a 4
DOCUMINT41ESOBEr)
ED 149 727 , IR 005,504
. . - TITLE AbstriOti of Research. 'July 1974-June 1975'. INSTITUTION Ohio. State, Univ.c'Cblumbus. Computer and Information Science Research Center. PUB DATE '75 :NOTE 92p.; For related documents, se ED 092 117, ED 097 .' 025, ED 139 394
EDRS PRICE MF-$0.83 HC -$4.67 Plus Postage._ DESCRIPTORS *Abstracts; *Annual 'Reports; Artificial intelligence; Cognitive Provesses; Computational Linguistics; *Computer Programs; *Compqter Science; Information
Processing; Ibformaticn Retrieval; *Informati?n . Science; Information Storage ABSTRACT
Abstracts'of research papers in computer and . information science aie given for 68papers in the areas cf information storage and retrieval; human informationprocessing; information analysis; linguistic analysis; artificialintelligence; information processes in physical, biclogicalpand socialsystems; mathematical techniques; systems programming; computerarchitecture and networks; joint programs; and computation theory.Abstracts are indexed by investigator and subject. Introductory materialincludes a -description of the Ohio State University Computer andInformation Science Research Center. (VT) -
r
S.
***************v*******4***************************4****************** Reproductions'eproductions Supplied by Eptts are the best that can be made * froi the original document.- * ***********************************************************************-
* wt.
U S DEPARTMENT OF HEALTH. x ..,1411, 'EDUCATION t WELFARE NATIONAL INSTITUTE OF I EDUCATION, iff:,-;e4ey'hi.S.tWral.' hAillitt4, REPRO OUCE EXACTLY AS RECEIVED FRAM THE P RSON OR ORGANIZATION ORIGIN. ATINIT POINTS OF VIEW OR OPINIONS STATED DO NOT NECESSARILY REPRE SENT OFFICIAL NATIONAL INSTITUTE OF EOUCATiCA POSITN OR POLICY --,
.
ABSTRACTSOF RESEARCH
JULY 1974 z-JUNE19,75
N O
I
;
\ , The Computer and Information Science Research Center
.'-,, k The OhioStateUnlversity
Columbus, Ohio43210 .4>
o:
FOR4WORD.
1. , The Computerand Information S-bience Research Center at The Ohio State Universityis a researcorganizatibn whiCh consists prfiarily of faculty, staffand graduate studeAts Of the,Department of Computer and Information Science. Some of the research activities are performed in conjunction with other University departments as well as off-campus organizations.
This publication contains the abstracts of research whidh has been carried on during the 1s974744(acad'emicyear. This research has been supported in part by grants from goverhinenal agencies as well as by The Ohio State University. Sponso4shig with goliernment agencies ,and with other units .on the campuS id!identified,at the end of an
4 abstract: v.>
. . A bibliography o?the'research reports published by the Cente,r_ 0 is included in this publication as Appendix F. Copies,of some of these reports are still available on a corkplimentary basis from the
. Cpmputer and Information Science Research Center, The Ohio State
'University, 2036 Neal Avenue Mall, Columbus,' Ohio, 43210:Titles With* PB or AD numbers may be obtainefrom The National Technical Informa- tion Center, The U.S. Department ofCommerce, 5285 Port Royal Road, Springfield, Virginia; 22151, in paper-copy, magnetic tap#4, or microfiche..There Is a nominal charg e for their service.-
Marshall Co. Yovits Director, Computer and Information Science Research Cent
>4
ii .
te,
7 TABLE OF CONTENTS
FOREWARD
I THE OHIO STATE UNIVERSITY COMPUTER AND INFORMATION-SCIENCE 1 RESEARCH CENTER ,
ORGANIZATIONALSTRUCTURE 1
OBJECTIVES OF THE CENTER , 1
SCOPE OF THE PROGRAM
ACADEMIC:PROGRAMSIN COMPUTER AND INFORMATION SCIENCE 4
Organiiation of ThekDepartment of Computer 4 and Information Science Objectives of The Department 4 Undergraduate Programs 4 Graduate Programs 6 Course Offerings 7 Faculty 7
INTERACTION WITHIN.THE UNIVERSITY 7
INTERACTION WITHIN THE COMPUTER AND INFORMATION SCIENCE COMMUNITY 7
DOCTOR OF PHILOSOPHY DEGREE 8
RECENT TECHNICAL REPORTS 9
II INFORMATION STORAGE AND RETRIEVAL lo .
. '. oe . AUTOMATIC PROFILE ENHANCEMENT BASED ON ASSOCIATION MEASURES DERINIEDI'- 10 FROMSDI,SEARCH PROFILESA. E. Petrarda, G. J. Lazorick, B. J.: Brinkman : ,,,.. I A CONCEPT SCATTERING 'FUNCTION FOR'EVALUATING THE-EFFEcITVENESS'OF 10 VOCABULARY CONTROL ALGORITHMS USED IN AUTOMATIC INDEXING A. --)E. Petrarca, W. S. Stalcup' ,
4 A DISTANCE MP SURE FOR AUTOMATIC DOCUMENT, CLASSIFICATION 1,::J.,.Whrte, 11
G. Kar 0
"t \ r / r
GENERATION OF UNDUE CONFIGURATIONAL REERESENTATIONS OF CHEMICAL > - 11 STEREOISOMERS FRA NON-UNIQUE UNAMBIGUOUS'3,DIMENSIONAL,INPUT DESCRIP,- ..; t TIONSA. E. Petrarca, D. G. Williams, K., J. Wells ., '
0
,, c t; iii 04 , IDENTIFICATION & MEASUREMENT OF 'COST & PRICING ELEMENTS OF COMPUTING 12 `RESOURCES D. Isaacs, T. G. DeLut\is
IMPROVEMENT OF AUTOMATIC VOCABULARY CONTROL IN INDEXES DERIVED FROM 12 NATURAL,LANGUAGE TEXTA. E. Petrarca,.W, S. Stalcup .
THE INFORMATION PROCESSING SYSTEM .SIMULALOR pss -T. G. DeLutis 13
A METHODOLOGY FOR A MULTIPLE GOAL APPROACH TOCOMPUTER SYSTEMS DESIGN 13
J.S. Chandler, T. G. DeLuti. .
OCCURRENCES OF TRAILING N-GRAMS AS A FUNCTION OF DATABA05SIZE 4
A. E. Petrarca, W. S. Stalcup ,
,
hFORMANCE EVALUATIO, OF DATA BASE MANAGEMENT SYSTEMS J. A. Aitken-, r 00*, J4 T. G. DeLutis
III HUMAN INFORMATION PROCESSING 15--
HUMAN EXTRAPOLATION OF STRINGS GENERATED BY ORDERED CYCLIC FINITE STATE 15 GRAMMARS J.*L. Beug, R. L. Ernst
IV INFORMATION ANALYSIS 16
AN EMPIRICAL ANALYSIS OF INFORMATION AND RELATIONS AMONG ELEMENTS,OF A 16 DECISION SYSTEM M. J. Lee, R. L. Ernst
. . . . ".4 GENERAL THEORY OF INFORMATION FLOW AND ANALYSISM. C. Yovits, J. G. 17 / / .. .7 , 4bilock . ,r , STRUCTURAL MdDELS FOR DECISION SYSTEMSR. L. Ernst, M. J. Lee 18
V LINGUISTIC ANALYSIS
THE AFFECTS- OF PHYSICAL ENVIRONMENT ON THE 'INTERPRETATION OF LANGUAGE N. it. Sondheimer.
A FAMILY OF 'TRANSLATORS FOR PHRASE STRUCTURE PROGRAMMING LANGUAGES' 20 H. W. Buttelmann, A Pyster,
FORMAL GRAMMARS WITHOUT SYNTACTIC VARIABLESH. W. Buttelmanti, A. 21, Pyster, L. Reeker
A FORMAL THEORY. OF 'THE SYNTAX, SEMANTICS ANb TRANSLATION OF PHRASE- 21 STRUCTURE LANGUAGES H. W. Buttelmann
ON THE SYNTACTIC STRUCTURES OF. UNRESTRICTED GRAMMARS. I:, GENERATIV 21 GRAMMARS AND. PHRASE. STRUCTURE GRAMMARS' H. Buttelmann
; 0 ON THE SYNTACTIC STRUCTURES OF UNRESTRICTED GRAMMARS. IIrAUTOMATA H. W. Buttelmann . N
iv 4
4 PRACTICAL ALGORITHMS FOR COMPUTER COLLATION -OF LONG NATURALTEXTS FOR 22 SCHOLARLY USE D. J. Moore
SOME PROPERTIES OF SYNTAX-DIRECTEDTRANSLATIONS H. W. ButItelmann, 2,3 F. 'J. Dickey
.
SPS:" A FORMALISMFOR SEMANTIC INTERPRETATIONN. K. Sondheimet, R. '23
Pardo,oD. Perry , 4
STRUCTURE OF COMPUTER AND NATURAL LANGUAGE ALGORITHMS S. U. Zweben 24 s
. .
STUDIES IN,DECIDABILITY, EXISTENCE, AND EFFECTIVENESS OF TRANSLATIONS 24 ON PHRASE-STRUCTURE LANGUAGES H. W. Buttelmann, R. KrishnasWamy'
STUDIES OF THE REPRESENTATION AND MANIPULATIONOF LINGUISTIC , 24 ri , INFORMATION .N.,K. Sondheimer ' I . 4 .. . (, SYNTAX-DIRECTED AND SEMANTIGOIRECTED TRANSLATION OF PHRASE STRUCTURE, '25 , . LANGUAGES H. W. Buttelmann ' - .
VI !.ARTIFICIAL INTELLIGENCE , 26
. THE INTEGRATION OF GRAPHICAL AND SEMANTIC KNOWLEDGE D. C. Brown% B. 26 Chandrasekaran, A. P. Lucido, N. K. Sondheimer
LOW-FREQUENCY RADAR AIRCRAFT DETECTION AND CLASSIFICATION L. J. Whj.te 26 . '7. , . PROGRAM INFERE FROM EXAMPLE COMPUTATIONS REPRESENTED BY MEMORY 27
- .$NAPSHOT TRACES- F. E. Petry . . . _
A SYNTAX-DIRECTEDMETHOD OF EXTRACTING TOPOLOGICAL REGIONS FROM A 27
. SILHOUETTE FOR PATTERN RECOGNITION J. Ampss, K. J.- Breeding, B. 4 . kt Chandrasekaran , 0
VII INFOMATIKPROCESSES-IN PHYSICAL, BIOLOGICAL AND SOCIAL'SYSTEMS 28
: DISCRETE COMPUTER SIMULATION MODELS FOR ANTHROPOLOGICAL STUDY OF 28 SETTLEMENT AND SUBSISTENCE SYSTEMS D. J. Moore'
VIII,MATHEMATICAL TECHNIQUES 29
AN ALGORITHM FOR SUB - OPTIMAL SET COVERS WITH CONEJDENCE MEASURES OF 29 OPTIMALITYL..J..White, M. E. Doherty
4 . - , 1 BINARY STRINGS AND TO!'OLOGY J. Rothstein . 291
COMPARATIVE STUDIES TOWARDS THE,PERFORMANCE.EVALUATTON OF SOFTWARE Fdk: 29 SOLVING SYSTEMS. OF'NONLINEAR EQUATIONS D. L. Kalmey, D. S. Kerr, L. J.'White
MINIMUM-REDUNDANCY VAkIABLE LENGTHCODINGL. J., White, V. Santhanam 30
6 r v11
1 SELF-SOLVING QUASIGROUPSJ. Rothstein s 30
SOMELP-NORM MAXMIII7PROBLEMS: ALGfYRITJ1MS AND APPLICATIONS L. J.' 30 White, B. Dasarathy
IX SYSTEMS PROGRAMMING 12
THE ANIMA GRAPHICS PROGRAMMING LANGUAGEA. P. Lucido 32
COMPILATION OF SEQUENTIAL PROGRAMS FOR PARALLEL EXECUTION C. R. Foulk, 32 O. C. Juelich
THE COMPUTER EXPERIMENTATION FACILITYA. P. Lucido, C. Kearns 33
COMPUTER SYSTEM FOR SECURE DATA PROCESSINGM. T. Liu, S. Mufeic, 33'. J. Rothstein
CONCURRENCY IN REAL-TIME INFORMATION SYSTEMS H. S. Koch 34
CONTEXT PROTECTION IN'DATA BASE SYSTEMS D. Cohen, D. K. Hsiao,. 34 M.T.' Liu
COST - EFFECTIVE ACCESS CONTROL IR DATA BASE SYSTEMSJ. M. Hennings, 34 . D. K. Hsiao
DATA BASE ACCESS CONTROL IN THE PRESENCE OF CONTEXT DEPENDENT PROTEC- t 35 TION REQUIREMENTS C. J.Nee, D.K. Hsiao, D.S. Kerr aiow ; DATA GENERATION FOR COMPUTER GRAPHICS B. Chandrasekaran,'R. Parent 35
A DATA SECURE COMPUTER ARCHITECTURE R. Baum, 'D. K. Hsiao 3,6
DESIGN AND%IMPLEMENTATION OF A DATA SECURE SYSTEM' N.K'affen, L.1. ' 36 Breene, T. Rodeheffer, D. Schmaltz, D. K. Hsiao
DESIGN OF A DAT& BASE MANAGEMENT SYSTEM FOR CONCURRENT PROCESSING H. S. Koch
DETERMINING THE STABILITY OF DISCRETE SIMULATION MODELS D. J. Moore, 37 P. Miller , S
AA FAMILY OF PROTECTION LANGUAGES FOR SPECIFYING DATA BASE. PROTECTION . 38 % ,
REQUIREMENTS D. H. Wong, D. K.-Hsiao, . / . . LANGUAGES FOR SPECIFYING PROTECTION REQUIREMENTS IN DATA BASE,SYSTEMS7- ler- ASEMANTIC MODEL,H. R: Harrson, D. K. Hsiao
A MODEL ,FOR A SECURE DATA BASE AND A F ORK FOR THE APPLICATIONOF:. 39 CRYPTOGRAPHIC TRANSFORMATIONS'IN A A BASE SYSTEM E. Glides, H; S. ,Koch, J. Rothstein, F.. A.-Stahl '.
vi STRUCTURED DATA BASE GENERATION R. Knablein, D. K. Hsiao 39
STRUCTURED FORTRANT. Rodeheffer, K.-Hsiao 40 . .
X COMPUTER ARCHITECTURE AND NETWORKS 41
THE ARCHITECTURE OF A GRAMMAR-PROGRAMMABLE HIGH -LEVEL LANGUAGE MAC8INE 41
M. T. Liu, S. Fournier )
THE DESIGN OF-THE COMPUTER NETWORKM. T. Liu, 41 C.C. Reames'
A NOVEL MODEL FOR A MIXED VOICE/DATA TRANSMISSION SYSTEM FOR COMPUTER 42 COMMUNICATION.M. T. Liu, J.'T. Wang
OPTIMAL PROCESS ALLOCATION IN DISTRIBUTED HETEROGENEOUS COMPUTER 42 WORKS M. T. Liu, T; T. gbeng
A PREDICTIVE RESPONSE TIME MONITOR FOR COMPUTER NETWORKS S. Mamrak 42'
XI' JOINT PROGRAMS . 44
THE STATUS OF WOMEN AND MINORITIES IN ACADEMIC COMPUTER SCIENCE S. 44 Mamrak,R. G. Mentanell
COMPDTATION THEORY , T
THE COMPUTATIONAL COMPLE TY1OF ORACULAR ENUMERATION PROCEDURES , 45 D. J.,Moore
THE COMPUTATIONAL COMPLEXITY OF RECURSIVELY ENUMERABLE TOTAL ORDERS 45 D. 'J.' Moore
PARALLEL RECOGNITION QF FORMAL LANGUAGES J. Rothstein, M. Noshell 46
THE STRUCTURE OF THE SET OF FEASIBLY COMPT BLE FUNCTIONS 1.D: J. Moore, 4'6
W. Leggett , 1 e TOWARD AN ARITHMETIC FOR CELLULAR AUTOMATA AND ARALLEL COMPUTATION 46
J. Rothstein ' I
TURING UNIVERSALITYA4D PARALLELISMOF GROUPOID ST kNG ALGORITHMS 47 J. Rothstein A ,
APPENDIXES
A COMPUTER AND INFORMATION SCIENCE COURSE LISTING BY NUMBER AND) 48 TITLE
B COMPUTER AND1NFORMATIOOSCIENCEFACULTY 51
/10 J
C COMPUTER AND INFORMATION SCIENCE SEMINAR SERIES ' 54
DRELATED ACTIVITIES OF.THE STAFF OF COMPUTER AN1,INFORMATION 57 SCIENCE RESEARCH CENTER
, E' PUBLICATIONS'-OF THE COMPUTER AND INFORMATION SCIENCE 63 - RESEARCH CENTER STAFF
F 'TECHNICAL REPORT SERI S r '66
G DOCTOR OF PHILOSOPHY D GREE 73-
INDEXES -
INVESTIGATOR INDEX 75 .
. - SUBJECT INDEX 77
I
'; e 4
dr r
er
viii
.4 I
IV
.
I. THE bHIO ST UNIVERSITY COMPUTER AND INFORMATION SCIENCE RESEARCH CENTER.
... . , t
. - ORGANIZATIONAL STRUCTURE '- 0. .
1, -. . , . ,,,, '...0 The Computer and Information Science Research Center. at The Ohio State, University .is an interdisciplin ry activity involving the faculty, staff and graduate students ftom'the Depa tment of Computer, and Information Science. . Some of ,the research activities are performed in conjunction with other t University departments as well as off-Campus organizations.The Center,also interacts closely with Battelle Memorial Institute, Chemical Abstracts Service, and'the Ohio College Library Cent r, which are_adjacent to the Ohio State
Campus, as. well aswith a number of other organizations rocated in 'COlumbus, .I. Ohio which are engaged in computer and,ipformatiOn 'science research,activities, such as Bell Laboratories, Western' ElettriC Corporation,'ind Rockwell
IntespationalCotpotatim. Although the Research dehter,and the Department . utilize many of the same personnel,and have the same Director, they are two
separate and distinct entities'. .
.. . /
:dBJECI.IYESOF/THECENTER .
1 .
. .: The CoMputer and Information Science Research Center has the following
. atlectiyes:. (1) to develop a broad research\program ln'computer'and /, infbrmation_science; (2) to deVelop, teat, and evaluate practical, applications of research in Cimputer and information science; (3) to coordinate and integrate these functions with an academic,rogram in computer find information 6cience,at The Ohio State University, as well as with other-disciplines at the University. The Center is a focal,point for a number-of applied information
procesping activities on thecampus.- A .
SCOPE OF THE 'PROGRAM
The program in computer andinformation.science at The Ohio State University,has been defined broadly to encompass most of the analytical activities frequently_considered-to be part of this discipline.This approach has been chosenbecaude it is felt that in order to generate the needed Concepts, foundations, and generalized techniques, it is necessary to-examine analytically a number of different areas of computer and information science. In this way a firm empirical antheoretical foundation may he established for generalized computer and information systems. Such a view commits .the program not only to the theoretics/ study of computer and information systems, but also to the study of their realizatfbq and their impadt on 1,11e user. '
0 , t A C . 2 Jo r.
.. . \-. are emphaSized both in the academic program and those areas of studywhich $ in, the, research activities of the Center are as follows:. , I_ 1: (eneral theory of information 2. information storage andretrieVal 3. theory of automata'and theory of computation 4. 4,tif,kc-ial it* elligence, ®
r, 5. Pattern recognition '.. J 6. Compfiter programming, including systeMs programming.
J. Theory and processing of programming languages sk
8. Digitalt'cgmputer architecture and organization .
. '9. Nlitherical analysis -' - 10. Manrmachine,interaction and systems
11. Formal and computational linguistic4 A ,12. Management information and systems
13. Biolqgical.information processing . 14. Social, economic, and psychological aspects ofdnformation.production,. procesSing, and use
O FACILITIES
6 The Cqmputerand Information Science Research Centex' has a Digital Equipment Corporation DEC System-10 computer: It is a moderate size flexible time-sharing computer Which is dedicated to research and education in the fieldof computer and information,science. This computer provide's valuable hands-on experience for the faculty and. students of the Computer and Information Science Research Center did permits research activities involving non - standard and innovatiye .Some of applications of computers of both a hardware'and softeare nature..
these research activities which are currently underway e: .' . ... *. 4,,,,,v '"` 4 . . 1) Experimental and developmental research in time-sharing'and multiproamming systems. 2) Complexystems 'simulation research using graphical displaydevices: , 1) On-line information retrieval. systems studies. . 4) Hardware modificaticlf and interface studies./ 5) Software modification and development (e.g.,'PL/1,end cow compflees): 6) Man-machine interaction and psychophysical experiments.
7) Pattern recognition studies. . . . 8) Computer simulation of 'language learningik
- . i ..c0, Speech analysis andsynthesis. . , 10)iAnalysis and synthesis of human locomotion. 11) Computer data-base systems.
The Research Center alsokhas access to the.Unive?sity Computer Centers. , . They are: Instruction and Research Computer Center; Hospital Computer Center; and UniversitySystems Computer Center.
Included in these centers are an IBM 370/165, two IRM 370/158,an IBM 1620 for plotting, and an IBM 7 for analog to digital conversion, -aswell as several 0 3
- IBM 1130 machines, and a number of remote terminals. An TBM 370/168 will be
availableJanuary 1, 1976% -
,Many specialized facilities and laboratories of the Universityare also avNlable-to the staff andstudents of the Research Center. Some of these are
the Office of Computer Assisted Instruction, the InStitute for Researchin r Vision, TelecoMmunications Centel', Behavioral Science'Labolatory,Listening Center, Communications and Control Systems Laboratory; Mershon Center for Education in Nation's]. Security, andmany others. ... t The Research'Center also interacts With the Nohio College Libtary,Center' which is administratively,independent of-the University. The Center was formed. 'by the Ohio College Association, and operatesa common coMmyterized library network connects g the Ohio colleges and universities (both privateand state assisted) and many points outside the state ok Ohio. These include sixteen k, regions with approximately 500 college, university, public and speCial libraries paricipaPing in the systems Two well established national infor- mation systems have units 'On the campus of-The-Ohio Stateyniversity..:The REDLINE system is an automatedon -line service to access medical journals of the previous three year period. The ERIC (Educationaa.Resarch Information Center) system is an automated batch system toaccess research reports and journal literaturein the field of education.These systems are availableto the staff and students. Interaction has also been initiated with'he Academy. for, Contemporary Problems.
I The University has ebtsblisheda university-centered-information system. The infoimation system, called-the Mechanized Tnforpation Center (gC)operates,
as a departmentof the University Libraries. MIC has developed a multi-4 . disciplinary batch-mode information system from machine-readable'databases, . primarily for the campus scientific community. The interface to the MIC systemis decentralizedasmuchas possible thrOligh the. (existing,, system of twenty-three libraries ardund the campus which serve sepcialized publics: MTC a acquires data bases from commercial sources, as well as fromprofessional-'
societies and governmental agencies.,-Research aCtivities in MIC aree . , ' toward improving the services oP MIC to its users through softw'are refinement 711,r and development. This center works closely with the staff and students of-the Department and the Center. , 4 . rf
MIC provides 'two basic types of_services:,(1) computer-based current . . awareness 'to help peoplekeep up-to-date with currentpublications, and e . .- (2) is computer-based retrospective searches to. bring.people .up-to-datt with previously: 1,
published inform:stip. There are-three current awareness services: ,(1) - ' 4, multidisciplinariainly in physical, and biological sciences, and engineering, '(2) social sciences, -afict (3) education: Two retrospectiue-services, multi- - disciplinary and education, are also available.yol- eaChof the,fiye-services, a:specific dita,baseis searched to select bibliographic citations that'ate pertinent tq a person's interests. MIC.now provides current awareness searches to more than 4,000 people and'brganizations and has performed-more
than 25,000 retrospectIva-searches. .., . . . ., z ..
.,/ , i . . . . -.. . \ 12 . .
t 4k-
4
A ACADEMIC PROGRAMS IN COMPUTER AND INFORMATION SCIENCE
The.Program at The Ohio State University emphasizes education, research and the professional prd'aice and application of computer and information, science. The academic program is offered within the Department of Computei and Information Science and encompass.es undergraduate and:graduate degrees through the Ph.D. Statistics showing the growth of the department are found
in Table 1. 4
Organization of The Department'of Computer and Information Science
The Department of Computer and Information Science is a separate academic _unit located administratively in fhe College of Engineering; operating in part as an interdisciplinary grogram with the cooperation of many other departments and colleges throughout the University.
