Attitude Towards a New Technology. T.H.E. Journal, 1981, 8(1), 50-53.

14. Watt, D.H. Education for Citizenship in a Computer-Based Society. In R. J. Seidel, R.E. Anderson and B. Hunter (Eds.), Computer Literacy: Issues and Directions for 1985. New York: Academic Press, 1982, 53-68.

380 3J 7 THE MICROCOMPUTER AS A TOOL IN EDUCATIONAL RESEARCH: A CASE IN POINT*

by Scott W. Brown and Daniel B. Kaye

Department of Educational Psychology, University of Connecticut Department of Psychology, University of California, Los Angeles

Abstract and the availability of a number of micros to conduct a study involving a large number of subjects. Microcomputers have had a major impact in the area of educational research. The present paper With the stage set, we would now like to pre- discusses some of the major advantages of using sent a program of research that has been (and microcomputers in cognitive research and presents still is being) devoted to investigating various several examples of our own research programs uti- types of cognitive processing using microcomputers lizing the microcomputer. The discussion touches as research instruments.This discussion will upon the specific hardware configuration, the soft- touch upon the hardware and software, program ware, the task demands of the experiments and the development, experimental administration, data transfer and analysis of the data. Conclusions collection, data transfer and statistical ,analyses focus on the rationale for the use of microcomput- relevant to this research program. ers in educational research discussing such factors as accuracy, reliability and flexitlity of the The research has been conducted cooperatively hardware and software. at two different sites; the University of Connec- ticut and the University of California, Los Ange- Introduction les. Each site has been equipped with the identical microcomputer equipment; thus the same experi- The use of microcomputers for educational and ments can be conducted simultaneously at both psychological research has become prevalent in re- sites without concerns for differences in computer cent years. No longer are computers relegated to programs. the statisticians of each research group to crunch and massage data for analysis purposes. With the The purpose of our program of research is to introduction of smaller and cheaper computers, examine the cognitive processes of children and first the minicomputers and then the micros, many adults, specifically, the acquisition and use of a researchers saw an opportunity to use these ma- variety of automatic and controlled information chines as stimulus presenters and response record- processes related to intellectual development in ers, in short, as research tools. From a research- general and reading skill development in particu- er's point of view, what could be more efficient lar. Tasks were developed requiring letter than collecting the data on a computer that may searches within different letter configurations also analyze the data, or at least minimize data and arrays, lexical decisions (making a word/non- loss, by eliminating the procedures for having word decision based on the presentation of a let- assistants punch data intothe computers for later ter array), and case judgements (upper case versus analyses? lower case) -- typical experimental tasks used by cognitive psychologists, and until very recently Recently, Johnson has discussed some of the involving the use of slide projectors and tachis- important implications of microcomputers in re- toscopes. search]. The discussion can be summarized by listing the advantages and disadvantages of this posi- The microcomputer hardware used in this re- tion. The advantages of the use of microcomputers search program consists of: for research are: replication and extension, programmatic research, measurement, experimental con- 1. A North Star Horizon microcomputer equip- trol, administration and experimental economy. The ped with 64K of RAM memory, dual single- disadvantages include the time and the level of sided double density 5 1/4 inch floppy training necessary to develop a functional program, disk drives (180K capacity each); 2. A Televideo 950 black and white video monitor; Portions of this research were supported by a grant from the University of Connecticut Research 3. A Mountain Hardware 100,000 day clock, Foundation, grant. No. 1171-000-22-0401-35-899 accurate to within one-tenth of one awarded to the first author. millisecond;

381

398 After the data were "cleaned up" 4. A Livermore Star modem; mission errors. the data were analyzed utilizing a major statisti- 5. An Epson MX-80 F/T printer. cal package on the IBM system.The mainframe was This hardware configuration was selected as the used for the analysis because of the large amount most cost effective system available on the market. of data and the number of disks involved. In addi- In our opinion, the speed, power, storage capacity, tion, transferring the data to the mainframe freed expandability and software compatability of the the microcomputer for continued data collection and system surpassed its' competition. program development.

The software consisted of North Star Basic Experiment #2 CP/M, Pascal/M and two programs to interface the The procedures for experiment 2 are very North Star with an IBM for data transfer through similar to those of experiment 1, except for the the modem (one for the North Star Basic programs locations. Data were collected at both Yale Univ- and one purchased for the Pascal programs using ersity and University of California, Los Angeles. CP/M). All other software, i.e., task programs and This experiment examined the development of auto- stimulus lists, were written and developed in our matic word recognition in children and adults. labs in North Star Basic or Pascal/M. Subjects were presented with a configuration of words/pseudowords and upper case/lower case. Sub- The Research jects were requested to make a binary decision re- The following is a discussion of four experi- garding different rules for response using these mental paradigms that have been developed in our two dimensions of the stimuli. Subjects were re- labs. These paradigms have been used to investi- quested to make their responses as rapidly and ac- gate reading skill acquisition (experiments 1-3) curately as possible. and the relationship of intelligence to reaction time (experiment 4).These experiments have been Data were collected on both children and adults conducted succes ;fully with children as young as and transferred to a mainframe for analysis using the second-grade and up through college. All para- the same procedures, as those described in experi- digms involved the use of microcomputers to present ment 1. stimuli and record responses. All forms of coun- terbalancing, timing, and sequencing of events The data collected and analyzed from the first were programmed and under the controlof the North two experiments were collated into a study of read- Star. ing skill development and contextual facilitation in early and skilled readers.2 Experiment #1 The first program that we would like to dis- Experiment #3 cuss was used to investigate the effects of con- Experiment 3 was funded under a grant to the textual facilitation in children and adults. We first author by the University of Connecticut Re- used a dual task of lexical decision and letter search Foundation to investigate several of the search with words and pseudowords presented on the components of reading skill development in child- screen of the Televideo 950. The subjects were ren3. This experiment is actually a combination of required to respond to the letter search or lexical three experiments utilizing the equipment describ- decision by manually pressing one of two predesig- ed above and a single program. The three experi- nated response buttons as rapidly and accurately as ments involved a microcomputer and a paradigm deve- possible. The program stored the condition (which loped by the first author3,4. The paradigm, a of four tasks), the stimuli, the trial number, the primed lexical decision task, requires amicrocom- reaction time, the subject number and whether the puter to present a prime stimulus very rapidly on response was correct or not. the screen within a two degree visual angle (the window) of a fixation bar and then present target This experiment was conducted in two loca- stimuli in the center of the screen above the bar. tions. One research site was the laboratory school The subject is to make a lexical decision (a word/ associated with the University of California, Los nonword decision) on the target stimuli as rapidly Angeles. The other research site was a rural and accurately as possible. The experiments used school district near the University of Connecticut. semantic, phonological or orthographic primes in This experiment used students in grades 3 and 6 as three separate experiments. The study involved 20 subjects. Subjects from each of these grades were subjects from each of grades 2 through 4 and col- included at each research site. In addition, to lege subjects for each of the three experimental the data collected on the children, college sub- conditions. jects were tested at each university site. A total of 12 subjects from each grade were included in the The reaction times, response, stimuli, condi- study. During the experimental sessions more than tion (type of prime) and trial were recorded for 300 reaction times and responses were collected on each subject. Subjects were presented with 256 each subject. trials broken into 8 blocks of 32 trials each. Subjects were encouraged to rest between each block Once the data collection process had been com- to reduce any fatigue effect. The subjects were pleted for all subjects, the data were transferred able to initiate the task once they had completed to an IBM mainframe computer via a microcomputer the rest period by pressing any button.As an interface program and a modem. Upon completion of added safety feature, if a subject's fingers drift- the transfer, the data were inspected for trans- ed off Of the response keys and onto any of the

382 3 9j r.

other keys, the bell (built into the keyboard) was stimuli was eliminated. sounded to alert the subject and the experimenter. Also, instructions were printed instructing the The results of this experiment indicated that subject that his/her fingers were no longer on the as the amount of cognitive processing increases, correct response keys and the subject was request- the relationship between intelligence and reaction ed to realign them. The trial that was presented time does not increase6. Our results seem to in- during this process was scored as incorrect and not dicate that the relationship is fairly constant included in the statistical analysis. Through across these two tasks. Our next experiment in these procedures the experiment was under near com- this area will involve the use of a more difficult plete control by the North Star. task involving taxcnomic or categorical judgements and will use a hybrid of the program developed for All data were transmitted to the IBM mainframe the simple and choice reaction time tasks. and analyzed using similar procedures to those described previously. Summary

