Ie

Special Issue

on

Thesaurus Software

from

Darmstadt Seminar

August 14,1990 International Classification Vol.17 (1990)N.3/4 UDC 025.4:168:001.4 (05)

INTERNATIONAL Contents CLASSIFICATION Devoted to Concept Theory, Systematic Tennino­ logy and Organization of Knowledge The journal is the organ of the INTERNA TlONAL SOCIETY FOR KNOWLEDGE ORGANIZA­ TION. (General Secretariat: Woogstr. 36 a, D-6000 Editorial Frankfurt 50) Thesaurus Software, Knowledge Organization, and the Human Editors Interface .127 Dr. Ingctraut DAHLBERG (Editor-in-Chief), Woogstr. 63a, D-6000 Frankfurt 50 Articles Dr. Robert FUGMANN, Alte Poststr. 13, D-6270

Schmitz-Esser, W.: Thesauri facing new challenges . · 129 Idstein

Fugmann, R.: An interactive classaurus on the PC . · 133 Prof. Jean M. PERREAULT, The Library, Univer­ sity of Alabama at Huntsville, P.O. Box 2600, Hunts­ RitzIer, C.: Comparative study of PC-supported thesaurus software. · 138 ville, AL 35807, USA Ganzmann, J.: Criteria for the evaluation of thesaurus software 148

Rada, R.: Maintaining thesauri and metathesauri . · 158 Consulting Editors Kenneth BAKEWELL, Liverpool Polytechnic, School of Inform. Sci. & Teehnol., 79 Tithebarn Reports and Communications Street, Liverpool, L2 2ER, U.K. Documentary Languages and Databases (165) - 14th International Prof. Pauline A.COCHRANE, 6568 Clagett Ave., Online Information Meeting, London (165) - Meeting on Data and Fairhaven, MD, 20754, USA Concepts, and Courses in Darmstadt 1991 (165) - Cataloging Heresy : Nathalie DUSOULI ER, lnst. de I'Inrorm. Science et Challenging the Standard Bibliographic Product (165) - Making Techn., Chateau de Monld, F-54500 Yandoeuvre, Connections. ACH/ALLC 91 (166) - RIAO 91. Intelligent Text and Nancy, France Image Handling (166) - ISCCR 91. Call for Papers (166) - Inter­ Douglas FOSKETT, I, Dale Side, Gerrards Cross, Bucks. SL9 7.1F, U.K. national Indexers Conference, Yugoslavia (167) - 2nd Albany Colloquium 1991 (167) - CDMARC Names (167) - Dewey Decimal Jens B. FRTIS-HA NSEN, Hoesterkoebvej 7, DK-2970 Hoersholm, Denmark Classification (167) - New Abridged Dewey Published in June 1990 Alan G[LCHRIST, 38 Ship Street, Brighton BNl (167) - Satija, M.P.: Indian National Terminology Bank (168) - lAB, U.K. Ethnicity Glossaries (168) - Terminology Workshop, Kent, OH, 1991 Prof. DipJ. Math. Winrried GODERT, PH Bib!. u. (168) - ASTM Symposium, Cleveland, 1991 (157) -Terminology for Dok., ClaudiusstLI, D-5000 Kaln 1 Knowledge Transfer, Vienna, 1991 (157) - 3rd IFCS Conference, Dr. M. A. GOPINATH. DRTC, Indian Statist. Edinburgh (147) - Cluster Analysis in Chemistry (132) - International Inst., 8th Mile, Mysore Road, R.Y. College P.O., Conference on Symbolic-Numeric Data Analysis and Learning Bangalore-5600059, India (132) - Classification, Data Analysis and Information Processing (147) Pror. Eric de GROLlER, ISSC, I, rue Miollis, F-75015 Paris, France ISKO News 4 Prof. Dr. JosefHOLZL, lnst. f. Techn. u. Warenwirt­ schaft., Wirtsehaftsuniv., Franz-Klcin-Gasse 1, Satija, M.P,; Tools for Knowledge Organization and thc Human A-1190Wien Interface. Report of the First ISKO Conference .169 Prof. Krishan KUMAR, Dept. of Lib!". & Inform. Workshop Reports at ISKO Conference 1990 . .171 Sci., University of Dehli, Dehli-II0007, India Prof. Yukio NAKAMURA, 14-tl, Nisikata-l, FIDICR News 28 Bunkyo-ku, Tokyo 113, Japan FIDICR Matters (177) - Mcllwaine, I.C.: Task Force for UDC Peter PAULSON, Forest Press OCLC, 85 Water­ System Development (177) - New Publication (178) -MARC Format vliet Ave., Albany, NY, 12206, USA for Classification (178) - McGarry, D.: IFLA Division of Biblio­ Drs. Gerhard 1.A. RfESTHUIS, Universitetsbibli­ graphical Control, Section on Classification and Indexing (178) - otheek, Singc1 425, NL-IOI2 WP Amsterdam, ASIS/SIG/CR (178) 1st ASIS SIG/CR Classification Research Netherlands Workshop (179) - U.K., CRG 276 and 277 (179) Leon ROLLING, CEC, Dir. XIII, .Iean-Monnet Centre, Kirchberg, LU-Luxembourg Thesaurus Software on Offer in Europe . . .180 TermNet Seeks Partner to Sell 9-Languagc Thesaurus Software . 181 Dr. Winrried SCHMITZ-ESSER, Oderfcldstr. 13, D-2000 Hamburg 13 Pro!". Dr. Eugeniusz SClBOR, llNTE, Broniews­ Book Reviews kigo tl3-157,PL-01-tl76 Warszawa, Poland Ranganathan's Reprints in Library Classification (M.P. Satija) .182 Dr. Otto SECHSER, In der Ey 37, CH-8047 Zurich, Meiss, B.: Thesaurus, integrated, online, dynamic. A new way of Switzerland working with DOMESTIC. (R.Fugmann) ...... 183 Prof. Dr. Dagobert SOERGEL, College of Libr. & Meinl, F.: Subject Characteristics. Key to their technical formation, Inform. Serv., University or Maryland, College description and information. (J .Holzl). . 184 Park, MD, 20742, USA Gottstein, K. (Ed.): Dircctory of International Cooperation in Dr. Edouard SUKIASJAN, Lenin State Library or Science and Technology. (I.Dahlberg) . .185 the USSR, Pr. Kalinilla 3, Moskva-lOlOO, USSR Prof. Nancy WILLIAMSON, Fac. of Libr. & In­ form. Sci., 140 St.George Str., Toronto M5S lAI, Classification Literature 17 (1990)No. 3/4 · 186 Canada UDC 025.4 + 168 + 001.4 (05) INTERNATIONAL CLASSIFICATION Devoted to Concept Theory. Systematic Terminology and Organization of Knowledge

Editors: Dr. I. Dahlberg. Frankfurt Or.R.Fugmann, Idstein; Prof. J.M. Perreault. Huntsville, Ala., USA; Edjtorial Office:c/o Dr. I. Dahlberg, Woogstr. 36a, 0-6000 Frankfurt 50, Tel. 0 69 / 52 36 90; FAX 0 69 / 52 05 66 Issue frequency: 4x/ann. Annual subscription rates: Institution per volume OM 94,-; Individuals per volume 20 % less if directly ordered from IN­ DEKS Verlag. Single issues: OM 28,-, Back issues available. ask for special offer. Subscription included in membership fee for members aftha Interna­ tional Society for Knowledge Organization (ISKO). MwSt (domestic only) and postage are not included; for Ail Mail delivery an additional OM 4.­ per issue required. Advertising rates: Advertising rate card No.2. Advert·lsing fact sheet with details of mechanical requirements and closing date available "uponrequest. Publisher: INOEKS Verlag, Woogstr. 36a, 0-6000 Frankfurt 50. Tel. 0 69 / 52 36 90; FAX 0 69 / 52 05 66. Bank Account Nos.: I NDEKS Verlag. Postgiro-Konto Frankfurt. No. 151208-608 (BLZ 500 1 00 60); Frankfurter Sparkasse. No. 852 082 (BLZ 500 502 01 ) All rights reserved. Printed in the Federal Republic of Germany 1989 by Oruckerei H. Guntrum II KG, PF 180, 0-6407 Schlitz/Hessen.

Contents page Vo1.17(1990)N.3/4

Schmitz-Esser,W: Thesauri faciug new challenges. Ganzmann,J.: Criteria for the evaluatiotl of thesaurus software. Int. Classil. 17 (1990) No. 3/4, p.129-132. Int. Classil. 17 (1990) No. 3/4, p.148 -157,23 rels., 1 App.

The chairman of the Thesaurus Software Seminar held on The growing number of thesaurus programs, especially for mi­ August 14, 1990 in Darmstadt, introduces into the topic by crocomputers, calls for criteria which facilitate the decision on asking the following 10 questions and by providing his answers the appropriate tools for thesaurus construction, application to them: 1) What is new In the view? 2) What is the real point of and maintenance. Following a brief outline of the detenninants attraction? 3) Cannot Information Retrieval (IR) profit from of the criteria,i.e. the changing environment and the routines of machine-processing of language? 4) Can we do better now? thesaurus management and application,criteria are discussed in 5) How can we do better? 6) When does fully automatic IR detail.Only the most important requirements relating to central arrive? 7) Thesauri for machine-aided IR - how do we get functions of thesaurus work are touched upon, further criteria there? 8) Which is the right way, which is the model, what to being induded in a check-list attached. Applying· the criteria to standardize? 9) Can IR people do it alone? 10) Are there ad­ existing thesaurus software in a final short assessment, the vanced information servIces with a truly human interface? author concludes that some programs reveal substantial short­ (I.C.) comings. (Author)

Fugmann, R.: Aninteractive classaurus on the PC. Rada, R.: Maintaining thesauri and metathesauri. Int. Classil. 17 (1990) No. 3/4, p.133 -137,6 refs. Int. Classil. 17(1990) No. 3/4, p. 158-164,30 rels.

Both classification systems and thesauri have their specific Maintaining a thesaurus is a time-consuming task which should strengths and weaknesses. Through properly combining both go hand-in-hand with the indexing of information and should be approaches one can eliminate the latter and largely preserve the supported by software. To connect different document data­ strengths. "Classauri" which originate in this well-known way bases their respective thesauri should be related. The most are most effective if they are constructed and applied during straightforward way to support this by computer is to map the computer-aided indexing. A special variety of c1assaurus is de­ terms, of one thesaurus to those of another. Such a mapping scribed which is characterized by the employment of simple but creates one kind of meta thesaurus. As citation systems are ex­ highly effectiveconceptual and technical devices and by the re­ tended to include full-text online, a new thesaurus may be used nunciation of attempts to generate the wording of index entries to index individual paragraphs of a document, and a meta­ algorithmically. thesaurus may apply to a universe of paragraphs. To illustrate (Author) these principles several computer systems are described which help people maintain thesauri anc;l metathesauri. Particular success has been had by the National Library of Medicine with Ritzier, C.: Comparative study of PC-suppoded thesaurus soft­ its Medical Subject Headings and its Unified Medic,al Language ware. Int. Classif. 17 (1990) No.3/4, p.138-147, 1 ref. System. (Author) This article presents the results of a comparative study of three PC supported software packages (INDEX, PROTERM and TMS) for development, construction and management of thesauri and otherwordmaterial with special regard to hardware and software requirements, handling and user interface, and functionality and reliability. Advantages and disadvantages are discussed. The result shows that all three software products comply with the minimum standards of a thesaurus software. After inclusion of additional features distinct differences be­ come visible. Einem Teil der Ausgabe Hegt eine Werbung des INDEKS (Author) Verlags beL

This contents page may be reproduced withoul charge INTERNATIONAL CLASSIFICATION IC Scope Aims

The more scientific data arc generated in the impetuous present Thus, [NTERNATIONAL CLASSIFICATION is meant to be times. the more ordering energy needs to be expended to control a programme for the improvement of classification methods these data in a retrievable fashion. With the ahundanceofkllowl­ and processes, a forum for discussion for all those interested in edge now available the questions of new solutions to the order­ the organization of knowledge on a universal or a subject-field ing problem and thus of improved classification systems, scale, using concept analytical and/or concept-synthetical ap� methods and procedures have acquired unforeseen significance. proaches as weB as numerical procedures and comprising a�so For many years now they have been in the focus of interest of in­ the intellectual and automatic compilation and use of classifica­ formation scientists the world over. tion systems and thesauri in all lields of knowledge, with special Until recenlly. the special literature relevant to classification l:lltention being given to the problems of tenninology. was published in piecemeal I�lshion, scattered oyer the numer­ ous technical journals serving the experts of the various fields. [NTERNATIONAL CLASSIFICATION publishes original such as articles, reports on conferences and similar communications, the Newsletters of FIDICR (Committee on Classification Re­ philosophy and science of science search of the International Federation for Documentation), science policy and science organisation and COCTA (Committee on Conceptual and Terminological mathematics.statistics, and computer science Analysis), book reviews, letters to the editor,.. and an extensive library and information science annotated bibliography of retent classilication and indexing lit­ archivistics and museology erature, covering between 400 and 500 items in each issue. journalismand communication scicnce industrial products and commodity sciencc [NTERNATIONAL CLASS[F[CAT[ON should thererore be terminology.lcxicography and linguistics availl:lble at every university and research library of every country, at every informl:l tion center.at colleges and schools ofli­ brary and information science, in the hands of everybody inter­ Beginning [974, [NTERNAT[ONAL CLASS[F[CAT[ON ested in the fields mentioned above and thus also at every office has bcel] serving as a common platform for (he discussion of for updating information on any topic related to the problemsof both theoretical background questions and practical applica­ order in our information-flooded times. tion problems in many lHeas of concern. In each issue .experts from many countries comment on questions of an adequate [NTERNAT[ONAL CLASS[F[CAT[ON was rounded in structuring and construction of ordering systems l:lnd on the 1973 by an international group of scholars with a consulting problems of their use in opening the information contents of board of editors representing the world's regions, the special new literal ure, of data collections and surveys, of tabular works classification fields and the subject areas involved. From and of other objects of scientificinterest 1974-1980 IC was p.ublished by K.O.Saur Verlag, Miinchen. Their contributions have been concerned with Back issues of 1978�1985 are available from.INDEKS Verlag, too. (Thc 9 volumes of 1978-1986 are offered now at the highly re­ (I) clarifying the theoretical foundl:ltions (general ordering duced price of DM 300.-.) theory, science theorelicl:l1bases of classilicl:ltion,data anl:lly­ sis and reduction) As or August [989, INTERNATIONAL CLASSIFICATION (2) describing p"raclicl:l1operations connected with numerical has become the oflicial organ of the INTERNATIONAL taxonom y /classilica tion, as well as applicl:llions of classi lica­ SOCIETY FOR KNOWLEDGE ORGANISATION (ISKO) tion systems and thesauri, manual and machine indexing (founded on July 22, 1989) and is included for every ISKO-mem­ (3) tracing the history of classilication knowledge and ber, personal or institutional, in the membership fee ($25/$50). methodology (4) discussing questions of education and training in classilica­ Rates: From 1990 on for 4 issues/ann. (including indexes) DM tion 94.00 + postage ( = DM 6.00). For air mail delivery add DM (5) concerning themselveswith the problems of lerminology in 4.00 per issue. Membership rates see above. general and with respect to special fields. [NDEKS Verlag, Woogstr. 36a, D-6QOO Frankrurt 50.