Objectives of The Departmeut-
The program at The Ohio State University emphasizes education, research and the professiohal practice and application, of computer and information science. The educational program offers undergraduate and graduate degrees through the Ph.D. The research activities ulhich are a centiallmrt of the program conalltftf a broad conceptual lise supported by a number of contracts and grants aslell as by the university. The broad core research program and these other research tasks Interact.to form an integrated framework.
yngraduate Programs
Undergraduate degrees in computer and intormation'science are available to students in the College of Engineering,_the College of Mathematics and 'Physical SCiences of the College of the Arts-and Sciences, and the College of Administrative Sciences.- The particular program chOsen depends upon the student's interestsand career objectives. ,
. The undergraduate prpgram in the College of,Engineering leads to the degree of Bachelor of Science id`computer and InformatiowScience. This program is designed fox,the student who wants to specialize i ,computera0 AnformAtion r science froglaithin an engineering environment: H Vtill.vale pro am provides thestudent with.,a)reoreAt_computer Sad information science, M &m4ics, and . engineering science Both compute 'wand -informatiqn science, are assured by specific required courseseqdeetes in sOei:al aw's of engineering
r and science yeti, Sufficient flexibilit eXIsts so:that,a studentFcan'elect a
.portiOn of his techAcal course work in order-to develop his Individual interests.
0.
' There are two undergraduate programs- in the Colleat of Mathematics and Physical Scien0s. .These programs lead either_to_the degree of Bachelor of Science or thejlegree of bachelor Of6Arts with a major in computer and infor- matiascience.',The programs are cast in a'liberal arts setting and are -./ in content. The Bachelor of.Science program provides a somewhat more technicaT and thorough education in-Computer and information science and mathematics while the Bachelor,af Arta program is Romewhat more flexible nn provides A opportunity tospelate,Computereand information science 0 some , other discipline. ,
." 4,41'4 4 r
A Take 1:Growth of Department of Computer and Information Science /
SEPT'67 SEPT'68 SEPT'611 SEPT'70SEPT'71SEPT'72- SEPT'73' 1EPT'74 SEPT'75
A. Staff
1. Full Time 5 11 14 15. 18 18 18 20 -21
2. Part Time 5 9 . 10 11 12 14 16 12 12
. . B. Graduate Studen6 32 89 114 151 165 187 209 1Q8 209 (est)
-C. Undergraduate Students '100 143 300. 485 576 450, 510 475. .490 (est)
D. Course Enrollment , (Autumn Ouarter) 542 770 1059 -1293 1447 1676 1728 1925 2050'(est)
a
'67-'68 '68-'69 '691-'70 '70-'71 71'-72' 72'-73' 73t.-74' 74'-75' 75'-76
N - Students Taught 1977 2892 3933 4703 5174 56100' .612g 6876 7500(est)
M.S. Degree Awarded 7 17 35 44 47 49 67 58 55(est) Po
Ph.D. Degrees Awarded A . 4 8 ' 4 7 18(est)
., -, . Applications for . Graduate Study 181 190 343 425 400 . 323 290 355
Number of Graduate _,... ..
.Students Supported 27 72 .' - 78 .88 89 83 78' 81
4.4 15
wt. .6
The undergraduate program in the College of AdministrativeScience leads to the egree of Bache 'r of Science in Business Administration witha major in comput and informs on science. This program. is designed for the student 4 r that s business ornted and desires an eduCation in computer and information
science0, and a gene al education in the administrative sciences. The program!s objective is not to Make a computer specialist out ofa student, but rather to enable him to re ognize the opportunities touse the computer in his managerial activities, to /now what to expect frotitit, and to'know how to communicate effectively with computer specialists so that computer4ed,projects willbe properly handled from a technical as well asa managerial point cif 1:riew.
Graduate Programs , i 1 . _. The Department of Computer and Information §clence offers graduateprogrpms leading to both the Master's and Ph.D. degrees.The graduate program leading
to the Master's Degree is available in seven options, . ,
Option I'for the student desiring a theoretical foundation in computer
.and information science. . . .
Option II for the student specializing in infotImation systems.
Option III for the student specializing in computer systems. .
, - Option IV for the studentalieciKizing in numericalanalysis.
Option V for the student Specializing'in operationg research.
Option VI for the student specializing in biomedical information
procedsing. 4. 1. Option VII for the student specializing in administrative sq.ence.
% : . N , Each of these options'provides a background ip several aspects of computer'_ and information science, as well as additional mathematical sophistication appropriate to the Student's interest: Tach Of the options may lead to the Doctoral program in computer and information science, mid each may be taken with a thesis option or without a thesis optiOn. (See Appendix A for a listing lof courses by number and tit14'.) ,
All courses of study at the Master's level require completion of acore program in computer and information science, together with the required courses specified for one*of the options and additional courses as specified, the student' adviser. The core programs Includes courses on; Principles-0 Man Machine Interaction, Numerical Analysis, DataStructured, Advanced Computer Programming, Digital Computer Organization, Mathematical Foundations'of Computer and Information Science, Introductiow:to Linguiaticlhalysis, Modern Methods of Information Storage and Retrieval, and, Advanced SeMinar, in'amputer and Information Science.' AO
The graduate program leading to the DoctorL Degree in'Computer and Information/ ScienceScience is flexible in that it is tailored to the particular back grourrd and-Interests of the individual student. These interests mfy lie in any one of the research and.ihstructional-areas already listed as welOas' in Many . . otherYcognete.,ereas. kcognate field is defined as a CieldrsuppOrting or closely .related-to the, fourteen Departmental fields and is ordinarily specified f)ykenIfilleirateci prograM of 'study in other departments of the University.'
Course Offerings'.:-
CurrentlSrthere are about 81 courses (each one quarter in length) offered by the Department,' 22- of which are-largely undergraduate with the remainder being upper level undergraduate and graduate courses. In addition to these courses there are over two hundted courses offered by a variety of departments of the UniversityWhiCh are of interest to our graduate'etudents wbo'are enCOuraged to take these:oourses.
Faculty
The Department of tomputer and Information Science has a full time faculty of twenty-one members-at.the assistant professor level and above. They have a wide range of backgrounds and'experience.The above faculty is. supplemented by staff who have joint appointments with other departments; by staff from other departments who teach courses primarily for Computer and Information .* Science students; and by adjunct staff peOPle who are employed in off campus organizations who teach courses in the Department of Computer and Information Science (see Appendix,B). ThePe.Narecurrently a tbtal of about, 14 supplemental, staff in this category. I.
OV- 0
INTERACTION WITHIN THE UNIVcRSITY .t
41.0,, ts,4,Bot:t.). thelv Research Center and the Department of Computer and Information ,Science'Interact with other departments and research programs within the . University. This is essential because of the multi-disciplinary-nature of
. the activities' encompassed in this field. A number.of the academic faculty have, joint appointments in other departments. Staff members of the Department of Computer and Information Science have appointments in the following departments and organizations: ..
a.' Accounting g., Instructionand Rdaearch b. Allied Medicine Computer 'Center
c.-Art, h. Mathematics / d. Biophysics i. Psychology e. ElectriCel Engineering j. University,Libraries f. EngineeringGraphiCs k. university,,Syatems.Computer Center
INTERACTION WITHIN THE COMPUTER AND INFORMATION SCIE CE e.0MtliJNITY . Air 'Columbus, Ohio is one of the majqr ce nters fo information science and for the transfer of information in the United Stotts. A number of organizations 'are involved with the activities of computer andnformation science. This affords an opportunity for students and facultyo interact with.ippropriate personnel in these-organizations. *Some of thee are:
4 41 a
8
Chemical Abstracts Service h. Industrial Nucleonics b. Battelle Memorial Institute i. State lof Ohio Department' c. Bell Laboratories of Finance; Department of.
*ad City National Bank Highways e. Columbus and Southern OhiO jdColumbus Board of Education.
0'''° Electric Company k., Ohio College Library Center f. Western Electric Corporation g. Rockwell International Corp. s.
Therel':area large number of scientists who come to Columbus in order to visit with the Departmenr:and Center and who usually present aleminar. ,(The ,seminars for'the Periodeof this report are listed in Appendix C.)The people over virtually all phases of computer and information science.
In addrtion,,our people interact at most of the major technical meetings in this%country as participants givfng.papers, assisting on panels, as
, .attendees-, and as officials. Hardly a major technical meeting in the appropriate. fields is held without contribution from one or more of the = personnel from the OhiokState Computer and Information Science Research Center. A,,lisCof these activities can be found in Appendix D.
Research efforts of the staff are disseminated to the professional community through several publication channels.A list of current publications of the Research Center staff is included as-Appendix E. In addition, the. Research Center issues a technical report series .(see Appendix F). s 4
DOCTOR OF PHILOSOPHY DEGREE
The Doctor of Philosophy degree was awarded to the following students during 1974-75.- See Appendixp for a complete listing of Ph. D. dissertations.
Name - Dissertation
James L. Beug, Human Extrapolation of Strings Generated By Ordered Cyclic Finite State Cramilars
Mithael E. Doherty A Heuristic For,ginimum Set Covers thing. .Plausability Ordered Searches 7
-Serge Fournier .The Architecture of a Grammar-Program- mable High -Level Language
Oluwumi Longe An Index of Smoothness for 'Computer Program 4 Slowgraphs r, Edwin>J,'McCauley. A Models for Data Secure Syatets .
Frederick E. Petry Program'Inference From Example Computations Represented by Memory Snapshot TraceW.
Hui-Yang Su Pagination of Programs for Virtual Memory 4 Systems is '9
RECENT'TECHNICALREPORTS
The Computer and Information Science Research Center began publishing a technical report series in 1968. A list of recent reports of .the research center follows (ee Appendix F for complete list.)
BIERMANN,' A.W,.: KRISHNAWAMY, R. Constructing program$ from example computations. August 1974. [41]p. (OSU -TR=/,4 -5)
BUTTELMANN, H.W. Semantic'directedtranslation of context free languages. September 1974. 36p. (.DSU-CISRC-TF-74 76)PB-242 854/AS)
-Is- de, HARTSON, H.R.:.HSIA0,,D.K. Languages for-Specifying protection re4uirements in dAa base systems (Part I). January 19/5.. 61p, (OSU-CISRC-TR-74-10). AD/A-006 280/2GI .' ! IMRCER, Jr.,W.A. Data base modUle'verification - a certification method for data secure systems. June 1975. 125p.(0SUrCISRC-TR475-3)
HSIAO, D.K.: BAUM R.I. Information secure systems. November 1974. 55p. t I (OSU-CISRC-TR-74-9) ' I
HSIAO, D.K.:,McCAULEY III, E.J. A model for data secure systems (Pprt II). October 1974. 40p. i0SU-CISRC-TR-74-7) (AD/A 506 276/0GA)
. HSIAO, D.K.: KAFFER, N.E. 'A model for data-secure systems (Part III). November 1974. 32p. (OSU7CISRC-TR-74-4}.(AD/A-T4,676/3GI)
McCAULEY III, E.J. A model 'for datA secure systems. March 1975. 119p... (0SU-CISRC-TR-75-2) (AD/A-011 359/7CI) V f
PETg, F.E. Program inference from example computations represented by memory snapshot traces. February 1975'. 143p.% (OSU-CI SRC.4R-75-1)
1
t
AI 19' '10
-II. INFORMATION STORAGE AND RETRIEVAL
.) `AUTOMATIC PROFILE ENHANCEMENT BASED ON ASSOCIATION MEASURES ERIVED FROM SDI SEARCH PROFILES
.. The use of association measures to improve informations irieval per- 41, formance has been explored to a considerable extent in thepa t decade, ,Much of the work done to date has been basedon small, static docu ent collections where thee documents in the collection have been usedas a"basis for deriving the association measures. However, this approach appears to be impractical for large or dynamic document collections suchas those usedih many SDI (selective. dissemination of information), systems.
Anlinteresting alternative, which isvbeing explored in this research,,,makes use oftheeintellectual effort expended oft manual preparation of a fairly large, but relatively Stable group of profiles used foran SDI syStem prolAding current awareness and retrospective search services on both large and dynamic document collections. Accordingly profile-profile associations and term-term associations d&rived'from a profile -term matrixare being investigated to determine their usefulness foy automatic profile enhancement of both existing profiles andnew
profiles entering the system. , a ///!A. E. Petrarca, G. J. Lazoritk,B. Brinkman (Sponsor: National Science Foundation. GN, 27458)
A CONCEPT SCATTERING. FUNCTION FOR EVALUATING THE EFFECTIVENESS OF VOCABULARY .CONTRQL ALGORITHMS USED IN AUTOMATIC INDEXING
To improve vocabulary control in indexes' derived automatically ftoM natural language text it is desirable to have an objective and quantitative method of. evaluating the effectiveness of different vocabulary control techniques.
Automatic indexes produced without vocabulary control are generLly beset by ' a certain amount of concept scattering resulting from the inflected forms of words or phrases representing the various 09ncepts being indexed, and in extreme cases, some inflected forms pertain tng to.one concept are completely separated by inflected forms of words unrelated to that concept. This last ti panomenon is illtthtrai'ed by the alphamerically ord redset of keywords plan, planar, planarity,plczning,plans, where the words denoting the concept of planning are separate by those denotiing the concep of planarity.
The disorder assodiated with concept scatteriwrepresents a form of index entropy. Accosdingly,.a concept scattering function has been developed which ti is,intuitively,based on the d 'Ifference in index entropy betweenany, given - index and its corresponding ideal index (having zero index entropy). Appropriate idea's indexes have been constructed and computerprograms have been written to calculate-Values for this scattering*function on indexes subjeLed to different automatic vocabulary controltechniqqesi The results obtained thus far clearly o 13 11
4.-. --- ,
v . i. provide a quantitative and,objective evaluation of, the relative effe'Oti/eness . .of theparticularvocabulary control algorithms studied.rdThecoreptscattering . function is now being used for, selection of a superibr vocabulary control al- gorithm to be incorporated intoan automatic indexing system.
A. E. Petrarca, W. S. Stalcup
e ,
A DISTANCE MEASURE FOR AUTOMATIC DOCUMENT CLASSIFICAVON
This research has concentrated on the..esign efficient algorithm for autotatic sequential document 'classificatio It is assumed that a set of categdries and a set of keywords;'describingbfiese categories is given. Each _keyword is viewed as a k-dimensional vector re k is the number of categories. This enables a document to,be epresented zripet of points in a k-dimensional' space. A distance measure, called-th'Bay sianaistance, isthen,defined on this,space, and consists'of'a magnitude ada a "direCtioxt. It is shown that by -Studying'the variation of this magnitude anddirection as keyword are read from a document, noisy or inappropriate keywords can be isolated and clusters of similar keywords can be identified. By analyzing these clusters of keywords, primary and secondary classes for a documentcan be obtained. Ts algorithm h '\s been implemented on the SPIN data base andencouraging resul ts have been obtained.
L, J. White, G. Kar (Sponsor: National Science Foundation, Office of Science Information ServiCes CN 36340 1 Et
. , 4. . f GENERATION OF UNIQUE dbNFIGURATIONALREPRESENTATIONS OF CHEMICAL STEREOISOMERS. FROM NON-UNIQUE UNAMBIGUOUS 3 DIMENSIONAL INPUT DESCRIPTIONS
During the past feW decades, many' techniques forcomputer -based storage and retrieval of chemical structural informatidn have been developed,several of which are now being utilized for creation of large firesof structural data: The two mist prominent storeeof,sUch dataare those being treated and maintained ,byby Chemical. Abstracts Service (CAS) and the Institutefor S,Elentifip Information - . 0 . g (ISI). , a. 4 , 7!" e The basic objective of creating such Stores isto enable'researchers to interrogate them in, various way to obtain information whiph is:difficult to retrieve 12y'conventionalprinted indexes. Such.interrogationsway take the form . 0 requests for (a) in matron on compounds having A common set of,str tural . characteristiCs (e.g., Studies on structure-activity. relationshipS); (b)in- formation on Apmpounds ing a desired set of chemical, physitta,or biologieaI properties needed for some specific application;or (3).information, on preCursors needed for the synthesis of a desired chemical structure. The databases also. provide many interesting opportunities for patternrecognition studies.. A. There are several, prototype systems available,or under development, to 'provide:search.capabilities suchas the above, utilizing data.baSeS1such as those provided by CAS and ISI. These systems permit sophisticatedsubstructure searches searches for subgraphs'of chemical.structures)on most 2-dimensional aspects of chemical structure. . However, stereospecific iubstructure searches I N 21- 1 la -
- . _...../Cannot, be readily performed on most of the large data bases available because of the way in which the stereochemistry%(i.e., the 3-dimensional "structural re- lationships) is handled.:The CAS apdISI data bases, for example, handle stereo- chemistry in a manner suitable primarily for making global distinctions between .0 to molecular structures whose 2- dimensional representations.are isomorphic. /....4. cr. r - Becau4se of the importance of local aime111 as global Sterecithemistry in O mane areas of research (particularly in medicinal clpemiatrA biochemistry, etc.) . a prototype system for handling these relationships has'been developed. The system generated unique global and locil dedc'riptionsi'for each structure exhibiting certain types ofstgreoisomerism in terms of a global operator colined with a set of unique.configurational descriptorl--:one for each.local site contributing
to the overall stereochemistry of the molecUle. The configurational descriptors . are derived from non-unique, unambiguous input.descriPtio5,0. The prototype system; which is an adaptation of that used by CAS, is currently Wing tested and extended to cever-a wider range'of geometrical configurations encountered in chemical structures.
A. E.-Petrarca,D. G. Williams, P. J.. Wells,
St,
. IDENTIFICATION & MEASUREMENT OF COST &,PRICINq_ELEMENTS PF COMPUTING RESOURCES
, . , .t Dual economic measurement and allocation. cAtgriafor-completing,"cost" and "price", are necessary due to the tncrehsinFly MultithSkitg, shared resource, ctlaracteristics of computing. "Cost" relhtes to,resourceutllizedf"price" relates to the placement of value (uftlity) on a service render t' 'Resource utilized is not necessarily congruous to service rendered; Rost allocation and pricing may require sepatate, distinct measurement and allocition mechanisms. A methodology is being developed tokidentify dosi.andopricing centers and to ,, synthesize measurement and allocation methods on the basis of that identification. In conjunction with this methodology, declarative and procedural-capabilities...m(11 be added to,the Information Processing System Simulafor (IPSS) to provide the modeler with the capability to develop cost and service stpkisties it a minter
similar to response time and queuing. .
..' z. ir i'.. D. Isaacs, T. G. DeLutis (Sponsor:Natioteal ScienceFoundation. Mil6622). 7 IMPROVEMENT OF AUTOMATIC VOCABULARY CONTROL IN IND DERIVED,, FROM NATURAL LANGUAGE TE
v _ . One o the main purPo s of vocabulary control in inindenting' systemis to reduce he uncertaintylin lochting information otythparticular topic. This, effect is .generally achieved-by restricting the number of synonyms of near syrionYms used to.describe a particularparticular topic, atid,q'providing cross references J to the preferred synonym chosen to ean topic.
. ., Unfortunately most automatic indexing s4htems ire totally cking in vocabulary control so that information oh a given_topid ma5, he sAtt among 0 many infidcteeforms'of a i4ord.(or phrase)"as wepas...mu:ler synonyms hav tg to011trdlifprent word foots. In extreme cases, soMe'infleeted forme Re taining to obtd concept are completely separated from eachotil'erby words unrelated to , .., ,, a- ?2- .1
':,,that concept. This can be seen in the alphabetically ordered keyword'sequence rat, rate, rating, ratio, ratewhere the words rat and rats are scattered from ;each other by unrelated keywords. -
Automatic vocabulary control, like manual control, can be attained by replacement of the original natural language keyword by a preferred word looked up in-a thesaurus or dictionary. .However, there are many situations where thesauri are unavailable or too small-for adequate vocabulary control, and other situations where the thesauri are'too large to conveniently store, main- tain, OT use very efficiently. Many of these problems can be eliminated or minimized by stemming-recoding algorithms which not only recognize conceptually related keywords via removal of their inflectional endings, but create preferred/ index words by appending suitable suffixes to the word roots, obtained from the stemming process. The use of such algorithms for improved automatic vocabulary Control in an automatic indexing system is being explored in this research. The effectiveness of each algorithm chosen for study is being evaluated by a Concept-scattering function (see separate abstract) which was designed to provide an objective and quantitative evaluation of techniques for providing automatic vocabulary controll.
A., E. Petrarca, W. S. Stalcup-
4
THE INFORMATION PROCESSING SYSTEM SIMULATOR --.IPSS
Th6 Information Processing System Simulator (IPSS) will betherealizatioq of.a methodology developed for investigating the behavior of complexcomputer based information processing systems. IPSS is being designed to deal with the -problems associated with Information Storage and Retrievalsystem, and to serve three distingt.users: the researchertheinformationsystem analyst/ designer, and the student. One important rysearch,area fob which IPSS will be especially_uSeful is the investigation of luitable data structures/datamanipula- tion operations for generalized Data Base Management Systems (DBMS). The most important departbre from current.simulation packages is'its emphasison the - interrelationship'between the application, the system'software and the I/O process. ¶PSS can be used to describe I./0 activity fora wide variety of computer' systems and applications It is being designed in a modular fashion to assist continued deVelopment Of. both language and`-run control facilities, and to /permit 'the simulation' time system to be otterthan fDM S/370's. 4
. T. G. DeLutis (Sponsor:Natiortal Science FoundatOn, Office of Science Information
Services. ON 36622) , ti
4 A A METHODOLOGY F012 A MULTIPLE GOAL APPROAGR TO COMPUTER SY-STEMS.DESIGN
The computer systems design process is formalized formore effective -0 analysis. The formalization follows a normal problem solving approach (goal determination, etc.) but requires quantiffeation of goals and constraints at each level. Because the design of a computer system inhereritly has any levels of conflicting objectives, the muiti-level, multi-goalstructure of goal programangwas selected *for the metholodogy. This approach fOces a designer. to determine the value of his goals beforetesting his,design, rather than
4. 6
14
molding them to fit his results. The output of the,goalfA,41 programming approach yields more effective information to the decision44ker (designer) than other linear programming techniques..The goal programming appraoch will be extended to Utilize data generated by the'- Information Processing System Simulator (IPSS) about a given computer system Model, and then indicate which variables_need modification in succaeding runs.'
\ J. S. Chandler, T. G. DeLutis, (Sponsor: National Science Foundation.
. GN-36622). '
OCCURRENCES OF TRAILING N-GRAINS AS A FUNCTION OFDATABASE SIZE
Tto facilitate the development-ofgeneralized'automatic procedures for the .removal of suffixes from natural-language words, the occurrences of trailingn- grams in natural-language data bases are being-studied. Digram and trigram occurrence counts have been obtained. for 7 graduatedrsize samples of words taken from the nattifil=1finguage titles 'of docune ts cited in a multidisciplinary data bale. The, 7 samples contain aRprOximately 1000,133, 10000,...., 106 title. words. A plot of digram occurrences as a function of sample size clearly shows asympotic behavior, with likely occurrence of all theoretically possible digrans 'in a 107-word sample of natural-language titles. However, a similar plot of trigram occurrences suggests that a data base of approximately 109 title words would be requiied=for the theoretical upper limit of trigram patterns to be reached, if at all. The data collected in this study is being used for evaluation and imprOvement of stemming-recoffing algorithMs to.be used forauto- matic vocabulary-control in indexes derived fromnatural language text (see separate abstract).
A. E. Petrarta, W. S. Stalcup (Sponsor: National Science Foundation. GN.27458) ,
PERFORMANCE EVALUATION OF DATA BASEMANAGEMENT SYSTEMS- it
. A methodology forithe performance evaluation of data basemanagement systems id being developed. A data base management system (DBMS) is charaCterized in terms of: ,(1) the structure and contents of,the data}mac, and (2) the capabilities provided by the DBMS-for the creation, update, minipulation, and ;accessing of data in the data base. The methodology `consists of declarative and procedural domponentsfor the characterizatiOn of a DBMS. There are declaratives for.the logical description ofa data base (in terms'of the relational.: structure of the data and the desired.acdeas paths to the data). Declaratives and proceduresare provided for mapping. a logically - described. data base to physical files. Special data management 'primftive operations are providedtoo characterize in a procedural manner the data manipulation facilAles,of-aDBMS. The methodology is sufficient* general to model a wide variety of-DBMS designs.