Although this data are still in the process of Four experimental paradigms utilizing microcom- being analyzed, we are very hopeful that through puters to present stimuli and record data have been the use of this specific paradigm and the accuracy presented and discussed. The advantages of using provided by the microcomputer and the hardware micros as research tools as discussed by Johnson clock, the subtle patterns in reaction times re- are certainly truel. We have found that through sulting from the differential use of the primes in the use of this equipment we have been able to ef- processing the targets may be delineated and pro- fectively conduct a number of studies within a re- vide information regarding the development of auto- latively short amount of time and make a signifi- matic semantic, phonological and orthographic pro- cant contribution to the field of cognitive psych- cessing in children. ology. Without this equipment we would still have been able to conduct similar studies but the use of Experiment #4 this equipment has certainly facilitated the development of our research programs and allowed for This experiment involved the investigation of collaborative efforts to flourish over 3,000 miles. the relationship between the speed at which a subject may process some very elementary information As a final note we would like to comment on the and scores they may receive on portions of an in- two disadvantages of using micros as research tools telligence test. Recently, this topic has been highlighted by Johnson. The first concerns the discussed by Arthur Jensen utilizing fairly simple amount of time and skill involved in program deve- equipment5. Our purpose was to examine some of the lopment. While many of these tasks discussed may hypotheses presented by Jensen and utilize a micro- appear to be very simple, the initial development computer to display and record the stimuli. whether in Basic or Pascal requires a substantial amount of sophistication in programming as well as The study involved college and elementary school children as subjects and utilized two simi- the requirements of the tasks, i.e., the types of data required to store, speed of stimulus presenta- lar tasks. One task was a simple reaction task in tion, placement of targets. This is not to imply which the subject was instructed to press a single that one must be a computer wizard to use a micro button as soon as he/she saw anything on the as a research tool, but some programming experience screen. The second task required the subject to is important on a research team. However, the ul- watch the screen and press one button if an "X" timate payoff is worth the effort. For example, appeared and a different button if an "0" appeared. with our hardware and software configuration we are The reaction time from the onset of the stimulus to able to present stimuli within milliseconds of one the manual response by the subject was recorded for another and to record responses accurate to within each trial. Each subject was presented with 200 one millisecond. trials of each task. The second issue regards the limited number of In order to facilitate recognition of the subjects that may be included in a study if you stimulus when it was presented on the screen, a box only have one or very few micros. Both of us have was drawn prior to each trial. Subjects were in- only one North Star at our disposal at each site, structed that they were to watch the box because yet we each have been able to gather a substantial the stimulus was to be presented only within that amount of data from a relatively large number of box, thereby drawing the subject's attention to the subjects. The series of three experiments discuss- proper location. The box occupied the center of ed under experiment #3 examined over 320 subjects the screen and comprised approximately the central in grade 2 through college. 25 percent of the screen.

The stimuli were presented in a predetermined In summary, the use of the microcomputer as a order for the second task because complete randomi- research tool may revolutionize educational and zation would have resulted in an unequal number of psychological research. By using computer pro- X's and 0's over the 200 trials. However, a random grams research may be replicated precisely and ex- number generator was used to determine the inter- tended with minor software modifications. Program- trial duration, since in both tasks anticipation of matic research is facilitated through the develop- the stimuli could enhance responses. Through this ment of a master program from which hybrid programs process each intertrial duration was independent may be developed by altering such things as inter- of the others and the effects of anticipation of trial stimulus duration, sequencing of events, or

383 4UU instructions. The accuracy of the measurment of such things as reaction time has certainly been increased through the utilization of sophisticated hardware and software. Experimental control is certainly optimized through the use of the micro as a stimulus presenter and data recorder.The use of the micro to administer a task, present identical instructions, and control most aspects of experimental administration is without parallel. Once the computer program has been written the use of the micro is extremely economical. In conclusion, the use of a microcomputer as a research tool can be extremely advantageous to anyone interested in developing programmatic research in a variety of fields. References

1 Johnson, C. W. Microcomputer-administered Research: What it Means for Educational Researchers, Educational Researcher, 1982, 11, 3, 12-16.

2Kaye, D. B. & Brown, S. W. Contextual Facilita- tion in Early and Skilled Readers. Submitted for publication, 1983.

3 Brown, S. W. Reading Skill Development:A Ques- tion of Component Processing. A Research Grant submitted and funded by the University of Con- necticut Research Foundation, 1981.

4 Brown, S. W. A Developmental Examination of a Primed Lexical Decision Task, Unpublished Doc- toral Dissertation, Syracuse University, 1980.

5 Jensen, A. 'g': Outmoded Theory or An Uncon- quered Frontier? Invited address presented at the Annual Convention of the American Psycho- logical Association, Toronto, Ontario, Canada, 1978.

6 Brown, S. W. & Boretz, H. Cognitive Processing and the Speed of Processing Information in Simple and Choice Reaction Tasks. Paper presented to the National Association of School Psychologists Conference, 1983.

384 STRATEGIC CONCERNS IN ESTABLISHING AN ELEME7PARY SCHOOL MICROCOMPUTER INSTRUCTIONAL SYSTEM

by Ronald Bearwald and Theodore Bargmann

School District 74 Lincolnwood, Illinois Abstract best captured in the words of renowned psysicist and computer science educator Alfred Bork: "At The microcomputer has become a most viable the present time we see the formation of another instructional device. There can be no doubt great wave that involves the use of the computer that it can enhance and facilitate learning. in learning. The computer promises to revolu- However, we dare not lose sight of .our educa- tionize education more than any other develop- tional purposes as we become enamored with the ment of our time."2 captivating characteristics of the microcanpu er. Planning and foresight are essential to As one assesses the value of microcomputer - insuring the successful contribution it can based instruction, a healthy perspective is to make to education. This article highlights the view the vast potential and promise as well as strategic concerns of prior planning that lead the possibility of failure.The enormity of the to effective use of the microcomputer in an immediate impact of the microcomputer can be ex- elementary school setting. (1) Adopt a devel- plained in part by its many intrinsic qualities. opmental approach. (2) Have a broad base of It has the capacity to provide immediate rein- participation in planning. (3) Assess and so- forcement of response even beyond wildest Skin- licit Board support early. (4) Activate com- nerian expectations. It can allow a relatively munity support. (5) Build your program on a naive learner to perform complex learning opera- curriculum. (6) Begin with computer literacy. tions with a few simple commands. Furthermore, (7) Designate adequate personnel. (8) Provide it uses its memory of pre-programed information for a broad base of student participation. to engage any student in logical reasoning re- (9) Establish a multilingual program. gardless of ability. Finally, there is no draw- back in that it has the capacity to generally be colorful, noisy and just plain fun. These and other inherent qualities have caused us to become hnmediately enamored with and receptive to the potential of the microcomputer. It is clear that a vast number of private and public, elementary and secondary schools Despite many characteristics which have throughout the country are currently investi- encouraged many to jump on the microcomputer gating or using microcomputers. A recent study "band wagon," caution should be exercised. Re- conducted by the U.S. Department of Education's meMber, afterall, that microcomputers are only a National Center for Education Statistics indi- delivery system and even the most rudimentary cates that the number of microcomputers used student of the art is familiar with the GIGC, for instruction in public schools has tripled garbage-in, garbage out credo. As has been between 1980 and spring of 1982. The amount of noted, the. microcomputer can be viewed as a microcomputers now in schools, which is approx- "double-edged sword which can cut both ways. imately 96,000, is expected continue to grow While improving the human condition, the micro- rapidly.' computer can create problems and mistakes."3 Without appropriate planning and/or support, To be sure, technology has assumed an in- the value of microcomputers may, indeed, be il- creasingly more legitimate role in teaching lusionary and transitory.To coin and at the and learning during the past two decades. same time reverse a familiar phrase, without Taken in this context, the application of mi- a thorough and thoughtful approach microcomput- crocomputers to education is perhaps an under- ers used in education could become the "silk standable phenomenon. In fact, never before purse" that turns into a "sow's ear." has new technology created such a rapid and profound impact upon the educational communi- At this time, the Lincolnwood School sys- ty.' Never before has an instructional inno- tem is operating a successful microcomputer vation received such widespread support from instructional program. The following were key teachers, administrators, students, and parents. considerations in establishing this program. Never before have these groups been so unanimous in their resounding surety of the value ADOPT A DEVELOPMENTAL APPROACH of a technological system to the educative process. Perhaps the current sentiment is That is to say, allow for growth of your

385 40 2 program in stages or phases.The value of view- IMPLEMENTING ing the innovative process in this manner was clearly stated by Everett Rogers who described This stage places your program into full stages in the adoption of innovations as (1) a- opczation. Teaching and hopefully wareness, (2) interest, (3) evaluation, (4) trial, learning begins. Of course, this as- and (5) adoption.4 This approach provides two sumes that these prerequisite tasks have important aspects of developing a successful concluded: (1) Board of Education ap- program foundation: proval of your plan, (2) hardware and software purchased and in place, (3) ap- 1. A conscious recognition which real- propriate personnel appointed, (4) cur- izes that the gestation period of an riculum developed, (5) students grouped effective microcomputer program is and scheduling arranged, (6) profession- months if not years. This view al staff oriented and receptive. should prompt an appropriate, long- term commitment from those involved. ASSESSING

2. Phases or stages which can be as- This phase is on-going and absolutely sessed and enjoyed as milestones of essential to make your program a "sys- success along the way.This in it- tem." Your efforts here will provide self can be a motivating force since valuable data for continuation, expan- it provides a basis for progress sion or re-direction of your use of that is both incremental and contin- microcomputers. Care should be exer- uous. cised to develop a model which will assess both the cognitive and affect- Our experience has shown these phaes to be es- ive impact upon students as well as sential: the attitudes of faculty.