Instructions for Authors Journal Reference.I' should contain the names and initials of all authors, full title oflhe publication, abbreviation of the journal according to the ISO Standard 4, volume number, year of pub­ Manllscripts may be submitted in either English, German or lication in brackets, issue number and lirst and last page num­ French to the editor-in-chief in two sets (the original and one bers. Monograph References should give the name(s) of the copy) typed in double space, comprising between 1500 to 3000 author(s), full title, edition, place of publication. publisher, year words. They should be accompanied by an English indicative ab­ and the number of pages. stract of I 00-200 words. The contributions are refereed. Examples: Mater, E.: Human intelligence as a precondition for Criteria for acceptance will be appropriateness to the lield of the the machine processing of knowledge. Int.Clas� journal (see Scope and Aims), taking into account the tnerit of sir. [5([ 988)No.3, p. [25-[32 the content and its presentati�n. Papers a. re accepted with the un­ Hunter, EJ.: Classification made simple. Aldershol, OB: derstanding that they have not been published, submitted or ac� Gower [988. [[5p. cepted for publication elsewhere and that, ifthe work received of­ Illustrations should be restricted to the necessary minimum. ficial.sponsorship, it has been duly released for publication. Graphs and diagrams should be supplied as black and white Authors will usually be notified within 6 to 10 weeks. Unless drawings suitable to reproduction. Half-tone illustrations specifically requested, man4,Scriptsor illustrations will not be re­ should be sharp, well-contrasted glossy prints, Illustrations turned. should be numbered lightly with a soft pencil on the back, Num� The text of the contributions should be structured by subhead­ bcred legends should be attached on a separ

Thesaurus Software, Knowledge Organization, and the Human Interface We are happy to include in this special issue first of all almost all of the papers of the Thesaurus Software network of thesauri and classification systems and Seminar held on Au g.14, 1990 at Darmstadt, immediately providing a tool that might serve not only one center preceding the First International ISKO Conference; or one country but probably the entire biomedical Sponsored by the Committee on Classification and discipline of the world - if it will be completed success­ Thesaurus Research of the German Documentation fully. The final paper by D.H. FISCHER, GMD Darm­ Society (DGD/KTF) and organized by the ISKO Secre­ stadt, (Modelling thesauri on the basis offrame systems) tariat, Frankfurt, this seminar was keynoted by DGDI could not be included in Ie 90-3/4, but will appear in KTF Chairman Winfried SCHMITZ-ESSER (who does one of the next issues. - We have, however. also one not always look as serious as in the picture on the fol­ book review of relevance in this issue namely the one lowing page) with an address on the topic Th esauri by Robert FUGMANN on the DOMESTIC Thesaurus Fa cing New Challenges. He chaired the seminar as well Software as described by Brigitte MEISS as well as the as its concluding panel on Experts answer your questions note about the Term Net Thesaurus Software in pr�para­ which he started out with a survey of European vendors tion to include 9 languages. of thesaurus software, listed on a handout given to the The second item in the headline of this Editorial, seminar participants and reprinted in a slightly updated Knowledge Organization, refers to the two comprehensive version in this issue, too. Since one of the previously reports in this issue on the First International ISKO announced speakers at the seminar, who had intended Conference in Darmstadt, August 14-17, 1990, which to provide a comparison of European thesaurus software cover (1) the Conference as such and its 12 plenary packages, had withdrawn, this survey was considered sessions, and (2) the short reports on each ofthe 12 work­ very helpful for each of the numerous participants in shops, submitted by the respective chairpersons or the seminar. Some of the thesaurus-software suppliers their representatives. Besides this, everybody is invited had also been able to come to the seminar and exhibit to orient himself on the contents of these sessions and their thesaurus programs in the halls of Darmstadt workshops through the two Proceedings volumes, one Technical University, where the conference took place. of which was already available at the start of the con­ The second paper by Robert FUGMANN, Idstein, ference, while the second one will be out in November Germany, went right into the practice of thesaurus­ 1990, and can be ordered from INDEKS Verlag. With making (An interactive classaurus on the PC; practical these proceedings a new series is started: Advances in experiences) using the LIDOS software. whose function­ Know/edge Organization! ing was demonstrated by Mr. Fugmann on his own PC A most interesting statement at this conference was later on. In the third paper, Jochen GANZMANN, made by Roland H)ERPPE, Head of LlBLAB, the Frankfurt, presented a brilliant overview of Criteria unkoping, Sweden, University's Department of Com­ for the evaluation of thesaurus programs. In this connec­ puter and Information Science, in his excellent paper tion, Loll ROLLING, Luxembourg, mentioned the presenting A Framework for Characterizing Systems fo r work done in 1986 by the Bureau Marcel van Dijk Knowledge Organization. He pointed out that the (through Mr. Georges van Slype) concerning an analysis term 'System for Knowledge Organization' (SKO) of the characteristics of thesaurus programs. A copy of was strictly speaking a .misnomer, as knowledge itself the paper concerned is still available from his office at cannot be organized, only (re)presentations of it. But the European Communities. With the fourth paper - since language and perception might be regarded as which concluded the morning session - Claus RITZLER, some of the most basic SKOs, whatever is built on Hamburg, presented a Comparative study of (German) them might likewise be regarded as an SKO. One should thesaurus software in which two German and one be aware that knowledge takes many forms, and that English thesaurus program were described and their the focus in research on classification and indexing respective merits pointed out. has largely been on knowledge made explicit in verbalized The paper replacing the retracted one mentioned form. The special SKOs considered in his paper are above was presented by Roy RADA, formerly with the classification systems and thesauri. However, the pertain­ U.S. National Library of Medicine (NLM) and now at ing deliberations could also be applied to systematic the University of Liverpool, Computer Science Depart­ nomenclatures and taxonomies, to encyclopedias as well ment. In Maintaining Th esauri and Metathesauri he as to the conceptual schemata of data bases. Even the offered his insights gained and experiences gathered in knowledge representations used in the knowledge bases the establishment of the UMLS, the NLM's Unified of, say, expert systems, might be regarded as SKOs Medical Language System, thus enriching the seminar despite the fact that their emphasis - so far - has been with the aspects of thesaurus software needed for a more on representation than on organization.

Int. Classif. 17 (1990) No.3/4 Editorial 127 It is somehow strange with our terms: When founding them. This first ISKO Conference had many of such ISKO and looking for a name for our new Society, we surprises to offer its participants. But, God willing, there had the German term Wissensordnung in mind which are more of such chances to come. See only the an­ would be translated "into English as Knowledge Order­ nouncements and Calls for Papers in this issue! Already ing. But we learned that one could not use the term in IC 90-2 we had the Call for Papers for the FIDICR ordering of knowledge because of the connotation Conference in Toronto, ]llne 1991, on the back inside to purchase knowledge. Thus, fully aware of the fact cover page. But there will also be an ASIS/SIG/CR that knowledge cannot be organized, but only knowl­ Workshop in Toronto on Nov.4, 1990 and another ledge units, viz. concepts, we agreed to use Know/edge conference in Rome, Italy, on Dec.2-4, 1990 on Organization in order to be understood in the English Documentary languages. Dean Halsey from the School speaking world. Now, this term Knowledge Orga nization of Information Science and Policy of SUNY at Albany, is also used by the Artificial Intelligence community, NY, plans to organize another Albany conference on maybe in a somewhat different Sense according to the classification in November 1991. With this issue, how­ understanding of their members. Nevertheless, we ever, we prepare already somehow for the next ISKO probably mean the same thing and must try to cooperate Conference in India, 1992, where the Ranganathan birth in order that both directions will profit from their centenary is to be observed and celebrated. The big UBS methodological knowledge, viz. concept analysis, con­ Publishers's Distributors company in New Delhi started cept relationships, and concept representation in adequate recently to reprint Ranganathan's books. A review on concept systems on our part and computerization of these reprints in classification, -�lready available, has knowledge bases and its technology on the part of our been made by M.P.SATI]A and is likewise included in AI colleagues. this issue. Thus, with Ranganathan's books again available The HUman Interfa ce, our third topic here, although for everybody, we w'ill all get a chance to return to the a technical concept and ,term, acquired a new connota­ beginning of modern classification theory and see and tion at the conference - as one participant wrote: "I am evaluate for ourselves how it can teach us to deal with looking forward to our human interface !". Is it not so: our present problems in possibly more adequate systems just as a microfiche or a computer screen will never for the organization of knowledge than we have had so replace the good feeling of holding an interesting book far. in one's hand and the joy to read its amusing, enlighten­ Thus, the Darmstadt events and their follow-ups will ing, exciting, stimulating, or enriching contents, in the keep us busy - which is good and necessary. Let us hope same way a computer conference will never replace the and pray that no international conflict will disturb or joy of the very personal eye-to-eye encounter of colleagues interrupt what has been started with so much enthusiasm, often knowing each other so far only through the good will, and promise for a better future of work and literature and now suddenly seeing the cherished object understanding in our field of concern! of their imagination in a human face and person before Ingetraut Dahlberg

Tools for Knowledge Organization and the Human Interface

Proceedings of the 1st International ISKO Conference, Dannstadt, 14-17 Aug. 1990, organized by the International Society for Knowledge Organization (ISKO), edited by Robert FUGMANN. Vol.I, 280 p., IS8N 3-88672-020-9, PM 56.- (for ISKO members 25% less)

Vo!.1 contains the first part of the almost 60 papers presented at the said conference with the abstracts of papers published in proceedin9s vol.2. Among the plenary session papers of this volume are those by B.C.VICKERY; R.HJERPPE, K.Markey DRABENSTOTT and H.ALBRECHTSEN. Workshop topics cover: General issues. Algorithmic text analysis. Terminology. Knowledge organiza­ tion by universal and special systems. Thesaurus issues. Online retrieval. �permedia. Retrieval technologies. and Indexing. VoI.2 will include the rest of the wotkshop papers and the plenary papers by Ch.R.HILDRETH, M.FUJIKAWA, A.GILCHRIST, and O.SECHSE� as well as the confer­ ence summary by E. de·GROLIER and the index to both volumes.

With this volume the INDEKS Verlag starts a new series: 1JAdvances in KnOWl­ edge Organization" (ISSN 0938-5495). If you are interested in subscribing to this series you are invited to indicate this

Int. Classi!. 17 (1990) No.3/4 128 Editorial Winfried Schmitz-Esser Information Systems Consultancy, Hamburg

Thesauri Facing New Challenges

First Generation of Software Finally Hit the Market, Second Needs Specification - 10 questions, 10 answers to a topic of renewed interest -

Schmitz-Esser, W: Thesauri facing new challenges. Of course, it is the micro to which we owe the (1990) Int. Classif. 17 No. 3/4, p. 129 -132. first bunch of cheap, readily available thesaurus software The chairman of the Thesaurus Software Seminar held on packages. It is the micro which opened a hitherto August 14, 1990 in Darmstadt, introduces into the topic by unknown perspective of a thesaurus software which asking the fo llowing 10 questions and by providing his answers could oe used as a tool for JR in the same way as text to them: 1) What is new in the view? 2) What is the real point of · attraction? 3) Cannot Information Retrieval (IR) profit from processing packages are used as tools fo� computer machine-processing of language? 4) Can we do better now? editing. And it is the micro which, by its universal use 5) How can we do better? 6) When does fully automatic IR and ubiquitious availability, underscores the need for arrive? 7) Thesauri fo r. machine-aided IR - how do we get application in much broader fields than just STI. there? 8) Which is the right way, which is the model , what to standardize? 9) Can IR people do it alone? 10) Are there ad­ Some of the main lines of micro-based systems are vanced information services with a truly human interface? being reflected in the few samples of exhibits on show. (I.C.) On a whole, it can be stated, that, as a result of the micro, thesaurus software has become as ubiquitious as the micro itself, and that prices of software are low. Micro-based thesaurus programs, as well as some After a period of rather slackening attention, thesauri systems demanding a more powerful IR environment as used in Iriformation Retrieval OR) are back in the will be demonstrated and discussed in this meeting. focus of interest again. The Seminar we are going to Later on in the Congress, some more special questions stage today intends to pay tribute to that finding. will be discussed, like superthesauri and cognitive Much has changed since the early days of computer aspects. application in the field of information when the Thesau­ Nevertheless, all the world-wide, ever-growing activities rus appeared as a key problem solver in IR, and was together, based on such systems, do not expl.lin the regarded - and usually kept - as some sort of Holy Grale current thesaurus renaissance. There is more about it. in the respective systems. These thesauri typically dealt with Scientific and Technical Information (STI), were constructed and maintained by manual procedures, printed on paper, and at best could be integrated in the retrieval systems proper by customized, non-portable software. Question No.2: What IS the real point of The fact that today a seminar on thesaurus software attraction? It is Conceptual is taking place shows at least that such software must be Structuring. This is needed for available now. Ten or fifteen years ago we could at best AI , and Language Understand­ have had a.seminar on thesaurus construction and maintenance. In the middle of the seventies, the only ing in particular. special software program on sale for such purposes was Principles applied in thesauri increasingly are becom­ an IBM program called TLS (Thesaurus and Linguis­ ing of interest to solve problems in computer li;1guistics, tic System), a rather heavy and - at that time - horribly or, more precisely, Linguistic Engineering (LE), which is expensive system which could only be run on IBM a special branch of applied Artificial Intelligence (AI). mainframe machines. LE is the now fashionable name for machine or machine­ aided language processing. In this context, IR represents just one (and possibly not the most important onc) out of a series of target areas of LE R&D, such as Natural Language Interfacing, Question No.1: What is new in the view? It is Machine Translation (MT), Machine Abstracting, Speech Recognition, and Language Understanding. not just the Micro. This does not mean that existing thesauri as of today It would be too simple to say that, at a later stage, it can be used right-a-way in a single one of these new areas was the advent of the micro-computer and its stunningly of machine or machine-aided, language-related problem growing capabilities which changed this all. solving.

Int. Classif. 17 (1990) No.3/4 Schmitz-Esser: Thesauri -new challenges 129 Yet, in traditional thesauri applications for IR it Before turning to my 4th question, the rather rough could be observed that thesauri tuned to the needs of sketch of the situation outlined above needs some more very special domains of knowledge, and very special IR detail and refinement. tasks, performed better than those in broader domains When I referred to "a term's meaning", this'may have and general-purpose environments, been typical of an earlier state of the art, suggesting that The real point of the newly-risen interest in thesauri it is viable that a given term can be defined as the bearer are some of their main functions, such as: of all its possible meanings, and that these meanings can -- Mapping of a term's meaning (semantics) be adequately expressed by other terms of the natural Mapping, and selective definition, of semantic and language. other rclations between different terms of a natural After more than two decades of discussions on language "uniterms", composite terms, noun phrases, preposi­ (Both is needed to establish a conceptually defined tional and other logics, as well as all sorts of term space in terminology-based systems - a prerequisite to frequencies in texts and text collections, it has become language understanding and ordering systems) clear, that properties of terms and term use in more Establishment of Preferential Terms, and, along with complex statements like phrases must be given much it: more attention, and that considerably higher grades of 1) of term use in closed language definition must be applied, if the goal of machine­ systems operability of natural language text in IR is to be a­ 2) Improvement of predictability of term use in such chieved. systems The requirement then draws very close to what is 3) Definition of semantic and other relations be­ needed in MT, and consequently, a thesaurus then tween preferential terms (system language) and probably is about to look very much like a machine­ terms of the natural language operable dictionary. At any rate it" is clear that the old 4) Definition of semantic and other relations between type of thesaurus would not fit for the purpose, preferential terms (system language) and terms or It should not be forgotten that IR since its very values used in artificial languages, like query beginning (and this remains as a heritage from the languages, or ordering systems, such as notations pre-computer era), has always been working with the and classifications, statistically or otherwise assumption that an abridged terminology only was defined terms and expressions, needed to "just find the document", and that it was _. These functions with regard to terms of more than definitely not anything comparable to the termin'ology one single natural language, of the much more refined one-by-one�procedures to In a rather rough statement one could say that it is be applied in other fields of computer linguistics, like now the linguists, who on the one hand, need thesauri, MT. or thesaurus-like conceptual structures, to solve their It went without saying that indexing then also should problem of meaning, i.e. the problem of language be less complicated, and cheaper, than lingustic one-by­ understanding, whereas the IR systems designers, in one-processing. turn, finally came to know that basic LE is required As an example, classic thesauri do not make a dis­ in their systems to corne up with more efficient, intelli­ tinction between plurals and singulars, they just know gent, machine-aided IR systems. the singular form (with some few exceptions, as a rule). Moreover, thesauri up to now have been designed Verbs, the very lifeblood of natural language systems, for use by human indexers and data bank searchers. don't occur in them. Among the many semantic rela� This concept is being challenged by the machine since tions that exist between single terms and the different some time, ways of expression, just some of the more important ones are considered, like hierarchy, partitive, resem­ blance. Only very few among the thesauri applied ?n Question No.3: If LE can profit from conceptu­ practice feature sy,ntactic rules in a proper sense, the al structures as are thesauri, majority of them being geared to the rather poor con­ can't IR profit from machine­ ditions of coordinate indexing. It is true, that, for the sake of better predictability - processing of the language? on this we will have to speak later on in this Congress - The answer is: Yes. with all this stripping in our thesauri, we effectively Computers have become capable of handling much brought down the articulative power of the artificial larger and much more complex relational systems, and IR languages to almost zero. therefore can be expected to brush up themselves with Of course, it is understandable, why we took this good or fair chances of reasonable results. This suggests approach. It earned us mllch criticism, and left Free that it may now be possible to leave to a human specialist's Text procedures appear more attractive. However, intervention only the more tricky cases, - interventions and this is: like disambiguation, word selection, etc., which then would have to be done by interaction. Question No.4: Can we do better now? The So, the question being posed today is this: How must a machine-operated thesaurus-look like, and how can it answer is: Yes, we can. be built and maintained , if it is to meet the needs of In the light of the new requirements and systems such machine or machine-aided natural language process­ possibilities, this approach must and can be corrected. It ing - among others: IR? can give us a better grip on information and warrant an