J. A. Aitken, T. C. DeLutis, (Sponsor: National Science Foundation. .(.N 1422).
- ,0
24 , III, HUMAN INFORMATION PROCESSING'
,
HUMAN EXTRAPOLATION OF STRINGS GENERAtBD BY ORDERED CYCLIC FINITE _STATE GRAMMARS ,
A The problem of systematically investigating human .behavior An relation to the complexity of sequential concepts can be decomposed into that of devising a suitable measure of human behavioi and a Aasure,of complexity for sequential concepts. In this study,_human learning was measured in, teplis of errprs.made
in extrapolating character,strings. The strings Used were, ordered in "terms of . structural parameters of the grammars used to generate the strings. The structural parameters used were the side of the terminal alphabet and, the level of embedding or chain length of the generating gramMars.Two experilents were performed using human subjects'in acomputerized implementation pf a string extrapolation paradigm called the charadter prediction task.'
. . In the first experiment, it was found that increases in the 'size of the terminal alphabet and the level of embedding,of the generating grammars produced . statistically significant increasesin the number of trials to last er r ilid .- the number of errors made in extrapolating the strings. In the seconJ7experiment, it was found that increases in the chain length of the generating.grammars produced statistically significant increases in the number ofierrors,made in extrapolating the strings, corrected for those trials where there was insufficient information to correctly predict the strings. Finally, for all rammars with.12 characters in theirterininal alphabets, a correlation of 0.695 was found between the corrected number of errors and the chain length of the Operating grammars. . . . These results show strong agreement betwen difficulty of learning the strings and the complexity of thd strings defj.ned by the generating grammars. Cyclic finite state gfammars ot,the kind used in this study may thus be considered as reasonable, represegpations of tasks involving the attainment of, sequential concepts where difficulty is a parameter.
'J. L. Beug, R. L. Ernst
.4 16 ,
I t . YY
IV. INFORMATION' ANALYSIS .
is Ak AN EMPIRICAL ANALYSIS OF INF'ORMA'TION AND RELATIONS AMONGELEME7§.9:4.A DECISION .., SYSTEM- i
' ,' The relationship among elements of adecision'systedis beingftelalyzed by % measuring .the effect of clifferefff information displays on aecision'task. In
. the- decision task, a decidion maker (DM) was placed in the,mleof director of an information retrieval sysegiVith the objective ofmaximizing the number of users of the system. The DM specified, relative weights of outcomes, action- : .outcome pair probabilities, and course of action selection for thetime frame of
twelve, consecutive decision periods -.The information' display wasamedium for . . Na. feedbackrof,results of DMresponse after each decision period. Differentiation - amqpg infomation'displays wis made oti*the basis of level of information piesented and combinations of these levels. in information display and the associated performance is viewed asas a transformation of data into different' Dtvels
. of information. . - , . .- ._... Three system functions transformedDIresponses into three levers'of infor- mation'defined as follows: operational - information to aid the DM in operational or procedural control; management -'informationb aidth4 DM in managerial or 0 policy control; and strategic -,information to aid the DM in top level control. The system functions so defined imply that the informatiod content of-operational level information is contained in management level which in turn is"contained in _ .
. strategic level. . . ..: , . .
. . . ..- -Three sets of values related to D11responseswere examined: a) the1actul._ pre- defined ordering of courses of action as-specified by the-system (i.e., state of nature), 13), the DM- specified ordering of courses of action, and c) ti4,DM course of action selection. System and DM ordering of a..cts. were based On expected value ,' compUtations which were determined from EA-outcome pair piobabilities and.V.clativ41
weights for, the outcome set. . t . 0. . . . 0- 11' .: , . Several techniques were, applied in theqmalysis of DM responses. DM perfor- - matrce measures were developed to determine the deviation from the optimal declsjon state as specliied by the system; In addition; a mithematice1 modeltof &incept learning was used to define and categorize DM responses for two methods of analysis.: For the first method, this mddel evaluated DM responses'overall twene decision periods ad'steady-state transition probabilities. For the second,,Bayesian .,
techniques were used analyze thetrate Of convergence-to the optimal de-cision , state from one decision period to the next. Thus, the impact of information,dis- . - play on DM responses was in terms of Maikovianand Bayesiari probabilities respect' tively. BOth methods. analyzed DM responses with respect to the prevailing state otnatureNs deffned'by the system and to the DM's perception ,of that -state of ...... , \ 0, , 1.- ( cloture, .
,: ----,\ T,hetti g ion task and theinformation displayswereinco. rporatvdinto an , , 9 t , 17 \
interactive simulation. A prototype model of this simuia n wasdeveloped earliir in the form of a management game supported by theechanizedInformation Center of The Ohio State University and was reported elseWhere.:
M. J. Lee, R. L. Efnst
A GENERAL THEORY OF INFORMATION FLOIi":"MD-ANALYSIS
A generalized framework for the development of a theor) of information flog which permitsrthe analysis and quantification of information has been suggested. In addition to its' theoretical and conceptual interest, there are major and immediaq,...implications fof the development of inflDrmation systems and networks aslaell as for the general understanding, of information flow, retrieyal, and transfer..
Numerous intuitive notions exist about the interrelationship between infor- mation and decision- making..At the pragmatic level, information has value to the extent that it is useful as a resource for purposeful activity. The primary "purposeful.activity" in life ia decision-making. Hence, information and decision-making are inextricably tied together In our formulatiOn, in fact, information is defined in just that way, as being data of value in decision- \ making.
As a eqUence, a measure of the amount of information in a data set or messagejs defined.in terms of a quantity called the decision state of a decision- maker. The decision sta6e. is a function of-the determinism of the decision - maker, i.e., how easy it is for.him to make a decision in a particular decision situation. We suggest a way of evaluating the decision state quantitatively-.e
We assume that the decision-maker can compute expected values for each alternative.' The relative expected value of each alternative Can then be computed and normalized so that the sum of althe relative expected values is The value oftlie decision state is then definedAisfthe summation of, the expected values of all the possible courses'of action'weighted by the relative expected value of each course of action. A new measure for the information contained in a particular decision state is develoi*d:
= m E {REVlai2n - i=1, ,c)
where m is the number of courses of action available to a decision-makeiWnd REV (d.) is the relative expected value of each course of action.The informa- tion is defined in terms of a two - choice deterministic situatiowhich we call a "binary choice unit". This measure is universally appliCalate for all informa- tion -that is)concerned with the effectiveness of the data upon the recipient. t c A mdasure of the amount of information in a data eel or message can be Artived'at by computing the difference in the amount of information in the - decision state before and after receipt of the data. ,Th.sit is, the amount of information is arrived at by considering the impact this new data has on'ihe
27 o
18
decision-maker's decision'state. In symbolic terms, L(D), the amount of in-
. formation in data set D, is
I(D) = I - I t#1 t'
where I and Iare,the amounts of information in the decision stdte after t4-1. t and beforereceipt'of the data set.
The suggested information measure leads to a measure of the.pragmati6' information content of a data set for a particular decision-maker at a partic- ular point in time., The data acquired, processed, stored, and disseminated by .an information system may be used, however,, as a resource by various decision- makera at edrious points in time. Hence; in the design and development of information systems, there exists a PI-oblm whose level of complexity isan'order of magnitude aboye that of the primary prOblebtaderaed in this study the problem of quantifying the information contained in a data set in terms of its overall usefulness for a range- of decision-makers Over a period of time.
One possible approach to this problem or'assigning a number to a data set to indicate its composite information content would be to start by determining the relationship between the effectiveness of a decision-maker and the informa- tion content of the data set. Since what is really desired is,some indication of the information content of this data set for this decision-maker over a period cif time, one may determine some 4ndex I(D) of the average information contained in data set D over some period of time. Then, if one were to develop a lqasure of the effectiveness of each of the decision- makers fcT whom this data set serves as a resource, it would be possible to formulate an information profile for the data set. Such an information profile would indicate the average information content of a data set as a function of a decision-baker- effectiveness. ts%
If such a profile.could be determined forevery data set tobe stored in an information system, then some namber derivedfrom this profilecould serve AS an index of the composite value_of this dataset. This methodwould be of ' major importance for the development of a soundprocedure for thedesign of more effective inforination systems.
M. -C. Yovits, J. G. Abilock (Sponsor:National Science Foundation. 'GN 31628)
A STRUCTURAL MODELS0R-pECISION SYSTEMS
This research is a continuation of'previous effOrts concerned with the development and application of gerMralized information systems model, Its major focus is on the structural properties and organization of elements-of decision systems. TO date, the research has been confinedto the formalization of such systems. 4.
A decision system is a Special case of an information system that possesses connected information acquisition and,dissemination,decisiow-miking, execution, and transformation functions as in the Yovits apd Ernst Generalized Information :System model. The maj r distinguishing characteristiscs of-a decision system are - that the decision-makg situation is well defined such that the elementsof the . 28 19
decision task may be represented as elements of the infoplation system. For a standard analysis of a decision talk into acts, states, events, prObabilities of states, values of act-state pairs, and consequences, correspondences .between the decision elements and the Generalized Information stem are: (a) infdrmation acquisition and dissemination- -under enVironme al and prevailing decision-making policy partitions of the set of consu.uences) inputs, an'act state matrix whose values are pairs of state probabi ies and values of acts conditional upon the occurrence of a state is generate., (b) decision-making--given the act-state matrix and its associated values, an a chi:dee rule (relation), a subset of pdirt within the matrix is defined and ou utted as the ' course(s) of action, (c)exe- utiongilen the course(s) of'action,-,a system consisting of a subset of the environment produces a subset of states (events), and '(d) transformation--under the prevailing decision-making policy and environmental context, the event(s)-. are associated with a particular act,generating an act-event pair. Partitions over possible act-event pair lets and the associated measures of,the partitions definb the decision state(s) for a particular decision. system. The states so defined also specify the prevailing context of the environment and decision- Making policy (consequence set partitions).
These correspondences have been cast as a system of reticulated linear stochastic seqttential Machines whose kernel is the act-stare_ matrix. Preliminary investigation of the system has shown interesting tentative results, bearing on information acquisition and utilization (learning), and uncertainty'in'decision- making when approached ihPeugh mathematical models of learning and Bayesian probability analysis. Moreover, the regearch show's considerab144 promise for developing various design criteria for some kinds of command and control decision systems, and developing their simulation.
R. L. Ernst, M, J. Lee. A
vt.
Cr
1. .
z
29 20.
. LINGUISTIC ANALYSIS,
THE AFFEOTS OF PHYSICAL ENVIRONMENT ON THE INTERPRETATION OF LANGUAGE
The aim of this: work is to explore the wsys in which the spatia ontext of a spedch act is_involved in the assignment of an interprela,ion t he A utterance and to allow for such involvement in patural language.und anding systems. Linguistics have studiddospects,of.the interaction of y Acal environment and' meanftig under the eitle of space-deixis.' Stall °work has centered on accounting for the affect the location of t e speaker and addressee
can have the ifterpretation of an utterance. Howev , the location Of other objects can algo have influence. This is seen when t...eorientationof one object affects the interpretation of spatial reratibnp with respect td another, such as when a theatrical stage defines what is to the left of at-actor without
consideration of his orientation. . it The study is also considering the ways in which objectsstake'on spatial decriptions:This includes such_conventionseas assuming orientation based on position when .utilized by humans and assuming orientation based on object t , . movement. . ,
The phenomena studied bear importantly on natural language man-machine' interaction in such application as 'machine control and computerized consultants for physical tasks: One way to a11nw for them,is to sufficiently restrict the linguistic forms such that the point of view and the. spatial descriptions are well defined. One such set of English forms with appropriate rules of inter- pretation has been identified.'. A possibly4preferable means of operation would be to allow the machine acdess tesome of the nonlinguistic sources of informatiOn which humans utilize'. Work to this enthis now in progress in connection with other projects here at The Ohio State Universityr
1
N. K. Sondheimer °. .
A FAMILY OF TRANSLATORSFOR PHRASE STRUCTURE PROGRAMMING LANGUAGES ti
° Formal properties of ranslation betOten'high:level.programming languages
are studied. The basis for thiastudy is a formal model of programming languages '
called a "phrase-structure system" which has both a formal grammar. and a formal = semantics. The sett of'sentences q phrase-structure aysteint are plat the recur- sive sets, and the meahings of the sentences are the. recursively enumerable sets. We defille and study a family. of "phrase-structure translators". Trans- lation can either be "syntax- dirette4 ", or "semantic-directed".or a combination of both. Phrase-structure translations fall into the latter category.Yor each Member of the family, we give 'sufficient conditions for the "correctness" of 'the translation. We showin certain cases that or a ,phrase-structure-t-rans----7-..- lation from language A to'B, an "inverse" phrase:-structure translation from B 30- I.
4 to A may be effectively constructed. We pr ve a kind of "pumping theorem for phrase-strupture translations and usit torshow that far certain members of the A family and certain languages (eg,, ary. b ternary notation), nOttranslators exist. .Next we prove that for any rec rsive phrase - structure, languages, A and B, where there is an ef4ctive ansla ion 7from A toT, if there is a constant k such that for all w in k*length (w) < length(T(w).), then there is a phrase-structure,trarislator whic computes T. Finally, we investigate the relative time complexities of the bet of the famkly of translators.
H. WT-ButtelMann, A. Pyster (Spon4 Air Force Office of Scientific Research. 75-2g1)
"F0RMAL GRAMMARS WITHOUT SYNTACTIC TABLES-
We examine the-languages of puo grammars, first studid by Gabrelian (unpublished). PUre grammars are ormal grammais with no.nonterMinal symbols (except the axiom, which may not ap ar on the right hand side of any production).
I, Aohierarchy of grammars and lan Oges, parallel to.that of Chomsky's, is formed and investigated. The propert os of the pure hierary are markedly different from those of the Chomsky hi'rarchy andmany resu4Aoon these differences are proven. In addition, we discuss same of the functions of nonterminals.and'why not having them in pure grammars affect the generated languages. 4 aft a- H. W. Buttelmann, A. Pyster,.L. Reeker r A FORMAL THEORY OF THE. SYNTAX, SEMANTICS AND TRANSLATION OF PHRASE- STRUCTURE. LANGUAGES 4. A formal definition for a semantics for phrasestructure grammars, called a phrase, structure semantics, has been developed. It is a model of the following semantic philosophy: 1) it is phrases which have meaning, and 2) the meaning of a'phrase is a'funttion of its syntactic structure,and of themeanings of6its constituents.A pAir (G,S) where G is aphrase structure grammar and S isa r--; phrase structure semantics, is called a phrase structure language description. A language is not just a set of sentences, buta set of sentences with meaningsk assigned to ithm.- The language of.a psld is the set of all pairs.(x,m).such that x is a'sentence of the grammar and it,ls'a non -empty meaning assignedto x by the language definition. The sets 'of sentences of-the languages of,,pald's are just .the recursively enumerable sets. It .has been shown.that for any psld with a type 0 or.type 1 grammar, there existsa psld with a context free graMmar that defines the same language. Translation is defined on the languages of paid's, and it has been shown that the translation function 4iiit the languages of arbitrary .pairsof Psld's is effectively computable.
H. W. Buttelmann (Sponsor: Air Force Office of Scientific Research.75r2811)
ON THE SYNTACTIC STRUCTURES OF UNRESTRICTED GRAMMARS. 1: GENERATIVE GRAMMARS AND PHRASE STRUCTURE GRAMMARS
4s s Formalciefinitionslfforthe syntadtic structures of unrestrictedgrammars are given. The traditional foims for grammar productions give riseto "generative 31
'.. c / .... 14 c 15
22
, grammars" with "deTivation ,structures" ,(where productions have theforth' a - a), and "phrase structure grammars" with"phrase structures" (where productions have....., the form A a,/p-0,,two distinct notions of ,grammar and syntactic structure which becomeinalstinguishable in the context free case, where the structures are trees. 'Parallel theories'aredeveloped for both kinds of grammar and structure. We formalize the notion of Structural equivalence for derivations, extended to unrestricted grammars, and we prove that two derivations arestruceurally equi- valent if and only if they have the same structure. Structural, equivalence is anqequiwaldnce'relation over t,e derivations of a grammar, and we give t simpler proof of a theorem by 'Griffiths that each equivalenceclass contains -a rfghtmost'derivation.We also give a proof for the uniqueness of the rightmost
. derivation, followi',4 a study f some of the prOperties of syntactic structures. Ne);t, We investiga*.relation'ship between derivationstructures and phrase Structures and -show thaethe two concepts, are non - isomorphic. There is,a natural correspondence between generative productions and phrase structure productions, .and, by extension, b9tween the two kinds of gramMars and between, their derivations. structural - But we,ihow that the' correspondence does not necessarily preserve equivalence, in either direction. However, if the correspondence from the productions of a phrase structure grammar to the productions of a genera4tive ,grammat is abijection,.. then structural equivalence on the generativederivations 'refines the image-under the correspondence of structural equivalence on the phrase, -structwie derivations.
4.. 'H. W. Butte \mann AV
ti .
ON THE SYNTACTIC STRUCTURES OF UNRESTRICTED GRAMMARS. II: AUTOMATA
We,define a generalization of. the finite state acceptors for derivation structures,and'for phrase structures. Corresponding to the Chomsky hierarchy of grammars, there is a hierarchy of acceptors, and for both kinds of structures, the type 2 acceptor; arse tree automata. For'iu, 0, 1, 2,.3, the sets of structures recognized by-the type i acceptors are just the sets of projections of the structures ofthetype i grammars, and the language's of the type i acceptors are i languages.' Finally, we prove that thelset of syntactic structures ofarecursively enumerable language is recur94ve.
H. Buteelmann`
PRACTICAL ALGORITHMS FOIL COMPUTER COLLATION OF LONG NATURAL TEXTS FOR 'SCHOLARLY USE
A collation of.a text and its variant, say-of a first and revised second .
edition of a novel, consists ()flan ind4ation of the changes. that weremade , from one text to the other. It is evidently desirablethat the tedious task of preparing a collation be automated if possible; the scholar it moreinterested in using. than in preparing collations. Unfortunately there are several prOblems which arise in attempting to automate the process. The data base in large,. say twice 100,000 English words, and anyineffiCient algorithm could be expensive to use.The definitions of what constitutes.a collation are somewhatsubjective. Also the direct finear scan approach can easilyagee lost on texts that have 'received any but the simplest revisions. f 23
We attack this problem by making precise'definiedonsof global and local collation. We-are developingan algorithm which realizes bath definitions and uses newly developed techniqdes:for fast pattern matching in attempting to produce useful collationsat.reason4lecost.
. D. 4. Moore t
SOME PROPERTIES OF SYNTAX-DIRECTED TRANSLATIONS
We are investigating problems relating to the existence and:.computabiliti of syntax-directed translations. The properties oftheaetof syntax., rees generated by a fte generating set of trees, and maps on-those sets, of trees, are being studiedand, We are investigating the following conjecture:Let Lbe a phrase- structure language of numbers Written in radix a and Lb a phrase-sta ructure language of numbers written in radix b. Then a finitely specified syntax-directed trans-
. lation frOm L to Lb exists if ldbb is rational. a . a
;- H. W: Buttelmann, F. 3. Dickey'(Sponsor: Air Force Office of Scientific Research, 75-2811.)
SPS: A FORMALISM FORJEMANTIC INTERPRETATION
This'project is experimenting with a formalism, called SPS;,for writing, semantic processors for natural language understanding systems. SPS is intended for use in turning underlying syntactic structures in the form of constituent structure trees into underlying semantic, structures in the.form of nets, composed of PLANNER like assertions. The formalism is based on Woods-style "pattern action" rules. "The pattern element specifies free fragients and various types of seledtional restrictions. Mn the action sidea variety of devices, including the use of registers, ,allow common reference to entities in the assertions produced. The registers used forteference'can also be used to specify selectional restrictions acros rifles and for establishing default conditions for handling semantic elli s. Finally, SPS provides a control structure for the ordering of,the app cation of the rules that interpret .constituents and to control, in part, where the tree fragments:are matched.
The power of SPS is seen in its unique ability'to allow for the develop- ment of Case structures,'especially'the structures connected with the English. prepositions that reference location, orientation and motion inspace. These forms have always been troublethothe for Case systems.ParticUlarly difficult-
4.4 are the facts that 1) more than one Of these prepositions can appear in a sentence in the same rule, 2) their Ippasarance can correspond to-the need for m(xtiple predicator semantic structures, and 3) they exhibit complex distri- butional and semantic relations among themselves and with respect to other sententi eleMents. Sk'Scan allow for each of these phenomena. ti An interpretation system for SPS,has been implemented in LISP 1.6.on-the-
DEC System-10. ,
N. K. Sondheimer, R. Pardo,D. Perry r 33 24 0 '40 STRUCTURE OF COMPUTER AND NATURAL LANGUAGE ALGORITHMS
The ultimate aim of the research is to provide better, techniques of storing
and retrieving information in a computer data base; by developing e theory of . the manner in which the information contained'in such data bases is distributed. The contents of such data baSes might be voiumes of text written in some natural language, such as English, or, algorithms written'in some computer language, such as FORTRAN or P171.:Th basis,of such a Aistributiowtheory for computer algorithms has already,b'een proposed, but' the area of applicabilfty.of the theory needs to be extended:.Thus, 'the technique t be followed will include (1) extending the area tf applicability of an existing distribution theory of computler algorithms', (2) developing a parallel theory Applicable to natural language, and (3) imple enting that thecry via computer program and testing it against current natura language data samples.
2. . S.H., Zweben
C
STUDIES IN DECIDABILITY, EXISTENCE, AND EFACTIVENESSQF TRANSLATIONS ON PHRASE- STRUCTURE LANGUAGES
V This researchrisdirected towards,an analysis of techniques of translation of formal languages. A formal model of semantics, from earlier work, is used to study various types of translation: Properties'of the translators are studied, and conditions under which they translate are investigated. A set of conditions is provided in.order to gdarantee the-C....the translators produce correct trans- lations. It is proved that these'conditions are a minimal set in order to 'provide this guarantee.Two types of translators'arestuci\ied and-compared. Finally, a study is made of a limited clad of translators\.that can be algorith- mical/y generated. An algorithm is developed to provide a'translation mechanism which is a 'best possible' translator. It is then proved that no other trans- lator of this type sari translat.et any source language sentences that the translator produced by the algorithm cannot.
H. W. Buttelmann, R. Krishnaswami (Sponsor:Air Force Office of Scientific -Research 75- 2811.)
''\ .
\
c STUDIES OF THE REPRESENTATIONAND MANIPULATION OF LINGUISTIC INFORMATION
...
. f The Computation and, manipulation of information abdtraCtec. from linguis tic forms for use.ia such tasks as Question-Answering or Roboticspresentsmany interesting pioblems. This projedt is frtely attacking whichever of these problems appear tractable. The representation of quantification, and spatial, and temporal refeience have))een the prohlems,most considered. The area Of spatial reference has been*thi\mos4Successfully studied with results including a proposal for extensive revision of the Case deep structure represetation._. Work is also in protress on a. theory Of interpretation for semanlid nets which would permit the nets tb be considered as procedural id-formation. 4 N. K. Sondheimer 3 4' 25
SYNTAX-DIRECTED AND SEMANTIC - DIRECTED TRANSJATION OF PHRASE STRUCTURE LANGUAGES_
. se" The basis of this theory is a model, from earlier work, of language denrip-- tiod called a phrase structure language description, which contains both syntax and semantic information. Certain translation algorithms for these languages are being studied, which promise to be more efficient than the general algorithm 'of .earlier theory. These algorithms are either syntax directed or semantic-directed, depending on whether they are controlled by syntax in-formation exclusively, o'r hy' both syntax and semantic information. A particdlarly fast translation algorithm ,is a slight generalization of the syntax-directed translations, called a "phrase-. structure translation'," where the syiltax control information can be specified in a simple, _finite way. It has been shown that it is possible-to use a semantic;a directed translation scheme as a translator generator, to'produce the finite ,specifications of a very fagt phrase-structure translation,'if such a'translation:.
; ' exists. . . r ,-/ t . :
. Running computer programs of the translator and translator generat9r have , been developed, and the programs are being used to develop and est definitions and translations of very iimplp languages. 0
H. W. Buttelmann (Sponsor: Air Force Office of Scientific Research 75-2811)
Six
,
^
35 , 26
VI. ARTIFICIAL INTELLIGENCE
TIEINIKGRATION OF GRAPHICAL,AND.SEMANTIC.KNOWLEDGE
This work addressee the problem of integrating(Graphical and,Semantic Knowledge into a-data-base for theouppori.of 'Graphics and Natural Language Processing Systems The goal of'the project is to allow . Graphical Information to be used, in the interpretation of Linguistic forms, and Semantic information to be used in the interpretation and construction: of 'Graphical Structures. 'Achieving thisgoal should facilitate man- machine interaction.