PLANNING HAVE A BROAD BASE OF PARTICIPATION IN PLANNING

This phase allowsfor the gathering, ana- Whenever possible, use the comnittee ap- lysis, and synthesis of data from a va- proach which enables a high degree of teacher in- riety of sources such as readings, con- volvement. It should be taken as good advice sultants, teachers, and the community. that "we must rely very heavily on the intuition Site visitations to assess on-going proof good teachers" in order to determine how "to gram are also a valuable source of in- best proceed and as to where the computer will be formation. The planning not onlypro- most effective in the learning process."6 In vides a format for gathering informa- assembling a committee, insure that representation and formulating goals, but also tion includes broad general areas (reading, math, serves as an in-service training compo- science) as well as more specialized teaching nent for the planning body. operations (gifted, library/learning resources, special education). It is also advantageous to PROPOSING include the sophisticated as well as the naive in terms of each person's prerequisite knowledge During this phase we should seek to in- of computers. As an example, our committee conform and motivate the ultimate decision sisted of the following representation: making authority, which in educational settings is often the Board of Education. - administration Broad program parameters should be clar- - mathematics ified and feedback solicited. - science - primary teaching (grades K-2) INITIATING - intermediate teaching (grades 3-5) - upper grade teaching (grades 6-8) This is equivalent to a trial stage - library/learning resources wherein the innovation is presented and - reading used on a small scale to determine its - gifted education usefulness when the program is fully - special education implemented. In our situation, this involved providing a structured in- In providing for such broad-based represen- service orientation program as well as tation, you encourage dialogue and broaden per- placing single microcomputer bySLeMS in spective about possible applications. At the each school building for teacher and same time, you are establishing a cadre of edu- selected student use.Testing a proto- cators that have initial ownership and invest- type and engaging in preliminary eval- ment in the program and will, no doubt, provide uation has proven to be invaluable to a base of on-going and future support. introducing microcomputers into the educational environment.5 ASSESS AND SOLICIT BOARD SUPPORT EARLY.

There's no need to keep any secrets from your Board of Education which will ultimately

386

It) 21 take action that will either launch or doom your dividual basis. intended program. Don't wait for THE FINAL RE- PORT OF THE MICRCOOMPUM COMMITTEE. Keep them The same opportunities for involvement in informed along the way by taking interim reports both planning and utilization should be directed and encouraging dialogue. "Yes, we believe mi- at other public bodies and/or institutions and crocomputers have potential for our instruction- civic organizations. Library boards, school al program." "Yes, we are currently investigat- boards, village boards, zoning boards, chambers ing this in a systematic manner." "Yes, we will of commerce, etc. often provide an interfaced be asking you to support the impending microcom- networking of people whose support is crucial. puter instructional program with both hardware When you can work in concert with these groups and software purchases." to establish common goals and utilization patterns, your efforts can be complimentary if not By approaching information sharing in a advantageous. straightforward manner as this, expectations and support are clear while hidden agendas are In our situation, we were able to obtain kept to a minimum. financial support from our local business community by demonstrating a viable instructional ACTIVATE COMMUNTTY SUPPORT. plan for using microcomputers. This type of enterprise has been effeCtive in numerous situ- An important compliment to Board interest ations throughout the country. However, will- is that of community support. In attempAng to ingness of the community to contribute seems to amalgamatesupport from the citizenry it is es- hinge on their awareness of the relevance of sential to consider both the focus and timing of microcomputers to the curriculum.? your efforts. In targeting one's energies, one must realize that community support may come from BUILD YOUR PROGRAM ON A CURRICULUM. a number of sources: The ultimate success of your program will - parents with children now in vour schools. hinge upon how educationally sound it is and/or - parents whose children once did but no whether or not learning occurs. Many micro- longer attend your schools. computer programs are often doomed to failure - residents who do not or never have had from the start as a result of not being firmly children in your schools. based Oh a curricular plan. We must never for- - former students now in high school or get that "the major problem with oomputer-based college. learning is not a hardware problem, but a learn- - other public institutions or bodies ing problem. To focus on the hardware is an who are directly or indirectly connect- error, drawing attention away from the major ed with your schools (parent /teacher issues that should .be considered."8 associations, public libraries, etc.) - civic organizations (Chamber of COm- Initially, when the microcomputer hardware merce, Rotary, etc.) is in place, the tendency is often to use com- - the business community. mercially prepared software as the main framework of the instructional program.This ap- All of these "publics" would have a differ- proach can be dangerous on a number of counts. ent vested interest in either supporting or not First, there is a substantial belief that the supporting your microcomputer program. Of course, majority of electronic courseware is "intellect- parents of children currently in your schools ually bankrupt" and, indeed, "mental chewing gum will be the greatest single source of support for instead of protein."9 Using commercially pre- your program once they realize the potential of pared software as an expedient way to get people microcomputers to increase their children's to use computers is at best a weak rationale. learning. However, it is important to remember In doing so, we merely allow ourselves to con- that parents of school children are often far tinually sidestep crucial curricularissues by outnumbered by residents who have no direct con- adopting selected software in a wholesale manner nection with schools. As an example, in our without judging its value of applicability to community, households without children in attend- our situation. In fact, evaluation and selection ance at school currently outnumber homes with of microcomputer software should always be done school children by a five-to-one margin. Under with a =Oman of teacher involvement who base these circumstances, it becomes clear that the their decisions on an accented set of criteria tendency of the community in general to support and an established curriculum.1° a microcomputer program increases proportionately as citizens perceive uses for the microcomputer To be successful any instructional system beyond the school.. itself. During initial phases, using microcomputers must be based upon a cur- the school must provide for community input in riculum - a sequenced set of learning objectives planning as well as for participation in orien- and activities. Tb be sure, the curriculum doc- tation. As the initiation and implementation ument provides a plan for learning. However, phases begin, opportunities which personalize it also addresses cost effectiveness and serves microcomputer use must be sought by offering as a strategic management tool.11 Microcomput- evening and weekend workshops as well as in- ers obtained with a considerable expenditure of stances to utilize the microcomputers on an in- capital outlay should not lie dormant nor be

387 404 subjected to capricious and haphazard use. activities and deal with (rather than While some random and exploratory use by stu- avoid) other important curricular is- dents should be allowed, the core instructional sues. program should be constructed in order to maximize microcomputer use while insuring that 2. provides a basis for uniformity and every minute of student participation is produc- continuity of instruction. t ive. 3. insures mastery of basic computer and After expending considerable money to pur- reasoning skills which have universal chase microcomputers, it makes little sense to application to all areas of learning. be parsimonious in establishing a curriculum. While curriculum development may be initially 4. presents an instructional program time consuming and even costly, it will not only which is accountable to varying skills provide direction but a necessary framework for of students. assessing the success of microcomputer-based instruction. 5. prompts teacher orientation, acceptance, involvement and interest. BEGIN WITH COMPUTER LITERACY. Definitions of computer literacy remain Essentially there are two broad uses for varied. Some choose to think of it in terms of microcomputers in an educational setting: two major components: (1) computer awareness and (2) oanputer programmimg.IXA consensus re- 1. to supplant or supplement instruc- view seems to include these cannon components: tion through application of comput- (1) how computers are used, (2) what a computer er-managed or computer-assisted can and cannot do, (3) what a program can and instruction. cannot do, (4) how computers work, (5) how to use a computer, (6) the impact of computers on 2. to teach the elements of computer society, (7) how computers can develop skills literacy including programming. of decision making and coping with change and, (8) an introduction to or awareness of pro- Both uses, of course, have merit so it is not gramming.13 The Lincolnwood Schools have adopt- suggested that deciding how to apply microcon- ed the following broad curriculum goal: "To puters be looked upon as an either/or issue. develop computer literacy by teaching important However, when beginning, it is important to computer relatedconcepts, increasing awareness nrA.row the focus of one's energies in order to of the values and applications of computers in eralance the possibility of success. Ultimate- our world, and providing opportunities to attain ly, the question becomes not "which approach ?" a certain level of competency in perforting but "which approach first?" fundamental computer operations."