Int. Classif. 17 (1990) NO.3/4 130 Schmitz-Esser: Thesauri - new challcngcs improved organization of knowledge contained in our are needed for which purposes of IR, e.g., indexing, documents. query formulation, abstracting, text or statement Almost twenty years after its introduction to the selection, etc. An interface must be agreed upon to information world, it has become obvious that the early, enable the IR machine to successfully brush up in the naive "Full Text" approach has not kept its promise. thesaurus entries of the dictionary. Both, the IR Nevertheless we have to state that it, has made heavy machine as well as the natural language text processing inroads. Highly specialized, ever more expensive labor machine, would refer to the same dictionary, whereas needed for tiring indexing tasks had come under heavy the IR machine would start working on the basis of pressure from ever cheaper, fully automated text index­ the text analysis results rendered by the natural language ing procedures, with electronic typesetting as the real text processing machine. The IR machine then would trigger. The result is known. The tacid reserve, if not control the different interactive processes necessary to open rejection of the electronic "Full Text" collections solve the IR task. by their prospective users is a matter of fact. Garbage Or, (2), an interface to the LE machine is agreed databanks widely have discredited the young informa­ upon to condition both, the handling of the thesaurus tion services market. entries by that machine, as well as the processing needed for the different functions of the IR system, in con­ Question No.5: How can we do better? We must junction with other dictionaries needed by the LE ' machine-process the natural machine. language text to condition it for Another (3) way may consist of an integration of a thesaurus in other more special text processing systems, JR. like Hypertext. The thesaurus then would guide the Indexing has to be reconsidered. The integral text in a system in choosing or proposing the appropriate terms machine-readable form is there. The computer power to for referral to the source texts. This also raises the swiftly process even very large corpora of text is there. question of where to instal the interface, and which Since expensive, highly-trained intellectual labor remains properties are needed. scarce, it is a must that it be concentrated on true That (4) a thesaurus structure can be formulated problem-solving in machine-aided, interactive procedures. in a frame system as used in a more general AI environ­ No longer can it be wasted for routine jobs and repetitive ment will be demonstrated later in this Seminar. work. The machine is powerful enough to take over these functions. In the now-emerging post-"Full Text" Question No.8, Which is the right way? Which era, the integral text, or parts of it, remain the basis of is the model? What to stand­ processing in so far as it is natural language text that ardize? If we only had the is treated, but this natural language text must be processed by methods of true LE. Playing around at random with answer to this! single words the candid way should stop. The language Most of the thesaurus software on sale over here engineered programs must get the appropriate tools to in Europe conforms with the traditional pattern of do the basic indexing in the best possible way. The tools thesauri as outlined in the classic works by Soergel, in question, whether integrated in other dictionaries or Wersig, Aitchison/Gilchrist as national and international not, are the thesauri, and these will be the Th esauri of standards, which, by the way, were only recommenda­ the Second Generation. In them, the IR language will be tions. They were available at a relatively early stage of much more refined. computerization in the information scctor. Such thesauri will also be useful to solve other It was this Tb esaurus Committee, the sponwr of this essential functions in IR, like machine-aided query meeting, which - as early as 1965 - started out with understanding and formulation, machine-aided abstract­ defining a thesaurus and drafting a guide for the con­ ing, machine-aided text evaluation and selection. struction of thesauri which later on was compiled by D.SOERGEL in his fi rst book on thesaurr(1969, in Question No.6: When does fully automatic IR German) whicb successfu lly laid the ground fo r arrive? Probably never. today's standardization. Above all, we have to give in to the evidence that the After some earlier work, the German pre-standard DIN 1463 was open for public discussion in 1972 and best we can achieve in this effort is machine-aided IR. We should write off all hope to see "Fully Automatic" emerged as a full-fledged standard in 1976. "Concepts IR systems .successfully at work, at least for the next ten and Terms, General Principles" (DIN 2330) were pre­ sented in 1974, and "Systems of Concepts and their years or so ... So, thesauri must be made fit for that purpose. Presentation" (DIN 2331) in 1976. All major standards, including AFNOR Z47-100, BS5723, 1979. ISO 2788, Question No.7: Thesauri for machine-aidedIR - all on monolingual thesauri, had been issued up to the how do we gft there from second half of the seventies. Thc discussion on multi­ lingual thesauri was on in the middle of the seventies bere? (ISOITC46/WG5). and "Rules for Building Multilingual There are several different ways to achieve this goal. Thesauri" (ISO 5964) was out in 1977. Either (1)' the thesaurus with all necessary definitions Nevertheless, it took some years for the first packages and relation work is made part of a larger lexicon which of isolated software to appear on the market. Admittedly, serves the broa.der and more basic LE tasks, like parsing, the overall development in computer hardware had some or MT. In this case, it must be defined which entries influen:::e in this, but the process shows something lnt. Classif. 17 (1990) No.3/4 Schmitz-Esser: Thesauri - new challenges 131 about how much time it takes for a standard to come to all appropriate groups or bodies dealing with the question application in readily available market products, and it of how to normalize dictionaries, terminology formats, also testifies the overall importance of standardization in etc., and it intends to discuss with them what thesaurus this very special field. theory and practice can offer to machine-aided IR in LF. A ncw round of standardization is now required, if environments. second generation, machine-aided IR one day is to be It is obvious that thesauri of that kind will also be triggered off. The process of elaboration of the new useful for other tasks of LE, such as speech recognition, thesaurus standard could follow the lines exposed above. and text understanding. With our knowledge of what can Since the IR procedures to be affected by such standards be achieved by means of thesaurus systems, we may will be highly automated, and will cover large areas of be in a position to stretch out a helpful hand to what application, substantial investment is at stake. The until now was at best a neighbouring discipline, - in money would not flow before the standards are there. exchange for other basic language technology which we feel is badly needed on our side, and of which we know Question No.9: Can the IR people do it alone? it is obviously at hand. The answer is: No. Teaming up with specialists from other Question No.10: Advanced information serv­ disciplines is necessary. ices with a truly human inter­ face? Not without thesauri. But before working on standards, we must know what we arc going to standardize. The Thesaurus Com­ Most certainly, machine-aided LE will play a major mittee wants to discuss this with all interested and role in overcoming the language barriers in our future, potentially important partners. Somebody would have European Single Market, which, by the addition of the to come forth with a model, or a set of possible models, Eastern countries, will appear even more Babylonic from which practical work can start. We know that the today. The availability of native, natural language computer linguists are faced with a similar problem in information services in the different member countries, their search for machine-operable, standard lexica. We including also the smaller ones, will become a vital suggest teaming up with them to discuss both our issue in this context, and it is safe that this cannot occur matters. We are eager to present our case and not to let without thc availability of appropriate multi-lingual that chance pass by. thesauri of the second generation. Early in the seventies, in �ur fo otnote on the validity As the present chairman of the German Committee of our Th esaurus Guidelines, we wrote: "This stand­ for Classification and Thesaurus Research lam particularly ard is not valid fo r the constrllction of thesauri in the glad you all came here to participate in this Seminar, sense of Linguistic Science (e.g. synonym diction­ and in the name of all our members I give you a warm aries)". welcome to this meeting. We will be most pleased to We arc now discussing the new challenges which lie hear your comments and suggestions, and we are hopeful ahead, and we are prepared and willing to help elaborate that at the end of the day we will know better how and and promote the models and standards needed for lR in how best to reach our common goal. more complex LE environments, being fully aware This common goal, needless to say it, is: Adequate of the fact that all this, of course, is linked to, or part tools for improved, intelligent, machine-aided, mono­ of, the Linguistic Sciences (as well as it is part of other lingual and multilingual IR. disciplines, like logic). Dr.Winflied SCHMITZ·ESSER ' ' The Committee will be trying to establish contacts to Odel'felder Str. 13, 2000 Hamburg 13

International Conference on Symbolic - Numeric Data Analysis and Learning From 17-20 Sept. 1191 The Institut National de Cluster Analysis in Chemistry Recherche en Informatique et en Automatique (INRIA) will hold its next conference at the Universite Paris The 1990 meeting of the British Classification Society Dauphine with English and French as conference languages. will be held at the AFRC Institute of Feed Research, Proposals for papers (in four copies) of 12 pages max. Shinfield, Reading, on Oct.2 3, 1990. It is a joint meet­ should be submitted by Nov.30, 1990. The topics include : ing with the UK Chemometrics Discussion Group and Data Analysis, Machine Learning and Modelling; Cluster­ the Multivariate Study Group of the Royal Statistical ing ordering, distances; Representation, analysis and Society and covers the topic. "Cluster Analysis in Chem­ synthesis of symbolic and numeric knowledge (structur­ istry". The following five papers will be presented: Nick ed, noisy, uncertain, etc.); Symbolic - numeric induction, BRATCHELL: Review/Tutorial. - Simon PACK: Applica­ knowledge acquisition from data; Formation and recogni­ tions of cluster analysis. � Mandy PARSELL, Steve tion of conceptual struCtures: discovery of laws, rules, ELMORE: Cluster analysis for sample selection. - Mike inheritance trees, decision graphs, lattices; Neural ADAMS; Application of cluster analysis to infra-red. - aspects ; Coherency, stability, and validation of results; Dave LIVINGSTONE: Applications of cluster analysis Software, and Applications. For further information in QSAR and molecular modelling. - For further informa­ contact: INRIA. Service des Relatlons Exterieures. tion contact: Dr.S.E. Hitchcock, Secretary, Brit.Classif. Domaine de Voluceau - BP 105 - Rocquencourt, F-7815 3 Soc., The Open University, Faculty of Mathematics, Lc Chesnay Cedex, France. Walton Hall, Milton Keynes MIG 6AA, England.

Int. Classif. 17 (1990) No.3/4 132 Schmitz-Esser: Thesauri - new challenges Robert Fugmann

An interactive Classaurus on the PC

Fugmann, R.: An interactive classaurlJs onthe PC. out on a search for literature on general concepts. The Int. Classi!. 17 (1990) No. 3/4, p. 133 - 137, 6 refs. most prominent feature of this "index language" must Both classification systems and thesauri have their specific be that it supp lies modes of expression for concepts which strengths and weaknesses. Through properly combining both are more predictable than those of uncontrolled natural approaches one can eliminate the latter and largely preserve the strengths. "Classaufi" which originate in this well-known way language. We base this claim on a more recent definition arc most effective if they are constructed and applied during of "in exing" a d " �dex lan uage", in which re re­ � � � � 2 conlputer-aided indexing. A special variety of c1assaurus is de­ sentatlOnal predIctability constitutes the core , 3, p. . scribed which is characterized by the employment of simple but B highly effective conceptual and technical devices and by the re­ Such an index language always consists of a vocabu­ nunciation of attempts to generate the wording of index entries lary , which may be a thesaurus or a classification system. algorithmically. An index language grammar, and expecially the syntax (Author) of such a grammar, is rarely used although its employ� ment could serve to keep the vocabulary small and trans­ parent. Grammar could thus contribute much to the re­ 1. Introduction liable use of the vocabulary and, hence, to the quality When searching for literature relevant to a certain con­ of retrieval. We shall refer to this topic later. cept, one must know in advance how this concept might have been expressed in the search file. All these modes 1.1. Strengths of classification systems of expression must be phrased as alternative search par­ It is inherent in classification systems that they use sys­ ameters, regardless of whether one uses a printed index tematic notations as descriptors Figure 1). Similar or a mechanized information retrieval system. (d. notations are thus assigned to related concepts. Hence, In our natural, uncontrolled language general con­ the phrasing of a general query, which is to include all cepts are encountered in an unlimited variety of expres­ the subordinate concepts of a concept in question (e.g. sions and, hence, it is unpredictable, which of the COI1- more than one million species of insect) is facilitated or ceiveable ones might have entered the search file (cf. Fig. even made possible. 1). Either Furthermore, it is possible in classification systems to * they are encountered there in the definition-like,par­ coin descriptors for important concepts, namely nota­ aphrasing, "non-lexical" mode of expression I tions, long before a corresponding lexical expression or emerges in natural language. In this respect, classifica­ , a general concept comprises so many subordinate tion systems can be more up to date than a thesaurus. concepts , that they cannot possibly be enumerated For example, we know several causes for the slipperiness as search parameters and be processed by a computer of a road. But besides "aquaplaning" there are so far no program. lexical expressions such as "oilplaning", "lcafplaning" An example of the first phenomenon is "fear of enter­ or "c1ayplaning". In a classification system such descrip­ ing any kind of sea vessel" or " ...felt a pathological fear tors could easily be introduced. of sea voyages ..." etc., which is equivalent to the lexical Classification systems, especially the faceted ones, are expression "thalassophobia". A country, a continent or conceptually transparent and easy to view because they a landscape can be implied or expressed by so many ex­ are categorized according to a set of predetermined, con­ pressions (spe.cific towns, rivers, mountains, people, ceptual categories. Text analysis is thus rendered more landscapes etc.), that it is again impossible to enumerate reliable (predictable), and a useful guideline for the fac­ them comprehen1iively and to use them as alternative toring of composite concepts is provided, apart from search parameters. several other significant advantages. In our natural language only individual concepts are Furthermore, in a classification scheme the concepts normally expressed in the lexical, i.e. non-paraphrasing, that are subordinate to a more general concept can be manner and in lingual monotony and, hence, in a fairly grouped most lucidly according to which character of predictable manner. subdivision is reflected in them, Characters of subdivi­ These are the reasons for which we need a language sion render a vocabulary particularly transparent and other than the natural uncontrolled one, when we set thus facilitate (or even make possible) the search for the

Int. Classi!. 17 (1990) No.3/4 Fugmann: Classaurus on the PC 133 LANGUAGE

Uncontrolled Language Controlled Language (technical and colloquial) (Index Language)

• Lacking in representational • Improved representational predictability, especially predictability for general concepts for general concepts • Well-suited for retrieval • Unsuitable for retrieval of general concepts of general concepts

• Suitable for retrieval of individual concepts

Vocabulary Grammar of descriptors esp. Syntax (complement to the vocabulary in an efficient index language)

Classification Thesaurus

• Systematic notations • Associative concept relations

• Lexicalization tndependent of • Mnemonics natural language • Timeliness • Perspicuity by categories and by • Hospitality

• logical subdivision

Classaurus

• •

• •

• •

• •

Fig. 1: Typical features of various kinds of language

AppateL . P5elldoc:lc,cr�rto.... - apparel by useR �<. c:------tCl."" "O"'IC; -- Apparel for chIldren ki.r... rch :t -- Apparel for hldles -- Appacel for men

apparel by IIliltetiaL ., ______-- Appatel In cotton Appatel In llynthe ticS --- Appl --- ApplS --- App155 Puvooduuipto'j --- Appl6 - apparel;pattS t------po.�tail/. ---- lIpp16 by charactetIstlc A -- Belts hic.f'o-r<:k,Y App161 -- Buttons ----- Appl615 -- Collats ----- App16155 -- Lining ----- Appl616 -- Zips ----- AppH2 lIpparel/ptoduclnG --- App2 -- Hanufacturets --- Apparel in othet lIynthetic material -- Sewing machines Apparel In wool -- Tailors

- p ______�� by body parT _____ ���! �< _-'- - apparel/sell inG •• ;---,--- " -- Apparel pr izes Gloves -- Appatel sellets ShirtS -- Appatel IIhops IIhlrtll by length of sleeveS � Short aheved shirtS - apparel/CleanInG f(------T- shirts -- Laundries " ----- Pull- on shltts Stain removers ----- Le Isute shJrts Appatel washinG ---- Long sleeved shJds --- Apparel Detergents Stockings Apparel drying