. As part of this project, current programs and structures in the areas of Computational Semantics, Computeg Graphics .and Computer Vision are being considered with the intention of isolating suitable structures and techniques or identifying the need rot new ones..Suitable domains for use in devglopment of a prototype system are also,being considered.
I D. C. Brown, B..Chandrasekaran, A. B. Lucido, N. K. ndheiMer
4 6
t - LQW- FREQUENCY RADAR AIRCRAFT DETECTION AND.CLASSIFICATION'
44 , Previous research has demonstrated the feasibility of using multiple low-
frequency radar returns for target clasdification.. Simple object shiapes have , been successfully clagsified by linear techniques; but,aircrafdata poses greater.difficulty, as In general such data are. not linearly separable. Since two parameters specify the aspect angle, this data lies On a 2-dimsional surface in n-space, whekre 9 is the number of frequencies utilized. A bill ear approxi-
mation of this Vilacd has provided interpolation capability:, impr ved proximity 00 information, and an intersection algorithm which determinesyhether the airCraft data is separable. Improved separability together with deareaied/susceptibility
to additive noise wasNtainen by mixing horizontal and,yerticalipolarization Y0 data, at an optimum polarization angle. Studies.haveindicated te,importance of a phenomenon known'as'bias,'where in, the idenrificatiOn of 'a to t signal as one of,two given aircraft, the probability of misclassification forione'1 aircraft can
be substantially larger for one aircraft than the other for high noise levels'. .
Thus classification algorithms are currently being developed which,jointly mini- . mize the probability of misclassification and the bias effect. Ano,therAppkoach % being investigated is multilmint classification, where the inpA is a sequenrce
1 of independent radar measurements of the aircraft target to be identified. It has been shown that this approach cansubatantiallreducey, the effects of noise even ,for a modest sample size, and a numbern of classifiers are h'pliag investigated
for -this purpose. . ,. . o > L. J. White (Sponsor: Air Force Office of Science Research. Grant 69 -1710)
-36
'1"1. 27
PROGRAM INFERENCE FROM EXAMPLE COMPUTATIONS )ESENTED BY MEMORY SNAPSHOT TRACES
Example computations are siudi an easy means of.communicating A person's idea of an algorithm to the computer. A computatiOnal environment is developed
in which"the user can carry out a computation needing toexpress hiMself only by . the changes in contents of the memory variables produced by the calcdlations. The sequence of.these memory'changes is captured by thelystem ina memory snapshot trace. Using the trace and the specification of the computational environment in which it was produced, two steps, decomputation and synthesis,are performed to infer the program from the given computation. Numetous experiments on the inference of various progtams are described illustrating the use of the system.
F. E. Petry
"Jo
4 A*NTAX-DIRECTED METHOD OF EXTRACTINGTOPOLOGICAREGIONS FROM A SILHOUETTE FOR PATTERN RECOGNITION
The overall objective of ttie_research is to developa method for decom- posing, hierarchically relating, and extracting topological regions from pictorial data. That objective is accomplished inawo .steps both pf which necessarily depend on a binary- valued pictorial pattern--a silhouette.
. .: In step 1, the picture decqmposition, a silhouette, definedas A picture, can be decomposed into two type's of topological regions:.aurrounda and °onager;
(multiply-connected images). Each collage is tbin considered a picture unto . itself, and it is, in turn, decomposed into surrounds and collagga.Thee . decomposition is directed by a top-downparser. The topological regions at any level of decomptisition are related by the Parserthrough a context-free grammar. Sueh a decomposition.enables any picture to be expressed in terms of surrounds and collages where the'collages are saidss, (simply-connected images).
, .
In stet') 2, the solid decomposition, a solid is further decomposed into ' topological regions by Separating it into disjoint subsets, which are formed by selecting' points of separation that break the solid into meaningfdl subsolids. The problem is formulated in terms oI a graph-theoretigAl problem of recursivelx selecting the weighted median and/or center nodeh of a tree and its Subtrees. A solid is first reduced to a tree by forming its skeleton or medial axis. This, tree is then expressed as a weighted connection matrix from which the weighted median. and/or Tenter nodes of the tree canbe.calculated. The tree is then ---16roken at ,the selected nodes- forming subtrees (and submatrices)on which the decomposition procedure is repeated.As in' tep 1, the decompositi'skl\is directed by,a top-down parser and the topological regions a.to,any ltve of . decompogition ere related by the parser through a context-freegrammar.;,grammar.; 'The N output of the parser is a string description which can be used to :extract any.'
..... indicated subset of the solid. , . . .. . , Software. has been writteg to simulate the array processor and to fortc group variant euler skeletons. . _,...- .,.
, . , . 'J. Amoss, 4 K. J. Breeding, B. Chandrasekaran (Sponsor: Air 'FOrce"_Office of Scientific- Fisearch, Grant 22 -2351) -- 4 37 8. 28
4
4
"\--- VII: INFORMATION PROCESSES IN PHYSICAL, BIOLOGICAL ANDSZALSTEMS \
... . .
DISCRETE COMPUTER SIMULATION MODELS FOR ANTHROWGICAL STUDY Of SETTLEMENT AND SUBSISTENCE SYSTEMS 2- 0
We attempt, to apply the techniques'oI computer-modelling using, disCrete simulation'to the problem of understanding patterns of settlement and subsistence of human groups. We_concentrate on prehigtoric populations which are small and isolated, and for which the archeological data are abundant.
A model is being developed for a particular aboriginal tribe'that lived in the Glenwood locality of southtrn,Iowa from 900 to 1300 A,DI During this period a transition occurred-from dispersed to nucleated communitiesi.and there isample archeological evidence that the 'manner of subsistence changed concurrently. These transitions accompanied -a continent-wide climactic transition.
The simulation parameters includeclimatic, ecologica and kinship patterns. Hypotheses.are being formed concerning the effects, of these parameters cm the settlebent and subsistence patterns. They are to be tested on G,SIMSCRIPT program for the model.
The main reason for the research is to study the usefulness of this type of modelling in anthropology, and to develop a methodology for constructing and validating such models.Thus careful attention is being paid to the problems of proper utilization of.artifacts in the inference' step, and 'theeplace of simulation models in this process. .
D. J. Moore 4 *0
4 N \,
38
.4 .46
VIII. MATHEMATICAL TECHNIQUES
1 - AN ALGORITHM FOR SUB-OPTIMAL SET COVERS WITH CONFIDENCE Mik-SURES OF OPTIMALITYI
. An effective heuristic hasbe\en developed for the minimum set cover problem: ran cx such.that Ax > 1,-x = 0 or 1; A binary. Two algorithms have been developed: one for the general costs problem (arbitrary c vector), and one pir the cardinality
problem(all ci equal)'. Both algorithms are implemented in computer pi grams .n: FORTRAN and assembler., ,Test results for the cardinality problem exhibivery good performance on test pioblemis from the literature. An analysis has been completed which can be applied to give a measure of confidence of,obtaininga global optimum solution. ' . i .
4 _ L. J. White,'M. E. Doherty . ,e4,.'
BINARY STRINGS AND.TOPOLOGY a 4.6 As little time was available fer this research duriftg the pastyear only. onesignificant result was obtained. The procedure for generating strings . (binary decimals) over (D.4) corresponding-in the limit to..,points ina space admitting both metric and treasure, was applied to the following case.'- Consider a rectangle in the Euclidean plan whose diagonal divides it into 300-60:-900 triangles, and dr9p perpendiculars on that diagonal, from the-other two vertices. Identify hypothenuses of equal length for each of the two pairs ofcongruent triangles, whereby the original rectangle is split into a larger anda smaller, rectangle, both similar to it. The points in tIfe larger recta gle have their strings start with 1,the-others with 0,, Continuing thisproce sad infinitum associates a unique point with every-string.'" Atevery step dimen- sionality is. preserved'but connectivity increases. In the limit e have an 3 infinitely connected "supertoroid",a 2-D manifold whcise geometrynd topoloty are Mind- boggling. The discovery Of such spaces and very prelimina y studies of their properties- comprise the yearis activity in this area.
J. Rothstein
COMPARATIVE STUDIES TOWARDS TIE PERFORMANCE EVALUATION OF SOFTWARE FOR SOLVING
SYSTEMS OF NONLINEAR EQUATIONS .
. .
4 The success of algorithms for solving systems of nonlinear equations'-is, very dependent an the nature of the particular, sYstem being solved. This research
wps directed towards parametrizing nonlinearsystema of equations so that one . 4 could more effectively select which algorithm is best suited for a given'system.
The parameters investigated included tthe size of the system,'the goodness of the - initial approximation* the nonlinearity of the system, and the huMber of potentisl nonconvergent regions inherent in the system. Compdrative studies were conducted t 3D 11.
foesix different algorithms. These studies show'ed ae,the above patameeers did influence the efficiency and likelihood of suc of each of thi algorithms. Based on 'these comparisons, a user:oriented guidel e for the choice of algorithm relative to. these parameters was suggested, and a model for comparing algorithms,
' f #.solving systems of'nonlinearequations was developed.. -
In.addition to parametrilistion, two add.tional related topics-were persued. First, Tropert es fora measure of the degree of nonlinearity inherent, in a sy'tem of nonli ear equations were established andone such measure was teveloped and.tested. Secondly, a'special type.of nonconvergent region was discovered and describedLthen, appropriate steps for4dentifying and dealing with such regions were established.
.D.L. Kalmey, D. S. Kerr, L. J. White O ./
!MINIMUM -REDUNDANCY,VARIABLE LENGTH CODING
. j eie' This research is aimed, at obtaining efLicien.t solution techniliues for solving prefix encoding problems involving cost associated code alphabets,,..An efficient algorithm has been developed for finding a 'most economical encodin,forg equipro-- '. bable messages using arbitrary size alphabets. For the case ornon-equiprobable messages and a binary'aiPhabet,wo_techniques have been developed - -a selective enumeration technique that outperforklathe previously known techniques, and an algorithmic technique that is better suited for large problems. Based on.a preliminary study of the latter technique, ,1t is conjectured to obtain optimal solutions under general conditions. 'Development of a proof of this conjecture is
. - currently under study. - i 6- Ss. L. J. White, N. Santhanam *i --) c6"
SELF-4OLV,ING QUASIGROUPS' .. A Properties OTV poids satisfying the condition that ab=c'iMplies a=bc, have been investigated.: hey are necessarily quasigroups. The squares of all elements areboth two -s4edf identitieg for the elements and idempotents. They ,e--, exist for all n (n an integbr, the order of the quasigroup), and for all n, are at least as numerous as groups of .the same order. '1141th any group (G, ) we can introdtice the associated self-geving quasigroup (G,*) defined by A.* b= c if lis
andonlyif alp = c7-1.. The are self.-solling ivasigroups which are not -
associated with any group for alln 3. Only for n-a 2m, can-a.sele-solving .. quasigroup be group, and-it must be.a _direct product of cyclic groups of order9
. 2. It is believed that these quasigroups can be useful in, cryptography and daia
security. , .. Kr , - ,/ , - J. Rothstein - '51
SOME LP-NORM MAXAIN PROBLEMS:IIILGORITHMS AND APPLICATIONS'
This research examins a few closely related maxmin nonlinearprogramming problem -.that have appii tions in such diverse areas as location theory, infOr- x.
40 0 31
mation theory and pattern recognition.One investigation involves the location of a point in agiven convex polyhedron which maximizes the minimum Euclidean' diStance from a given set of m pOints in the polyhedron. ,This problem is also equivalent to finding the minimal radius of m hyperspheres centered at the given points such that aoh poineln.tHe polyhedron is "covered" by one or more df the hyperspheres. The'existence cif a finite candidate set solution for the problem is demonstrated. The computational growth of an algorithm which generates this finite candidate set solution is shown to be 0+1 where m and n are the cardinality of the set of given points and the dimensionality of the space respectively. However, efficient algorithms are suggested for the problem in the plane, and in three dimensions when theconvex polyhedron corresponds to the convex hUll of the given points. Several heuristics were employed in achieving this efficiency. The potential relevance of the abqve problem to pattern classi- ffcatton is shown. In particular, algorithms are suggested to geprate,nearest- neighbor.d6cision surface& for dnarbitrary number of diajoint piftpern sets in two and three dimensions-as an application -of the above maxmin problem.
L:J. White, B. Dasarathy
O
I
41 32
IX. SYSTEMS PROGRAMMING
THE ANIMA GRAPHICS PROGRAMMING LANGUAGE°
In conjunction with Profedaor C. Csuri, Department of Art and Director of the Computer Graphics Research Group; the graphics, programming language ANIMA is being developed. ,ANIMA users will be able to:
1. Treat pict re descriptions in a fashion analogous to the treatment' given other tructured data types, such as numeric array;
2. program in a device independent manner, in that a file and a graphic input device may be used interchangably to assist in, say, pictuie definition;(
-3. , have a compiler available for the language. Also, the intent is to
base the language implementation on a minicomputer. ' 4 ith the in4usioa of a data type specifically for the storage of pictures escriptions, and verbs in the language that are patternerfor efWtivemanip- lation of these data types, ANIMA becomes a true graphics prbgramminp language.
A: P. Lucido (Sponsor: National Science Foundations.)
40 r I
COMPILATION 0E SEQUENTIAL PROGRAMS FOR PARALLEL EXECUTION
4 , This research derives techniques for the practically optimalcompilation of a subclass of the class of acyclic programs for execution on.an idealized' . parallel computer. The computer, partially described by Hell4manin 19b6, has a single central processing unit controlling several arithmetic units. Each arithmetic uniecan perform any of the computer's arithmetic instrGtions in one time quantum. Completion of a design optimal for a given work load would require information on the number of arithmetic units required for minimal time execution of-programs representative of that work load, iormation.
. derivable from optimal compilation of Programs,. Even in the,abs cc of a "Hellerman computer" compilation of sequential programs for it could give insight into the deCreeof parallelism inherent in the programs, setting 'apart % array paralleliiM. . .-_
, Heileman offered two techniques for the compilation of single algebraic 'expressions: earliest stage assignment and latest stage assignment; preferring
the latter to minfmize simultaneously the 'limber of timed quanta and the number . of arithmetic units required for execution of the expressions. In this r4earch, an algorithm has been developed, called "balanced assignment", intermediate 4 between Hellsrman's techniques. Balanced assignment is proved to be optimum for single expressions and for maximal sequencesof branch free code, "basic
v 42 13
blocks;' "Prosser's 1959 defnitioni of dominance is employed to generalize this algorithm to-some branching programs. The notion of the "compound block" is further defined as the set of- program statements dominated bysome program statement. It lop ree programs these compound blocks are proved fo..be!tesiod like programming language blocks if and only if the.program hasa flowchart that is%a series-parallel network with two -way branches and junctions. This property is named "acyclic smooth% The property is independent of the pro- gramming language used, but it is possible to specify acyclic smooth programming languages; i.e., languages that can express only acyclic smooth programs.
The balanced assignment rule is generalized to acyclic smoothprograms, but ". tHe'optimality of the generalizationie proved only withina subclass of the clagi'"tf all permissible, arrangements of thepro am, a class the.class of linear level assignments. Generalizati f balanced assignment to looping programs appears possible once a suitable,definitiom-if cyclic smoothness has been formulated, a topic left to future research.
C. R. Foulk, O.C.* Juelich O
. THECOMPUTER EXPERIMENTATION FACILITY
A computer experimentation facility is'being adeembled that will support research into Areas such as computer'graphics, programming liftuages and their translators, opetating systems; and.comp ter architecture. (Currently available equipment intludes'an Applications Group, Inc., model AG-60 plasuro panel bas,ed
graphicsterminal, a touch pane for the AG-60, a sonic tablet, Hewlitt-Packard - 2115Aminiconputer, data commUcations equipment to allow long distance 2400 based rate communications between the AG-60/A-P.2115A and the DECIO of the CIS Depart- ment, a plotter, and more. The main idea is o provide an environmentthat-will foster and aid research into some of t led areas of computer science.
A: P...Lucidb, C. Kearns (Sponsor: French Fellowship of the Engineering Graphics, Department,;)
. 1 .
COMPUTER YSTEM FOR SECURE DATA PROCESSING
The ork investigates four log cal levels Of.computer,protection structure: systeOr tection, data communication ,and processing, protection of stored data and progr and user"identifftatien ,and authorization procedures. The complete secure co uter system is designed which is-general enough so that it enables all operations that present day computer sistems offer. The protection mech- anism is realized in hardware for better security and it is independent Of any internal/program, machine or data structure. The main problets in such a system (execution of prograds; data set manipulations, hardware modifications and security procedures). are-all given in detail..The system is evaluated in terms ofihardwake and 9perational overhead.,
,In additidn, cryptographic4techniques for computer use'are invesZigated, classified and compared. -The design of the generalized homophonic cipher ( system, its secrecy and efficiency are given too. Finally, necessary hardware 43
Ai 0
ho" 34,'
. e . modificationsare.all designed and their operations in the system are all.
described. .
"M.-T. Liu, S. MuftiC, Rothstein Y
CONCURRENCY Di REAL-TIME INFQRMATION SYSTEMS
. A, method 'of synchronizing RioCesses of a real-tiMe'information system is developed. Two requirements of the system are that if a query Q is submitted: at time Ti, all queries- ubmitted before. Ti. are not affected. by Q and all queries submitted after Ti are affected by Q. Many 'current systems satisfy these requirements by sophisticated"-scheduling algorithms and data base lockout mechanisms; our synchronliation method guarantees these requirements wishouf either of.these mechapisms. We belie also` identified they system design,iequire- merits that will allow the maximum amount of multiprotessing. 'Assuming there is only one process that.can change die-Value Of a variable, it isdesirable far variables to have common reference sets?It also has been shown that if, for each variable,,the only process that referencesa variable directly foliows the, process that changes that variable, concurrent processing.of all processes can occur.
H. S. Kodi
rr
CONTEXT PROTECTION 'IN DATA BASE SYSTEMS.
The goal of,thisreseaTch'is the development of a newtype protection mechanism known as the context protection meEhanism for 3 interactive data secure systems.
n all currently known protection mechanisms access control 7 decisions, are made a priorin the basis of access rules related to users, data and data base eratiosis.. In this research,an ,attempt will be made to include protecon niedhaniams'inwhl.ch access control decisions depend upon additional factors suchas 5ast'access hi'story and the content "df data units.
Plans far implementing the.new type of:mechanisms in an xperimental data base system will be cpnsidered.'
I w D. Cohen, D. K. lisido, M. T. Liu (Sp'onsor:. Office of Naval Research, .N00014475-C1573)
, .
COST-fFFECTIVE.ACCES'S CONTROL IN DATA BASE SYSTEMS
' In every.datalhse systet'thae.has been proposed,'an.unavoidable consequence hal. Uten the introdUctiOn of system performance .'overhead' attributed to the logical access controljmechanisme provided by the' data base system..This overheap.pessesses the'characteribiics that the cost of' provVing, adequate data security for each user of the system is directly proportional to the effectivenes-s and sophistiCation of , . 44
Aka. ./7=m117.
35
that security. A data secure system that can provide the user with his required level of security and still keep thecost of information' processing and transfer at an acceptable level is termed "cost-effective". The objective of this research is to conductan examination of the
. requirements for providing cost-effective data security in data base systems.
Q. .1 J. M. Hemmings; D. K. Hsiao (Sponsor: .Office of Naval Research. N00014-75-. C-0573)
*
DATA BASE ACCESS CONTROL IN THE PRESENCEOF CONTEXT DEPENDENT PROTECTION REQUIREMENTS '`
Although the dapability of theaccess control mechanisms 'to regulate field, record, and file:security has beenrecognized-as indispensible in advance data base systems, thereis the need ef more subtle-protection which cde shall callcontext dependent protecaon.