It is, of course, easier to begin a pro- Regardless of the operational definition gram by relying upon a CAI (computer-assisted of computer literacy to which you adhere, one instruction) approach using prepared software could not disagree with the position of Daniel as its basis. Often this results in purchas- Watt, Director of the Conputer Resource Center, ing and "plugging in" commercially prepared Cambridge, Massachusetts who argues the impor- software either adopting them in total or tance of computer literacy. States Dr. Watt, adapting them as needed. The idea of using "Universal computer literacy is a basic skill commercially prepared software is not in it- of the 1980's and deserves a major role in the self abhorant. However, we must never lose school curriculum."14 Maxine Greene of the sight of the fact that CAI should never be an Teachers College, Columbia University, sees end, but merely a means to an end. The most literacy as a way "to learn how to think con- legitimate rationale for adopting a CAI approach ceptually, to structure experience, to look should be to improve ;earning within the context through wider and wore diverse perspectives at of the curriculum. Consequently, such an ap- the lived world."1-Furthermore, Dr. Greene re- proach should be implemented only after consid- minds teachers that "literacy ought to be con- ering such key questions as "In what way can ceived as an opening, a becoming, never a fixed microcomputers teach something better than we end."16 are currently doing it?" In this manner, care will be exercised in integrating CAI into the Perhaps this is the most useful perspect- accepted framework of curriculum. ive of all. For, if teachers can grasp the universality of computer skills to learning In our judgment, initiating a program on without feeling they are about to be replaced the basis of a sequential curriculum of com- by a machine, they are more likely to engage in puter literacy skills provides a viable be- training, evaluation of software, development ginning as well as a lasting foundation. The of courseware and curricular integration of reasons are numerous. Using microcomputers to software - all a prerequisite to success in any teach computer literacy skills: computer-managed or assisted approach. When one views literacy, as Dr. Greene does, in terms of an 1. forces us to specify our goals and

388 4 u5 "opening," the case for beginning with computer corporation into the educational program, then literacy becomes quite clear. After all, what all students should have access to than and the better place to begin than at the beginning. related skills that they teach.After all, if acquisition of computer literacy has the capa- DESIGNATE ADEQUATE PERSONNEL. city to improve learning in all areas, then access to this skill improvement should not be ex- In considering what resources are necessary clusive. Approaching the initiation of such a to implementing a microcomputer program, school program as based upon broad student involvement systems must never underestimate the value of serves to underscore the magnitude of one's com- providing adequate human resources. Serious con- mitment to the belief that microcomputer litera- sideration should be given to the appointment of cy skills are a valuable component of the curric- an individual who can coordinate the microcomput- ulum and important to each student's learning. er instructional program on a full-time basis. Someone must provide continuous impetus to the In our situation, once we agreed that the program, especially during its early stages and teaching of computer literacy had a legitimate to focus efforts to integrate microcomputers place in our curriculum, we proposed to provide into the educedonal setting. relevant learning experiences to heterogeneous groups of students at the primary; intermediate In addition, this person will need to be a and upper grade levels. In doing so we hope to: trainer, curriculum developer and teacher. All of these activities will not only benefit the 1. encourage and obtain a serious initial program, but also will enable this person to re- commitment from the local Board of main in touch with the substantive issues of mil- Education rather than expanding the crocanputer learning thereby enhancing his/her program with a series of add-on pro- credibility in the coordinator role. It goes posals. without saying that this person will also need to be a program advocate, ombudsman, promoter 2. provide computer skills access to all and overall Iiampion of the micromaguter cause. students thereby engendering a widespread feeling of ownership and in- When selecting a coordinator, emphasis volvement by all students and subse- should be given to appointing an educator rather quently their teachers and parents. than a technician or programmer. Someone who approaches the task with a clear understanding 3. create a substantial hardware base of the educative process will not only enable and mainstream instructional program your learning objectives to be achieved, but will from which additional specialized also have a better chance to be perceived by col- programs and microcomputer uses leagues as a credible leader and therefore ac- would derive and evolve. cepted. It is even more desirable to designate, whenever possible, someone who is currently em- ESTABLISH A MULTILINGUAL INSTRUCTIONAL PLAN. ployed in the local school system, as their credentials as an educator will have been establish- By this we mean, make a serious commitment ed. to expose students to more than one computer language. Commonly, students have been intro- PAN= FOR A BROAD BASE OF STUDIENT PARTICIPA- duced to microcomputer programming through the TION. BASIC language. Though this may be the most universally used language throughout elementary The tendency of many schools has often been schools, it may not be the most appropriate for to approach the initiation of microcomputer in- the young learner. struction in a piecemeal fashion. The idea is often to get a "foot in the door" by obtaining Before we proceed perhaps it would be ad- one or two microcomputers and place them where visable to reiterate our belief regarding the it can be presumed they will be accepted and importance of programing experiences at the successfully applied. By building a showcase elementary level. To be sure, elementary stu- program with gifted students or perhaps in math dents should engage in programming as part of or science, it is hoped that we can trigger a their computer literacy' instruction. Too muCh catalytic reaction from which the program will is to be gained to ignore the potential of this grow and expand. This premise as a beginning type of learning. As indicated by Seymour is really not all that bad. The problem is Papert,"The child programs the =muter and in that many programs never get beyond this point. doing so, both acquires a sense of mastery over What is thought of as a beginning becomes per- a piece of the most modern and powerful technol- petuated as a series of unrelated and often ogy and establishes an intimate contact with isolated programs with no common thread of con- some of the deepest ideas from science, mathe- tinuity or curriculum holding than together or matics,and the art of intellectual model build, providing direction.

If a microcomputer instructional program Furthermore, Alfred Bork notes that when is to be successful over time, it must be com- we allow students to engage in programming, mitted to provide learning for all students. "we are giving the student an important in- Assuming that microcomputers are worthy of intellectual tool, an increasingly critical mode

389

406 for all areas of the future." 18 REFERENCES After reaching consensus regarding our in- 1. "CPR News Briefs," Curriculum Product Review, tent to include programming in the computer December1982, pp. 6-7. literacy unit, we determined there would be a 2. Alfred Bork, Learning With Computers (Bedford, great deal of merit in providing a development- Mass.: Dijital Press, 1981), p.3. al approach which would introduce students to 3. H. Dominic Covvey and Neil H. McAlister, more than one computer language. Logo, for Computer Consciousness: Surviving the Auto- example, and its "turtle" graphics component is mated 80s (Menlo Park; California: Addison"* highly motivating, does much to eliminate syn- Wesley, 1980), p. 6. tax errors and allows primary age students to 4. Everett M. Rogers, Diffusion of Innovations engage in programming with a minimum of train- (New York:The Free Press, 1962), p. 81. ing.I9 The use of such child- appropriate lang- 5. William H. Pritchard, "Introducing Instruc- uages as Logo, can enable children to learn to tional Computing into the Educational En- control a microcomputer in tiw same manner that vironment," Electronic Learning, September they learn to read or write. Users of the 1981, p. 24. programming language PILOT have praised its 6. Bork, p. 7. simplicity and clarity which allows the student 7. Robert Neumann, "Doing Business with Busi- with no prior computer experience to interact ness: How to Raise Money in Your Commu- with the machine in a human way. It provides an nity," Electronic Learning, September alternative to algebraic languages and is com- 1982, p. 43. posed of powerful and nearly syntax-free conver- 8. Alfred Bork, Learning-Not Hardware-is the sation-processing instructions. '1 We chose to Issue," Electronic Learning, September couple these two languages with the widely used 1982, p. 13.

BASIC to create the following microcomputer 9. Epiegram, May 1982, p. 3 . language sequence used in our computer literacy 10. J. D. Gawronski and Charlene E. West, instruction: "Computer Literacy," ASCD Curriculum Up- date, October 1982, p. 6. Grade Level Computer Lang' 11. FenwiCk W. English and Betty E. Steffey, "Curriculum as a Strategic Management Primary (1-2) Introduce: Logo Tool," Educational Leadership, p. 277. Intermediate (3-5) Review: Logo 12. Gary G. Bitter, "The Road to Computer Introduce: PILOT Literacy: A Scope and Sequence Model," Upper Elementary Review: PILOT Electronic Learning, September 1982, (6-8) Introduce: BASIC p. 60. 13. Gawronski and West, p. 3. CONCLUSION 14. Daniel H. Watt, "Computer Literacy: What should schools do about it?" Instructor, The application of both formal and informal October 1981, p. 87. procedures to effect change will, of course, vary 15. Maxine Greene, "Literacy for What?" Phi in accordance with the attitudes and resources Delta Kappan, January 1982, p. 329 present in the situation.Each school or school 16. Greene, p. 326. system must approach the introduction and adop- 17. Seymour Papert, Mindstorms (New York: tion of new technology in a manner which is Basic Books, 1980) suitable to the local educational setting.The 18. Bork, p. 6. introduction of microcomputers for learning may 19. Molly Watt, "What is Logo?" Creative Clan- require that we give attention to a unique set puting, October 1982, pp. 112-113. of considerations. We have attempted to focus 20. Watt, p. 86. issues which were important. to the integration 21. Rita May Liff and Keith Vann, "PILOT: A of microcomputers into the educational program Programming Language for Beginners," of the Lincolnwood Schools.The relevance of Interface Age, September 1978, pp. 64-67. any innovative procedures which you choose will 22. Covvey and McAlister, p. 12. be tested in light of the impact that microcomputers make upon the system at large - in fact, upon learning.