Fig. 2: Model hierarchy with taxonomic, partitive, and associative relations

Int. ClassiL 17 (1990) No.3/4 134 Fllgmann: Classallflls on the PC most appropriate descriptor, which should always be mitted and in part already implemented. Examples are ,, performed by the indexers and questioners. "Thesaurofacet 4 and "Classaurus"s,6. A specific var­ Vocabulary transparency (and the indexer's com­ iant of classaurus is described in the following. It is re­ petence in the subject field) is crucial to reliable indexing stricted to simple but particularly effective conceptual and, hence, to the quality of the searches. If, for ex­ devices and it utilizes an efficient program for an efficient ample, we are informed of a case of "illegitimate use of personal computer. a pharmaceutical of the roborantia group with the aim of winning physical superiority in sport competition", of then the descriptor "doping" is more appropriate than 2. A new variant classaurus the descriptors "superiority", "sport" or "competition", It is specific to our variant of c1assaurus that , it is grouped by conceptual categories, ' although just these descriptors are even suggested * by the wording of the text. it works with characteristics of subdivision and these It is the task of an orderly, systematic arrangement of appear in the vocabulary as "pseudo-descriptors", , the vocabulary to lead the indexer and searcher from it contains systematic notations at those locations in less appropriate, less specific descriptors to the most ap­ which this is desirable, propriate ones, even if the less appropriate ones may be * it permits the presentation and mechanized usage of suggested by the wording of the original text. In other any kind of associative relations. words, A section of a model classaurus of this kind is depicted any successfu l search in a file of documents is preceded in Figure 2. by another search, namely by one in the file of de­ These features make possible "mechanized grouping" aip tors. Any fa ilure in or even omission of the of related concepts even in those cases in which the corre­ search on this level will inevitably impair sponding descriptors have no string of characters in com­ the quality of the subsequent search on mon, as is usual in natural language. No "manipulative the document level. grouping" is necessary, i.e. the (sometimes very elaborate or even impossible) exhaustive and explicit enumeration of all the subordinate descriptors under a 1.2 Strengths of thesauri more general descriptor for a concept of the inquiry. One of the strengths of a thesaurus is the possibility Any kind of relationship, the partitive and associative of expressing the entire set of the associative concept re­ included, can thus be expressed and used as a search lations and not merely the hierachical relations. parameter. Furthermore, an indexer or inquirer is more familiar To consider a characteristic of subdivision as a kind of with the natural-language descriptors of a thesaurus than descriptor is justified because such a pseudo-descriptor with the notations of a classification system, the meaning in fact represents a general feature of all its subcrdinate of which must in most cases be looked up. concepts, namely the kind of difference prevailing among In those cases in which natural language has already them. For example, "apparel by user" collects all con­ coined a term for a new and important concept, it can cepts in which some specific statement about the kind of immediately be adopted in the vocabulary, and no user is expressed or implied. special decision about the phrasing of such a descriptor The kind of notation depicted in Figure 2 permits the must be agreed upon, in contrast to classification sys­ insertion of any number of concepts at any location in tems. an arra/. This prevents hierarchies from becoming This possibility, however, should only be used with chaotic in the course of time, when new concepts have caution. It is the precombining descriptors, i.e. descrip­ to be inserted and when no appropriate location is avail­ tors which comprise the meaning of two or more descrip­ able forthem (the end of the array is not always the most tors, which are often too readily introduced into the appropriate location). This notational device is similar vocabulary. The transparency of the vocabulary will de­ to that used in Dewey'S Decimal Classification and is cline concomitantly with the proliferation of these de­ easily understood if the notations proposed for this pur­ scriptors, as will the reliable use ofthe vocabulary by the pose are considered as decimal numbers with the pre­ indexers. ceding (and here omitted ) string "0." Through the re­ It should be kept in mind that any vocabulary and a currence of a common string of characters in all concep­ classaurus, too, constitutes only half of an index lan­ tually related decriptors they are also susceptible to guage . If an inforniation system is large and is incessantly mechanized grouping. growing with respect to file size and use frequency, then Practically unlimited hospitality in chain (i.e. hospi­ an index language grammar, especially the syntax of such tality with respect to the number of hierarchies available) a grammar, shoulp be available. This would largely elim­ should be provided by the capability of the software. In inate the necessity of introducing descriptors of the com­ the case of LIDOS, up to 30 hierarchies can be used. posite type and help to preserve the survival power of Within these limits any number of hierarchies can be ' the information system. inserted at any location in the hierarchy. Again, this is This enumeration of the essential features of both a valuable device for preserving a meaningful systematic classification system and thesaurus shows that they can structure of the vocabulary over the course of time. well complement each other while eliminating their typi­ Analogously to the characters of subdivision, any cal weaknesses. Proposals to this end have been sub- other kind of relationship can be phrased as a pseudo-

Int. Class;!. 17 (1990) No.3/4 Fugmann: Classaurus on the PC 135 descriptor. This makes it possible to represent these re­ Hence, subject indexing by no means consists merely lationships in a predictable way and, hence, in a way of extracting meaningfu l words from the texts, useful for retrieval and for mechanized grouping. as is often erroneously assumed. In contrast to other, similar classaurus approaches it is not intended to generate the wording of index entries algorithmically. Rather, they are phrased intellectually, 3.2 Structuring the vocabulary and descriptors are assigned to them so that they will In the first step of scrutinizing a book for the first time, occur at all locations in the index where they can be ex­ the indexer also obtains an overview of the conceptual pected. categories according to which the subject headings should be systematically grouped. During the progress 3. The interactive use of the classaurus of indexing such an overview will soon become necessary The possibility of interactively working with such a in order to insert the new accessions of subject headings c1assaurus facilitates correctly, to avoid unintentionally introducing synony­ , its construction, mous ones and in order to have an overview of the con­ cepts that are not yet (or should not be) represented in * descriptor choice for indexing and retrieval, , the vocabulary. Only after the indexer has attained a presentation of the vocabulary in print. good overview of the entire contents of a book, will he The assistance provided by a computerized classaurus or she be able to make the best choice of categories. is greatest in the construction of a book index. Here, Frequent re-grouping of the subject headings, i.e. simultaneously with the progress of indexing, the vo­ frequent shifting of subject headings (see below) within cabulary of subject headings must be established from their systematic arrangement, will therefore be neces­ scratch. Its structure and contents are constantly subject sary. to change, depending on the concepts which are newly encountered or which with hindsight prove important

enough to be represented in the vocabulary, and also 3.3 Shifting of concepts depending on the emerging desirability of changing the Additional vocabulary transparency can be achieved characters of subdivision and, hence, the arrangement through the introduction of characteristics of subdivi­ of subject headings. All these procedures can be per­ sion. Here, too, it will only be after some experimenta­ formed interactively and with relatively little expendi­ tion and after considerable revision of the work already ture of time and effort with the aid of a good computer performed that one arrives at the best choice, so that the program. Some of the desirable manipulations even re­ class of residual concepts, which could not yet be located quire this computer assistance. under a characteristic of subdivision, has become as small as possible and so that the characteristics of sub­ 3,1 The compilation of the vocabulary division can be named in a lingually acceptable form. In the first step of scrutinizing a book for the first time, Here again, the shifting of subject headings to more and those concepts are collected which might be of interest more appropriate locations is necessary. for a searcher. They are phrased as subject headings for In the categorization of articles in a catalogue of a de­ the index. In the catalogue of a department store this partment store, subject heading shifting will often be­ might apply, for example, to "shoes", "watches", "bicy­ come necessary for still another reason. For example, cles" etc. one may have begun with the categories APPAREL, Sufficiently concise and expressive terms will not al­ TOOLS, JEWELLERY, FURNITURE. Later on, one ways immediately come to mind when one encounters finds that there is no place for lawn mowers, bicycles i..nd paraphrasing expressions for a concept in a text, i.e. ex­ books. When one introduces corresponding categories, pressions which are not suitable for inclusion in an index, one will have to check whether some of the already es­ due to their non-lexical (1) nature. For example, in one tablished subject headings ought to be located in the passage a "spray for conibatting plantlice on indoor newly established categories. For example, a book on plants" may be mentioned. Later, one becomes aware tools for carpentry, originally located among TOOLS, of expressions such as "insecticides" and "plant protec­ will then, at least additiona!ly, have to be located in the tion" or still more appropriate lexical expressions and category BOOKS, as well. would assign these subject headings to this concept when In such a shift, not only the postings of the subject it recurs and when its importance has become more ob­ heading in question, but also all subordinate subject vious. headings, together with all their appertaining postings, Only in a second perusal of the text, after the vocabu­ will have to be transferred. The machine prograrrimust lary has J"argely been established, can one go about con­ perform these shifts without requiring the indexer to sistently assigning the subject headings of the index to all enumerate all the corresponding subordinate subject passages, to which they apply. This is true particulary headings and all the appertaining postings. for the early passages, read in the first stage, when the Since only trial and error will lead to a satisfactory vocabulary was still largely incomplete. For example, it solution as to the best choice both of categories and of is only in this second stage, that the above-mentioned the characters of subdivision J one will frequently have passage would receive the subject headings "insecti­ to re�organize the vocabulary by repeated and extensive cides" and "plant protection". shifting.

Int. Classi!. 17 (1990) No.314 136 Fugmann: Classaurus on the PC 3.4 Subdividing general subject headings phabetical version, for each subject heading at least its with hindsight immediately superordinate one should be displayed for In the initial stag� of indexing textiles in a catalogue better navigation. one may have contented oneself with fairly general sub­ ject headings, for example with "synthetics". Later on, 5. Conclusion it becomes apparent that the various kinds of synthetics Conventional, approved conceptual devices such as (polyester, polyurethane, acrylics etc.) should be differ­ categorization and logical subdivision can effectively be entiated. This is demonstrated in Fig. 2 by the notations used with the aid of simple computer technology. This "Appl", Appl5" etc .. However, one of the newly intro­ facilitates and accelerates the look-up procedure in an duced subject headings may be implied by other subject index, renders the intellectual indexing procedure more headings already in use. For example, Spandex™ always reliable and thus drastically improves retrieval, regard­ means textiles in polyurethane. Then it is desirable to less of whether it takes place in a printed or in a com­ retrieve all the postings under this brand and to assign puterized index. The mere extraction of meaningful to this group of postings in one single step the subject words from the texts cannot lead to good subject index­ heading "polyurethanes", too. ing because it has not been shown so fa r that the equiv­ 3.5 Collecting specific subject headings alence of paraphrasing expressions with their under a more general one corresponding subject heading or descrip tor During the development of the vocabulary the indexer can be algorithmically recognized. may also perceive that it will eventually become too large and specific in certain sections. For ex'ample, initially he Acknowledgement: The author thanks Software Land may have differentiated the various kinds of bicycle for their prompt advice in the author's adaptation of (mountain, racing, children's bicycles etc.). He may then UOOS™ to the task described here. decide to bring them all together under the more general subject heading "bicycles". In this case, he will retrieve all the specific postings with one single query, assign Notes and references: them the more general subject heading in a single step (1) Lexical expressions are deemed those which are commonly agreed upon and in which the sequence of the characters and finally erase the more specific headings in the vo­ (spaces included) is fixed. They can therefore easily be located cabulary. in an alphabetical arrangement of words. Non-lexical expres­ sions are consequently those which lack this property. (2) FlO/Classification Research, News No. 2: "By 'indexing' is 3.6 Lead-in terms meant the description of the essen"tial contents of a document, During indexing, one will encounter again and again by extraction and/or assignment of significant terms with or without syntactical relationships with a sufficient degree of new lexical expressions for a concept that has already fidelity and predictability for retrieval demands". Int. Classi­ been represented by a subject heading with different fication 8 (1981) p. 96. wording. One common example is acronyms or near syn­ (3) Fugmann, R.: "Indexing is the translation of the essence of a document into an indexing-lingual mode of expression." and onyms i.e. expressions for concepts so closely related to "The task of an indexing language is to represent concepts and another one that they need not be differentiated in the statements with a sufficient degree of predictability and fidel­ index. In the alphabetical part of the index all these syn­ ity." Int. Classification 6(1979)p.3-15. (4) Aitchison, J.; Gomersall, R.; Ireland, R.: Thesaurofacet: a onymous ter;msshould be listed and lead the searcher to thesaurus and faceted classification for engineering and re­ the preferred subject heading (or be directly available lated subjects. English Electrical Company Ltd.; Whetstone, Leicester, England. Cited from Aitchison, J.; Gilchrist, A.: as search parameters in a mechanized index). Thesaurus construction. A practical manual. ASLIB 1987. (5) Bhattacharyya, G.: Classaurus: Its fundamentals, design, and use. Studien zur Klassifikation 11 (1982), p. 139-148. Frank­ 4. Printing the index furt: INDEKS Verlag. It is much ill. the interest of good indexing that the (6) Devadason, F. 1.: Online Construction of Alphabetical latest stage of vocabulary development should be avail­ Thesaurus: A Vocabulary Control and Indexing Tool. In­ form. Process. and Management 21(1985),p.11-26. able also in print. Display on the screen normally lacks (7) Descriptors with a capital letter at their end indicate on the the degree of transparency which is desirable for index­ screen that they are generic to other descriptors. ing work. This holds true for the alphabetical as well as for the systematic version of the vocabulary. In the al- Dr. R. Fugmann, Alte PoststraBe 13, 0 6270 Idstein

Int. Classi!. 17 (1990) No.3/4 Fugmann: Classaurus on the PC 137 Claus Ritzier

Comparative Study of PC-supported Thesaurus Software

Ritzier, C.: Comparative study ofPC�supportcd thesaurus son� 2. The different thesaurus software packages ware. Int. Classif. I7 (1990) No.3/4, p. 138 - 147, 1 ref. and their producers This article presents the results of a comparative study of three Three products have been compared: one British and PC supported software packages (INDEX, PROTERM and TMS) for development, construction and management of two German thesaurus software programs. thesauri and other word material with special regard to hardware The British one is called TMS (Thesaurus Main­ and software requirements, handling and user interface, and tenance System) Version 1.0 and was developed by Py­ functionality and reliability. Advantages and disadvantages arc discussed. The result shows that all three software products ramid in Reading. The TMS thesaurus software package comply with the minimum standards of a thesaurus software. is used, mainly in com.mercial libraries and information After inclusion of additional features distinct differences be­ centres, for constructing and maintaining thesauri which come visible. ' (Author) will be used eventually for indexing company docu­ ments. The user can purchase this product as part of the Computer Aided Library Management System 1. Background (CALMS) or as a standalone thesaurus construction sys­ This article presents some results of a comparative tem. TMS is priced at 1.840 DM. study (I) made a year ago as part of the author's thesis. One ofthe German products is PROTERM-TVersion The essential part of this work was the comparison of 2.5. It is a software package from PROGRIS in Berlin. three PC-supported software packages for development, The software PROTERM-T was especially designed for construction and management of thesauri and other constructing and maintaining vocabularies. PRO­ word material with special regard to TERM-T is a standalone system, but it is possible for Hard- and Software requirements the user to buy up to 5 additional modules. These mod­ - Handling and user interface ules are not included in the comparison. The price for - Functionality and reliability the main software PROTERM-T is 1.200 DM. The other German package is named INDEX Version The author has drawn up 8 different tables of parame­ 4.1. It is a product of ERNST LUKAS. INDEX is a ters to evaluate the software. universal package for developing and maintaining thesauri, vocabularies and classifications. For INDEX The 8 tables and their headings had been as follows: are also up to 8 modules available, which are not in­ Software producer information cluded in this comparison. With INDEX the user re­ - Hardware and software requirements ceives a Dataflex Runtime License or a Datafle}; Software installation Development Licen"se from Data Access. The costs for Operation and user interface INDEX are 6.200 DM plus the Dataflex Runtime Li­ - Functionality: Input of terms and classifications cense, which is approximately 300 to 2.000 DM. Functionality: Creation of term relationships Batch functions (involving e.g. different input and output functions) 3. Software and Hardware requirements Management and system functions PROTERM-T and TMS run under MS-DOS 2.1 up­ (involving e.g. user access control, reorganisation of wards. INDEX requires MS-DOS 3.3 or another Oper­ the data etc.) ating system such as PC-DOS, UNIx/XENIX, OS/2 and others. All three programs can be installed on a PC­ With this catalogue of evaluation criteria it was XT or AT, with a hard disk and 512 KB memory. For possible to identify the differences between the products INDEX the producer recommends an Enhanced providing a basis for discussion of the salient pros and Graphics Adapter with a suitable display. TMS requires cons. Additionally it was possible to determine whether one disk drive to be free for the program disk, because all three software products comply with the minimum it is not possible to copy the program to the hard disk. standards of a thesaurus software and under which con­ Thesaurus data, however are saved on the hard disk. ditions these products are suitable for establishment, TMS also needs a printer, because it is not possible to development and maintenance of individual thesauri. display all kinds of reviews on the monitor.

Int. Classif. 17 (1990) No. 3/4 138 Ritzier: Comparative Study of Thesaurus Software 1. Software Producer Information IHOEX PROTERH-T THS Version 4.1 Version 2.5 Version 1.0

AwA Softwore Producer Address ERNST LUKAS PROGRIS GtflH PYRAMIO SOFTWARE lelpziger R1ng 16 Auguste-Vlktor1a-Str. 64 PROOUCTS lTO_ 6054 Rodgau 3 1000 Berlin 33 9 Church Street Reading RGI 2SB United Kingdom

AwA Status of the Producer Subcontractor estob lished Company established company

AwA Company Foundat 10n 1984 1977 1982

AwA Number of Emp loyees altogether 1 Employee 8 Employees 7 Employees

AwA Custom Servlces Consu 1tat 10n Consu ltat 10n Consu 1tat 10n Course of 1nstruction Course of instruct10n

AwA Number of References 60 Single-User Installat10ns 24 Installat10ns 40 Insta1 1at 10ns 5 Multi-User Installat10ns

AwA slgn1fles: Generally 1mportant product 1nfoT'1l18.tlon. These repr�sent the necessary requ1rements of a given product .