, Context dependent protection of data enablepus to'chavge-the accessibility of adata unit (field, record,or 'file) dynamicailywhen some other data units have been .accessed.,A context dependent protection requirement specifies how and under what conditionsuch a change is,to be made for a particular data unit. , I In this research,two basic types-of context dependent protection requirements will be identified and studied.. The firsttype of require- ment, called an access decreasing requirement, Will reduce the'actesai= bility of ad unit when the condition specified by it-is satisfied. The other of requirement, called an access' increasingrequirement, will make'a datunit more accessiblewhen the condition specified is satisfied.- Our research shows that, bothtypes ofrequirements can be studied/by meansof cer:tain built-in telationships among the data units involved. Therefore, the first goal of this research istckAnyestigate these'relationships and their. enforcement. Since the introduction of -context dependent protection will definitely havesoke,effectsion the access control problem as aWhole, the second goal isto studytheorecti:- A tally the overall _impact of context dependentprotection on protectin datin data base systems. A graph-theoretic ipproaCh will be used i directed graphs are good,candidates for representingsimple relations.. 14* C. J. Nee,D. K. Hsiao,, D. S. Kerr tf
DATA GENERATION FOR COMPUTER GRAPHICS
In this project, whiAl is performed-inconjunctron with e Computer Graphics Research Group on+campus (Prof. Charles,Caufi,Direc or), the problem of simplifying what'is'curreritlya chore and a drudgery, name the-generation 'of three-dimensional data specifically foruse in Comppter Ani tion,'is'being investigated. Some of the problem,consiraints*are: the use o a minicomputer --- e, 45.. 36 A scenario we are working towards can bf described as follows; The user sketches in free. hand (using, say, the sonic pen in its 2-D mode) two or three orthographic projections of the object he has in mind. The system constructs the 3-D version of the object, which'will necessarily be onfy apprOximately the object desired. This, however, invorVes not only the interpolation of the Object in threespace from two-dimensional information; but also the preproces- sing required in cleaning up the inaccuracies and inconsistencies associated with hand -drawn pictures. The user,Might also, for classes of polyhedral, Ojecta, use the'systemrs pr6jected capability to obtain, plane edges from free- ; . form lines drawn in orthographic projeAions. Notice that these capabilities involve formation hypotheses by the mactine and drawing upon stored know- ledge about different kinds of objects./ Once the approximate 3-1? version'is created, then the user might sculptthe)finished Product by. calling upon a variety' of 3-D warping and sculpting routines: / B. Chandrasekaran, R. Parent (Sponsor; National Science Foundation. Grant HCR,, 74-00768 A01 ) A,DATA SECURE COMPUTER ARCHITECTURE .,- ..A I ' . A new computerarchitecture' tosupport a very large, highly secure attribute- . i oriented data base management system (DBMS) has been developed. A secure DBMS 4 "must be designed from the outset with security in mind.The system's security mechanisms must'be an integral part of itsarchidcture.The complex software currently' necessary a secure DBMS is a primary'cause of unacceptable performance, and reliability in many systems. The architecturedevelopedhere overcomes this by providing unconventional hardware that inherently simplifies --ihe algorithms of a secure-DBMS The proposed architecture contains four major components: the dirpctory memory,. the intersector, the mass storage and a pair of interface propesiOs. The directory memory maintains ettiibutes of the data base, the intersector executes setmanipulation operations, the mass-storage- contains,the data 'base, and the interface processorg6control the otherkomponents. The diredtory memory, intersector and mass storage are implemented wit% segment associative men4ries. This new kind of memory does not ascribe to the extremes . of easy-to-use "16cation-addressed" memiries. Segment associative memn4ies of .. many different sizes and speeds-- including ones of sufficient capacity to contain , / a very large data base,are.wel/ within the capabilities of near-termtechnology. By using sqglent associative-memories-it is possible to build an inherently more .reliable anbetter performing-DBMS that can support a wide. range of contem- porary data management activities, \R. Baum, D. K. Hsiao (50onsor: OffiCe of Naval Research... N00014 -75 -C- 0573.) DESIGN ANDIMPLEMENTATION OF A DATA SECURE SYSTEM The.g al of this research Is to'designand implement an experimental data base, management.`system with Multi - level access control capabilities. The design work has been completed. ;The implementation work is being done on the DEC-10 Computer. Langl6ges:ffpr they4mplemeatation include both MACRO-10 and PRESTO, whichli's-a7tarm-bitstruCEUred FORTRAN that .was created for this effort.The three levels of loglcdlitc-cess control in the system, the files, records, 4, 46 Oak 1,,/ 37 # and record.fields, Aare each partitioned into those which theuser can access and those which the user cannot.access. The, system is based on some known research.results such as the Generalized File Structure, the ParallelAccess klgoriChM, and the Securlty Atomconcept. User queries to the astern are in the f2rm of BOolean combinations of keywords. The systent.also sfipports full update capabilities. When finished, be used as a test bed for other concepts in datalsecurity. N.Kaffen, L. Breenek T.Rodeheffer; 13. Schmaltz, D. K. Hsiao (Sponsor: Office of'Naval Research., N00014T-75-C-0573) DESIGN OF A DATA BASE MANAGEMENT SYSTEM FOR CONCURRENT PROCESSING 4 Ode have.described how a DBMS can be designed to be structured to allow concurrent processing of that system as it is Orocessing various queries.That is, we want to resolve queries b4i'multiproceiSing the components of the DBMS. It is assumedthat the DBMS will be operating in a real-time environment, which puts further constraints on the system. NormallY, in a concurrent processing environment, variables are shared between processes and the crittcal sections are monitored to solve the mutual exclusion problem. However, this'Aolution not*syfficient for, a DBMS that operates in a real-time environment.It has been shown why this solution is not sufficient. A model has been developed that'will be able to satisfy real-time constraints. , H. S. Koch . 1 bETERMINING THE STABILITY OF DISCRETE SIMULATION MODELS b When modellers write discrete simulation programs using SIMSCRIPT or GPSS they usually do not evaluate the stability-of 'the model,directly. This is _a very,difficult analytic task and the best they cam.do is to obtain an empirical estimate of the stability.-They execute the simulatiOn program many timeg, varying slightly the parameters,ofthe model dhd changing the seed of the pseudo-random number generator, determining how much the output is affected. This can'be a very expensive process, and it is not as informative ass many would like. .-- ---- We are attempting tQ develop a way to aMalytically determine the stability properties of such models.' Consider a differential equation (d.e.) and a numerical integration scheme for obtaining the approximate solution of its initial value problem. There are substantial theories dealing with the stability properties of the (i.e. (Lyapunov's for example) acid the nUiercal '1 stability, roundoff and turncation problems of the integration method.We'use these theories as follows: let a simulation modelling program be written in a normal form which.is so designed that one may conveniently view it as a numerical integration scheme.for some d.e., and from that ormal form program extract the the.'in question. Analyze the stabilitybehavior of the d.e. and the integration scheme and apply the results to the-al nation model. To facilitate the study we define a simple simul tionlan uage, SSL, which has tire two important properties that any discrete simul ( tion model can' 47 38 4' be expressed in the language, and that any program in the language can be easily transformed into thenormal form. D. J. MoorerP. Miller A.FAMILY OF PROTECTION LANGUAGES FOR SPECIFYING DATA BASE PROTECTION REQUIREMENTS In recent years,, the issue of security and privacy in large .--...,.. integrated data bases has attracted much attention.Along with the installation of access control systemg td protect stored information,' there is the need to develop a protection language for the users to, communicate to the system their security requirements. Due to increasing demand for more effective communication and sophisticated security requipments, thcscope of protection languages has evolved from simple commands that result in alimited range of features.to a language with sophisticated constructs; prbviding highly discriminating security ep control. ThiS research develops a set of protection language component parts (PLCP!s) that embodies access control features at many different levels of sophistication. The PLCP approach "comptrtmentalizes" the language constructs, allowing concise language descriptions wi h eJhasis on salient protection features. We show that the vari s features in the PLCP's can be selected as the basis of a family of roeectidn languages 4 for usekln different system environments. An interactive protection language for use in an attribute - based, data base is implemented. Expefience with the system reveals extensions for further,study, especially, in the area of enforcement dynamics. P r . D. H. Wong, D. K. Hsiao LANGUAGES FaRSPECUYING PROTECTION REQUIREMENTSIN DATA BASE SYSTEMS- -A SEMANTIC MODEL ' This work develops a model t8 provide a semantic base for constructs of 0 "I protection languages at many levels'of sophistication, accommodatipg a wide_ . range of protection,`policies.policies. ,. .c, 4 ) t The basic sets of-the model are presented_ and subsets of states are defided'hy.Boolean expressionS (conditions) of ariEhmetio relations bn the / resource values. 'History keeping is used to allowa sensitivity to the context of access history sdthat access decisions can be 'based on the occnrrence'or non-occurrence of previous operations. `Provision is made for the apedification of auxiliary program invocations to achieve additional.protection measures . before, during, and after,the access decision making process. The concept of . extended resources allows.a user to invoke procedures having greater access . - 4 / rights than dOes the user himself. , / . , ., .C-7,iP : . . . A family of protection languages areprdpose\Ias the key to'expreasion of tflicies andrules'of access by authorizers. Translation of protection languages 44, . . ., , '48. .. , le .. is . . ' ^ / r . 39 consists of lexical (syntactic) translation into entitiesof the model and semantic tranalation by the authorization and enforcement processes of--the model. Protection language constrdcts fall into four categories: definitional, operational, advanced protection services.,. and semantic parameters. The , deLnifional category is for specifying-definitions_ofusers, user groups,, resource units, and. conditions. The operational categoiywaYs who can do what operations, on what resources, and under what conditions. The advanced pro-: tection services contain protection constructs suchas history keeping, auxiliary program invocations, and extended resources. Semantic parameter specification provides the means to define protection policies at the-highest levels of sophistication. . / The work is now completed and has resulted/ ina Ph.l& di..ssertatio. H. Rex Hartson, David K. Hsiao (Sponsor: 'Office of Naval Research. N 00014-75- C-0573)/ A MODEL FOR A SECURE DATA BASE AND A FRAMEWORK FOR THE APPLICATION Of CRYPTOGRAPHIC TRANSFORMATIONS tN A DATA BASE SYSTEM In our research we introduce a top-down approachto data security. Starting with the general subject area'we show why the security problemsin data base systems should be discussed separately, from those in operatingsystems. 'We have defiltd'some basic concepts such as: protection specification, protection mechanism and protection binding. Using these concepts vIt analyzed the current '-/ models of data base system (mainlyethe CODASYL model) and pointedout some . serious disadvantages. related to security, in particular. tile, problem ofcentra- lization. We then developed a levels structured model whicir.incorporates in it many.6f the security models known today andYt haVe shown that with this model 'decentralization can be achieved if desixed. We develbped some notation for the levels model and using this notationwe were able to deal with the question of the application of cryptographic transformations ina data base. J More research has stilltobe done 9n the problem of how the cryptographic transformations are going to be.used. More research has to be done also on the. 'security engineering problem, the evaluation ofdifferent protections mechanisms and in particular the evaluation of cryptographic transformations. Z. Gudes,.H. S, Koch, .J. Rothstein, F. A. Stahl 19, STRUCTURED DATA BASE GENERATION. 1 As 'part of data base management system futctions,*.the generation' ofa 'Variety of new data badea is a large effort. In order to minimize the amount ,.-1-mPtieffort for each data 'base generation, we pyopase tostructure the process into four basic steps. 1a, . The first step.equiree a special-Turpose program that converts raw data . into records in the form of attribute-value pairs. During the conversion from raw data to attribute-value pairs, a standard statistical package is'used to gather information about the raw data,for the'data base administrator. Since the conversion must be tailored to the raw data format and is used onlyonce '40 , for each data\base generation, a one=timeprogram is needed. Subsequent steps can be standardized, for they depeid only on well defined inplit, formats. Once the data base adMinistrator'knows the required record organizationand file structure of the new data base, he can proceed with'the second step'by building record templates and specifying the physical sizes, of the attribute dictionary eegmet, the atom directory segment, and the Storage cell. This:step is' dote inter\C etively in order to aid the data base administratorto review his work.' Step.three involves converting re,carsIs froth groups.of attribute-value pairs'to the system's physical organization.Thfsprocess involves the use of record . templates to validate attributes and values and toiverify the hierachtical Organization of the records. The final stepis to insert the records (in. system's physical' organization) intoa cell of data base stoiage.. This process includes updating attribute dictionaries, and atom direttofies to A reflect the presence of new records. R. nablein, D. K. Hsiao (Sponsor: Office of Naval Research. N00014-75-C- 0573 STRUCTURED FORTRAN C' 3. 1 PREST4' is a pro ram on the DECSxstem710 that takes an input file containing source code writtein the PREST41anguage,,and generates an outpiit'Ille in "ft sten,- = a C FOR As an option, it all& generates a listing Basically, . the la gagels'an extension of FORAN withchangesmade to facilitate the writing of str Utured programs. Thus name PREST4 - PREprocessor for- ' STructuied FOR(4)TRAN, isocoined. , ,: . ° 0 J . . ! PRnc.re2F0g112eS, fourtypes of stfUcture s :_anIF -THEN-ELSE'structur4forr specif/ing,ditilimalfeieeution; aDO -END structureforgrouping.a.series.4" statementsintwa'unlitand-for.specifygOga apeti4tioo criteriaa READ. structure 40'14eading dat4 an tOrhandlinethe end of file by specifying an assignment, dt!i subrouVinecall;7and a simplestructure which consists of a ,f, ' : . single stateAtt. ., -.10 .ecz , 1 .. , ° f S., .Ake q 6,, , y Besidespgramstatements, Ice dgey may-eupplypreprogessor control 14 statements in Ilie PREST4source e. ,TheSe contrOlitatements tausg PREST4 , , to perform some specific'action (e.g., ejectapage in.the.listing) and do not cause themselves to be generated in the rtpUt file. Elsvery-control statement ' hat a percent sign (%) in Column 1. 0 . ° o PREST4 can' be told 'to TRACE a ,program, in whiefi case it places a subroutine call to the tracing package (passinglthelline number as an argument) into the output file before examining each statement of the,program. The tracing pack:- apse records at run,time the last 500 ca nd oan list the line number's upon request. Thus In the'eliefit of4an error theuser-cam deterfnine,whiCh statements have been recently executed. T. Rodeheffer, D. K. Hsiao (Spo6or: 0 Tice of gs;a1 Research.N00014-75-C- 0573) - O 41 X. COMPUTER ARCHITECTURE AND NETWORKS 411 e., . /'. - . , THE ARCHTCTURE OF A GRAMMAR-PROGRAMMABLE'HIGH-LEVELANGUAGE MACHINE . , , This research is concerned with system destgn of a grammar-programma-1d''j high-lever language machine. The term-grammat-program is applied to the context of Aigh-Level Language Machine sand defines an intermediate revel be- tWeen the basic hardware/firmware functions ofa computer system and its-9oft7, ware language translatoisiThrough,grammar-programs, .the syntax and semantics olvarious.programming languageseCan b4 spe'cified to the Grammar-Programmable Machine which uses these specifications to process directly the users.' high- . . , . 1pvel(lafiguage programs. - . , . Taking advantage of its intermediate pOsition between software compilation and hardware interpretation of high-level.language's, the grammar-programmable machine emphasizes the str ng pointof interpreteri to overcome and even, enhance the best feature of compilers, anekvice-versa.. For this reason, it enables the realizatio of the potential which neither a compiler nor ari'inter- preter can achieve invidually. T. Liu, S.-f Fou ler THE DESIGN OF THE DISTRIBUTED LOOP COMPUTER NETWORle This research isconcermed-with-the design 'of the Distributed Loop 4" ° Computer Network (DLCN) using a new _transmission mechanism proposed in previous work. ahis new transtission mechanislpis much more efficient.and sophisticated than those in current uselnd has the folloWing advantages: concurrent transmission of variable-length, messages is allowed, 2) nearly immediate loop :------access is guaranteed, regardless of load, and,3) automatic regulation of message traffic is provided; all accomplished in ECcompletely.distributed network. 6 : The design ruiremente properties arie hardware implementation, using' .--.0 standard 16gic crsuits,of the loop interface are investigated. Following that, several novel solutions to message types and Irmats, message licknow- )! ledgement, error detection and retransmission,- lost message detection and 1 1 'removal, prioritx message insertion'and lockout prevention are added"te the 1 interface design.. In addition, software system reqfffrements for the Distributed Loop Operating System (DLOS) are'studied, including network protocol, interprOciAs , communication, resource allocation and manageMent, and*eri-Or detection and recovery. -4 , I . ..-, . J .., ,. s--I The main goal of DLCN is to develop an'integrated.hardware/software4 cOmmunication system with distributed control that provides efficient, felexihle, reliable and fail-:soft service at low cost in the fate of constantly changing <. 5 1 V 42 tr uset demands.-- M. T. C.Reames A NOVEL MODEL FORA MIXED'VOICE/DATARANSMISSION, SYSTEM FORCUTER . --\ .. COMMUNICATION . . 0 . - . Thid research is concerned with the development of a novel' model for a mixed model for a mixed voice /data transmission system for computer communication v)hich opeiates by,embedding digital data signals into a. set of voice channels during the silent periods of voice conversation. Several previous works are 'briefly reviewed. ''Mini /micro computer technology is used in the control of the digital signal embedding procedures. Design consideritions such as interfacing I to the existing telephone channels, voice detection, voice/data-recognition and separation,,routing strategies,modems, message protocols, data security,tetc. are discussed in sufficient detail. It is expected that' the propose' model .will not'only lOwer communication costs for computer networking but wil'also ',improve the efficiency of Channel resource utilization. Q.M. T. Liu, J. T. Wang OPTIMAL PRbCESS,ALLOCATION IN DISTRIBUED HETEROGENEOUS COMPUTER NETWORKS 0 . . . . , , , - / r This research is 'concerned wi'th optimal process allocation indistrioblited heterogeneous computer networks so as to minimize hardware/softwa're redundancy. _ A mathematical model is proposed for analysis using integer and dynamic pro- * . gramming which allows different service rates for each propess.A simulation .. mgdel/islto be implemented to verify `the results. It is expected that the goal* of minimiz n eduRdancy can.not,only increase everall system efficiency, but , can also se f-tune the systeni according to the constantly xhanging request . [ . . pattern.,/ , N:0, e Cheng M. T. Liu, T. :T. 4 ' ( 40/ aku A PREDICTIVE. ASPONSE TIME AbglITOk FOR COMPUTER NETWORKS . . , . - , to assist users of a heterogeneous network ,,,,,,.., A predictive responbe time , of computers was designed. The netwqrk user is able to query this dynamically updated* software moduli to obt'a'in current i*formation relating to the busyness of tine sharing systems qn the network and the amount of_time required on these systems to run var.ioug classep Of computing. applications. I 4,4.&.. ' 0 ' '...... A, a' The imrlementa111on of'the module is proposed-as an extension of a Network - Access Machine 04AM)developed-by the Computer'Networking Section, at the National - -Bureau of Standards. The NAM is a minicomputer which assists the network user in obtaining network, services. Aeraponse time module WU). acq4pt from users the designation' of time. staring systems on which they wish toy do work and will , return-either a narrative description 'of the general response time characteristics .... ' 52 Olo 43 of those'systems or comparatiye lists of response times for running short and- long FORTRAN, COBOL.,..and BASIC jobs on those systems. . 4k- cc The feasibility of the dq,6ign a'd implementation planwas verified by 'response time measurement experimes conducted on a DEC System-10. A measure - of system busyness called-"PercenCPU Idle Time" was foundtp,_ bg_a satisfactorily accurate parameter on which to be reponse time predictions for the FORTRAN, COBOL and BASIC jobs mentioned above. S. Mamrak (Sponsor: Ty National Bureau of Standards) / .4 0- 4 XI; JOINT PROGRAMS THE STATUS OF WOMEN AND MINORITIES IN ACADEMIC COMPUTER SCIENCE' A survey of women and minority students and faculty in computerscience during' the years 1971 to 1975 was conducted.Analysp of the data indicated, that effective affirmative action programs for recruitmelt.into graduate degree programs are needed to enlarge the number of women and minorities qualified for I- later employment in computeescience. Alsd, po'ssib.le diAcrimination in employ- ment of women and minority' raduate students was revealed. S. Mamrak, R. G. Mentanelli (Spopsor: The-University of Illinclisat Urbana- e, Champaign). . . rt r 9 ft> L. . ' 4 e 4 r g t V . 54 IA* 45 XII. COMPUTATION THEORY O THE COMPUTATIONAL COMPLEXITY OF ORACULAR ENUMERATION PROCEDURES We study enumeration procedures in whichan oracle is used which supplies a stream ofinforkation at no cost., The model develope:gsuitabayor a study of the abstract complexity proPertigs of the sftu Er' .--. , , # ,,, The questions asked concern the existence of oracularsets which are very helpful, and output sets which can be helped byno recursively enumerable, oracle. We investigate the, influence of the order of thdstream of oracular ,,-t 'slate on its,usefullaess. . ../ . 