"A computer system does not exist in a vacuum. It is always a part of a larger human

A computer system should never be an end in itself. Its success or failure is measured by the success it effects in the situation in which it is used.Therefore, if computer systems are to serve people's needs, they must be carefully integrated into the human and procedural domain they are intended to improve.,,22

390 4 u EVALUATION OF MICROCOMPUTER SOFTWARE: HOW VALID ARE THE CRITERIA AND PROCEDURES?

Robert M. Caldwell, Ph.D. The University of Texas Health Science Center at Dallas Department of Allied Health Education Dallas, Texas 75235

This paper presents three major issues related to the are loaded with graphics and flashy screen displays might evaluation of educational software. The first relates not be as effective as those that are quite simple in to the nature of criteria used to evaluate software. appearance and presentation. Few of the criteria used to evaluate software have In short, the problems inherent in evaluating educa- been validated through research and experimentation; tional software can be traced to three major issues. instead, they often find their basis in speculation and The first relates to the nature of the criteria used to intuition only.Criteria for evaluation are frequently evaluate software. Few of the criteria used to evaluate highly inferential in nature which, of course, makes software have been validated through research and ex- them highly subjective.Subsequently, subjectivity perimentation; instead, they often have their basis in can only serve to lower reliability among separate speculation and intuition only.Criteria for evaluation ratings. are frequently highly inferential in nature which, of A second problem related to the evaluation of course, makes them highly subjective. Subsequently, sub- software is the qualifications of the individual raters jectivity can only serve to lower reliability among sep- themselves. Recent studies have demonstrated wide arate ratings. vatiances among ratings by individual raters. A second problem related to the evaluation of soft- A third problem in evaluating software stems ware is the qualifications of the individual raters them- from a general lack of knowledge about how com- selves.In a recent study by Blum (1982) wide variance puters should be used in instruction. Many was found in ratings of software by three or more re- school districts are acquiring micros without first viewers.This variance was attributed to designing classroom models for instruction. There- A.Error in scoring due toinadeuqate training or fore, much software is purchased without regard to background of the reviewers. how its objectives fit into existing curricula and the B.Subjective judgments of the evaluators on items overall goals of the school. that were highly inferential. A third problem in evaluating software stems from a general lack of knowledge about how computers should be used in instruction. The emergence of the inexpen- INTRODUCTION sive microcomputer has prompted many school districts In the past several years there has been a prolifera- to purchase them. Unfortunately, many of these districts tion of software developed to be used with microcomput- are acquiring these micros without first designing class- ers in elementary and secondary schools.Unfortunately, room models for instruction. The result is that much of much of this software varies greatly in its quality and the software is used and therefore reviewed within con+ scope.In an effort to distinguish quality software from textsthat might be inappropriat e.In addition, most "overnight" or "instant" software programs, journals software is purchased Without ard.to how its objec- devoted to this new field of "electronic education" have tivities fit into existing curricula and the overall goals initiated columns where software is reviewed. These of the school. reviews typically describe the purpose of the software, These issues, therefore, seem critical to the estab- summarize its content, and usually review its perceived lishment of valid criteria and procedures for determining effectiveness as a teaching medium. These reviews can effective use of computer-based instruction.In the serve an important purpose in helping set standards of following paper each of the issues cited above is dis- quality in software development but can also have the cussed and recommendations for developing more effect- deleterious effect of condeming a product that might be ive criteria and methods are offered. useful in certain contexts. If the product is terrible, then it certainly deserves Evaluating Evaluation Guidelines criticism.Terrible software is usually easy to evaluate. As more software programs become available, It contains inaccuracies in spelling, content, and struc- the number of individuals and organizations who-recom- ture; is a victim of ineffective lesson design, and suffers mend guidelines for evaluating software seems to in- from poor screen display and appearance.This type of crease proportionately. As Chairman of the National software, however, represents only a small portion of all Council of Teachers of English Committee on Instruct- the programs available. The major problem with eval- ional Technology, I have been currently involved in uating software programs beyond those which are deafly developing a set of criteria for evaluating computer- "terrible" is that it becomes difficult to discriminate be- based materials for the English language arts.In the tween the effective and ineffective ones. Programs that process, the Committee has examined at least ten (10) separate sets of criteria.In all these sets of criteria,

391 408 three basic problems seem to emerge. The first problem then, "1848," and so on. Each time my effort was re- is that many of the criteria on the evaluationinstru- warded with, "No. Try Again." ments are presented as if they are absolute indicators of Now, this program is certainly "interactive" and by program effectiveness when in fact they arebased more the definition implied ir+ most of the evaluationguide- on conjecture and intuition than onevidence of instruct- lines I have seen, it would receive a positive evaluation ional effectiveness. Few, if any, of these criteriaoffer on that basis. However, the program isclearly terrible any support to show that they are theresult of a rein that it provides the learners with absolutely noinform- search effort to establish a relationship between a spe- ation to help find the correct answer.Instead, it only cific feature of a program and its effect on the achieve- serves to increase frustration, inhibit cr:ativity, and ment of the learner. stifle the learner. Another problem with the criteria we examined was The validity of other criteria is less subtle. One that they invariably represent each separate criterion as author offers this criterion, "Does the program offer paging (not scrolling)?"In my opinion, a frame present- having equal weight in the evaluation of a particular visually than piece of software. None of the guidelines seem to ed in a page format is much more appealing acknowledge that certain features of program design text scrolling up from the bottom of a CRT. But might contribute more to the overall effectiveness of whether that feature contributes more to the effective- the program than some lesser features. ness of a segment of instruction thanscrolling does is A final problem was that none of the evaluation dubious. But it is this exact type of criteriathat is guidelines took into account variations in teaching offered on evaluation guidelines as absolute standards developing strategy.Instead the criteria were offered to be ap- of excellence. As individuals interested in plied to all software materials regardless if they used and encouraging the development of goodsoftware, we a simulation, drill/practice, tutorial, or gameformat. must look carefully at these criteria and encourage more Logically, one might reason that each of these strate- validation and research. This can only be accomplished gies would utilize the capabilities of the computer in a by finding positive correlations between variousfeatures slightly different way and therefore certain criteria of programs and learner achievement. would be less important or not important at all in an 2.. Weighting of Criteria evaluation of that strategy. A second problem which seems to reoccur oneval- In short, there are many problems with these uation guidelines is that in almost all casescriteria are guidelines that require more thinking and certainly given equal weight in the evaluation process.That is,, the problems each criterion is given equal importance inevaluating more research. Taken separately, each of piece of mentioned above might be examined and alternative the software. Obviously, not all features of a software contribute equally to the effectivenessof the procedures recommended: logical 1. Validity of the Criteria total program. With this in mind, it would seem Probably the most significant problem plaguing the that evaluation guidelines be structured so that a com- development of evlauation criteria is that, for the most posite rating put more emphasis on a featuresuch as, "interactiveness" than on something like, "Feedback is part, few if any of the criteria have been validated in only any sort of well conducted research study. Wedo know personalized." Evaluative criteria, therefore, not from research that some features of an instructional need validation,they need to be categorized into program can affect learner performance, but these are a hierarchy which will reflect the degreeof contribu- few and are often poorly documented. Most of the tion each criterion makes to the effectiveness of the criteria included on evaluation guidelines are based on overall program. what the developer THINKS contributes to an effective program.This same problem was inherent in most 3.Variation in Teaching Strategy teacher education research until correlational studies A final problem seen quite often in evalutiveguide- began to establish clear relationships between specific lines is that criteria are usually designed tobe applied teaching behaviors and student achievement. Until these to all software programs regardless of theteaching same methods are applied to instructiondelivered strategy used to deliver the content. Classroomteach- through a computer-based medium, the problem of valid- ing is very often evaluated without regard for the con- ity of criteria will remain a significant. one. text of the objectives taught in the lesson. Few educa- One need not look far to find examples to illustrate tors would argue, however, that very differentskills this point. One criteria which invariably appears on are needed to make individual strategiessuch as inquiry, evaluation guidelines is, "Is the program interactive," lecture, drill, and group discussion effectiveindependent- or "Does the program provide interactions." Few educa- ly.It would seem to follow, then, that the same criteria tors would argue that interaction is a valuable charac- applied to evaluating a drill/practice software program teristic of ANY form of instruction, but It is particular- might be inappropriate when applying themto evaluate ly important when using the full power of the computer- the effectiveness of a computer-based simulation. based system. Unfortunately, however, there is wide variability in the definition of interaction.In a program Qualifications of Reviewers I evaluated recently the designer posed the question, "In In the Blum (1982) study a wide variance in ratings what year was Texas admitted to the Union?" As my of software were found when a single piece ofsoftware response, I was supposed to enter a date. So as an ex- was reviewed by three or more raters.Blum attributed periment I entered "2,000,000 B.C." The feedback in- this variance to two major factors: formed me, "No, Try Again."I then entered, 1.Subjective judgments of the evaluators on items "4,000,456 A.D." The feedback urged me to "Try Again." that were highly inferential, e.g. "the presentation was Finally, I got serious and entered, "1843." The computer boring." told me, "No. Try. Again." For some time I really tried This conclusion only serves to verify the need for to get the right answer; in fact, at one point I just be- objective criteria .which are validated empirically gan entering numbers consecutively, "1846,"then, "1847," through correlational research models. Blumwrites,