Table 1: Software producer information

2. Hordwore and Software Requ 1rements INDEX PROTERH-T THS Vers10n 4.1 Vers10n 2.5 Verslon 1.0

AwA Hardware 8asfc Equ1pment PC-XT. PC-AT IBM kompatible PC-XT, PC-AT IBM kompatfble PC-XT. PC-AT IBH kompatlble PS/2-Serles and others 512 KB Memory 512 K8 Memory 512 KB' Memory Hard d1sk Hord d1sk Hard dfsk Mln1mum 1 Disk drive Enhanced Graph fcs Adapter Pr1nter w1th a sultable dlsploy recorrmended

AwA Operoting System MS-DOS Version 3.3 MS-DOS V. 2.0 upwards MS-DOS V. 2.1 upwards

AwA On other Operatlng Systems usable e.g. PC-DOS. UNIX/XENIX. Ho Ho OS/2 and others

AwA Standard-Software Bas lc Equ 1pment IHDEX PROTERM-T THS Dataflex Runtime license or Dataflex Development l1cense

AwA Number of add1tional Standard Modu les B Modules 5 Modules Ho

AwA Costs of the S1ngle-User Basic Versfon OM 6.200 OM 1.200 OM 1.840

AwA Costs of llcenses for other Softwllre OM 300 to OM 2.000 - -

AwA Costs tor add1tlonlll Standard Modules OM 150 to OM 3.320 OM 190 to OM 380 -

• Capllble for Mult1ple Places Yes, on all conmon PC-Mu ltl- No Ho User-Opera t lon-Systems

• Capable Jor Networks Yes. on all cOlTlllOn Ho Ho Network -Operat fon-Systems

AwA Deve lopment System used Oataflex Cllpper Turbo Pascal

AwA Data Structures Semont1c Network S lmu lated Re lat 10na 1 Mode 1 Relational Hodel wlth purely hierarchlcal structured elements

• Source Code ava ilob le Katter of Hegot111tfon Ho Ho

AwA Quant1ty of De l1very 35% 1nch Diskettes a 1,2 I1J 2 5% 1nch D1skettes a 360KB 2 5% 1nch Diskettes a 360KB 230 Pages Manua1 100 Pages User-Hanuo1 27 Poges Operoting Guide Bas1c Set of Test Oatil

AwA Documentation and literature HanuIIl User-Manua1 User Introduction - Qua l1ty � Very good Good Satisfactory - languages Germon GerlMn English

AwA signif1es: Generally important product informat1on. These represent the necessary requirements of a given product.

* signifies: Here are features. requfred 1n applicatlon cases or by the system's environment.

Table 2: Hardware and software requirements

Int. Classi!. 17 (1990) No. 3/4 Ritzler: Comparative Study of Thesaurus Software 139 4. How are the programs presented to the user? - by entering different key combinations The initial presentation of a software product is com­ the user can choose different options from each menu. parable with the first interview between a personal The interactive possibilities are not unified in PRO­ manager of a company and the person who is looking TERM. If the user wish to jump to another (e.g. a two for a job. The first impression must be good. or more levels higher) menu, he must path through all The first menus the user gets are in all three programs intermediate menus sequentially. It is not possible to similar. One is more detailed the other less. jump between the different levels as it is in INDEX be­ In Fig. 1 with the main menu ofTMS there are mainly cause the function keys in PROTERM are not used at two parts called construction and review. On the con­ all and only for statistical operations in TMS; the opera­ struction side we see the input components like the op­ tion is cumbersome. tions to enter terms, relationships and scope notes, the TMS is not menu-driven, it-is driven by commands, option to change or delete terms, the option to add facets cursor keys or partly by function keys. Alternative in­ or even the option Re-index what is normally a manage­ teractions are not possible. ment or systems function. On the other side the user will Unlike INDEX, PROTERM and TMS have no help see more or less all output functions. facilities. In Fig. 2 with the main menu of INDEX we see a clear menu with 7 options, which contains completely differ­ ent functions. The normal user needs option 1, to carry 5. Illput of terms out all necessary operations for developing his This part describes the input and search facilities of thesaurus, classification or source register and naturally the different programs. option 2, with the data output functions. The other func­ The data entry mode in TMS is accessed by typing the tions are more or less for the system manager, they in­ number 10. In PROTERM and INDEX it is selected by volve features like reorganisation, data backup, the defi­ choosing option 1 from the main menu. The user is pre­ nition of relationships and control parameters etc .. The sented with a new display layout, which is differen! for background is coloured in blue, the line 1 is coloured in each program. grey. All colours are optional and may be changed by In TMS the user may enter a term by typing the specific the user. term or by typing the number code for the term, if In Fig. 3 with the main menu of PROTERM we have known. Terms are limited to a maximum of 40 alpha­ three areas as well. First the dialog functions to enter, numeric characters only. delete and list terms, than the batch functions for reor­ The word can be described with a scope note or a facet. ganisation ahd output. The last bit contains the different To register a scope note the user has to type the word management function. again, after pressing the function key F3. A field with However all three thesaurus programs are presented maximum 1.110 character space appears. To locate a in a different way to the user. facet the user has to go back to the main menu to activate INDEX works with different types of objects. First option 12; making operations inconvenient. The facet there are forms, these forms are structured into different can be 20 characters long. The facet feature is primarily fields and look like record cards. And there are lists. See intended to relate to a classification scheme. However, Fig. 4. as yet, no facilities exists for sorting and printing the INDEX provides all kinds of interactive possibilities; thesaurus by the facet. the screen layout at the different working levels is fully In PROTERM data entry, data display, data cOrrec­ coloured. In spite of the complexity and the different tion and data deletion are all available at one level, but functions, INDEX is completely driven by using the ten it does not include term correction or term deletion. T�lis function keys Fl to FlO and the cursor keys. Additionally is included in option 2 of the main menu . Each term can the user gets a menu line on the upper part of the screen be 60 characters long. The user has the possibility to by using the function key FlO. This main menu line with eilter a scope note with a maximum of 150 characters various submenus is partly equivalent to the function and a notation with 6 characters. The program proofs keys. At some stages, especially at the systems manage­ each term simultaneously to the data entry. A small dis­ ment level the user is presented with pull down menus. play shows a changing tern: with the same sequence of All these interactive possibilities and the different usage characters as inserted. The program also displays a list of colours are unified at every stage of the program. Each of terms by pressing function key F2.The user can choose colour has its own meaning, aiding the user in learning words from this list. and understanding the software and its working It is not possible to insert terms successively. The user processes. always gets the menus for the different relationships first. PROTERM also has menus at different working Interrupting this process is possible by typing the levels. The options for each level are numbered up­ "Hatchmark"(=H=) and some carriage returns. This is wards. The monochrome display of PROTERM and again a very time consuming pro.cedure for users who TMS operate in conjunction. want to enter lists of terms before they start with the relations. Especially at the beginning stage of a thesaurus With project, where the word material will be collected first. different cursor keys, or It is not possible to enter short terms who are identical by entering a number, or with the first characters of already existing terms. In the

Int. Classif. 17 (1990) No. 3/4 140 Ritzier: Comparative Study of Thesaurus Software Pyramid Software 0734 595633 • • • • • 7/8/1990 Thesaurus development Construction Review 10. Terms, relationships and scope notes 50. View th�saurus 11. Change or delete terms 12. Facets 13. Language alternatives 60. Term lis t 14 . Indentation levels 62. Single level term list 20. Re - index 63. Single term systematic list 64. Alphabetic systematic list 65. Alphabetic NT1, NT2 list 70. Create language index 80. Change print parameters Please enter your choice Fig. 1: The main Menu ofTMS

INDEX-Datenbank Datenausgabe Definitionen Systemverwaltung Dateiverwaltung Datensicherung Aktuelle Datenbank wechseln

Fig. 2: The rQain menuofINDEX PROTERM-T Verso 2.4 (c) PROGRIS 1987, 1988 .Dienstag, 07.08.90

Dialog-Funktionen 1 > Beziehungen anzeigen/erfassen/ltlschen 2 > Terme korrigieren/ltlschen (Termnummer erforderlich) 3 > Term-Listen anzeigen Batch-Funktionen 4 > Term-Listen drucken 5 > Reorganisation: Ltlschen/Sortieren/Sichern Verwaltungs-Funktionen 6 >. Grundstruktur anzeigen 7 > Erf_8ssungs-/Anzeige-Formate anzeigen/!lndern 8 > D�uckformate anzeigen/andern/erstellen o > ENDE

Mit CUrsor auswahlen, dann oder Ziffer wahlen Fig. 3: The main Menu of PROTERM

Int. Classif. 17 (1990) No. 3/4 Ritzier: Comparative Study of Thesaurus Software 141 1------Klassifikation nach Notation------* Aachener und MUnchener Beteiligungs-AG allgemein. AfAa0010 : :FA * Aachener RUckversicherungsgesellschaft al1gemein. AfAa0015 : :FA * abc-Autorenbuchhandlung a1Igemein ...... AfAa0020 : :FA * Ackermann-G�ggingen allgemein ...... AfAa0025 : :FA * Arr=�------�------==-� 0030 : :FA *. A SSIFIKATION Nr: 5231 S: M: 1 0035 : :FA A [iCU 0040 : :FA * A Notation B T Eb U 1 2 3 F 0045 : :FA * A AfAa0005 ::FA K 1 0050 : :FA * A 0055 : :FA * A * Aachener und MUnchener Lebensversicherung AG 0060 : :FA * A allgemein 0065 : :FA * A 70 : :FA * AG WORT Nr: 8428 H: 0 S: M: J 75 : :FA * AG -I 80 ::FA * Ag Aachener und MUnchener Lebensversicherung AG 85 : .FA • Ag 90 : :FA • AG Sp Sgr Fa Wo G Gm Sch 1 2 3 Zl Z2 F T 93 ::FA * AG In J P5 6 D X 95 : :FA * AG 00 : :FA -> Objekt Zoom Verbinde Liste Suche Form Mark Hilfe Weiter Ende

Fig. 4: The objects in INDEX; forms for words and classifications and a list in the back.

Pyramid Software 0734 595633 * * * * * 10/8/1990 F4 F5 F6 F7 F8 F9 Rl No. Term US UF BT NT RT SA 53 Volkswagen 2 0 0 0 1 0 SN US 54 VW 0 1 0 0 0 0 US 55 VW Kafer 0 1 0 0 0 0 RT 56 Audi 0 0 0 0 1 0 Fig. 5: Data entry mode in TMS

PROTERM-T Verso 2.4 (c) PROGRIS 1987, 1988 Dienstag, 07.08.90 oc==== I I ERFASSEN ANDERN L�SCHEN J Bitte Term eingeben: a

D-Zug Damenfahrrad Dampflokomotive Diesellokomotive E-Lok --->Eilzug Eisenbahn Kfz Kraftfahrzeug Motor �ffentliche Verkehrsmittel

AuswHhlen mit Cursor oder Term Ubernehmen mit Liste ausschalten mit Fig. 6: Data entry mode ofPROTERM with a list

lnt. Classif. 17 (1990) No. 3/4 142 Ritzier: Comparative Study of Thesaurus Software German language we have often the problem with COITI­ onymy, Hierarchical Relation and Associative Relation posita: e.g. Eisenbahnschaffner; Eisenbahn; Eisen. In are involved. this case the user has to enter the shorter terms first or All relationships have been designed to conform to he has to insert a blank at the end of the shorter term. British, German and international standards. The "See Homonyms and other words with identical character Also Relation" in TMS is identical with the relationship sequences cannot be entered in PROTERM and TMS. "Identical Use" in INDEX. As mentioned, it is possible In INDEX a homonym term can be marked in a special to define more relationships in INDEX, but not in PRO­ field with an H. In the search mode of PROTERM no TERM orTMS. PROGRIS provides additional modules distinction is made between upper and lower case which for additional relationships. can be very helpful. Furthermore all or parts of the control ro'utines may In INDEX the user is presented with a blue formula be switched offin INDEX as if you want to ignore special to enter the words. This blue formula represent a 3by5 routines. INDEX has completely user controlled con­ card from the card index (Fig.4). The formula is used sistency checks. In this part the user is completely free for recording, searching and relating the data. You have in his decisions, but it is worth to think about every different formulas for words, classifications and sources. change of the default values very carefully. INDEX pro­ In the formula for words you enter a new term and in vides more control routines than the other programs be­ the different 16 smaller fields you may add specific in­ cause a different description ora term can be controlled formation about the term. Special information can be a as well. For example the user can forbid to relate terms language code, a subject code, a facet and others. In this of a subject A with terms of a subject B. case the user is completely free in his decisions, he can In all three programs the reverse relation is assigned define the categories by himself. Some of the fields are automatically after saving a relation between two terms, indexed. This is a very helpful feature for sorting and but the control functions are not as strict as in INDEX, listing the material for various aspects. specially not in TMS. For instance: once assigned terms This is possible in PROTERM only partly by combin­ as non-descriptors can be related to all other terms, just ing the term with a notation. But each notation may have like a descriptor. In fact everything can be related with only one combination with a term. So, if one wants to everything. There is only one exception. A descriptor add more specific information one should contact PRO­ cannot be combined with itself. If so, the program will GRIS, they provide an additional module for that. stop. Recorded with each word saved is the entering date, In PROTERM the consistency cheeks are done per­ and the date of the last change is saved together with the fectly well. If you want to enter more than one relation name or code of the user who did the entry/change. The to a non-descriptor, the term has to be assigned as pol­ maximum size of a controlled term is 50 characters. The ysem first, otherwise it is not possible. term itself can be as long as wanted, but it is not control­ Automatic reorganisation of very complex relation­ led any longer. That means, if one has two terms which ships after changing a descriptor to a non-descriptor will differ at character 51, INDEX will not accept it. only be done in INDEX which is very comfortable. In During the recording of terms it is possible to define a the other two programs one has to delete the rel!ltion­ master record for such terms which have e.g. the same ships first before assigning the non-descriptor to a de­ descriptions. In this case one needs only to enter the scriptor, thereafter one has to build up the relationships term into a 'Copy of the master containing the fixed in­ again. This can be very ti�e consuming, especially ifthe formation. The user has various possibilities to search, descriptor had a lot of relations on various levels. correct and delete terms. He can do this in a list or in In all three programs there are no limitations regarding the formula. It is no problem to switch between a list the number of relationships per descriptor. and a formula. In both types of objects he is able to browse forward and backwards in the term material to 7. How to enter the relationships make further selections. The user can also mark words In PROTERM the user gets, as already mentioned, in a list for creating a separate working list. In INDEX each kind of relationship after having entered the term. scope notes have no limit in their length. To enter a scope He can affect the order by typing the minus sign (to get note you have to open a text field. This mode is sup­ the previous kind of relationship) or by typing an addi­ ported with wordprocessing features. They enable the tional carriage return (CR), than he gets the next kind user to delete and insert lines, to search for a word or a of relationship. The order and the abbreviation of the phrase and more. The user has the possibility to define relation'ship can be changed by the user under the man­ different kinds of notes (e.g. Definitions, History Notes, agement functions. In the same mode, relationships per Scope Notes etc.) by himself. term can be displayed and deleted. If he wants to delete all relations�lips of a term at once, he has to delete the 6. Different relationships and consistency control whole descriptor under menu point three and two. INDEX provides the user with the choice between 24 In TMS it is similar to PROTERM, to enter a new different relationships. relationship the user has to type the abbreviation for the In INDEX it is possible for the user to create and de­ relationship followed by CR, followed by the term or its fine more relationships if required.TMS and PRO­ number. Just as many relationships may be entered as TERM provide 5 different relationships to the user. In are required. Whilst manipulating relationships of a cer­ all three programs the minimum standard relations: Syn- tain type, it is convenient that the user can exclude other lnt. Classif. 17 (1990) No. 3/4 Ritzier: Comparative Study of Thesaurus Software 143 ==t Textfelder zu Wort Text zu: Arabische Staaten SCN 1 DEF Definition Staaten des arabischen Sprachraumes und der SCN Scope Note Arabischen Liga (ausgenomrnen der PLO) , sowie HIS History Note dem Iran alB nicht arabische Ausnahme am -- PersiBchen Golf --

--

I WORT Nr: 7186 H: o S: M: I Arabische Staaten Sp Sgr Fa Wo G Gm Sch 1 2 3 Z1 Z2 F T Ge S 0 D f T -> Objekt Zoom Verbinde Liste Suche Form Mark Hilfe Weiter Ends Fig. 7: Scope note with its term in INDEX.