7 . -rd' We attempt to study questions analogous to those answered by NancyLynch in the casg_of recursive Oracles used by oracular Turing Machines. Using techniques developed by Paul Young," Albertleyer and Pat Fischer we hive shown that the effectiveness of an dhumeration oraclecan depend very ,heavily upon the order-in which the oracular'inforniatlonis presented. We 'conSecture that. theme are interesting situationsin which the o* raou-lar information is so redundantly coded that it is usefulno matter what the order of presentation might be. We dfscuss the applirability of this work to theorem proving. D. J. Moore 1 I THE COMPUTATIONAL COMPLEXITY OF RECURSIVELY ENUMERABLE TOTAL ORDERS. In:the study of abstract,complexity theorymany disturbing pathological ) characteristic of recursive functions have been discovered. Here we investigate the hypothesis that sets wiiichare graphs of total orders, and thus hig4ly -:structured, may posses6 les§ severely patholqgical.,properties:We study in 'particular thecomput&ional complexity of the problemof recognizing membership . in the graphs of recursively enumerable (r.e) totalorders. '-We.show that, independently of the ordertype and underlying field, there ,arel,11) r.e: total orders for whichit is arbitrarily difficult to recognize infinitely many elements, and (2)r.e. -total orders Nil- which any membership recognition procedure 'can be spedup,by anlirbitrary amount oninfinite y' .-- _,manyiarguments. Also"we show that whether or not there are r.e. total dors is 'possessing these properties,on all butfinitely many arguments fs strongly dependent-on order type. Total orders' With a-least or greatest element possess the Infinitely oftenbut not the almost everywhere complexity properties;- however 'those total orders with no leastor greatest element possess, the almost A ., . 4 6 , everywhere complex ty Troperties.' The proofs that the infinitely often compleixity,proper_ies/ hold for order type w make use of finite injury priority arguments. D. J. Moore PARALLEL RECOGNITION OF FORMAL LANGUAGES , We have generalized cellular automata and iterated switching networks to 'what we call bus automata (BA). These can switch Trogrammably variable commun- ication paths between automata (cells). This permits speed-up'in computation by working on different parts of e problem in parallel at different locations. Imerird,ate forma/ languages (IML) are defined as those accepted by BA in a finite number of state-change intervaf independent.pf length of input,atring. IIL includeS all regular languages'and important classes of CF and CS languages also.,, Closure_properties of IML languages were investigated. SeVeral relationships were established to previously investigated language systems including the,"Chamsky Hierarchyvarious parallel acceptance language families previously studied, and the polynomial time complexity families PTIME and NPTIME. A hierarchy of languages amenable to parallel 'acceptance emerges. J. Rothstein, M. Moshell THE STRUCTURE- OF THE SET-' OF` FEASIBLY COMPUTABLE FUNCTIONS 'is ..... Functions whose values can be calculated in a time which is a plomomial function of the length of the input are cons,idered feasibly computable: We investigate the structure of the set of feasibly computablefunctions in . , . several wayS.., . , , ) . __4 ID We define a polynomial hierarchy analogous to the arithmetic hierarchy . of recursion theory. This notion has been.linked to the concept of oracular polynomial computations. We use both characterizations in lookingfor patpral problems which lie a* th.various levels of the hierarchy. We define a polynomial T-prediCate, and analOgies of Rice's Theorems., recursiontheorems, creative and simple sets. Any successful developments of this nature can help in attempts to solve the problem of whether P.!NP, D. J: Moore, W. Veggett y TOWARD AN ARITHMETIC FORiCELPLAR AUTOMATA AND. PARALLEL-COMPUTATION The base 2 number system suggested bygroupoid patterns (which was reported, on last year).has been further investigated and found to have properties at once fascinating, frustrating, and encouraging. It is frustrating in the sense that the siMp.iestArithmetic,operations can-so far only be done in a complicated 56 f 47 manner, fascinating in the sense that many complicated combinatorial relationships can be dealt with'simply, and encouraging because the operations exaiinedto date can be done simultaneously, in_large numbers, and.withoutmutual inter- ference, by cellular automata. An earlier generalizatioW of ase-2 . to an infinitude of integer represent- tations using 0 and 1 also grew out of the groupoid research. They correspond, in the limit of large numbers, tomu tiplication by afilted number between 1 and 2 when the number of digits is reased by one. It appears.likely that all of these systeis will have count art "parallel" systems in the same sense' that .ordinary base 2 has its cdunte art in the system of the previous para- graph. . I J. Rothstein TURING UNIVERSALITY AND PARALLELISM OF GROUPOID STRING ALGORITHMS t The chief"theorem to emerge during the pastyear describes a general speed- up of Turing machine 'calculations by an I =D, cellular automaton.It builds on the groupoid ' arent_string" parallel computation earlier reported. We simulate . a given Tur.4.n chine, augmented.by the parallelism capability of the I-D CA. Its cells control switching paths running.parallelto the CA; this device is, called a bus automaton (BA). The theorem is that a bus automaton , can perform exactly thesame computatiodas a Turing machine in a number Of state-change intervals exceeding the number of tape reversals of theoriginal, Turing machine by unity. Furthermore, if the distance between reversal's is bounded forthe class of computatLons that Turing machine does, thenthe. entire computation can be done in one interval.* J. Rothstein V APPENDIX A' COMPUTER AND INFORMATION .SCIENCE COURSE LISTING BY NUMBER AND TITLE 100 Computl n Society 549 Numerical Analysis for High School Teachers 201Elementary, Digital Computer 1 Programming *550 Introduction to Information Stbrage, and Retrieval 211 Com puter Data Processing I" 555 Survey of Programming Larlm Iges 212 Computer Data Processing II 594Group Studies (Discontinued 7/75 221'Programming and Algorithms I 610 vrjinciples of. Man-Machine- 222 PrograMming and Algorithms II interacition 294Group,Stadies' 640 Numerical Analysis 294/223 Tntroductioil to Computer 641 Computer Systems Programm'ing I Systems 4 642NumeticAl, Linear _Algebra, 311 Introduction-to File Design and Analysis 643Linear,bptimization Techniques in Imfdi-Mation Processing . 411Design of On-Line Systems g 644Advanced Computer Programming 422 TopiCs in Computing 'for 7 a Engineers 652Modelie,g of Information Systems 494 Group Studies,(Discontinued 7/75) 675 Digital Computer Oganitatiori 505Fundamental Concepts of'Computer 676, Minicomputer Evaluation and and InformationN. Science Selection 509 Survey of Computer &Information 67 -7 Computer Networs Science .for High School Teachers 680Data Structues 541 Survey tIoNdmerivf Methods 693 Individual Studies 542 Intfoduction to Computing in the Humanities 694 Group Studies 543Intermediate Digital Computer 705 Mathematical Foundation of Programming Computer and Information Science r. . 548Digital Computer Programming-for .706 -1.eformntion Theory in Behavioral Aigh School Teachers 4 Science. .. 5© 49 . 712Wan-Machine Interface'- 780 File Structures . i720 Introductio to.Linguistic ...... 781 Aspects of ComputerGraphiCa--:------ Analysis Systems , 726 Theory of Finite Automata 788 InterMediate Studies in Computer & Information Science 727 Turing Machines and Computability. k., 788.01 ThebrYof InfoiMation 728Topics in Theory of Computing 788.02 InformatiOnStorage & Retrieval 730Basic Concepts in Arilficial Intelligence 788.03 Theoryol Automata 735 Statistical MathOds inPattern 788.04 Artifi,Cial Intelligence Recdgqgion 118.05 Pattern Recognition 740Computer Systems PrOgramming II r 78a.06,Computer Systems Programming I 741. Comparative Operating SyStems 788.06A.OS-MUT 745 Numerical, Solution of Ordinary 788.07 Programming Languages differential Equations . 788.08 Computer Organization 746Advanced Numerical Analysis 788.09 Numerical Analysis '750 'Model-xi:Methods of Information Storage & Retrieval 788.10 flan-Mathine Interaction 751;Fundamentals of Document-' 788.11 Fornal .Languages Handling Information SyStems 188.12 Management:Information Systems 752 Techniques for-Sifnulation of .4 -Informat-ionSyhtens 788.13 Biological Inprmation -Processing) 753 Theory of Indexing 788.14 Socio- Psychological Aspects of ; Informhtion Processing 754 Language ProcessingNr Infor- mation.Storage and Retrieval 793 Individual'StudieS, 755 Programming Languages 794 croup Studies (Discontinued 7/751 )) 756Compiler Dehign & Implementation 797 Interdepartmental Seminar" `"1 -760 Sele Sed Topi4p in the Mathe- 805. InformationThegry in Physical mati s of Information Handling Sciende -T765--Theo y of Management Information 806 Cellular Automata & Models of 'Syst ms Complex Systems., A . .775 'Adv ced Computer Organization 812 Computer & Information Science ReseWrch Methods 59 50 820 Computational Linguistics 888.13 Biological Information Processing 835 Special Topics in Pattern . 888.14 Socio-Psychological Aspects of Recognition Information Processing 846 Operating System ImpremehtatiOn 889 Advanced Seminar in Computer & `Information Science 845Numerical S tatioa_of Partial , Differential Equations 899 Interdepartmental eminar .6.- 850 Theory of Information Retrieval I 399 Resea'rcA . 851 Theory of ;nformation Retrieval II 852 Design and Analysis of Information Systems Simulstiqns 855 Formal Theory of Programming Languages 865 Seminar.on Sdcio-Psychological Aspects Of the Information Sciences 880Advanced.TheOry of Computability '888 Advahcedtudies in Computer & ,Information Science. . 888.01 Theory of Information 888.02, Iriformation Storage & Retrieval el 888.03 Thepry of Automata '888.04 Artificial Intelligence 888.05 Pattern Recognition Net 888.06 Computer Systems Programming 888.07 Progamming Languages a88.08', Computer Organization 4 888.09 Numerical Analysis '1 888.10, Man-Machine Interaction .888.11 Formal Ldnguages 888.12, Management Information Systems,; .0 60 51 APPENDIX B COMPUTER AND INFORMATION SCIENCE.FACULTY Marshall C. Yovits, Ph.D.,. (Yale University).. Professof and Chairman Of-Department,of Computer and InformationScience and PrOfessor of- Electrical Engineering. L 'Director, C.F.S..Research Center. Information systems, theory of the Plow of information,self-organizing systems. 4 Thanko Bojanic, P1.D., (Mathethatical Instituteof the Serbian Academy of Science). Professor of Computer and Information Scienceand Professor of.Mathematios. Mathematical analysis, theoryof approximation. Charles A. Csuri, M.A., (The Ohio State University) Professor of Computer and Information Science and Professor ofArt. Advancement of computer graphics technology in software and hardware-." (language algorithMs, data generationor inputs)', use of computer technology in telecommunicatiOins. Richard I. Hang, M.S., (The Ohio State University). Professor of Computer and Information Science'and Professor of Engineering Graphics. CoMputer graphics, engineering application of computers. Clyde H. Kearns, M.S., (The Ohio State University). Professor of Computer and Information Science and Professor of Engineering Graphics. Computer graphics, engineering application of computers: Robert B. McGhee, Ph.Di., (university of Southern 'Oalifornia). rOfessor of Computer and Information SciencF and Professor of Electrical it .Engineering. Control theory, switching theory, lOgical design. Harold B. Pepinsky, Ph.D., (University of Minnerioia); Professor of Cdmputer and Information Science and Professor of Psychology. 40 inical and socio- cultural psychology. r- Roy.FReeves, Ph.D., (Iowa State University). . Professor of Computer.and Information Science and Professor of Mathematics. Director,'Instruction and Research Computer renter. Numerical andyais and programming. .). Jerome Rothstein; A.M., (Columbia University). Professor of Computer and Information Science' and Professor of Biophysics% d. Informational problems in science, methodology, biocyberneiics. Charles Seltzer, Ph.D.,,(Brown University).. Professor of Computer and Information Sciehde and Professor of Mathematic Coding theory, numerical analysis, automata theory. ) Hugh.Atkinson, M.A., University of Chicago). Associate Professor of Library AdministratiOn and Director of Libraries. Library on-line automation, data processing for libraries, technical serviceprocessing models. Kenneth Breeding, Ph.D., (University, pf Illinois). Associate Professor of Computer'And Information Science and Associate Professor of Electrical Engineeting. Computer organization and switching theory. , 61 $ . 52 . H. William Buttelmann., Ph.D., (University of North Carolina). +Associate ProfessorOf Computer and Informati9n Science -(Summer 1975). Automata theory, computer, architecture and programming languages. Balakrishnan Chandrasekaran,.yPh.D., (University of Pennsylva9p). 44,, Associate Professor-of-Computer and Information Science.: Pattern recognition and artificial intelligence, learning automata theory, finite memory decision . theory and game thedry. Ronald L. Ernst, Ph.D., (University of Wisconsin). Associate Professor pf Computer and Information Science and Associate Professor of Psychology. Human performance theory and engineering, complex information processing and systems evaluation. . . , . .4 Clinton R: Feulk, Ph.D'., (University of Illinois). *a C-71 Associate Professor of Computer and Information Science. Programming .: languages, systems programming, programming heuristics. ) [ David K. Hsiao, Ph.D, (Universitrof Pennsylvania). - Associate Professor of Computer'and Information Science. Systems programming, information storage and retrieval systems, file systems, data base manage- ., ment.systems, access control and privaCy protection of data, data definition language and processor,'systet architectures.' -- , -Douglas S. Kerr, Ph.D., (Purdue Uni,versity)-. .- . ,, AssoCiate Professor of CoMputer and_Information Science. Numerical and sis and programming. . ,Ming-Tsan Liu; Ph.D., (University of Pennsylvania). - Associate Professor bf Comprer and InforTation Science. CompUter organi- I zation, switching and-automata t eory, mathematical programming, comput4r architecture; pseudo-Boolean pro ramming, 'threshold logic. i Antony E. Petrarca, Ph.D., (University of New Hampshire). Associate Professor ,of Comput r and Information Science. Autoildtic indexing, chemical str,uctural inf ion processing, automated search systems; other! aspects of information storage and retrieval. 1 1 James B. Randels, Ph.D.,'(The Ohio State University).( Associate Professor,of Computer and Infofmation Science and'Assistant Computer operating systems i Director, Learning Resources Computer Center. 1 aa'utilities telecommunications applications, subroutine libraries, P . v programmlfig languages. r . * '1 k James E. Rush, Ph.D., (University of Missouri);'` '7 3 Adjunct Associate Professor of Computer and InfOrMatidn SCienCe. Ddexing theory, automated language processing, organization,of information, and , parallelpallel processing. 1 % I 4 Ceiianna I. Taylor, EtS.I..§.,'(Craduate School of LibraryScience,lase-WeRtern *Reserve University)i . Senior ResearChAssociate and Associate Professor of Library Administration.t Information dissemination and utilization systems, information centers, library systems an& management.. , '. . . Lee. J. White, ph.pj, (University of'Aichigan): ' '''k Assdciate Professor of Computer and tnformation Science and Associate' Professor of Electrical Engineering. 'Mathematical programming,., data. structures, organization of information. , i ' ,, - 62,., ' kJ. 0 53 Ronald L.Wigington, ph.D., ('UniVersity of Kansas). Adjunct-Associate Professor of Computer and Information Science and Director of R. & D., Chemical Abstracts -Service. Computef-system design. . " . a Thomas G. beLueis, Ph.D., (Purdue University) . Assistant Professor of Computer and InformationScience. Design and evaluation of infOrmationsystems, syste programming. Donald L. Kalmey, Ph.D., (The Ohio StateUniversity). Appointed Summer 1975. Assistant Professor of Computer andInformation ScienCe. Numerical analPis, computer architecture and organization,programming. , ) " Harvey 'S. Koch; Ph..D., (Pennsylvania State Univeisity).. . Assistant Professor of Computer and InformiltibnScience. Data definition language, data base management,.programminglanguages and compiler design. Frederick S. Koehl, Ph.D., (The OhioState Univer's'ity); Adjunct Assistant' Professor of Computer and Information Science and Math., Analyst, Instruction, and Research-ComputerCenter."' Sortingttechniques, topological groups, compiler design. I 4 Anthony'''. Lucido, Ph.D., (Iowa StaEpUniversity). Assistant Professor of Computer AndInformation Science. Computer architec- ture,:compiler design,,,-interactivecomputer- graphic's. Sandra Mamrak, Ph.D., (University of Illinois). Appointed' Autumn 1975. Assistant Professor of Computer andInformation Sci-ence. Petformance 'evaluation, computer networks,systemp"programming. ,,Robert F. Mathis, Ph.D., (The Ohio'State University). Assistant P;ofeasor of Computer andInformation Science and Assistant Dean and Secretary of the Graduate School. Programming languages, numerics] analygia. Daniel J. Moore, Ph.D.; (University of Kansas)'. Assistant Professor of Computer an#Information Science.Complexity.theory, recursion theory, artificial' intelligence. . ) i Laurence L. Rose, Ph.D., (Pennsylvania State. University). AppointedAutualn 1975. Assistant Professor.of Computer andInformation Science. Information ) retrieval, simulation, data structures, progi-amMingJlanguages Norman K. Sonc6eimar, Ph.D., (Universityof Wisconsin). ., v Assistant Professor of Computer and InformationScience. Natural language processing, artificial intelligence,information stprpe and retrieval.. Frederick A: 'Stahl, Ph.D., (University ofIllinois). Assistant Professof of Computer and InformationScience. Computational Ilecuftyj,crypt graphy, information .reviev computers. in the human' tiea, and nthe lawartificial intelligence. Stuart H. Zweben:, (Purdue University). Assistant Pro sor of CompUter and Information Science. Programming languages, compilers, data structures, operatingsystems. Ernest Staveley, B.S., (U.S. Naval Postgraduate School); Administrative Assistant and Assistant Di'recto'r,C.I.S. Research Center 63 54 '' APPENDIX C' CQMPUTER AND INFORMATION SCIENCE StMINAR SERIES Octolzer,a0, 1974 "Real-time Computer Animation," Cliarles A. Csuri,' Profedsor of ,, Art and Director, Computer Graphfcs Research Group, The Ohio State Univer- - qrty% October 17, L974"Banyan -Trees and Live Research," Jerome Rothstein, Professor of Computer and Infoiahon Science, "The Ohio State University: .% . I October 24,'1974"A HeuristicAlgorithm for .the Minimum Set Cover Problem Using Plausibility Ordered Search Neighborhoods," Michael E. Doherty, Ph.D. Candidate', Computer and InforMatton Science, The Ohio State University. . ,October 31, 1974"An Overview of DebuggiWg Tools and Techniques," Robert F. Mathis, *Assistant Professor of Computer and, Information Science, The Ohio State University. Alp November 7,T974 "Microprocessors - andMicroto*uters, Fred A. Hatfield; President, Computer Data Systems.- 47. November 14; 1974"What is Industrial Computer Science -One Man's View," Phillip S. Dauber, IBM Thomas J. Watson Research. Center. November 21, 1974"A Model for Data Secure Systems,"Edwin J. McCauley, Ph.D. Candidate, Computer and Infoimation Science, TheOhio State University. December 5 1974i, "Interpretive Structural Modeling,"John Warfield, Battelle, Columbus," Ohio. January 7, 175, "comparative Response Times of Time-Sharing Systems on the ARRA NetwOrk,"'Sandra'Mamrak,University 'of,llinois. January 9, 1975, "Where Did All the Computer PowerGar Herbert. R. J. Grosch,, Computerworld. . . . .4 January 16,' 1975"Emulation of Computer Networks by Microprogrammable Micro-. computerSY! David Cohen, Ph.p. Cand)ate, Computer and InfOrmation Science, *The-6hio State University. A ' . . January 23, 1975"ProtectionStructures inLanguages and Systems," Anita Jones, Assistant.PtofiessorComputer Science, Carnegie- Mellon University. *7 r . 1. 1 1 lk January 28,1975 "The.Proper Role of: 'AletentiOn'in Programming Languages," HowardRosis Hale, Jr., University of Delaware, Newark, Delaware, .0 January Yo,1975 "Some Thoughts On the Care andFeeding of Research Sponsors Foundation ' MclvlerW. Woody, K4sociate Direct.o? for Development,'Research The Qhio State University.,, 64 55 /'- February 6'; 19j5 ,.."Parformanee...tieasurq..s ofDerived Search Keys, *Edward T. O'Neill, Assistant Dean, hool of Information and Librag'Studied,State ... Universityof New'York .. .,, ffalo. , February 13, 1975"Careers' Romputing," Robert Porter, Jr.,Corporate.Collegd Recruiting, IBM Corporationl.,end DavidKnojt, Programming Manager, Office Products Equipment, IBMTCoriioration. , , - 'i, , February 20, 175 "Automatic Prdgramming: Inference of Program froM Memory Snapshots of,Sample Calculations," FredPetry, Ph.D; Candidate, Computer and Information-Science, The Ohio'StaSe University. February 21, 1975 "Organization of Stored-ProgramControlled Switching Systems-°- An Overview," Santanu'Das, Head,Maintenance and Diagnostics Research Group;' Paul H. Henson Research Center,North Electric Company. March 6, 1975 "Adaptive Systems Modeling,"Daniel Howland,,Professor, 'Manage- , lent Science, The Ohio State University. , h 10, 1975 "Automatic Keywords Selectionfor Reference Retrieval Systems," an -Shu Hubert Yang, Ph.D. Candidate inComputer Science, Cornell Univer- sity. March 121975 "SERAPS: A Self-Repairing Automatic Programming System," Michael J.'Ktssler, Instructor of Computer Science,IState.Universityof New York at Buffalo, as March 13, 1975 "Computer SChedulingStrategies and Their EffeCts,'4 Walter, Doherty, Manager of ComputerUsage Measurement . and Evaluation, IBM 'Corpora- tion. 0 ' 4 'I March .31, 4.975 "Automatic File Management Heuristics,"Lawrence L. Rose, State- Univevicyof'New York at the BinghamtonUniversity Center. 4 ' April 9, 1975"The Design of Programs for AsynchronoulMUltiprocessorsi" Philip'H.Mason, Carnegie-Mellon University. cri , , Q 1 April 10, 1975 "Teacher' Control in Computer AssistqdInstruction," Peter Calihgaert,d,:, ,Professorof Computer Science, University of NorthCarolina. . 4 ) 0 April 14, 1975' ,l'AilInteractive, AnalysisSystem for Execution -dime Errors," Alan M.'Davis, Ph.D. Department ofComputerScience; University of'Illinois . at Urbana-Champaign. April 16, 1975"Graphically-Enhanced Data Base Management.System Design," Wayne D. Domini*, Vogelback Computing Cdat'er and Department of Computer Science NorthwesternUniversity. .,, ' , 4 April 17, 1975 "Parallel Recognition of formalLanguages by Cellular Automats" J. MichaelKOshell, Ph.D. Candidate, 'Computer and InformationScience,-The Ohia,StateUniversity. ( 65 .01 56 4isw April 24, 1975 "Real-Time Procesi Control--The Programmer's Cloud Nine," James P. Shaffer, Senior Analyst Programmer, Industrial Nucleonics.. " May 1, 1 "Compilation of Sequential Programs for Parallel Execution," Otto C. Juelich, Ph.D. Candidate C,ompgter and Information Science, The Ohio State University. "High Capacity Optical Storagg'Systems," Carl Verber, Research . May 8, 1875 Leader, Solid StateandOptical Sciences Section Battelle Memorial Institute.' May15, 1975"Data Security a Privacy," Robert H. Courtney, Jr., Manager, Data Security ma, Privacy,-IBM Corporation. r 1 May 22, 1975"The Ohio &liege Ltbrary'petiter' Network;" Larry L. Learn, Director; . Computer Facilitiesyivision,:TheOhfo College. Library Center. May 29, 1975 "Extended Syntax= Dir4ctedTranslation of Programming Lan. uages," Arthur B. Pyster, P .b. Candidate,Computer ant Information Scieae, The Ohio State University. 