392 `BUJ "There is basically little that can be done to reduce vari- al software.Tens of thousands of dollars are being in- ance in evaluation studies when judgements have to be vested in software development.It is important that made as they are based on personal opinion and retrain- the production of that software is guided by character- ing would not alleviate the problem." (p. 27) To a great istics which have a relationship to learner achievement extent, personal bias will never be eliminated from any so that the consum ers of that software can make evaluation process.However, if we can find and vali- reasoned choices about the software programs they buy. date objective criteria which rely on low inference (e.g. "Feedback helped shape behavior in the direction of the REFERENCES desired learning outcome.") rather than high inference Blum, B.L. "Evaluation of Educational Softward for (e.g. "The program provides adequate feedback."), much Microcomputers: An ANalytical Approach," Paper of the bias in evaluation that Blum found might be represented at the American Educational Research duced. Association Annuual Meeting, New York, 1982. 2.Error in scoring due to inadequate training or background of the reviewers themselves. In the Blum study, an evaluation instrument was developed which systematically analyzed software in terms of its intent, contents, methodology and means of evaluation.Unlike many others, this instrument also evaluated the software in terms of the strategy it used. Blum found that what often occurred was that one rater would respond to the software from the perspective of his/her backgrou nd while the other would respond from a different background. A mathematics teacher, for example, would look carefully at content in his/her evaluation whereas a rater with instructional design expertise would assess the software from that perspective. Blum's conclusion was the "There is no way to ensure that training and retraining will acount for varying background factors so that this (variance) can be reduced." (p. 26) The solution to this problem seems so obvious that I feel that I have perhaps missed the significance of Blum's conclusion.It would seem that one way to achieve greater reliability among raters would be to have raters from similar backgrounds evaluate the courseware. That is, allow three reviewers who have designed software for the English language arts or for mathematics or for whatever each evaluate the same program.This might provide a certain degree of reliability that is currently lacking. Another solution to this problem would be to choose raters who had actually designed software so that they had that experience in common as well.Individuals who have tried the laborious and frustrating experience of actually creating software often have a deeper appreciation of what an author or designer was attempting in the program. A final solution might be to have each software program rated by three individuals and then report the composite rating rather than let the review fall on the shoulders of just one reviewer. Lack of Instructional Models A final problem in evaluating software stems from a general lack of knowledge about how software should be used in a classroom. As students we have had the benefit of hundreds of hours of instruction.Each lesson was in itself a model of how (or how not) to teach a lesson. As teachers we emulated the models we found most useful.In using computer-based education, however, few such models exist. Few of us have ever experienced a computer-based course so have little insight into potential problems. As a result, we often evaluate a software program on the basis of the diskette or computer component itself instead of looking at it as a total pystem which includes print materials or other media integrated into the system. These issues, therefore, seem critical to the establishment of valid criteria for evaluating education-

393 41 MICRO-NETWORKING: SOME PRACTICAL APPLICATIONS

By David R. Rieger

Department of Spacial Education/RelatedServices Johnson County Public Schools, Buffalo and Kaycee,Wyoming

ABSTRACT disk system connected to just one microcomputer. The second part will deal with some of the appli- This paper deals with the practical aspects of cation problemsthat were encountered in setting up A detailed account will be given using a networking system for microcomputers. Ad- such a system. vantages and disadvantages of networking aswell as to the process of acquiring software for a network the author's reasons for selecting a disk-sharing system. system over the standard floppydisk format for school use are delt with. The Corvus Omninet The Corvus Hard Disk Network System is only Networking System is described at length and the one of several disk sharing systems on the market positive aspects of hard disk technology are at this time. For the purposes of this paper only the Corvus Omninet System will be discussed. This discussed. results from the author's opinion that the Corvus is a superior product and from the author's belief INTRODUCTION that the Corvus Hard Disk Network System has the best practical application of disk sharing for pub- For today's educators, the use of a disk shar- lic education, Grades K-12. (Note: The Apple II+ ing system for microcomputers can be bothconveni- was selected as the type of microcomputer for use due to the large amount of software available for ent and cost efficient. Some advantages and disadvantages have been identified by Fisher(1982). it. The Corvus Hard Disk Network System and Apple Advantages include: Microcomputers are completely compatible.) * The use of disk sharing can save the educator The author's experience with disk sharing time and energy. Since all data is located on one disk, management becomes very simple. stems from the use of a Corvus Hard Disk Networking The system was purchased for a Title IV-C * The teacher has more control. With some systems, System. Study to investigate the effect of computer assist- one can restrict access to partsof the memory. This allows the educator to manage program orfile ed instruction (CAI) on the learning achievement of handicapped students, Grades 1-5.A method was usage. needed that would allow special education teachers * A great deal of money can be saved. Instead of providing memory devices (or disk drives) for each to have access to a vast number of computer pro- microcomputer station, a central drive is shared. grams. The standard method of individual floppy Also, other peripherals can be shared. disk drives and floppy disks for use with each microcomputer was found to be less than desirable Disadvantages include: for several reasons: * Disk sharing systems are not portable. The microcomputer stations have to be wiredinto the sys- * The teachers would have to share the programs on floppy disks so that only one program could be used tem. by one microcomputer at a time. * Not all systems restrict access to memory. Some students might cause damage to files or com- * More teacher control would have to be used to pletely destroy some data. insure that the disks would not be managed by the handicapped students. * Some disk sharing systems are not easy to use. Additional training to allow teachers to learn how * The goal of having handicapped students control be the system works would have to be provided. a great portion of their own education could not reached due to the increased teacher control of the * Technical limitations might cause considerable problems. Some systems allow only one student ac- software. cess to a file at a time.Other systems have dif- The desirable aspects of using a disksharing ferent limitations. system include; In this paper, two aspects of disksharing * All programs would be available to every compu- will be discussed: the first part will deal with ter on the network. Teachers would not have to act as policemen to how a hard disk network works. In this study, * careful attention will be given the Corvus Hard safeguard the floppy disks. Students could access programs by themselves. Disk Network System. A complete overview of the * The general availability of a vast libraryof system will be presented and explanationwill be * given on how it differs from a standard floppy software on the network would allow the teachers