INDEX 4.1 PROTERM 2.5 TMS 1.0

1. Synonymy Oeser 1ptor Non-Descriptor X x 2. Abbrev iat10n full Descr1ptor Abbrev1ation 3. Ident 1cel Use Term Same Hean tng X 4. Quas1synonymy Descr1ptor Quas 1-Synonym 5. Alternat1ve Descriptor (alternative) Same Hean1ng 6. Comb tmlt1on Stngle Descriptor Combined Term X

7. Use BlISit Term Examp le of Use 8, Hierarch1cal Relation Broader Term Narrower Term X X 9. Generic Relation ,Broader Term Narrower Term 10. Part1t1ve Relation Who le Port II. Cous.l1ty Cause Effect 12. Appurtenance Relat10n Generft1 Term Instllnce 13. Field Relation FIeld Element

14. Assoc18t1ve Relat10n Term Rellited Term X X

15. Tempora l Sequence Predecessor Successor 16. Spl1ttlng Predecessor Successor 17. Un10n Predecessor , Successor

18. Oppos lte Term Oppos lte Term 19. Use Warn1ng Term In Contrast To

20. Product 1on Producer Product 21 . Re18t1on of Mater1al MIIter18l Object

22. Other l8ngu8ge Language Fore1gn llinguage 23. Tr8ns18t1on Source lllnguage Tnget language

24. Systemat1c Oeser 1ptor Notat1on Xl Xl one .Notat1on/Facet per Term only

Table 3: The different kinds of relationships in INDEX, PROTERM and TMS relationships from the screen during editing. This can be The input of relationships in INDEX is different from done by using function key F4 to F9. These keys are used the other two software products. Here the user gets two as toggle switches to turn on or off the display of each record cards on the screen after pressing two function relationship type (see the right side of Fig. 5). To delete keys. The contents of each card may be switched to one a relationship is quite easy. One has to move the cursor another, so that the upper content appears in the lower next to the relationship to be deleted and to press CRTL card and vice versa. After pressing a function key again and D. It is not possible to delete more than one relation a red clamp appears with the kind of relationship used once. last, in it. Now it is possible to choose another kind of

Int. Classif. 17 (1990) No. 3/4 144 RitzIer: Comparative Study of Thesaurus Software Wtlrter (G) alphabetisch M Sgr Beziehungsarten Offentliches Verkehrsmittel 1 sbnonymie 01 1.1 A kUrzUn� , Oltanker WAS 1.2 Identisc e Verwendung Paddel WAS 1.3 Quasisynonymie Paddelboot WAS 1.4 Al terna tive Pannen ion Parkh veSkr ptor Nr: 482 H: o S: M: g Parkp e Hierarchie Parks Offentliches Verkehrsmittel e Hierarchie Parku lit Parkv Sp Sgr �a Wo G Gm Sch 1 2 3 Z1 Z2 F T gkeit Passa I I I I I I ehung Passa haft Passa N chtdeskriotor I 1 Synonymie I Asp: iolge Pedal Pendler Kraftwagen Pe.rsonen Personen Sp Sgr Fa Wo G Gm Sch 123 Z1 Z2 F T a Personen STR G Personen I I I Wort VERBINDEN mit: teXt Wort Klasse Quelle Fig. 8: Building a relationship between two terms inINDEX.

KLASSIFlKATION Nr: 5231 S: M: Notation B T Eb U 123 F AfAa0005 : :FA K 1 . * Aachener und MUnchener Lebensversicherung AG allgemein WORT Nr: 8428 H: 0 S: M: Aachener und MUnchener Lebensversicherung AG II tz

[��;g :======, I =< " :�>11

Wort VERBINDEN mit: teXt Wort Klasse Quelle Fig. 9: Building a relationship between a classification and a term in INDEX.

I WORT Nr: 482 H: o S: M: I Offentliches Verkehrsmittel Sp Sgr Fa Wo G Gm Sch 123 Z1 Z2 F T G QUELLE Nr: 5 S'x M: I

Meyers Enzyk1.opl!disches Lexikon in 25 Bllnden -- Zitcode Jahr DT E Sp Geg L Sgb 1 2 3 Standort MEYERS 1980 Ix de xyz Wort VERBINDEN mit: teXt Wort Klasse Quelle Fig. 10: Building a relation between a source and a term in INDEX.

Int. Classif. 17 (1990) No. 3/4 Ritzier: Comparative Study of Thesaurus Software relationship either from a list (by pressing Fl) or by typ­ Import and export of structured material ing the number of the relationship into the clamp and Output on different media pressing a function key. Different kind of output At every stage the user is able to zoom (means: to Before starting to construct a thesaurus it is more or display) the relations to a descriptor. He gets a list. The less common to implement a list of words from other list he receives can be posed everywhere on the screen. existing specific thesauri. These data exists normally in It is possible to delete relations from the list, but not machine readable form, hopefully in ASCII format. more than one relation at once. All three producers provide to convert the data into It is also possible to connect terms to a classification the special thesaurus software format. But big problems card or to a source card by defining a special kind of come up, if the data are already structured, because no relationship. standards for data exchange exists. ERNST LUKAS has sct up a program to convert the already structured 8. Limits and data structure AGROVOC-Thesaurus of the FAO into INDEX for­ Regarding the limits of data input, there is a limit of mat. Furthermore ERNST LUKAS SOFTWARE pro­ 65.535 terms in TMS. The TMS data structure is some­ vides a data exchange program to convert data files from thing of a hybrid. Probably the nearest formal structure INDEX to the Data Bank Management System BASIS to which it could be compared is a relational model, how­ and they plan to do one for the system TRIP. ever there are elements of the data which are purely hier­ The author is not absolutely sure if PROGRIS provide archically structured. The program becomes slower as similar exchange programs. Maybe it is possible with more terms and relations are entered. In PROTERM PROTERM-K (for Communication), a special modu!. and INDEX there are practically no restrictions regardM ing to the number ofterms. PROTERM is designed with 10. Output functions a developing tool called Clipper. The data structure is a In all three programs the user gets highly developed simulated relational model. the program becomes much output functions, especially in the very sophisticated slower as one enters more terms and relations. INDEX INDEX and in PROTERM. In INDEX for instance the is a purely semantic network. The access time to the data user can produce an output file of his thesaurus,- ready is very fast, it will not be reduced by the quantity of data for electronic publishing. entered. Apart from TMS all standardized lists are possible, for displaying on the screen, for saving as a file or for the 9. Batch functions output on the printer. The batch functions contain: In addition, PROTERM and INDEX provides various Import and export of word material sorting parameters.

4. Management and System FUnct10ns INDEX PROTERH-T TMS Vers10n 4.1 Vers10n 2.5 Version 1.0

User Password and Restrict10ns Yes No No

Display/Change of - Menus Yes No No - Error Messages Yes No No - System Messages Yes No No - Colours Yes No No

Change of M Recording Order Yes Yes Yes - Record1ng labels/Display labels Yes Yes No

Draw Up/Change/Display of Print Parameters Yes Yes No

Reorgan isat 10n - S1ngle F1les Yes No No - A 11 Files Yes Yes Yes

Display of the Statlstical Structure Yes, Module Yes, for all terms Yes, per Term and Relatlon

IIIIplemented Backup Rout lne Yes Yes No

Management of IIIOre Thesauri Yes Yes No

HanageLl'lE!nt of a Mult111ngual Thesaur1 Yes No Yes, up to 2 additional l.

Change to Operat 10n System leve1 without leav1ng the Program Yes No No

Change to II Word Process1ng system without leav1ng the Program Yes No No

Table 4: Management and system functions

Int. Classif. 17 (1990) No. 3/4 146 Ritzler: Comparative Study of Thesaurus Software The different output lists are: program turns out to be more efficient than PROTERM (I=INDEX; P=PROTERM; T=TMS) and TMS. PROGRIS with the demand to serve their users with - Alphabetic term list I P T an acceptable product for less money is on the right line. - Systematic term list I P Regarding the relation between price and performance - Minithesauri/Partthesauri I P they produce a good software. - Termlists with/without relationships I P T How does higher cost compare with high perform­ - Termlists selected by different criteria I ance? That is not always the case. TMS is much more - Synonymlists I P expensive than PROTERM but its performance is only - Systematic thesaurus I P T half of PROTERM. - Alphabetic thesaurus I P T Naturally there is a third factor: the suitability of the - Thesaurus with different BTINT levels T software, All three programs are reasonably good for - Hierarchical BT/NT lists P the establishment of small thesauri, but if one needs a - different classification lists I powerful multiuser system for the management of ex­ - different source register lists I tensive archive material or for constructing a broad mul­ tilingual or even a monolingual thesaurus one should 11. Management and system functions take PROTERM rather than TMS. In this case INDEX All three programs provide various management and with its highly sophisticated features would be the pref­ erable alternative. system funttions, as shown in Table 4.

12. Conclusion References In this article the author attempted to present some (1) Ritzier, C,: Vergleichende U ntersuchung von pe-Thesaurus­ programmen. Dipiomarbeit. Fachhochschule Darmstadt, results of his evaluation of the functions of three Fachbereich Information und Dokumentation. Darmstadt thesaurus software products. Each program complies 1989. 156 ref. 200 p. more or less with the minimum standards of a thesaurus software. If ergonomical factors of software and addi­ Dipl,Inform ,wirt Claus Ritzier, tional features are taken into consideration the INDEX Liinkenweg 6, D-2000 Hamburg 60

3rd IFCS Conference, Edinburgh A second announcement and Call for Papers has been Klassifikation, Datenanalyse und released for the Third Conference of the International Informationsverarbeitung Federation of Classification Societies to be held from The 15th Annual Conference of the Gesellschaft fUr Aug.6-9, 1991 at the Heriot-Watt University, Edinburgh, Klassifikation is planned from Febr. 25-27, 1991 at Scotland. C@ntributions from all areas of statistics are Salzburg, Austria with the topic Classification, Data expected, theoretical and applied subjects will be cover­ Analysis and Information Processing. It is understood ed. Invited speakers will include E. Diday (France), as being a challenge for interdisciplinary research. R,C. Dubcs and D. Pregibon (USA), B.S, Everitt and Relevant application fields have been pinpointed, such P.I1.A. Sneath (U.K.), and N. Ohsumi (Japan). Other as archeology, biology, medicine, linguistics, geography, invited Sessions will cover: law, technology, economics, social science on the con� Tree theory - Cluster analysis - Combinatorial ception of knowledge and databanks, expert systems and optimisation � Diagnostic keys - Phylogeny - Classifica­ data analysis packages, aswellas terminology, documenta­ tion in psychology - Geological classification - Multi­ tion, subject analysis and library classification. 12 areas variate longitudinal data - Sequencing - Classification of with 48 subtopics are outlined in the Announcement living organisms - Computing methods. which calls for papers until Oct,3D, 1990 to the Con­ Abstracts in 3 copies are due by lan.15, 1991. They fe rence Chair, Prof.Dr.H. Goebl, Institut fUr Romanistik, should be sent to Prof.David 1, Hand, Faculty of Math­ Universitat Salzburg, Akademiestr, 24, A-5020 Salzburg, ematics, the Open University, Walton Hall, Milton Austria, Keynes MK7 6AA, England.

Int. Classif. 17 (1990) No. 314 Ritzier: Comparative Study of Thesaurus Software 147 Jochen Ganzmarm Lehrinstitut fUr Dokumentation in der DGD Frankfurt am Main

CR ITERIA FOR THE EVALUATION OF THESAURUS SOFTWARE

Ganzmann, J.: Criteria forthe evaluation ofthesaurus software. thesaurus software integrated into retrieval systems Int. Classif. 17 (1990) No. 3/4, p. 148 -157, 23 refs., 1 App. (e.g BASIS, DOMESTIC, LIDOS) The growing number of thesaurus programs, especiallY for mi� crocomputers, calls for criteria which facilitate the decision on the appropriate tools for thesaurus construction, application Stand alone software systems, very often for microcomw and maintenance. Following a brief outline of the determinants of the criteria, i.e. the changing environment and the routines of puters, can facilitate the construction and maintenance thesaurus management and application, criteria are discussed in of thesauri to be used in main-frame computers where detail. Only the most important requirements relating to central the handling can still be laboursome and ihe updating functions of thesaurus Work are touched upon, further criteria being included in a check-list attached. Applying the criteria to procedures are often time-consuming. Of course this im­ existing thesaurus software in a final short assessment, the plies some transfer of the thesaurus data into a retrieval author concludes that some programs reveal substantial short­ system in order to use the thesaurus in indexing and re­ comings. (Author) trieval. For the end-user it has nowadays become easier and also less expensive to develop, maintain and use thesauri even in smallwscale institutions and, equally important, A. The Need for Criteria with the advent of the micro-computer his independence Over the last ten years we have seen a renewed interest from computer experts has considerably increased. At in the thesaurus which is reflected both in the literature the same time, however, the rise in number of programs on the topicl as well as in a steadily growing number of and the above specified diversification of thesaurus soft­ thesauri (4). ware types have more and more imposed the burden of The reasons are manifold and complex. Most imporw decision-making on him. tant have been the impact of artificial intelligence resp. Knowledge of the requirements caused by the chang­ the concept of knowledge representation, dissatisfaction ing conditions in which thesauri are used nowadays and with the results of free-text retrieval and finally advances of the criteria that are to act as a measurewstick'in the in computer technology which entailed decreasing costs, decision which software is best suited to one's specific easier hand.ling and better performance both in respect needs has become more important than ever before. to software and hardware. These advances, esp. the ad­ Checklists of criteria can be helpful since they can iden­ vent of the micro-computer, have greatly facilitated the tify possible pitfalls and help reduce decision-making previously tedious task of development, management based on uncertainty as to the essential requirements of and application of thesauri (5). a software type. That there is an increasing need for Over the last decade the number of thesaurus pro­ orientation has been my experience at several seminars grams (i.e. software for both development/updating and on thesaurus software held at the Lehrinstitut fUr Qoku­ application) in general and especially for micro-comput­ mentation (LID) for which the check-list to which this ers has been rapidly increasing. The programs available paper refers has been originally developed. The check­ today can be classified list had these end-users in mind and consequently it is either above all concerned with defining criteria for micro-com­ by the type of computer for which they have been puter applications but can to a great extent also be ap­ developed: plied to mini and mainframe computers. software for microcomputers (e.g. INDEX, PRO­ I wiIi outline those criteria from the check-list which TERM, CICADE, LIDOS, TMS) seem to be most important taking into account the func­ software for mainframe computers (e.g. DOMES­ tions of thesaurus software and the changing environ­ TIC, BASIS) ment ofthesauri already mentioned. These determinant factors for the development of the criteria will be pre­ or sented (chapter B) before a discussion of the criteria by the functions which they fulfill: (chapter C). A short look at the present state of the art stand-alone software for the construction and main­ of thesaurus software in which I will try to make an tenance of thesauri (e.g. PROTERM, TMS, assessment in how far programs measure up to the de­ INDEX) fined requirements will conclude the paper (chapter D).