1 r ) 7. ' 4 .10 *NI t. 6 5.7 6 4 1117 Yr APPENDIX D "RELATEDACTIVITIES OF THE STAFF OF. =MUTER AND INFORMATION SCIENCE 'RESEARCH CENTER J. G. Abilock presented an invited paper entitled "A Semidtic Framework for Information Science Leadingto the Development of a Quantitative Measure Information" (Co- author: M. C. Yovital\at the 37th American Society for Information Science Annual Meeting, Atlanta, October 15. Titepaper appears 4 in Information Utilities; Proceedings of the 37th ASIS AnnualMeeting.; Vol. 11. J.. A. Aitken presented a paper entitled "PerformanceEvaluation of Data Base Management Sxstems4Co-author: T. G. DeLutis) at the Computer) Science Confeience in Washington, D.C.,` February 18-20,1975. 5. J. 'A. Aitken presented a Paper entitled "An On=line Interactive DatilBase: Management System" (Co-authors:J. S.-Chandler and T. G. DeLutis) at the Computer Science Conference in Washington; D.C., February 18-20,°197.5 ° . B. J.Brinkman presenteda paper entitled "Use of-Associatipn,MeasuresBased on SDI/ Search Profile0 (Co-authors: G. J. Lazorick and Anthony.E.Petrarca) attheComputer Science Conferetice in WashiLgtm, b:c.,February18720, 1975°, H. W.°Buttelmann presented a:hell:Liar on ."'ApPlied Research in AutomatedEanguage Processing at Chemical Abstract's Serviceon Octpher.24, 1974:4 . . r ) J. . Chandler,presentedia paper entitled "An On-Line Interactive DataBare Man- agement System" (Co- authors: J. A. Aitken and T. G. DeLutis),at the Computer Scienc& Conferenceln .Washington, D.C., February 18:10,4975".' .. it, 1 J. S. Chandler presenteCa paper entitled "A Methodbloliy fora MatipleGbal 7 Approachto Computer Systems Design" (Cp-author:'T. G. DeLutis) it ehe Computer Science Conference? in Washington, D:6., .February 18 -20,1975.\ . o . .-,. i B.Chandrasekaran pressOled a paper entitled "Distance FunctiOns fogIndependent Measurements .andFinite Sample Size at the II Internationagoint Confer- . . ; ence on Pattern Rei Ognition, Lyttgby, Denmark, August 14:1974i, Co- author and co-presenter was A. K. Jain, Department of'Marhematics,Wayne°Stata University.. The. piper appears in the proceedings of the confirence. / . . c . . . . E.,I. When presenteda paper entitled "Goals and,DireCtionslor Research in ' Program Verificatibri" at the Computer Science Conferencein Witshington, 6.c.,. .- 07 , February 18-20, 19/5. ''° s . . C f e ' e V i . . .o B. Dasarathy presented apaper entitled "SomeMaxmin Location andPattern Classi- fier Problems: Theory and.Algorithms" (Co- author: -1- J. White) atMt , . Computer Science Conference in Washington, D.C., Febettary 8-',40, 1975. :\...... :: . f 67 So a ,... ... .' '58 0,...... L.; . , p jr.. G. peL,tis was the speaker'for a session on"Simulation Techniques" at.the Association for .Computing Machinery East Central egionel Conference, Troy, Michigan on November 1-2;197.4% . --,I, 2 ' , 4' \.-G°. DeLutis piesented papers entitled "Simulation of Data Base Files:. . DeTinition Independence" and "Simulation hmploy ng,MultizleCONputerk" (Co-authors: J. A. Aitken and.J. S. Chandler) pt tt'ComputeeScience Conference in Washington, 15.0..,Fehruary,18=20,'I:95. C.. R.Folk was the chairman of a,panel discussion o 'structuredyrogramming". at the October 9, 1974, meeting of t-.Centre hio Chapter or the Association $br.Computing Machinery Columb. ti M., S. Friedmam presented a paper entitled vflbsign andImplementation ora Method Output" atthe Computer Science . for Controlling PostmorteM Diagnostic. Conference in'Washington,.D.C., February 18-20,1925. E. Ludes presented a paper entitled "Cryptography and Data ifdse Security" atthe,, Computer Science Conference in Washingt0 on,'D.C.,February 18-20., 1975. D. K. Hsiao presented a paper entitled "Minicomputersd's Communication Subnets on Computer Network" at theInternational Meeting on Mini-Computers and Data' Communication, Liege, Belgium, on January 1,1975. Co-authors are Roy Reeves and Thomas Wyrick. ModelsAttribute -Based . D. K. Hsiao prented an invited talk entitled "Data Base .-4 and Relat Onal".atthe*Computer Science ColloquiUm held atParyrd tniver- e sity on 1'anuary 13; 1975. - . * 4 , . . v , . :/ . . . 1 D.. K. Hsiao presented an invited talkentitled "Information Secure Systems" a the Computer Science Colloquium,Rutgers University on April 3, ,,(015.._ .., , . . .. ;i . Programme'Committee of IF P onrgess O. D. k.Hsiao has accepted invitations from thb 74 to servelAs Chairman ofthe'sesslonlon Applications -of Interactive Computing arid as a panelist in the dession on Privaey_andCompUters. The Congress will be held on August 5-10, 1975, 11Stockholm:Sweden.: . - . D. K. tHsiao has been appointed a memberof the-Publications Committee of the - Institute of EleCiricaa and Electronics Engineers(IEEF) Computer Society of'ths/ditortal Board of the Computer Magazine, a monthly and as a member % , publication of tte same Society. . ' . % a . . D. K. Hsid9taccepeed an invitation-to serve 'on the DataManagement Pagel of the Cammitt e on 'computer Science And Engineering Research Study chaired by Profess ; r Bruce Arden,'Princeton University. The committee is-preparing a definitive volume on'comotter sciencevand engineering research of which -- , 4 . dWta management is.one of .the ien areas of research specidlizalion. , . - D. Isaacs presented a paper entitled "Tde5tIfication and Measurement of (dust and Pricing Elements of Computing Resources" (Co-author T. C.'DeLutis) at the, -CompVier Science Conference 'yin Washialtton, q.C.,Tebruary l8,--1,1 1975. . ) , C. Kar presented a paper entitled "Craletatics1 Inference Problem" (Co-authov:, e. L. J. White) at the Computer Scfencf Conference, in Washiftion,, n. February 18-20, 1975. ov. 68 :as lee 59 D. S. Kerr was an invited patitipent in a Symposium/Workshop on "Planning for :* Action - The Privacy, Mandate" co-sponsored by the Institute for Computer Sciences and Technology of the National Bureau of Standards and by the MITRE -1 Corporation, held in Washington, D.C., April 2-4, 1975. D. S. Kerr was elected Vice-Chairman t. of the Special Interest Groupon Computer Science Education of the Association forComputing Machinery. H. S. Koch presented a Paper entitled "AMissing_Componentof Current Data Base Management Sy-items- Data Base Reorganization" at the 37th Ameiican Society for Information Science AnnualMeeting, Atlanta, October 17,1974 The. paper appears in Information Utilities: Proceedingsof the 37th ASIS Annual Meeting, Vol. 11. , --% H. S. Kock- served as Setrdtarytof theSubschema Task Group of the CODASYL DDLC April 74- April, 1975: eR. Krishnaswamy,and Alan W. Biermann', Computer Science; DukeUniversity,co- euthored'apaper entitled "A System 4 for Program Synthesis from Examples" which Dr. Biermann presented at theInstitute of EleCirtcal and,Electfonics Engineering Systems, Man,(*d Cybernetics1974 International Conference, Dallas Texas, on October 1974. M. T. Liu presentecLa paper entAltig"Emulation of Computer Networks by-Micro- programmable MicrocoMputers" at the 7thAnnual Workshop on Microprogramming held in Palo Alto, California on September 30-October 2, 1974. Co-author was 15avid Coheri-: The-paper is'published/in the Proceedingsof the Workshop pp. 159-167. M. T. Liu co- authored a paper "Variable -ngth Message Transmission forDistributed Loop Computer Network " presented by C. C. keames at the 2ndAnnual Symposium on Computer Architec e, Houston,Texas, January, 1975. M. T. Liu presented "Minicomputers as Communications Subnets in Computer Networks" (Co-authors: D.K."- Hsiao, R. F. Reeves, and T. P. Wyrick) at the Inter- .° national Meeting on Mini- Computers and Data Communication, Liege, Belgium,, January, 1975 A. P. Lucido was 1-1e-Session Chairman for graphicslanguages at the First Annual Conferente on Computer Graphics and InteractiveTechniques. A. P. Lucido was the TechnicalProgram Chairman, and 8essioh,Otganizer, ( Second Annual for the Conferenceon Computer Graphics and (SIGGRAPH"7.5) Interactive Techniques A. P. Lucido is Editorof the Proceedings'of SIGGRAPH'75. In addition, he is Editor of SIGGRAPHITI,the newsletter of Group SIGGRAPH, the Special.Interest on Computer Graphics, andInteractlye Techniques, for.Computing Machinery. of-the Association, D. A. Marik presenteda paper entitled"Grammaticai 1,, J. Nate) at the Inference Problem"(Co-author: Computer 54$14.2eConference in February 18-20,3$75. Vishington,D.C., .89", 60 P.F. Mathis presented an invited paper entitled "An Overview of Debugging Tools and Techniques" at theAssociation for Computing Machinery Fast Central_ , Pegional.Conference, Troy, Michigan, November 1, 1974. R. F. Mathis presented a paper entitled "Pre-Execution, Batch, Interactive, and Post- Mottem Debugging" at the Computer Science Cqnferenct in Washington , DtC., February 18-20, 1975. ' D. J. Moore presented a paper entitled "An Algorithmto Collate Long Natural Language Texts, for Scholarly Use (Co-author: F. R. -Horowitz) at the Computer Science Conference, Washington, D.C., February 20,1975. D.J. Moore presented a paper entitled "Simulation of Settlement/Subsistence Systems: A SItISCRIPT Model for the Glenwood Locdlity" (Co-autho: L.' Zimmermann) at the Thirty -First Plains AnthropologyConference, November, 1974. M..Moshell presented a paper entitled "Parallel Recognition of FormalLanguages by Cellular Automata". (co-author: J..Rothstein) at"the Computer Science Conference in Washington, D.C., February 18-20, 1975. J, Rothstein was invited to 'speak ata Symposium on Parallel Proceasing in.Arti- ficial Intelligence at the Courant Institute ofMathematical Sciences Department of Computer Science), New York, January 17,1975. The symposium was co- sponsored by the Office of Naval Research. Professor Rothstein .spoke on "Universality and Hierarchiality in Parallel Processing :Sr 4 and Their Importance in Pattern Recognition and Artificial Intelligence." J. Rothstein was elected secretary of the local Institute of Electrical& Elect- ronics Engineers (IEEE) chapter of the Professional Groupon Computers. J. Rothstein and Carl Weiman,'New York University, co-authoreda paper entitled. "Fast Algorithms for Generating; Translating, and RotatingStraight Line Paths on Grids for Computer Graphics' which Dr. Weimanpresented.at the. Computer Science Conference, Washington, D.C., February 18, 1975. J. Rothstein presented papers entitled "Groupoid Strings, Turing Universality, Sand Parallel Processing in Cellular Automata." and "Parallel Recognitiono Formal Langpages by Cellular Automata" (Co-author:M. Moshell), at ;the Computer Science Conference in Washington, D.C., Februaty'18-20, 1975. V. Santhanam presented a paper entitled "An Algorithm for Efficient GnerAtion of Optimal Prefix Codes " k0o7author: L. J. White) at the Computer Science Conference in Washington; b.C., February 18-20, 1975. J. Smith presented a paper entitledSequential Document Classification Techniques: Experiments 'and Experimental Results" (Co-author: L. J. White) at the Computer Science Conference in Washington, D.C., February 18-20, 1975. S. N. Srihari presented a paper entitled "Statistical Classifier Design-for. Aircraft nenttfIcation" (Co-author: L. J.4White) at the Computer Scierice 4 Conference in Washington, D.C., February 18-200%1975. 10 -I 61 , . ..s! . tt ' F. A. Stahl accepted an invitationwfrom UkESCO,the United Nations Educational, Scientific and Cultural Organization,to advise and assist in the development of,graduate studies and research in'Computer Science-atUnivereidad Simon Bolivar, Caracas, Venezuela, under the NationalSystem of Engineering' Education for IndustryProject, from September to December 1974, W.S. Stalcup presented a paper entitled "Automatic Vocabulary Control in Printed Indexes: Evaluation Through Use of, a Quantitative Measure of Concept Scattering" ,(Co-author: A. E. Petrarca) at the Computer Science Conference in Washington, D.C., February 18-20:1975. L. J. White presented an'invited paper entitled'"An Efficient Algorithm for Maximum k = Matching in Weighted Graphs" at the 12th Annpal Allerto Con- ference on Circuit and System Theory hponsored by the Department of, . Electrical Ehgineering and The Coordinated Science Laboratdry,UniVersity Of Illinois, Urbana, Illinois, October 2-4, 1974. Th,paper appears in the Proceedings og the Conference. M. C. Yovits presented an invited paper entitled PA Semiotic Framework for °Information Scien5e. Leading to the Development ofa QuaAtitative Measure of Information" (Co-author: °J. G. Alidlock) at'the 37th AmericanSociety for .Information Science Annual Meeting, Atlanea, October 15,1974. Tge-paper . appears in Information Utilities; Procoedings of the 37t11 ASIS Annual Misting, Vol. 11. Dr. Yovits also chaired a session at the ASS Meeting. , . M. C.Yovits was Chairman of a session on Education at the Computer Science Con- . ference in Washington, D.C., February.18-2G,.1-9-7-5. M..C. YoiVits presented a paper entitled "Graduate Educatien in CO4uter 4.146ce and its Relationship to Industry" at the National Computer Conference',' Anaheim, California, May 20, 1975.41 M. C. Yovits was a'panelist on"Informition.Transfer Today and.TotporrowPat the American Society for Engineeringlducation National Conference,a; Rensselaer Polytechnic Institute, June 17-20, 1975. M. C. Yovits is Co-editor of Advances in Computers, Volume 13,Acidemic PleiS) 1975 with M. Rubinoff. , M. C. Yovits is a member of the American-Committee for the SecondWorld Conference in Education (WCCE '75), sponsored by the International Federation :far Information processing, September 1-5, 1975, Marseille, France. \ M.C. Yovits is a member of the Accreditation Committee of the Association for -Computing'Machinery beginning March 1975. M. C. Yovits is.a member of the Curriculum Committee on:pmputer Science of the , Association for 6omputing Machinery, beginning May 1975. S. H. Zweben presented an.invited paper entitled "'A Recent Approach tthe Study of Algorithms" at the Association for Computing Machinery Annual onferance, San Diego, November 11, 1974. The paper appears in the Proceedings of the ACM Annual Conference, San Diego, 1974, p. 747-748.Dr. Zwehenewas alsoa member of a panel on "Software Physics" at the Conference. 71 62 H: Zweben will be a reviewer for the World Conference on Computersfn Education, Marseille,'France, to be held in September, 1975. S, H.,Zueben will be a reviewer at the ACM 75 National Conference,Minneapolis, Minn.', to be held in October, 1975. S., H. Zweben was awarded a $4,000 University Research Grantfrom the Craudate ' School for his research on "Structure of Computer and NaturalLanguage Algorithms." . k tam- O V ft---^ U 4r 72 63 APPENDIX E PUBLICATIONS OFTHE CMPUTER`AND INFORMATION SCIENCE RESEARCH CENTER STAFF1 ABILOCK, J. G.: YOVITS, M. C. A semiotic framework for information science leading to the development of a quantitativemeasure of information. In: ' Information Utilities; Proceedings of the 37th ASIS Annual Meeting,Vol.:11, Atlanta, October 15, 1974. ATKINSON, H. C. Extension of new services and the role oftechnology.Library Trends, October, 1974. ATkINSON,'H, C. Priorities for new librarian of congress.Library Journal, July, 1974. ATKINSON, H. C. The Ohio State University mechanized information center. In: Proceedings of EDUCOM 1973 Fall Conference. BAMMEL, S.E.; ROTHSTEIN, J. The number of 9 x 9 Latin squares. Discrete Mathematics, Vol. 11, 1975, pp. 93-95. EUTTELMANN, H. W. Semantic directed translation of context free languages. American ,Journal of Computational Linguistics, microfiche no. 7, _ November, 1974. e - -,.- CHANDRASEKARAN, B.; and YOVITS, M. C. Artificial intelligence. In: Encyclopedia_ of Computer Science and Technology, VoJit.2. Marcel Dekkei, Inc., New York, 1973.'. f 4 .' 14. : 1 CHANDRASEKARAN, B. Artificial intelligence - The past decade. Ini Advances in Comjuters, Vol. 13, 1975., pp. 170...225. 1 CHANDRASEKARAN, D.; JAIN, A. A. Distance functions for inde endent measurements and finite sample size. In: Proceedings of the II Intnational Joint Conference on Pattern Recogation, Lyngby; Denmark, Augu CHANDRASEKARAN, B.; JAM, C. C. A finite memory deterministic algorithm for the symmeir,ic hypothesis testing problem.: In Transactions oh Information Theory, published by the Institute of Electrical and Electronics Engineers," Inc., January, 1975. HSIAO, D. K.; BAUM, R. A Semantic model for piotection mechanisms in data baie systems. In: The Proceedings of the atirthrwei-i--Errternational Conference on System Science, Honolulu, January 6, 1975. H..S.. A missing component of current data base management systems- data base reorganization., In:' Information Utilities; Proceedings of the 37th ''ASIS Annual Meeting, Vol. 11; Atlanta, October 15:1974.. . See Appendix F for publications issued4te part of the Computer_and Information Science Research Center te ical.reporte.reries., M 7 3 4 F A -64 LIU, M. T.; COHEN, D. Emulatiorof computer networks by Micr0Programmable micro-computets. In: Proceedings of the ACM 7th Annual Workshop on Microprogrammin, Palo Alto, California, September7OctoP6r, 1974. 4 159 -167. LIU, M. T.; REAMES, C. C. Variatle-lengthmessage transmission for distributed loop computer networks. In: The Conference Proceedings of the Second Annual SymposidM on Computer Architecture, Huston, Texas, January, 1975, pp:7 -12.. tai . MOORE, D. J.; TUGGLE, D. Computer solution of verbal analogy prOblems. Computer Studies in the Humanities and Verbal'Be avior, Vol..VI, No. 3, October, 1973'. PETRY, F. E.; BAUM, R.I.; BIERMANN, A. Spee ing the synthesis of programs, from f traces. IEEE Transactions on Computers, Vol. C-24, No. .2, February, 1975, pp. 122-136. STAHL, F, Cryptography in the ageof automa tion.14In:The Proceedings of the IEEE Internatio4a1 Symposium on-Information TIleory, October, 1974'. STAHL, F. Cryptography and computation. In: Encyclopedia of computer science. Edited by A. Ralston.Auerback, 1974. WHITE, L. J.; KSIENM, A. A. Aircraft identification using a vilinear surface . reRyesentation of radar data. Pattern Recognition, Vol. 6, 1974, pp. 35-45. WHITE, L. J.; GILLENSON, M. L. An efficient algorithm for minimum k- covers in weighted graphs: Mathematical Programming, yp].. 8, 1975, pp. 20-42. WHITE, L. J.; KSIENSKI, A. A.,and REPJAR,ii. G. Object Identification'from4 mufti-frequency radar returns.The ladio and Electronic Engineer, Vol.N- f 45, No. 4, April, 1975, pp. 161-167. WHITTEMORE, B. J.;YOVII1S, M. C. A generali;ed concept for the analysis of infolmation. Infoimation Science: Search for Identity; Proceedings of the 1972 NATOAdvanced Study Institute in Information Science, A. Debons, ed Marcel' Dekker, Inc.,*New York, 1974. WHITTEMORE, B. J.; YOVITS, M. C. The Quantification and analysis of information used in decision pfocesses. Information Sciences, Vol. 7, April,1974, 14. 171-184. YOVITS(M. C.; RUBINOFF, M., EDS. Advancgs In Computers, Vol. 13. Acade mic f'Pred'i", 1975. V YOVITS, M. C. Graduate education in computer science'and its relationship to industry.- In: AFIP8 Conference Proceedings National Computer Conference, Vol..44, Anaheim, California, 1975. YOVITS, M. C. A Theoretical framework for the development of information scieue. I In: Information Science, Its Scope, Objects of Research and. Problems,' edited byA. I. Mikhailov. Publication of- International Federation of Documentation Committee on Research on the Theoretical Basis of Infofmatlon, MoscoW, 1975, '74 r 4. 41. PAPERS ACCEPTED FOR PUBLICATION BUTTELMANN, H. W. On the syntactic structures of unrestricted grammars. Generative grammars and phrase structure grammars. Infofmation and ,Control, Vol. 28, July, 1975. 52 pp. BUTTELMANN, H. W. On the syntactic structures of unrestricted grammars. II. 'Automata., Information and Control, Volo..28, July, 1975. 21 pp. CHANDRASEKARANr8.; JAIN, A. K. Independence, measurement complexity and . classification performance. Transactions on Systems,. Man & Cybernetics. LIU, M. T.; REAMES, C. C. The design of the distributed loop computer network. In: Proceedings of the 1975 International Computer Symposiu Taipei, =Taiwan, Republiyof China, August, 1975. ) LIU, M. T.; WANG, J. T. A novel model for a mixed voice/data transmission system for computer communication. In: Proceedings of the 1975 International Computer Symposium, Taipei, Taiwan, Republ/c of China, August, 1975. -444, w tunic, S. Social aspects of computer networks. Management Information, Journal of IFIP. November, 1975. ROTHSTEIN, J.; WEAAN, C. Parallel and sequential specification of a context sensitive language for straight lines on grids. Computer'Graphics 'and Image Processing, 1st issue. 1976., , . , \ Q PAPERS SUBMITTED FOR PUBLICATION - KERR, D. S.; KALMEY; D. L. Theobachelors and masters computer science graduate. Communications of the-ACM. a KERR, D. St Recent .vesults on the attribute based data model a tutorial. International Conference on Very Large Data Bases, Boston, Mass., September, 1975 ," LIU, M. T.;.REAMES, C. C. Simulation of the distributed loop computer network. The 3rd Annual Symposium on Compute.r.AAxchitecture, Clearwater, Florida, January, 1976. 'LIU, M. T.; FOURNIER, S. System aesign of a grammer-programtable high-level. language machine. The 3rd AnnualSymposium on Computer Architecture, Clearwater; Florida, January, 1976: 4 o V I 4 MOORE, D. The computational complexity of recursively enualerable total orders. =t Theoretical Cbmputer Science. / MUFTIC, S.' Social aspects of computer newtorks. Management Informatics. MUTTIC, S. SegfA157 systems for numerical data. Journal of the Association for Computing Machinery. . a 75 0 4 411 66 . APPENDIX F a. a TEAICALREPORT SERIES, 1968 A YOVITS, M. C.; ERNST,'12:41. Generalized information_aystems: Some conse- quences for information transfer. October, 1968. 47p. (OSU-CISRC-TR- 68-1) (PB-180 929) FILLMORE, C. J.; LEHISTE, I. Working papers in linguistics no. 2. November, 1968. 128p. (OSU-CISRC-TR-68-3) (PB-182 596) 401 FRIED, Jr%B.; LANDRY;. B. C.; LISTON, JR., D. M.; PRICE B. P.; VAN BUSKIRK, R. C.; WASCHSBERGER, D. M. Index simulativ feasibility and automatic document classification. October, 1,968. ,21p.OSU-CISRC-TR-68-4) (PB -182 597) ROTHSTEIN, J. Thermodynamics & information: Before, in and beyond quantum Mechanics. DeceMber, 1968. 21p. (OSU- CISRC- TR -68 -5) (PB-183 738) - ,FINLEY,6JR., M. R. The dev- elopment of a basic language for art.ificial .intel- ligence. January, 1969. 24p. (OSU-CISRC-TR-68-61. (PB-182 305) . a f 1969 COLOMBO, D. S.; RUSH, J. E. Use of word fragments in computer-based retrieval systems. February, 1969. -i-(93p: (OSU-CISRC-TR-69-1) (PE-184 104) WHITE, L. J. .Minimum covers of fixed cardinality in weightedugraphs. March; 1969. 14p. (OSU-CISRC-TR-69-2) (PB -1.83 737) . JACKSON, D. M. The construction of retrieval environments and pseudo-classi- ficationglbasedon external relevance. April, 1969. 74p. (0S9-,CISR-TR- 69-3) (PB-21.84 462) $ ELLIOT, D. E.; HUANG, S.; LANGENDOEN, D. T.; LEE, P.d.; LEHISTE, I. Working papers in linguistics no. 3. June, 1969. 181p. (OSU-CISRC-TR-69-4) ?, 1(PB-185 855) (ED-060 689)' . . BRIGGS,.G. E. Reaction time and uncertainty in human information processing. March, 1969. 36p. (bSU-CISRC-TR-69-5) (PB4-184 135) WHITE, L. J.; RUSH: .J.E. Linear lists for spiro graphs. June, 1969..69p. (OSU-C,RC-.TR-69-6) .,(PB-194 402) 44/11 PETRARCA, 4. E.; LAY, 4. M. The double KWIC coordinate index. Anew aOptoach for preparation of high-quality printedindexes by automatic indexing.tectl-. niquep. April, 1969. 12 +[17]p.- (00-7CTSRC-TR-69 -7)* *Journal'of Chemical, Documentation, 9, 256 (1969) 76 67 YOVITS, M. C. Information science: Toward the development of a itrol/ scientific discipline. June, f969. 27p. (OSU-CISRC-TR-69-8). (PB -187 983) s, PETRARCO A. E.; LAY, W. M. The double KWIC coordinate index. II. Use of an-automaticalry generated authority listto eliminate scattering caused by some singular and plural main index terms. August, 1969. 13p.. (OSU- - CISRC-TR-69-g)* 00 MCCUjvLOUGH, J. L. The acquisition of information across cultures: I. Per - suasjve role play, counterargument and attitude change.August, 1969. 18p. (OSU-CISRC-TR-69-10) (PB-197 568) . or' ERNST; R. L.; YOVITS, M. C. InforMetion science as an aid' to decision making. September, 1969. 22p. (OSU-CISRC-TR-69-13) (PB-189 666) (ED-054 782) LANDRY, B. G. An indexing and re-indexing simulation model. Jdne, 1969. 