394 and students much greater latitude in using the This slot is accessed automatically when the Apple computer. is turned on, so the user is booted into the dmni- net system by simply turning everything on. (This For the purpose of this paper, the author assumes that the hard disk is already on and ready assumes that the reader has little or no experi- to receive data.) ence with networking systems and has therefore presented the workings of the Corvus Omninet System in The network asks for a user name and password detail. before a user can have access to any part of the hard disk. With different names and passwords, HOW A NETWORK SYSTEM WORKS some security is provided. With an application of the network system in an elementary school, a sim- The normal configuration of a microcomputer ple user name/password was developed that allowed station (Figure 1) includes the microcomputer the user "read access" to most everything on the (which also includes such parts as the central pro- hard disk. This name/password was "work/go". cessing unit - CPU, read only memory chips - ROM, random access memory chips - RAM, and support cir- With the basic network system by Corvus now cuitry) a monitor, and memory storage - such as a explained, an explanation of some of the practical floppy disk drive. The basic difference between applications, problems, and positive uses will this normal configuration and the use of a network follow. is the memory storage; the network uses a shared memory device. Rather than access memory on sep- MICRO-NETWORKING: SOME PROS AND CONS arate devices, the network uses a single memory device. In the case of the Corvus Omninet System, Negative Items a hard disk is used (Figure 2). All of the independent microcomputers are wired together on a A variety of problems developed with the in- common network (Figure 3). stallation of the Corvus Hard Disk System. Other problems developed in securing software to place The Corvus Omninet System is controlled by on the hard disk. a device called the disk server (Figure 4). Each computer is linked to the disk server by wires. One of the problems that took a great deal of Software on the hard disk allows each computer to time to solve was faulty hardware. Over a period call into the network via a user name and.,password. of a month and with much hair pulling and with un- The disk server acts like a traffic controller in necessary trips to computer dealers and with many that it allows each user access to oily certain telephone calls across the country it was finally parts of the hard disk and to use those parts in found out why several Apple II+ Microcomputer-i, either read, write, or read/write modes. If more would not work with the Omninet System, yet would than one computer calls for access to the hard work perfectly well with floppy disk drives. disk at one time, the disk server will serve the The Apple Computer Company has a habit of improving first user while the rest wait. When the shared their hardware during the actual production of the wires are free, the next user is served. Although product. The changes in the hardware do not effect this might seem to take a long time, the hard disk he compatability of the Apple components. But works so fast that all users can be served very in this case, newer Apple Microcomputers would not quickly. The disk server also remembers the loca- work with Corvus Transporter Cards. tion from the "scratch" if another user logs on to The resolution of this problem came with an up-grading of the the system.Therefore, each system user seems to transporter cards. The problem was very simple to have private use of the hard disk. solve, but sales people at local computer suppliers (where the equipment was purchased) provided little The wiring, mentioned above, is very simple if any help. to install and is very inexpensive. Unlike the multi-strand cables that have been used in the Another major problem was the time it took to past to connect computers, the Omninet System uses learn how the system worked. As the system man- a simple twisted pair of wires. All the wiring is ager, or trouble-shooter, I spend many hours to configured in a line with each computer tied in become familiar with the hard/software of the net- at whatever point is chosen. The total distance work. The manuals were not very helpful in provi- the system can serve is 4000 feet - 2000 feet on ding this help and most skills were gained through each side of the disk server and hard disk. This trial and error.Again, with little or no support, distance is most adequate for most K-12 schools. the process of setting up a network can be difficult. Each computer has a "tap" off of the main wire with a "tapbox" (Figure 5). The tap box con- The most difficult problem with the system tains small wire clamps that allows for the easy developed through the manner in which the system installation of "tap cables" (Figure 6). These was to be used.The Title IV-B project called for are the wires that connect the computers to the the system to use commerically available software. main wire. I found out that most software producers are very new to the business and what I was asking them to For use with Apple II+ Microcomputers, a do had never entered their minds. The Corvus Om- "transporter card" (Figure 7) - shown here (on the ninet System must have uncopy-locked software in top) with a standard Apple Disk II Controller Card order for it to be placed on the hard disk.This - is installed in slot #6 of the Apple (Figure 8). is due to the fact that a copy must be made of the

395 412 Figure 1. Normal Configuration of a Microcomputer

- . ,4f .444.11.

Figure 2. Corvus Hard Disk

396 4 1 MONITOR MONITOR MONITOR MONITOR 1 _L

COMPUTER COMPUTER COMPUTER COMPUTER

UP TO 64 STATIONS

DISK SERVER COMPUTER COMPUTER COMPUTER COMPUTER

MONITOR MONITOR HARD MONITOR MONITOR DISK

riwArp Shnued Memory Vie. Corso, thrmitmt Ard Dint. NotHOrk

Figure 4. Disk Server

397 414 Figure 5. Tap Box

Figure 6. Open Tap BoxShowingeonnections

398 4 1 5 Figure 7. Transporter Card (on top) with Apple II+ Disk Controller Card.

Figure S. Transporter Card in Slot *6 Inside Apple II+ Microcomputer

399 416 program from a floppy disk to a volume on the hard Disk has few moving parts and normal servicing is disk. Most producers are very concerned that rated in years of continuous operation rather than their software might be stolen by unscrupulous in hours, weeks, or months; and the Winchester computer users so that to them my need was almost Hard Disk does not wear the surface of the disk: heresy. I prepared a form that stated my needs the head floats just above the surface on a cushion and my intentions to preserve the copyright of of air thereby extending the life of the disk intheir material. This seemed to help some software definitely. Let us look at each of these features producers in allowing me to purchase their mater- in greater depth. ials, but their prices ranged from standard list price to twice the list price. But whatever the Sealed Disk (Figure 9) - The entire disk/head price, I was allowed to place a copy of a partic- assembly of the Winchester Hard Disk is sealed in ular program on ONE hard disk system at ONE loca- a plastic case that has a super-clean atmosphere. tion. The disk assembly turns inside the case and the heads reach out over and under the disks by a ser- Another disadvantage of the networking system vo arm. This assembly is also enclosed inside the is the limitation of space on the hard disk. Al- case. With the use of the sealed assembly one of though the unit I was using had a capacity of 20 the major problems of floppy disks is eliminated: million bytes (approximately 134 - 5 1/4" floppy dirt and grime on the disk. diskettes worth), the space is not interchangeable. A floppy disk drive allows for unlimited use of Fast Access Speed - Since the networking sys- different software. The user simply places dif- tem is based on sharing the memory device, high ferent floppy disks in the drive. With a hard speed is very important. The fast access time of disk, the actual disk is captive and cannot be ex- the Winchester Hard Disk is also a characteristic changed or replaced with another.Data has to be of all hard disks. The hard disk operates some- coried to the hard disk. When it is recorded to where in the neighborhood of 4 to 6 times the the hard disk, this data is more or less permanent. speed of a floppy disk drive. It can be removed or replaced, but this again is a permanent change. I found that 20 MB is not all Few Moving,Parts - Since the only moving that much space. The Corvus Hard Dick Units allow parts are the disks and the heads, great care is for up to 4 - 20 MB units to be "daisy chained" taken to make these extremely reliable. With such together, thereby creating up to 80 MB of storage few parts to go bad, many Corvus Hard Disk users space on a hard disk network. simply leave the unit on all of the time.

I found the difference between a hard disk and No Disk Wear - Each of the heads that read a floppy disk system to be much like the difference and write information to the surface of the hard between a record player and a juke box! A juke box disk does not actually touch the surface as in a has a limited number of records that are installed floppy disk drive. An air cushion under or over in the device. A user can very easily "call up" the head floats the head about one micron off the any record. They are kept neat and clean. But surface: a distance close enough to transfer data the user has a limited number of "selections" for but far enough to eliminate wear to the surface use at any one time. On the other hand, the re- of the hard disk. cord player does not keep the records neat and clean. The user has to handle the records and keep One of the major reasons that attracted me to them someplace and in some sort of order. But the hard disk networking system is still very vi- there are unlimited numbers of "selections" that able: teachers who want to access a particular pro- can be made by the user; one has only to change gram can have almost instant access to it. Since them. all of the programs are contained on the hard disk and since all the microcomputers on the network The hard disk system works just like this have access to the hard disk; each microcomputer juke box. The user is limited in what he or she can have a catalog of over 130 volumes. may access from the system, but the data on the system is extremely easy to get at and use. In This feature is especially useful if there is order to replace data on the hard disk, quite a a "core of instructional and/or management programs bit of work has to be performed. that all users would need to use. Programs for the K-12 setting might include attendance, inventory, Positive Items student records, etc. for management; while those that support the curriculum by a relationship to a The advantages of the hard disk network in- school-wide text book or basic skills program. clude all of the positive working features of the Winchester-Type Hard Disk, as opposed to a standard One last positive feature of the network sys- hard disk. These features are especially good for tem is the ability to expand the system to include several reasons: the Winchester Hard Disk is seal- many computers (up to 63 total). In Figure 10 we ed from the atmosphere and is not affected by dirt, can see what a network configuration for a single dust, or damage from handling; the Winchester Hard school could look like. There is just one memory Disk accesses information from and to the disk storage device that is connected to all of the very quickly (as fast as 500,00 bytes per second computer stations. In addition, peripherals such can be up or down loaded); the Winchester Hard as printers could also be shared.One or two

400 Figure 9. Corvus Hard Disk Drive F:-.owing Sealed Disk Case at Top

HARD DI SK INDIVIDUAL SHARED MICROCOMPUTER PR INTER DI SK SERVER STATION STATION

In. 11,1 v:r rI ..... elof rr.s!yr.{ csI School

401 41S printers could serve a whole school. The hard disk serves as a buffer and can store data to be printed. If several items are to be printed at the same time, the first item is printed while the rest are ordered and held in the buffer. As each item is printed, the buffer allows successive items to be printed.

CONCLUSION

Although there are many different types of disk sharing systems (Fisher, 1982), the Corvus Omninet Hard Disk System has many positive features that make it very usable in a public school, grades K-12. Micro-networking has some negative features, such as the difficulity of obtaining uncopy-locked software; but these are more than amply offset by such positive features as providing a large library of software to all microcomputer stations at all times, expandability at low cost, and extremely good reliability of the hardware.