Int. Classif. 17 (1990) No. 3/4 148 Ganzmann: Evaluation of thesaurus software B. Determinants for the Defmition of Criteria a) The Construction and Maintenance of the Thesaurus Construction and updating of a thesaurus includes a B.l The Thesaurus and its Environment variety of tasks and activities: Over the last thirty years the thesaurus concept has Word-material is selected and recorded, sometimes undergone a change. Traditionally and even today the parts of existing thesauri must be integrated, information thesaurus has been foremost an indexing language as to the source of terms, language, systematic grouping developed by individual institutions for their specific has to be recorded. Terms are controlled in respect to needs. Due to this the main functions of the thesaurus clarification of meaning (homonyms, definitions) and re­ have remained relatively the same over the decades: its lations must be defined between terms. purpose was to optimize indexing and retrieval in a given Finally, thesauri must be updated regularly with re­ environment with a circle of users with specific needs. gard to relations and terminology due to changes in the There have, however, been factors affecting this tradi- terminology of a given subject-field, to changing index­ tional concept, especially over the last ten years: ing and retrieval practice, to fa�ling adequacy of the vo­ increasing national and international technical, cabulary's scope or specificity. scientific and political cooperation b) The Output of the Thesaurus increase in data communication relating to local net­ The output of the thesaurus (or parts of it) is necessary works as well as to online databases in all phases of thesaurus work, either on the screen, by closely related, office automation the printer or also into a file in case the data are to be growing importanc� of artificial intelligence and with transferred to a word processor or for integration pur­ it the concept of knowledge representation poses to another system (thesaurus maintenance or re­ renewed interest in machine translation. trieval system). The display of vocabulary usually follows standards These factors which cannot be seen isolated but rather and conventions which have developed in more than as interrelated determinants have resulted in various thirty years. Most thesauri have at least one alphabetical trends (6 - 11): and one systematic section, often KWIC/KWOC or hier­ (1) Growing importance of multilingual thesauri which archical displays, sometimes even graphical displays are facilitate cooperation among institutions from different added. The representation of three types of relations countries in the exchange of data and in indexing and (equivalence, hierarchical, associative relationship) in retrieval of documents in a common database. Multi­ accordance with the respective standards on thesauri and lingual thesauri can also help users of online data-bases their construction (14,15) has also become a common to search in their own language for documents indexed trait of most, though certainly not all thesauri (4). in a foreign language. Finally, multilingual thesauri have gained increasing importance since they can be inte­ c) Indexing and Retrieval with the Th esaurus grated into expert systems supporting machine transla­ All thesaurus programs support the construction, tion and retrieval (3,6). maintenance and printing of thesauri. (2) Renewed and increased interest in the compatibility It is possible to index documents and search for rele­ or integration of various indexing languages. The crea­ vant information in a retrieval program when a printed tion of so-called switching languages/intermediate lexi­ thesaurus is used. cons and mapped or integrated thesauri (12) admits, The integrated thesaurus can support specific tasks as­ above all, an easy integration of data downloaded from sociated with indexing and retrieval, by acting as an in­ other systems into one's own database, supports re­ teractive instrument for orientation about potential trieval in the search in various document collections index and search terms, a tool for consistency controls (either in-house' as with office automation or in online regarding the data used in indexing and searching and a databases (8,9» indexed with various indexing lan­ tool supporting update routines. guages (either of one or different natural languages), and can suggest amendments concerning the specificity d) Exchange, Integration and Compatibility and scope of the vocabulary (12). of Vocabularies (3) Increased importance of additional differentiation of This aspect refers to all of the above mentioned com­ relations between concepts in thesauri which can be in­ plexes of thesaurus work, it is, however, given special tegrated into expert systems as knowledge bases, for in­ attention here because it is gaining more and more im­ stance for improvement of automatic indexing and so­ portance in the near future and entails a variety of called intelligent information retrieval (8,9,13). specific tasks and functions, above all: These changes in the environment of the thesaurus must be taken into account when the functions of soft­ batch input and output of machine-readable data in ware are defined. a suitable format special types of relation for multilingual thesauri

B.2 The Tasks of Thesaurus Software (either with a dominant language or with equal lan­ guages) or the connection of various thesauri Considering what has been said so far, the specific functions of thesaurus programs can be related to tasks use of compatible vocabularies in indexing and associated with the following complexes: searching.

Int. Classif. 17 (1990) No. 3/4 Ganzmann: Evaluation of thesaurus software 149 C. The Definition of Criteria work, and desirable requirements which might be im­ With regard to the criteria for the evaluation of portant in more specific applications. thesaurus software the author could refer to various sources: C.l Structural Definitions a. the literature on thesaurus construction and main� Structural Definitions refer to those features that tenance which mentions some of the possible applica� determine above all the degree of complexity and differ­ lions and advantages of the computer (2,3,16,17). entiation which the software will admit in the construc­ b. the check-lists of three authors concerned with the tion of a thesaurus and therefore to criteria highly im­ development of criteria for the evaluation of thesaurus portant in the evaluation of thesaurus software, namely software (11, 18, 19). term resp. term related attributes and the number and Whereas the general literature on thesauri lacks sys� types of relations that can be defined between terms. tematization of criteria and mentions only some aspects relating to the evaluation of thesaurus software, the checklists have been more specific in their definitions of C.I.I Term and Term Related Attributes criteria. Still, these checklists have either been Any thesaurus software program must help the user developed for and applied to the evaluation of stand­ to create the terminology suited to his specific purpose. alone programs (11,18), been restricted to a limited set Two decisive factors determining the degree of flexibility of criteria (19), or more specific criteria are only made are length and number of fi elds. explicit in the discussion and evaluation of a thesaurus Basic requirements for any thesaurus program are, be­ program (11). sides the term field, a field for scope notes to define the The approach chosen here differs from that of the meaning of terms, and a field for notations to facilitate three authors (11,18,19) in that it integrates criteria for the structuring of vocabulary and a systematic display. stand�alone and integrated thesaurus software and adds Some thesaurus programs restrict the length of the criteria not mentioned in these check-lists. term field to 50 characters or less which can be problem­ atic for two reasons: The criteria in my check-list are categorized as follows: - the more specific the vocabulary of a thesaurus the (1) general criteria (i.e. criteria applicable to any type more likely are compound terms, but also p'roper of software): names (for instance of parties) which can have con­ criteria relating to technical data concerning hard­ siderable length. ware configuration, operating system and the soft­ - exchange of data can be made impossible when words ware package evaluated from a source thesaurus program with more charac­ criteria relating to "market data" (development data, ters than admitted by the target program are to be pricing, support, acceptance) transferred. general criteria relating to overall ergonomics with Since the selection of vocabulary should not be dic­ respect to software and documentation tated by restrictions in the length of the term field it should admit at least 50 characters and - ideally - be criteria relating to data security (access controls, user user-definable. views, backup routines, recovery etc.) Scope notes should preferably be of variable length to (2) criteria relating to the specific tasks of thesaurus soft­ provide for long definitions, and the field for notations ware should ideally be user-definable in order to admit even Since the purpose of this paper is the definition of more complex notations (like, for instance, the lon.g criteria for thesaurus software with respect to the specific MESH notations) .. requirements of a tool for the construction , maintenance Apart from these basic requirements there are further and updating it will concentrate on aspects relating to desirable features, depending on the size, planned dis­ these functions of the software. General aspects to be play and the languages to be used: considered with any type of software will be left aside in a field for the recording of sources of the origin (s) the following discussion though they have been included of terms duting the collection of terms in the checklist since they may well affect a user's deci­ a field for facet codes in case grouping of vocabulary sion, too.2 Ergonomic aspects have been considered in according to facets is planned the following discussion only where they can be referred finally a field for information on the language the directly to specific functions of thesaurus software . term belongs to (for muitilingual thesauri). In the following I will concentrate on the most impor­ tant criteria to measure the capability of tools for thesaurus construction and application under several C.1.2 Relations headings relating to the process of construction and ap­ With regard to the types of relations the demands of plication of the thesaurus and discuss the criteria with individual institutions can conflict· with the standards respect to their relevance concerning the different tasks which recommend the representation of three types of and functions. Taking into account that the various func­ relations (synonymy, hierarchical relation, relatedness). tions of thesaurus software mentioned are not equally Many institutions have chosen to ignore these standards important I have differentiated between basic require­ and/or have defined additional relations (4). To meet ments, i.e. requirements fundamental to any thesaurus these demands any thesaurus program should ideally let

Int. Classi!. 17 (1990) No. 3/4 150 Ganzmann: Evaluation of thesaurus software the user define freely the type of relations he wants to Display of data on the screen does also refer to these use. phases of work but will be treated in connection with the lf the relationships cannot be defined freely the fol­ general aspects relating to the output of data (cf. C.3.1.). lowing requirements apply: The mode of input should ideally be possible by key­ - The standard relations as defined in the standards board as well as by data capture from outside in batch (equivalence, hierarchical, relatedness/associative mode. Batch input of thesaurus data is becoming increas� relationship) are a must. ingly important since thesaurus construction can involve If the associative relationship between terms cannot the adaptation of parts of already existing thesauri and be defined individually by the user, as is sometimes additionally there is, as already pointed out, a growing the case in retrieval software, this means lack of a interest in the integration and merging of machine�read� structurizing device facilitating orientation in index� able thesauri. At present there is no generally'accepted ing and retrieval in general and especially in the social exchange format for thesauri3 so that the batch input of sciences where the associative relation plays an im� structured thesaurus data generated by another program portant role since the relation between concepts can­ depends to a great extent on the ability of the source not always be clearly defined in terms of hierarchies. system to create the format needed by the target system. Often more specific relations which can also be Consistency controls concerning the relation structure as well labelled equivalence relations are required: on the as the terminology are of prime importance in one hand the relation between compound terms thesauri. Any program for the Gonstruction and main� which are not to be used for the representation of tenance of thesauri must include consistency checks, to concepts and their semantic factors, on the other avoid illogical features in the thesaurus terminology and hand the kind of rel�tion which exists between a non� structure. The most important checks are those that pre­ descriptor and two or more terms to be used alter� vent: natively instead (e.g. homonyms or broad concepts multiple entries of the same term (duplicates), unless represented by different resp. more specific terms). the user wants to construct compatible or multilin� Especially semantic factoring can be very important gual vocabularies since it supports consistency in indexing/retrieval - multiple relations between two terms (either of the concerning the representation of complex concepts same or of different type) and is necessary when so-called "inexact equiv­ incomplete relations, like alences" between terms from different natural lan­ - relation from one term to another without rec� guages in the construction of multilingual thesauri proc- ity are to be treated in accordance with the respective - relation between only one factor and com­ standards (20, 21). pound terms For the construction of compatible or multilingual rejection of relation other than that of synonymy be­ thesauri the program should let the user define tween a non-descriptor and a descriptor special equivalence relations between different in­ - rejection of contradictory relations between terms dexing or natural languages to admit compatibility across several levels of a hierarchy. or multilingual vocabularies (either with equal status These consistency checks must of course be applied or a dominant language). when terms are entered and the relations are defined but - In many instances, especially where large vocabular­ any time a term or relation is modified or deleted these ies are compiled and displayed the differentiation of changes must be reflected' correctly and consistently in hierarchical relationships (generic and partitive rela­ the thesaurus, too. tionship) can be helpful. With regard to the number of relations individual terms C.3 Output of Terms and Relations can have, there'must be no restrictions (as long as they As has been already said, output of data is necessary do not affect consistency (cf. C.2.)). This applies espe­ in all phases of thesaurus construction and maintenance cially to the relationship between terms and their and must be possible on the screen, by the printer or into broader terms (polyhierarchy as a characteristic trait of a file. In addition it can be referred to the output of the thesauri) and to the more specific equivalence relations thesaurus on the screen when the thesaurus is applied in described above - semantic factoring into v'arious (three a retrieval system. The specific requirements of an inte­ or more) factors or representation of homonymous grated thesaurus concerning orientation and support of terms or broad concepts by alternative terms in the indexing/retrieval are defined in the section on indexing thesaurus shouid be made possible if the vocabulary re­ and retrieval (cf. CAl. quires it. The output of data must optimally support the routines of conctrucHon and updating by presenting the vocabu­ C.2 Input of Terms and Relations lary from various angles, in various forms and freely defi­ nable layout (for the print). Capture of data, modification and deletion of terms and their relations are the typical work routines, when thesauri are constructed and maintained with a computer C.3.1 Display on the Screen program. Most important are the mode of input and con­ In any of the phases of thesaurus construction and main­ sistency controls relating to vocabulary and relations. tenance, display on the screen is necessary, above all:

Int. Classif. 17 (1990) No. 3/4 Ganzmann: Evaluation of thesaurus software 151 in order to check if terms have already been entered consistency of large vocabularies with many hierarchical in order to check if all necessary relations have been levels and many compound terms, namely hi�rarchical established displays as well as KWIC or KWOC lists. These addi­ in order to check if deletions of terms or relations tional forms of display cannot usually be generated by have been performed correctly by the program microcomputers, except for one defined term and its upper and lower terms (hierarchical display) or a defined Selection of words for specific modifications or checks KWOC). according to certain criteria will la.rgely depend on how string (KWIC, Optionally the program should be able to generate a differentiated the description of a word can be made in graphical presentation of the thesaurus. The standard terms of its attributes, that is on the structural defini­ thesaurus programs do, however, not offer that option. tions. Output of terms according to the various relevant More specific questions of displ,ay relate to the details criteria (source, notation, facet, language, type of rela­ of presentation of the vocabulary. The user should be tion) should be possible as well as a combination of cri­ given as much freedom as possible so that he can print teria (including strings). It should also be possible to a thesaurus according to his own needs. select certain words that have been previously marked Especially the layout for the print of the thesaurus should for specific purposes (deletion, modification, change of be freely definable with respect to pagination, typo­ relationship etc.). graphic differentiation of descriptors and non-descrip­ The display fo rms on the screen must -for orientation tors, line pitch etc. purposes - include an alphabetical display (with or Free selection of the types of relations to be included or without relations resp. term attributes) and a systematic ommitted in the print (e.g. only synonymous terms, no display of terms. Desirable requirements are a KWIC implicit relations or relations not defined in the stand­ display for large vocabularies with many compound ards etc.) and free definition of their external repre­ terms and a hierarchical display to check the consistency sentation by abbreviations (either user-defined or of the hierarchical relations. by standard national or international reference codes) can As concerns the possible forms of interaction naviga­ be important, since it facilitates the output of the tion through the semantic net of the thesaurus should be thesaurus for a variety of applications (e.g. comparabil­ made as easy as possible, and edJting functions should ' ity of thesaurus structure for integration or compati­ ideally also apply to the term lists in order to facilitate bility) while admitting the maintenance of a very com­ the routines of construction, modification and deletion. plex thesaurus with a multitude of specific relations (e.g. for use as a knowledge base). C.3.2 Output by the Printer The output of word lists by the printer is necessary for several reasons: C.3.3 Output into a File the physical limitations of the screen can only offer Transfer of data to other systems will ask for a special a restricted view on the vocabulary at a time format of interchange which the program must be able to generate. The transfer of thesaurus data to a word controls of word-lists are easier with printed material processor or of word lists into a thesaurus software is usu­ not all institutions make use of an integrated ally no problem as long as the program can generate in thesaurus software but want to use a thesaurus standard exchange formats (e.g. ASCII) whereas the print. transfer of structured thesaurus data into a retrieval sys­ As has been pointed out the ability of a thesaurus pro­ tem or a thesaurus maintenance program is much more gram to present the vocabulary from various angles de­ difficult since there is no generally accepted exchange pends above all on how differentiated the term descrip­ format for thesaurus ,d ata and the input formats required tion (term related attributes) can be made. The selection by the individual thesaurus maintenance programs and should be possible according to the various criteria of retrieval programs vary greatly. As has been said, a selection as defined for the screen thesaurus. Again all thesaurus software should ideally be able to generate selection criteria should be freely combinable for the some or various formats necessary for this transfer, but output. usually these devices will have to be programmed on As regards the display form of the vocabulary (or demand. selected parts of it) a program for the construction of thesauri must, in accordance with the standards, at least be able to generate a display in an alphabetical array C.4 Indexing and Retrieval Functions (including the scope notes and indications of relations of Integrated Thesauri . between terms) and in a systematic display (with the no­ In addition to the requirements mentioned so far in­ tations necessary for systematic ordering and indication tegrated thesaurus modules in retrieval programs of the hierarchical level of a term, ideally also with in­ demand specific functions in order to gain an advantage clusion of node labels for a faceted display in a clas­ over the printed thesaurus. saurus). Additionally the multitude of known variations For easy orientation on the vocabulary and the seman­ and various additional forms of display (e.g. broad sub­ tic relations among terms it must be possible to display ject groups, ordering by facets) (17) should also be the thesaurus on the screen both alphabetically (both as generated by a computer program, above all the two a single word-list and with indications of relationships) types of display useful for structuring and controlling and systematically while the user is indexing and search-