50p. (0fU-CISRC-TR-69-14) (PB -198 115) SALVADOR, R. Automatic abstracting and indexing. June, 1969. 93p. (OSU- CISRC-TR-69-15) . "\k STEVENS, D. W. A computer program for the reduction of flow tables.' June, 1969. 97p. (OSU-CISRC-TR-69-16) '(PB-189 679) COLOMBO, D. S. Autdinatic retrieval systeins and associated retrieval lan- guages 1969. 69p. 4(DSU-CISRC-TR-69 -17) (PT-198 116) SCHLESSINGER, J. D.; WHITE, L. J. Optimum p efix encoding. August, 1969. 85p. (OSU-CISRC-TR-69-18) (PB-198 117.) , DAY, R.; agn, L. J. Hebhign neural simulation: Computer program docu- mentatia% (OSU-CISRG-TR-B9-19) (P3-204 003) REEKER, L. H. Extended finite state' representation of infinite machinea. September, 1969.',36p. (0SU-CISRC-TR-:69-101 (PB-187 949) WILLIAMS, N. T.; ST R: L A computer simulation of human short-term ) memSy. 1969 62p. (05U- CISRC- TR- 69 -22), (PB-197 874Y BEZDE , R. R. The' acquisition of information across cultures: II. .Social science researchin a different -culture. III. Cross - cohort activity and attitude change, by J. L. McCntllough.. January, 1970. 36p: (OSULCISRC- -,TR-69-23) (PB-197 876) _, '0101 . 1970 . , . ''..."'' ,',.;` UNKLESBAY-, M. K., A one step version of Younger's algorithm for bounded . .?... , . ., context grammars. . 1970. .p. .(0SU-CISRC-TR-70-1) (PB-197 6 3) .,... 4 1 . /' 4 , *Proceedings of. the American Society fOr Information Science, Vol. 6, 1969,,277 -2a2. . $ o .- tilt 7 7 . \ i 68 e . LI, Y. Information structure and optimal policy. September, 1970. 18p. OSU-CISRC-TR-70-2) (PB-197 605) I DILLON, S. R. Some procedures for finding substitution pfoperty partit ns, substitution property covers, and cover pairs,for finite state sequential machines. 1970. 79p. (OSU-CISRC-TR-70-3) B-197 643) MATHIS, B. A.; WHITE, L. J.; JACKSON, D. M. Stability, analysis of term crsimilarities lor info'imation classification theory. July; 1970. 79p. (OSU-CI SKT-70-4) (PB-195 376) , . , ...., . MCGHEE, R. B.; DILLON, S. R. A paull-unger procedure for substitution prop- erty partitions. April, 1970. 16p. (OSU-CISRC-TR-70-5) (PB-192 12.0) DAY, R. G.; WHITE, L. Study f-a random search method for function mini- mization. March, 1970. 77p. (OSU-CISRC-TR-70=6) (113-194.404) liAkts; R.;:WHITE, L. J. , Identification of circuits in chemical structures. June, 1970. 73p. (OSU-CISRC-TR-70-7) (PB-194 396) HARALSON', K. M.; WHITE, L. J. Opt mal prefix codes for ensembles of N Aa eqdiprobable messages using a b ry alphabet. May, 1970. 105p. (OSU-' CISRC-TR-70-8) (P8-197 642) PETRARCA, A. E.; LAZTINEN, S. V.: LAY, W. M. Use of the double KW C Coor46-' nate indexing technique for"chemical linefttations. 1970 14171p., (OSU-CISRC-TR-7,0-9) (PB-198 269) LAY,, W. M.; PETRARCA, A. E. Modified double KWICcoordinate index: Refine- t mentslm main term and subordinate termiselection. 1970. 11+1111p. (OSU-CISRC-TR-70-10) (PB -197 567) (ED-054.815)v , LYONS,J. J. '91e speed-accuracy:, qade7off processing different, Classes of Material 1970. 38p. (dSU-CISRC-TR-7Q,-11), (PB-198 114) 4 FILLMORE; C. J.J,LEHISTE, I.; MELTZER;:D.; TATHAM, M.A:; THOMPSON, S. A. Working papers in linguistics no. 6? September, 1,970. 132p. OSU- CISRC-TR-70-12) (PB-194 829) THSTEIN; J. Information generalization of entropy in physics. February; 70. 22p. (0St-CIsk-TX-70-24)'(PB-192 128) JACKSON, D..M. Basis for an improvabilitymeasure for retrieval performance February, 1970. 31p. *(0SU-CISRC-TR-70-25) (PB-197 812) DRACHMAN, B.; EDWARDS, M. L.; FLLMORE, C. J. E, G.;. LEE, T.; LEHISTE, I; ZWICKY, A.,M. Working papers in linguistic May, 1979. 164p. 4 (OSU-CISRC-TR-70-26)' (PB-192 163) . , _ 1971 y ROTHSTEIN, J. Patterns and algorithms.-January, 1971. .8p. (OSU- cTSRC -TR- 71-1) (PB-197 604) p 1 69 GROSU, A.; LEE, G. Working papers in linguistics no. 7. February, 1971.": [243]sp. (osu-cislic-TR-Z1 4).__Inl.198 278) (ED-060 688) CfiANDRASEKARAN, B.; KANAL, L. On linguistic, statistical, and mixed models,', for pattern recognition. Maich, 1501. 33,A5p,. (OSU-CISRC-TR-71-3) (PB-198 279) WHITTEMORE, B. Afilexampleof the application of generalized informition systems concepteto the quantification of information in a decision sys7em: The examination of quantified information flow in an industrial controX problem. May, 1971. 51p. (OSU- CISRC- TR -71 -4) (PB-202 621) a OSTROM, T. Me; STEELE,,C. M.; SMILANSKY, J. Ihfo;mationand attitudes: The effects of information context and pe'rceived discrepancy on attitudes-. May, 1971. 23p. (OSU-CISRC-TR-71-5) (PB-202 622) , Y.Equipment replacement models: A generalization and extension. Mayo 1971. r 17p. (OSU-CISRC-TR-71:-6) (P8-200 548) ELLIOTT, D.; GEIS., M.; OROSU, A.; NOBEL:1;*ZWICKY, ANN; ZWICKY, ARNOLD. Working papers in lingufstica no. 8\ June, 1971. 197p.. (OSU-CISRC-TR- `o 71-7) (PB-202 724) BOND, Z. S.; GREGORSKI, R.; KERE .;,LEHISTE, SHOCKEY,rL.; WENDELL, M. V. Working papers in linguist s no. 9. July, 1971. P2p. (OSU- CISRC-TR-714) (PB-204 002) (ED-00 686) A JOHNSEN III, A., M. Marformance6i,n memoir scan task under conditions of .fixed versus varied memory sets. Novem , 1971. 49p'. (OSU-CISRC-TR- 71-9) ,:(PB-210 432) . MEHAFFEY III, L.° Thi spectral'sensitivity of the'turtle Pseudemys scripca elegans. Novemberf 1971. 89p. (OSU-CISRC-TR-71-10) (PB-210 433) ll 6 WILLIAMS, J. D. Stimulus-encoding andAmaninformation processing. November, 1971. pp.. (OSU-CISRC-TR-71-11) /' SANDERS,,S. A. A modification of a method of generating iandOm numbers ° using a combination of two congruential generators. Decembet, 1971. 76p. (OSU-CISRC-TR-.71-12)' - 400 I 44 14. L4NDR1, B. A theory of indelting: Indexing/theory as a ;model fOr infor- Usionstorage and retrieval-. December, 1971. 14211p. (OSU-CISRC-TR-71-13) (PB-205 829) (ED-057 843) - . OSTROM, .;EDWARDS, J...D.; ROSENBLOOD, L. R:'Inte ation of discrepant 9 in interpersonal attitudes. Decembei., 191. 43. '(OSU-CISRC- ' TR-71- (PB-110 500) 4- . 1972 WHITE, L. J.;GILLPSON, M. L. Optimum center location. January; 1P72. 69,A6,B32p. (OSU-CISRC-TR-71-1) . MELTZER, D..Speech.,sy thesiaby haar'cfunctiOhs with comparison tb a terminal r t, analbg device. Janua y1972. 135p. (OSIICISRC=TR-72-2) am. CHANDRASEKARAN, B.; JAIN, A. IC. Quantization of independent measurements and recognition performance. March, 1972. 14p. (OSU-CISRC-TR-72-3) (b-747 706) CAMERON, J. S. Automatic document pseudoclassification and retrievalword frequencytechniqueg. March, 1972. 165p. (OSU- CISRC- TR -72 -4) (13B-210 435) OSTROM, T. M.; SLOAN, L. R.; MG CULLOUGH, J. LCInformation ancrattitudes: ,The effects of repetition and amount of information. April, 1972. 38p. (0SU-CISRC-TR-72-5) (PB-209 802) (ED-061 968) 4 e, , LEHISTE, MELTZER,..D.;SHOcKEY, L.;,GREGORSKI, It, Working papers in. ,linguistics no. 12. June, 1'972. 88p. (OSU-CISRC-TR-72-6) (PB -210 781) (ED-069 162) VP f k EKONG,. J: U. Rate of convergence of, hermite interpolation based on the roots of certain,Jocab polynomials. June, 1972. 5. (0SUCISRC-TR-72-7) (PB-211 237) . , ._ WEIMAN, C. F. R.; ROTHSTEIN, J. Pattern recognition by retina-like devices. July, 19.72. 154p. (O$U-CISRC-TR-72-8) (PB-214465/2) (ED7065.144) iV CHANDRASEKARAN, REEKER; L. H. Artificial intelligence - a;,icase for agnosticism. August, 1'972. 28p. (OSU-CTSRC-TR-72-9) (FD-069 296) ROTHSTEIN, J. Loschmidts''s and Zermelo's paradoxeg do not exist. October, 1972. 9p. (0SU-CISRc-Ti02-16) (PB-213 712/8) .(PB-214 666/0) . YOVITS, M. C.;CHADRASEKARAN,aB. Artificial intelligenCe. 4cember,,,1972. 39p. (OSU-CISRC-TR-72-11) (AD-760 HSIAO, D. K.; DENICOFF, M.; BERGART, J. C. An annotated and cross-reference bibliography on computer security and access control in computersystems. November, 1972. 57p. (OSU-CISRC-TR-72-14/(AD-755 225) "(ED 1378 879). tFOULK, C. R.; JUELICH, O. Smooth programs and,languages. NoVember, 1972. - 16p., (0SU-CISRC-TRIe2-13) (PB-218 805/0) (F0-071 683)1 ' . YOVITS, M. C.;WHITTEMORE, B. A generalizedco nceptual development for the December, 1972: lip.' (OSU-CMC-TR-72-14) analysis and flow of fnfornlation. . .. . (PB.-216 530/7) . .. MATHIS,, B. A. Techniques, for the evaluation and improvementOf computer 7' prOduced abstracts. December, 1972. 262p. (OSU-CISRQ-TR-72-15). (PB-214 675/1) (ED-071 686Y ' .973" /vc\ . , L.. , .. 1AY, W. M. The double KWICcoordiriate indexingtechnique: ,theory, design, 250p..(0SU-CISRC-TRz73'-1). (PB- and,,Amplementation. February, 197,3. , 228 181/4gI). (ED-072 826)* o80 71 1 KOUNG, C. E. Development of language analysis procedures with application to automatic indexing. April, 1973. 298p..(0SU-CISRC-.-TR-73-2)N (PB-227 088/2GI) (ED -078 843) , STRONG, 'S. J. An algorithm for generatinefiructural'surrogites of English :text. April, 1973. 140p. (OSU-CISRCTR.:73 -3) (PB-227 F71/8GI) ED-078 844) . . - HSIAO,D. K. LogicalAccess control mechanisms in computer systems. July, 1973., 24p. (OSU- CISRC- TR -73 -4) ,(AD-765 727/3) (E6.47e 880)' WANG, P. T. R;' Bandwidthtminimization, reducibility decomposition, triangularizitidn_of sparse matrices. September; 1973. 162p. (OSU1 CiSRC-TR-73-5) (PB-227 396/9GI) 0 BIERMANN, A. W.; BAUM, R:1 KRISHNASWAMY, Rfl-PETRY,'R.*E. Automatic progrik synthesi.yeporta. October, 1973,62p. (OSU-C1SRC-TR-73-6) (ADJ171 332/4GI) (ED-086 1 7) WRICK, T.; HSIAO, D. K. SystemConfiguration study on the interconnection of the IBM 370/165, 'DEC System-10 and MICRO 1600/21 computer systems. October, 1973% 26p. (OSU-CISRC-TR-73 -7) (AD -770 611/2GI) #11W =V. 1974 HSIAO, Di K.; KERR, D.,,S.; MC CAULEY III, E. J. A. model for data secure- , systems (Part 1).0,February, 1974.,3.9p. ,(0SU-CISRC-TR-73,43) (AD7-778 6577G1) (ED-090 983) HSIAO, D. K.; KERR,1,D. S.; NiE: C. J. Context protection and consistent control in data base'ystems (Part I). February, 1974. 28p. (OSU- CISRC- TR -73 -9) (AD-77e 658/5GI) (ED-090 982) ) BAUM, R. I.; HSIAO, D. K. A.data secure computer,architecture. July, 1974. 22p. (OSU-CISRC-TR-73-10) (AD-787 580/OGI) RUSH, J. E.; PEPINSKY, H. B.; LANDRY, B. C.; MEARA, NI M.4 STRONG, S. M:;_- V'ALLEY, J. A.; YOUNG, C. E. A computer language analysis system February, 1.574. .71p. (OSU -CISRC-TR-474-1) ED-090 948) REAMES, C. C.; LIU, M. T. Variable-length message transmission for distributed_ loop computer networks (Part I). June, 1974. 9p. :,(0SU-CISRC-TR774-2) (PB-242 913/2GIT (ED-092 152) J 4. KIRSCH, B. M. An improved error diagnositcs system for IBM System/360-70 assembler program dumps. Jule, 1974. 60p. (OSU-CISRC-TR-74-3) .(PB-235 06,0/1GI) (ED-092 113) . COHEN, D.; LIU, M. T. Emulation of computer networks by microprogrammable microcomputers. June, 1974. 17p. (OSU-CISRC-TR-74-4) (PB-242 912/401) 4* ,`" f , . BIERMANN, A. W.; KRISHNASWAI4Y, R.1Constrticting programs from example computations. AugVat,-1974. 41p. (OSU- CISRCrTR- 74 -5) 0 72 ... . 1) BUTTELMANN, H. W. Semantic directedtanslatiOnof context-free languages. September, 1974. 36 : (OSU-CISRCR-74-6) (PB-241 8t4/AS). :' ) HSIAO, D.. K.; KAFFEN,N. E. A model.for datasecure systems (Part II). October, 1974. 40p. (OSU-CISRC-JR-74-7)'(AD/A-006 276 /OGI) HSIAO, D. K.; KAFFEN,N. E. A model for datasecure systems (Part III). November, 1974. OSU-CISRC-TR-74-8) (AD/A-004 676/361)4' ...._ HSIAO, D. K.; AUM, R. I. Information secure.gystems:' November, 1974.AD . A 55p. (0 -CISRC-TR-74-9) .** 11- e 1915 4 HARTSON, H. R.; H'SIAO,-D.K. Laniguages for specifying protection requirements 1IN in data base systems(Part 1). January, 1975.---61p. .(0SU.-CISRC=TR-74-10) (AD -006 280/2GI) .7 PETRY, F. E. Program inference from example computations represented by memory snapshot traces. February,. 1975.ul/Op. (OSP-CISRC-TR 75-1) . . ... N , . . MC dAULEY III, E. J.'A model lor data secure systems.. March, 1975.,119p. ! (OSU-CISRC-TR-7572)- (AD/A-011 359/7GI) I HORGER, JR., W. A. Data base module verification--a certification method for data secure systems. June, 1975. 125p: (OSU-JtISK-TR-75-3) RODEHEFFER, T.; KAFFEN, N. fREST4--A highly structuredariTRAN language fbr systems programming. July, 1,975. 36p. (OSIJ- CISRC- TR -75 -4) , r V ot, e K 82 . . A 1 O 0 73 APPENDIX G o. DOCTOR )F PHILOSOPHY DEGREE 1971 4 LANDRY; B. S "t Theory of Indexing:' Indexing Theory as avMOdel for 4 Informatt torage and Retrieval. 7. tfi _ 1969 . CAMERON, JAMES S. Automatic Document Pseudoclassification and Retrieval by Word FrequencTechniques g , . lierMite Interpolation Based on:theR05,5ta 'EKONG, il4TOR J.La7"4 te ofdOnvergence of COiain J4bi Polynomials GORDON, ROBERTThe Organization and Control of a Slave Memory Hierarchy HARRISDAVID R: GOLDA: A GraptiicalOn-Line System for Data Analysis MATHIS, BETTY ANN Techniques for the Evaluation and Improvement of Computer- 'Produced Abstracts WEIMAN; CARL F. R. 'Pattern Redognition 'by- Retina -Like Devices - , WHITTEMORE, BRUCE X. A Generalized Decision Model for the Analysis of Information j 0,7 1973 A CHAN, PAUL SUI-YUENAn Investigation of-SymetricRadix- for Computer Arithmetic - DEFANTI,-.THOMAS A. The Graphics Symbiojais,SYstem - an Interactive Mini-Computer _Animation Graphics LanguageDeaigned for Habitabi(lity and-Extensibility i. 1 ClaUse( Deletion in RIpOlution Theorem Proving GELPERIN, DAVID.H. - . . . , . HEPLER, STEPHEN PHILIPUse of Probabilistic Automata as Models of Human Performance LAY, W. MICHAEL The Double-KWIC Coordinate Indexing Technique: Theory, Design, and Implementation . WANG, PAUL.TIING RENNBandwidth Minimiiation, Reducibility Decomposition,and Triangulation of,Sparse Matrices . .YOUNG, CAROL E. -Development of Language Analysis FrOceduras with Application to Automatise IndeXing 83 00' ,, ,-, 4.1 174 1 197-4 BUG, JAMES L. Human.EXtrapOlation of Strings Generated by Ordered Cyclic Finite Statetammars DOHERTY, MICHAEL E. A Heuristic For Minimum Set Covers UsingP1uSability Ordered Searches GILLENSON, MARKJ.. The Interactive Geleration.of Facial Images Dn'a CRT Using, a Heuristic Strategy . - LONGE, OLUWUMI An Index of Smoothness for Computer Program Fiowgr4phs SU, HUI-YANG Pagination of Programs' for, Virtual Memory Systems s k 1 I C. I A V V .84 -9- -2 . 75 :INVESTIGATOR INDEX 0 Abilock, J. G., 17 Juelich, 0. C.,'32 14 Aitken, J., 14 Kaffen,.N., 36 Amoss, J., 27 Kalmey, D. L., 0 Baum, ,R. 36 Kar, G., 11A Beug, J. L., 15 Kearns, C., '33 Breeding, B., 27 Kerr, D. S., 29, 35 -Breene, L., 37 Knablein; R. 59 Brinkman, B. J:, 10 Krishnaswamy,'R., 44 Brown, D. C., 26, 'Koch, H. S., 34, 37, 39 Buttelmann, H. W., 20, 21, Awl Lazarick, G. J., 10 23, 24, 25 Lee, M. J.;r16, 18 Chandler, J. S., 13 Leggett, W.,'46 Chandrasekaran5 27, Liu) M: 34, 41, 42 Cheng, T. T., 42 4 Lucido, A. 264 32, 33: Cohen,-D., Mamrak, 43; 44 4 Dasarathy, B., 1 Mentanelli, R. G., 44 ' DeLutis, T. G., 12, 13; 14 a' Miller, 11.,'37 Dickey, fs; J., 23 Mopre, D. J., 22,48, 37, 45, 46 .0 Doherty, M. E.1. 29 Moshello M., 46 4t Ernst, R.,,15, 16, 18 Muftic, S., 35'\ . Foulk, C. R., 32' Nee,, C. g 'ournier, i., 41 . 237 Gudei,,E., Parent, 11:,' 35 Hartson, H. Rex, 38 Petri Ca; A.' 10, 1.1, -4, 14 Hennings, J. M., 34 Perry. b., 23 Hspao, D. K., 34, 35, 36,' >/. .37, 39, 40,. F. .27- 4 -- . 0. 85 a # -: , , . 76 A a Pyster, A.1., 20, 21 Reames,C. Reeker; L., 21 Rodeheffer, T.,437, 40 Roth'stein, J., 29, 20, 33, 39, 46, 47 f Santheam, V., 30 if 5 Schmaltz, D., 37 t' ..--Sondheider, N..K., 20, 23, 24,-26 6 .Stahl, F., 39 italcup, 'S., 10, 12, 14 Wang, J. T., 42 0 Wells, X. J. 11 . White,1... J., '11, 29, 30, 31 Williams, D. G., 11 Wong, 6.,38 'Yovits M. C., 17 Zweben, S. H., 24 r A 0 $ a.1 , a6 r .t , IZ ...... 40 ii, Att , $ t 77 '"- SUBJECT INDEX 8 Academib Programs4 ,,Binary -Strings and Topology, '29 I Access Controlin"Data Bus Automata, 46 o Base Systems, 34; 35 1-__.4 Cellular. Automata-and,Pa allel Activities, -Related of Computation, 46.- GISRC Staff,, 57 , . .- Chemical Sterois9mers,-:Generation ircraft:Detection .and. :10f Unique' Ccirifi'gura.trona). Classifica'ti'on, 26- epreseniatrof; 11 ,. Algorithms-,for Coiputer Collation as icaploni:*Aut.omatic- of Long Natural Tex:is for -DocUment? 11 . Scholarly Use,_ 22 1, ,,i` - Coding, 4.traiim-R e d un 4.4 y Algorithms'Structure of Computer yartabia.:Length;_30 4 and Natural. Language7," 24 _CollationofLong Natural .Texts - Anima ,GraphicsPrograiaming far Scholarly-Use, Prar4ical Language, 32 _ Algorithfas-for,122 6, 'Anthrological,'Siniulation;po 28 Arcjiitecture, Computer, 41 Corkiuter'sna Information. .-Science Department, at . Atahitectpre,.pata Secure: . Systeia, 36 Cothpikter and -Nattaal Language/ .. "V. _ . .,,. .-. Algorith'ias r _1Struceure of, 24 ,:re 1 :. ---. Artificial Intelligence,k24, L '-' ". e- :ComputerArchitecture,-. Jo. Association Measures for Seakih (- .Profile' Enhancement, to Computer ;Cpllation of Lang , = , Nattiral Texts; -22 Automata, 25 - 1 .1. - : Computer COmmunicittiona.,..` Automata Theory-,_20,:te.1,, -7* . -Kixqd Dati/VOice 'rrans--. . miesion- §Ys tem fop', 42 Autoniatic Dodtiment, Classi7-. - - ftqation, 11 ComPuter-Exkrimentatfon . ), - - _,_.f:Pacility, :33 Automatic Indexing, 10; 12 _.: , - ". . 4 "Computer Gp..phics! 33; 3 Automatic-Searth Profile Enhancement, k10 Conuter-Networks417 Automatic Vocabularyontrol -Computer Systems .D66,811, in .Print *Indexes1.2 -ok Computihg ,Resuurces, -`tok4ii t ic a - . Automaticocahulary Codtts --tion 6.-tfeasurement 4,f,LCost Techni ues, Evaluation of, 16 & _Pricing Elements, 12 ro 3- _. /. !O.-if '1,4 , 4 3 78 ob, .Concept Scattering, Elimination Doctoral-DegreesA,73 of, in Automatic Indexing; 10, 12 Document Clasgification, Concurrent Procebging, 34, 37' Automatic, 11 Connecdivity2:29 Dynamic.Programming, 42 41: Context Protection, 34, 35 ,,Equationst, Nonlinear, 29 Control in Data Base Evaluation of Data Base Systems, 34 Management Systems, - _Perfgrmance, 14 .v Cost & Pficing Elements of C6mputing5Reaources, 12 Evaluation Of Software for lying Systems of Nonlinear= Course.Offerings, 72,48 Euations, Comparative Studies towards the Performance,. 29 Cryptography, 30 9 Evaluation of Vocabulary Cryptography. and Security, 39 Control Techniques, 10 DataBa*e Generation, .Facilities of CISRC, 2 Structured, 39 . ( , k Faculty, 7, 51 , Data BaseManagempt Systgm,'37 ior Formal Language Theory, 20, 21, ? / Data Base'Management'Systems, ?2, 25 ° Performance Evaluation of, 14 d . "Formal Languages;. 46 i. Dates Base%Protection, 33, 18' PoiRTRAN, Structured, 40 Data Base Size, Occurr,epces of -Trailing N -prams as a 'Flinctions, The Structute of Function of, 14 theSet, ot Feasibly [ 2 Computable, 46 1 - Data Base'Systems, 33, 34, 35, 36,. .37, 38, 31 , General Theory of Infot ation Flow apd Analysis, 17. 4 Data Secure System. Archi- / tecture, 36' Generalized Automata, 25 # , . . . . Data ,Secure Sy4tems, 13, Graduate Programs, 6 362.38, 39 Grammathwi h?out'syntactic . EF Data Trankaission,42 ',Var,iablqg,aormal, 21 .. Decision System,, an EMpirical Grammars, Iludan Extrapolation' . 'Analysis of Information and of:"Strings GenerpteA 'by, 15- 4 Relations Among Element's . 4r . . of a,,-16' Grammars-Programilable-High7 Level Language:Machine,°41 DeciSion Systems, Strnctilre'ls. Models for, 18 88 79 o r Grammars, Syntactic Structures Information Systems, Concurrency of Unrestricted, 21, 22 in Real-Time, 34 - Graphical and Semantic Knowledge, Integer Programming) 42 Integration of, Interface Design, 41 Graphics Programming,j,anguage; The AniMa, 32 t44 terfteter, 41 A ON, C group, 30 IPSS, 13 droUpOid:30- % Iterated Networks, 46 GrouptSid String Algo414s, Language, The Affects of Turing Universaliftand Physical Environment on Paralleli§m of, 470 the Interpretation of,, 20 t't 4 Growth of Department of Language 'Machine, Architecture Computer and Information of, 41 Scipce, 5 ,Language Processing for High Level Language Machine,41 utomatic Indexing,,12 _Human Extrapolation ofeStrit* Languages, Protection; 38 Generated by Orderea Cyclic Finite State Grammars, 15- Linguistic AnafiaisA 20 ' 7 Human Infbrmation Processini, 15 Linguistic Analysis of . Suffixes, 14 . p Idempotent, 30. v:w 62c Loop Networks,41. Iderifity Element,.30 I P 4 L -Norm MaxMin.Probiem, 30 Index Entropy, 10 "- .. 'Mathematical Techniqueli, 29 , . Indexing, Automatic,10,.12 1 . ' Measurement of Codt & Pricing Information Analysis,16 'Elements of Computing Rebources, ' Information Flowand Analysis, A General Theory,of, 15. /f Measure Spaces;. 29 . f . InformatIOn Processes in -Memory SnapshOtATrades, 27 Physical: Biolbgical,' and Social System0; 28 Mptric Spaces ;' 29 4 Information-Processing, Human, 35.- :Minicomputer Pictures, 32 Information-Processing Systems 'Mddels for Decision Systemsi, SiEulation, 13 Structural 18 . . Information SeOrageand to. . Multiple Goal' ftogramming, 13 Retiieval, 10 '/ ..va; , 89. t.. r 4.; o 4 80 1. Natural Language Algorithm,,24 Phrase St'r'ucture, 20, 21, 22, 25 Natural Language Proceising, Phrase Structure Grammars, 21 12, 20, 23; 24 c- PhraseStiuctureNLangrges,020, 21 Natural Language Processing for Suffix Atialyais, 14 . Pragmatics, 20 4. Networks, Computerl Pricing Elements'of,Computing Resou'rces, Identifi cat_ion and N-Gram 14 Measurement, 12 2 Nonlinear Systems of Equations, Profile Enhancement, Automatic, 10 Comparative Studies Towards The PerfOrmance Evaluation of Programming Language, The V Software foSolling, 29 - Aaima Graphics? 32 r Objectives of CIS Department, 4' PrOgramming Languages,AFamily of Trans ators for Phrase ObjectiVes.of CISRC, 1 Structu e, 20, Operating Systems, . Protect* languages, 38 Oracular EnuteratiOh Procedures,- Publication7,s of-'CISRC Staff, II 411 Computational Com131exity.Of;r45 61,,'66 --, Orgafization of CISRC, 1 QUasigroup, 30, Ordersf, mputaiional RadarnAircraft Detection and Zomplek of Recursively ClassiflicaAon, Low- Enumefable Total,.45 ,Frequancy Radar, %26 . Parallel Cdmputation, 46 Realtime information ys ems4-7 Concurr-.nt Process$ng, 34,37 Parallel'Execution, ComOilatio 4 ofSequentialPrograffis for, .32 Reports, Technical, 66. a o , - ' ,_./ , Parallel Recognition of. . Retrieval, Infdrmatiqen'Storage e Formal Languages, 46 and,1010 , Parallelism of.Groupoid Scope of.CoMputer and Stting Algorithms, 47 *Information Sciente -Program, 1110 Pattern Recognition, 27 ; SDI Search Erofile Enhancement, tot Per rmance Evaluation of Automatic, 10 Data BaseManagement Systems, l4 Secure Systems, Data, 33; 34, 35; 34 I r Performantg.'paluation,of Security, ', '36, 38, 39 -,-t sSof,tware for, Solvingilph 14 Systems of, Nonlineai Salantic-Directed Translatiori Apationi; 29. of Phra Structure 90 Languaget, 25 . 1 4 8.1 Semantic Information.Proces4g, 24. Switc ing Networks, 46 Semantic Knowledgd, The Syntac4C'S'truOtUes of Integration of. Graphical' Unrestricted G-rit.mmass and,-26 N Syntax-Directed and Semantics, 20, 21, 22, 25 Semantic-Directed Tranalation" of Phrase emantics and Translation Structure Languages, 25 :ft of Phrase-Structure Lguage's, A Formal Syntax-Directed Translations, 2 Theory of the Syntax, 21 Properties of, 23 ..".Seminar Series, CoMputer and Syntax, Semank s and Information Science054 Translati of Phrase- Structure Languages, 421 . Settlem en atterns, 28 ,5tems Programming, 32' Simulation, 42 (. . . Technical Repor 91 66 Simulation, Informatiqn. \, ,Processing Systems,: 13 TelecomMunicati , 42 Simulation Modelling,.28 Uxkfor Schq Practical Algdrithms for /2 -SiMi.hatiOn Madera, 37 a 0 ,,ITopology, 'Binary Strings Space-Deixes, 20 and, or / - SPS, 23' Toroids, 29 Stemming-Recoding Translatt,20, 21, 22, 23, Algorithms, 12 24', 25 'Sterqsiisomers, Chemical;.Generation Translation Of Phrase-Structure s of Uni9ue' ) Langua , AFoimal Theory of. Configurational 11. the S tax, Semantics-and, 21 Representations", 111. Translations, Some Properties of Stotage and Retrieval, Syntax Directed, 23'1 Information, 10 0 ' Translator:4483 . Strings, Binary, 29 r\ I Tsansiators for Phiase Stfucture' .Structure Automata, -25 Programming Languages, A \ , \ A Familyof, 20. NSA-Optimal Set Covers . with Confidence Measures: Tree Automata, 25 ( of Optimality, 29 Turing Universaliti f Subsistence Patterns, 28:- Groupoid Strint Algdrithmsi 47 ',.SUffix*Ialysis, 14 Undergraduate Progratia, 4 \ 4 A . . et or " r 0 82 fr Vocab-ulary Control in. . Printed Indexes, Automatic, 12 Vocabulary Control Techniques, Evaluation of, 10 ,Women and Minorities in ti Academic ComputerSG-nce, 44. 4 4 es . 0 ( -: * 0 4 V AA . 4 . 4 ' V s" e