REFERENCE

Fisher, Glenn. Disk Sharing: How To Make One Disk Go "Round. "Electronic Learning", 1982, 1, 46-51.

LIST OF FIGURES

Figure 1. Normal Configuration of a Microcomputer Figure 2. Corvus Hard Disk Figure 3. Shared Memory Via Corvus Omninet Hard Disk Network Figure 4. Disk Server Figure 5. Tap Box Figure 6. Open Tap Box Showing Connetions Figure 7. Transporter Card (on top) with Apple II+ Disk Controller Card Figure 8. Transporter Card in Slot #6 Inside Apple II+ Microcomputer Figure 9. Corvus Hard Disk Drive Showing Sealed Disk Case at Top Figure 10.Network Configuration for a Typical School

402 41 Computers in the Elementary and Secondary Mathematics Education Sheldon P. Gordon Suffolk County Community College

ABSTRACT This session will focus on the uses of by looking intensively at the structures computers in elementary and secondary designed and successfully implemented in mathematics education with emphasis on several different schools. developing and implementing such usage. In one school, the entire 9-11 The University of Delaware has developed Mathematics curriculum has been extensive programs to work with elementary restructured to allow infusion of computer and secondary schools in enhancing the use probl4q, solving and computer programming in of computers throughout the curriculum with BASIC '.,roughout the ., Students special focuson mathematics. As part of are introduced to o-,,,ous ,:ogram-ming the present session, a report on one phases techniques as needed tc solve the problems of these efforts involving a project to encountered in the traditional curriculum. provide computer literacy for preservice The philosophy behind this approach is that elementary education majors will be the school should teach mathematics and described. In this project, each student problem solving, not computer programming. is given the opportunity. to become familiar A second school, operatingunder the with the PLATO computer system and a same basic philosophy, has approached variety of microcomputers. The curriculum development in another way. The presentation will consist of program faculty chose to .keep programming entirely description and sample lessons developed separate from the traditional mathematics for the Apple II computer. curriculum. Their programming courses, In New York City, the high schools have however, are stillmathematics courses. developed a variety of curricula in Instruction is in the mathematical concepts computer literacy and computer mathematics. with computer solutionsgrowing out of the The varietyof curricula will be reviewed mathematical solutions.

PARTICIPANTS: William B. Moody University of Delaware

Neal Ehrenberg New York City Board of Education Project Director for Computers,Technology and Research

403 42u AUTHOR INDEX

Adams, J. Mack 342 Evans, H. 141 Adamson, Carl 55 Evans, Richard 144 Alexander, David 225 Anderson, Cheryl 290 Feibel, Werner 152 Anderson, Ronald 267, 367 Fisher, Glenn 147 Arons, Arnold 308 Fletcher, Lincoln 192 Fordham, Malcolm 91 Badger, Elizabeth 132 Forman, Kenneth 4 Baker, Herbert 12 Fosberg, Mary Dee Harris 368 Bargmann, Theodore 385 Frank, Ronald 122 Bearwald, Ronald 181, 385 Friske, Joyce 180 Beidler, John 320 Bell, Spicer 90 Garcia, Linda 148 Berztiss, A. T. 258 Garland, Stephen 1 Bigliani, Raymond 224 Garris, Barbara C. 364 Blank, Deborah 138 Geist, Robert 360 Bolick, Jerry 330 Giangrande, Ernest 231 Bonar, Jeffrey 239 Gibson, Bobbie 141 Bork, Alfred 308 Gilbert, Steven 270 Bray, David 326 Gordon, Sheldon 109, 345, 403 Bregar, William 231 Gottlieb, James 341 Brown, Scott 381 Gregory, Carl 68 Brown, Warren 32 Bryant, S. 224 Harrow, Keith 64 Bull, G. 141 Healy, Nancy 311 Burger, W. 294 Heller, Paul 371 Burk, Laurena 35 Henkins, Robert 225 Burtis, Eric 316 Hilberg, Barbara 142 Horan, Rita 32 Caldwell, Robert 391 Horn, Carin 109 Caviness, Jane 2 Hostetler, Terry 244 Chars, Sylvia 189 Hughes, Charles 68, 103 Cheyer, John 108 Hunter, Beverly 191, 316 Christensen, Margaret 180 Hyler, Linda 255 Christopherson, Jon 227 Church, Marilyn 272 Icenhour, James 0. 330 Clark, Carol L. 252 Ince, Darrel 146 Clark, James 149 Connelly, Frank 112 Jackson, Robert 90 Cornelius, Richard 91 Jawitz, W. 22 Cossey, David 194 Johnson, Dale 255 Coulson, Helen 184 Jones, Ken 99 Crist, Mary 317 Jones, Nancy 33. Crowther, Sandra 255. Juels, Ronald 279 Czejdo, Bogdan 220 Kaye, Daniel 381 Daiute, C. 22 Kelly, Pat 143 Dalphin, John 3 Kendell, K. 224 Davidson, P. 141 Khailany, Asad 48 Davies, Joan 110 King, M. 141 Dayton, C. Mitchell 336 Kirkpitrick, Carolyn 249 DeBoer, Mary 144 Klenovj, Carol 31 Denenberg, Stewart 253 Kurshan, Barbara 311 Derringer, Dorothy 229 Dietz, Henry 279 LaFrance, Jacques 126 Dove, L. 224 Landis, Marvin 342 Durnin, Robin 283 Lathrop, Ann 365 Leahy, Ellen 250 Edwards, H. 224 Legenhausen, Elizabeth 252 Eltschinger, Michel 255 Leinbach, L. Carl 364 Entwistle, John 186 Levin, James 302 Epes, Mary 249 Levy, C. Michael 386 Esty, Edward 111 Lewis, David 181

404 427 Liao, Thomas 273 Saltz, Martin 141 Liff, S. 22 Saluti, Dean 200 Lindahl, Ronald 7 Sands, William 12 Little, Joyce Currie 183,230 Schafer, William 336 Little, Joyce Currie 230 Schloss, Patrick 13 Lockheed, Marlaine 372 Schubiner, Marc 48 Loper, Ann 183 Schwartz, T.141 Luehrmann, Arthur 316 Sennett, Mary 143 Lundstrom, David 266 Shields, A. 22 Shimsak, Daniel 200 Markuson, Carolyn 141 Sieben, J. Kenneth 249 Maron, M.J. 42 Siegel, Martha 261 Matheson, W.S. 146 Simms, Dennis 99 Maurer, Stephen 263 Smaldino, Sharon 33 Maurer, W. D. 355 Soloway, Elliot 239 Mazur, S. 22 southwell, Michael 249 McGinnis, Richard 318 Spicer, Donald 180 McLaughlin, Brian 157 Spoeri, Randall 91 Mezzina, Maria 369 Spraycar, Rudy 27, 321 Mikiten, Terry 85 Starling, Greg 350 Miller, Clarence 54 Starnes, W. 92 Miller, Jon C. 365 Steinhoff, Carl 4 Mitchell, William 56 Stone, Meridith 372 Moore, M. 294 Strang, Harold 183 Moshell, Michael 68, 103 Streibel, Michael 214 Moshell, Mitchell 68 Swensson, Rochelle 142 Moxley, Roy 141 Muntner, J. 92 Taylor, Harriet 255 Murdach, Richard 143 Taylor, S. 141 Muscara, Carol 186 Thompson, Carla 180, 255 Thompson, John T. 193 NaditchiMurray 80 Tipps, S. 141 Nielsen,Antonia 372 Tobias, Joyce 141 Nielsen,P. 224 Trauth, Eileen 204 Trowbridge, David 283, 308 O'Brien, P. 22 Olivo, Richard 174 Updegrove, Daniel 270 Oviedo, Enrique 115 VanderMolen, A.M. 48 Parker, Janet 377 Verth, Patricia Van 208 Pelham, William 163 Peters, Alonzo 90 Wagner, William 107 Petitto, Andrea 302 Walker, S. 141 Piper, Karen 18 Wall, Robert 183 Pollock, Marilyn 253 White, G. 224 Poonen, George 370 Wholeben, Brent 7, 298 Ppancella, John 186 Widmer, Constance 377 Pritchard, William 180 Wiersba, R. X. 250 Pyka, Ronald 85 Wilcox, David 167 Williams, Joyce 311 Rafacz, Bernard 12 Winner, Alice Ann 184 Ralston, Anthony 115, 208,256 Wittenberg, Lee 68 Rampy, Leah 142 Wolf, Melvin 188 Rieger, David 394 Wolfsheimer, Joseph 366 Roblyer, M. D. 226 Wright, June 272 Rogers, Jean 268 Wright, Muriel 184 Rose, Shelley 31 Wright, P. 22 Ross, David 368 Rossien, David 225 Zeidman, Edward 187 Royster, Linda 90 Zgliczynski, Susan 254 Russo, Mary 33 Zuckerman, Dan 75.