Int. Classif. 17 (1990) No. 3/4 152 Ganzmann: Evaluation of thesaurus software ing. Orientation on the term best suited for the repre­ This means above all: sentation of a given concept is greatly facilitated when The software must provide for: adequate term length navigation through the thesaurus, i.e. the quick switch­ (not less than 50 characters), scope note facilities and ing from one term to semantically related terms and notation. browsing through both the alphabetical and systematic In the interest of effective"application in indexing and display is easy. retrieval every program must admit at least definition The function of control of input must be supported by of the basic types of relations, semantic factoring and the software as well, that is the program must check differentiation of hierarchical relationship being a whether a term used for indexing is contained in the elec­ desirable requirement. tronic thesaurus and eventually reject it (unless other­ Since the applicability of the thesaurus in indexing wise defined by the user). Furthermore the software and retrieval as well as the updating depend largely should enable the user to select descriptors directly from on the consistency concerning vocabulary and rela­ the online thesaurus without having to enter them. Of tions, the consistency checks as described above can course a program should automatically replace those be regarded as vital for any program. terms contained in a thesaurus but not admitted for the representation of concepts. This will be the case when­ Any program should at least be able to display the ever equivalencies between a given set of terms can be thesaurus in alphabetical and sys.tematic order on the represented by one term only whereas the others serve screen, transfer it to the printer or into a file. as lead-in vocabulary, e.g. non-descriptors (normal syn­ All the functions relating to orientation and control onymy, compound-terms - provided the program admits with regard to the integrated thesaurus must be avail­ semantic factoring) and also terms from a secondary lan­ able in a retrieval program, in order to guarantee guage in the case of multilingual thesauri. consistency and predictability in indexing and re­ The updating of vocabulary can be greatly facilitated trieval. Support in the formulation of requests and when the frequency of use of thesaurus' terms in indexing in the updating of the thesaurus can also be regarded and retrieval can be verified. Statistics on whether and as elementary functions. how often a term has been used for indexing and search­ A short glance at software packages available suggests ing can help identify failures of a vocabulary with regard that most programs fulfil the basic requirements for to specificity and scope of vocabulary. The system must thesaurus software. admit global changes, i.e. whenever a descriptor that has But in several of the programs some of the basic require­ been used in indexing (and retrieval) is replaced by ments are not fulfilled4: another term better suited to represent a concept (either - the maximum number of characters is limited to 40 a former non-descriptor, a completely new term or a or even less term of broader scope), the former entries in the index the associative relation is either missing completely must be replaced automatically by the new descriptor in or cannot be defined explicitly by the users the inverted file. semantic factoring is not possible or the number of One specific advantage of the online thesaurus in re­ factors is limited to two trieval is its ability to facilitate the fo rmulation of a search strategy . Automatic generic search (i.e. searches for all fields for scope notes and notations are missing documents indexed with terms belonging to a specific consistency controls concerning the vocabulary and hierarchy) is a function common to many retrieval sys­ the relations are either missing completely or only tems and can be regarded as a basic requirement of incomplete thesaurus software. The user should, however, also have presentation of the systematic display either on the the opportunity of automatic inclusion of all terms de­ screen or on paper is not possible fined as related terms of a specific search term to op­ control and orientation devices in indexing and re­ timize recall. trieval are missing (e.g.: no automatic replacement of non-descriptors or rejection of non-thesaurus terms), no alphabetic display of terms and difficulties D. The State of the Art in navigation As the discussion of criteria relevant to the evaluation automatic generic searches are not always possible. of thesaurus software has shown there are a variety of Considering this short survey of substantial shortcom­ features relevant to almost any application of thesaurus ings one can only assume that some thesaurus programs software whereas others will largely depend on the user's (esp. those integrated into retrieval software for micro­ specific environment !lnd requirements (KWIC, defini­ computers) have been developed for end-users who tion of layout, multingual thesauri etc.). apply their prC'grams to small-scale if not mini-thesauri Certain qualifications can be made as to "ital and basic and small-scale document collections. Whether a user elements any thesaurus software should have. Especially will - in the long run - be satisfied with a software of important are those features which determine the struc­ such limitations is doubtful. ture and complexity of the vocabulary and features re­ However, even thesaurus software which fulfills all lated to consistency control and display of the vocabulary basic requirements will, in many cases, not be sufficient. since the value and applicability of the thesaurus will Thesaurus construction still means above all creating an largely depend on these features. indexing language optimally suited to the specific situa-

Int. Classif. 17'(1990) No. 3/4 Ganzmann: Evaluation of thesaurus software 153 tion where it is to be used, i.e. to the potential or real Notes: size of the document collection, to the user's demands, 1 Cf. for instance thespecialissueofNachrichten fiir Dokumen­ tation (1) on thesauri and the second editions of Aitchi sonl Gil­ to the type of documents, to the subject field etc. christ (2) and Lancaster (3) Thesaurus features like vocabulary size, the relations de­ 2 For an extensive and comprehensive discussion of these fined and the display of vocabulary are interdependent aspects the reader is referred to Van Wyk's (22) checklist for retrieval software but also determined by these specific conditions of a 3 MATER (23) has not been accepted as an exchange format for given environment. thesauri so far and any feasible alternative has not been In order to meet these needs a good thesaurus pro­ developed yet 4 This judgement is based both on the author's experience with gram should leave as much flexibility and freedom in various programs and on the literature on this topic (11 ,18, 19) adapting the indexing language to one's requirements with respect to size and complexity of the vocabulary, the definition af relations and the display of vocabulary. References: Additionally, thesauri must more and more respond ( I) Nachrichten fUr Dokumentation 39(1988) No. 1 to needs and demands from outside the traditional in­ (2) Aitchison, J., Gilchrist,A.: Thesaurus Construction and Maintenance: A Practical Manual, 2nd ed., London 1987 dividual working environment. The new tendencies have (3) Lancaster, F. W.: Vocabulary Control for Information Re­ already been outlined: Office automation calling for trieval, 2nd ed. , Arlington 1986 compatibility or integration of thesauri resp. subsystems (4) Schondorf, P.: Nicht-Konventionclle Thesaurusrelationen als Orientierungshilfen fUr Indexierung und Recherche - of thesauri, the increasing importance of artificial intel­ Analysen ausgewahlter Beispiele. Nachr. Dok. 39(1988) ligence calling for additional relations, but above all a p.231-244 trend that will more and more affect the end-user (5) Burkart, M.: Neue Thesaurusansatze- frischer Wind in alten Segeln? Nachr. Dok. 39(1988)p.207-208 directly, namely the exchange and integration of data (6) Svenonius, E.: Compatibility of Retrieval Languages: Intro­ from different systems which demand for greater open­ duction to a Forum. Intern. Classif. 1O(1983)p.2-4 ness of the systems both in a technical and conceptional (7) Dahlberg, 1.: Towards Establishment of Compatibility be­ tween Indexihg Languages. Intern. Classif. 8(1981)p.86-91 respect, that is, import and export of thesaurus data must (8) Rada, R., Martin, B.K.: Augmenting Thesauri for Informa­ be facilitated by suitable import and export facilities of tion Systems. ACM Transact. on Office Inform. Syst. the software, and the software must support the con­ 5(1987)p.378-392 (9) Rada, R.: Connecting and Evaluating Thesauri: Issues and struction and use of compatible languages (multilingual Cases. Intern. Classif. 14(1987)p.63-69 or monolingual). (10) Panyr, J.: Thesaurus und WissensbasierteSysteme-Thesauri Taking a look at the available thesaurus programs with und Wissensbasen. Nachr. Dok. 39(1988)p.209-215 (11) Sick, D.: Autbau und Pflege komplexer natiirlichsprachig these additional requirements in mind it seems that a basierter Dokumentationssprachen (Thesauri): Aktuelle considerable portion of these programs do not meet all Tendenzen und kritische Analyse einer ausgewahlten autono­ of the requirements. Of course, it is the end-user who men Thesaurus-Software fUr den Personal Computer. Uni­ versitat Saarbrucken 1989 must decide which software is best suited to his purposes. (12) Aitchison, J.: Integration of Thesauri in the Social Sciences. There is no harm in choosing a software that fulfils only Intern. Classif. 8(1981)p.75-85 basic requirements as long as the functions of the soft­ (13) Schwantner, M.: Entwicklung des Indexierungsworterbuchs PHYSIPILOT. In: Deutscher Dokumcntartag 1987. Ed. H. ware meet the end-user's requirements. It is, however, Strohl-Goebel. Weinheim 1988. p.329-339 harmful if the user buys the wrong software product be­ (14) International Organization for Standardization: Documenta­ cause either he does not exactly know what he must ex­ tion: Guidelines for the Establishment and Development of Monolingual Thesauri:ISO 2788, 2nd ed., 1986 pect from the software or what he wants to achieve with (15) Deutsches Institut tilr Normung: Erstellung und Weiterent­ it. The design of a thesaurus should not be dictated by wieklung von Thesauri: Einsprachige Thesauri: DIN 1463, the software but the software should be made subservi­ r.l. Berlin 1987 (16) Soergel, D.: Indexing Languages and Thesauri: Construction ent to the requirements of thesaurus work. and Maintenance, Los Angeles 1974 The actual purpose of this paper has been to provide (17) Wersig, G.: Thesaurus-Leitfaden: Bine Einfiihrung in das the end-user with criteria in order to help him select the Thesaurus-Prinzip in Theorie und Praxis. 2nd ed.,Munchen 1985 appropriate software for his specific needs. If, however, (18) Ritzier, c.: Vergleichcnde Untersuchung von PC-Thesaurus­ one takes into account the present state of the art of pfiegeprogrammen, FHDarmstadt 1989 thesaurus software, it seems that the check-list might (19) Rohou, c.: La Gestion Automatisee des thesaurus; Etude Comparative de Logiciels. Documentaliste 4(1987)p.103-108 well be useful in more than one way: Like check-lists for (20) Deutsches Institut fur Normung, Erstellung und Weiterent­ other types of software, it may also - hopefully - stimu­ wicklung von Thesauri: Mehrsprachige Thesauri: DIN late some of the developers of thesaurus software to re­ 1463,T.2, Berlin 1988 (21) International Organization for Standardization, Documenta­ vise their software conceptions and improve their some­ tion: Guidelines for the Establishment and Development of times insufficient programs. Multilingual Thesauri: ISO 5964. Genf 1985 (22) Van Wyk, J., Tonsing, R.E.: The Evaluation of Interactive Information Storage and Retrieval Software. South African J. of Libf, and Inform. Sci. 55(1987)p.34-50 Dr. lochen Oanzmann, (23) Deutsches Insti tut tilr Normung: Format fUr den maschinellen Lehrinstitut fUr Dokumentation in der DOD Austausch terminologischer/lexikographischer Daten - Westendstr. 19, D-6000 Frankfurt 1 Mater: Kriterienkatalog: DIN 2341. Berlin 1986

Int. Classi!. 17 (1990) No. 3/4 154 Ganzmann: Evaluation of thesaurus software CHECK·L1ST FOR THESAURUS SOFTWARE - examples - training disc A. GENERAL CRITERIA - tutorial

1. Technical Data 6.2 Software Ergonomics Hardware Compatibility: * Language of User Surface - computers on which software runs * Complexity of Screen Layout: - storage required: - structure of information - RAM - colouring - external storage devices - window technique - operating systems - single user * Dialog Forms: - multi user - command driven - menu driven * Software Package - hybrid - programming language - mouse - single user - multi user * Help Functions * Messages: 2. Development Data - self-explanatory * Developer - explained in manual - error messages * Versions: - feedback messages - recent version - first version - alert messages - confirmation messages - overall number of versions * Provision for Different User Levels 3. Prices 7. Data Integrity * Software Package: - single user *' Access Control: - multi user - password - restrictions for individual users * ExtraslModifications - restriction to specific databases * Updates: - restrictions to specific functions - single user * Backup Procedures - multi user - automatic * Support: - storage device - Installation * Reorganisation Features - updating - application * Recovery Features - hotline * Training B. CRITERIA RELATING TO FUNCTIONS * Discounts OF THESAURUS SOFTWARE 4. Support 1. Structural Delinitions * Supporting Institution 1.1 Term and Term Related Attributes * Forms of support Predefined Fields for: - hotline telephone - Term - consultation - maximum number of characters - training - Scope Noterrext - newsletters - maximum number of characters - active support - Notation - installation - no differentiation - updating - differentiation for: - modification - broad categorization (subject groups/facets) 5. Acceptance - systematic categorization - maximum number of characters * Number of installations - Source of Term * User groups - maximum number of characters

* Reviews in articles - variable length - Information as to Language of Term 6. Ergonomics - maximum number of characters 6.1 Documentation - additional fiefds - maximum number of characters * Types of manual: - operations manual * User Definitions - user manual - number of fields * Parts included: - length of fields - table of contents - sequence of fields - documentation of: 1.2 Relations - technical specifications 1.2.1 Among Tenus of One Vocabulary - installation (Monolingual Thesaurus) - application - error messages * Definition of Relations: - backup and recovery - predefined relations - index - relations user-definable * User Friendliness: Number of Predefined Relations - structure of manual * Types of Relations: - completeness of information - equivalence relationship: - correctness of information - normal synonymy (non-descriptor(s) � descriptor) - clarity: - semantic factoring (nondescriptor � descriptors) - style - alternatives (non-descriptor � alternative descriptors)

Int. Classi!. 17 (1990) No. 3/4 Ganzmann: Evaluation of thesaurus software 155 - hierarchical relationship: 2.4 Consistency Controls - no differentiation * Definition: - differentiation of partitive and generic relation - predefined - definition of dividing principles (categories) - user-definable - associative relationship: - no differentiation * Term and Term Attributes: - differentiation of various types (e.g, predecessor- - rejection of duplicate entries of the same term successor, appurtenance relation etc.) - modification of control possible for input of several - which relations? natural or indexing languages - definition of admissibility of characters for attribute fields * Number of Relations between Individual Terms: (language codes, notation etc.) - equivalence relationship: - normal synonymy (max. number of non-descriptors * Relations: per descriptor) - control of reciprocity of relations - semantic factoring (max. number of factors - rejection of more than one type of relation between per non-descriptor) two terms - alternatives (max. numberof alternative descriptors - rejection of incomplete relations (e.g. semantic factoring per non-descriptor) with only one factor) - hierarchical relationship: - rejection of duplicate relations of one type between - number of lower terms per broader term two terms - number of broader terms per lower term (polyhicrarchy) - rejection of hierarchical or associative relationship - number of hierarchical leve1s between descriptors and non-descriptors - associative relationship - control of illogical relations across hierarchical levels - other controls 1.2.2 Among Tenns from Different Vocabularies 3. Output >I'- Type of Vocabularies: 3.1 Display on the Screen - multilingual thesauri - compatible vocabularies. * Mode of Search for Terms: - browsing * Connection Between Different Natural Languages - scrolling (Multilingual Thesauri) - other possibilities - max. number of different lan�uages - status of individual language(s): * Display of Individual Terms - equal languages - with attributes - dominance of one language - with relations * * Connection between Different Indexing Languages: Display of Word-Lists - max. number of indexing languages - criteria for selection of terms: - types of indexing language: - alphabetical section - classifications - strings - thesauri - attributes (language, notation, source etc.) - status of individual language - types of relation . - words marked for specific purposes * Mode of Connection: - combination of criteria - reference of terms to a switching language - forms of display ofword�lists: - direct translation of different vocabularies - alphabetical array: (mapping of vocabularies) - word-list - word-list plus relations and attributes 2. Input (Thesaurus Construction and Maintenance) - other variations 2.1 Capture of Data - KWIC-display >I'- Mode of Capture: - hierarchical display - batch input from other system - systematic presentation (sorting by notation) - keyboard: - detailed system - mode of input of terms and attributes - without reference to relations - mode of input of relations - with reference to relations * Ease of Capture: - broad categories (subject groups/factes) - complexity of input of terms and relations - graphical display - separate steps? * Interaction Possible in Thesaurus on Screen: - fixed sequence of input routines? - scrollinglbrowsing - display of entered terms (and relations) on screen - navigation to semantically related terms - automatic derivation of implicit relations - selection of terms for editing and deletion - direct modifications and deletions in lists 2.2 Modification * Mode of Modification: 3.2 Output by the Printer - global changes possible (of language codes etc.) '" Definition of Output Formats: - keyboard - standard formats predefined - mode of modification of terms and attributes - user definable formats - mode of modification of relations - storage of user defined formats * Ease of Modification: * Criteria for Selection of Terms: - complexity of modification - alphabetical section - ease of changes affecting the status of terms - strings descriptor - non-descriptor) - attributes (notation, facet etc.) - display of terms (and relations) on screen - types of relation - combinatio'l of criteria 2.3 Deletion * Forms of Display: * Mode of Deletion: - alphabetical array - global deletions of terms/relations - without further information - keyboard - with relations - mode of deletion of terms and attributes - with attributes - mode of deletion of relations - KWIC-display * Ease of Deletion - hierarchical display - complexity of deletion - without relations - automatic deletion of relations of a term deleted - with relations

Int. Classif. 17 (1990) No. 3/4 156 Ganzmann: Evaluation of thesaurus software - systematic presentation (sorting by notation) - user-definable for use of candidate terms - detailed system - replacement of thesaurus terms not admitted for indexing: - without relations - replacement of compound terms by semantic factors - with relations - replacement of non-descriptor by descriptor - with attributes (for thesauri with preferred terms) - with node labels - replacement of terms in s�conda!r language by �erms - broad categories (subject groups/facets) from dominant lang�age 10 multtlmgual thesaun - graphical display . * - display in columns for multilinguallcompatible vocabulanes Representation of concepts; - preferred term (descriptor) * User-definable Features: - no preferred term - information added to terms: - relations * Updating: - attributes - global changes in index - presentation of the rel tions: - statistics on use of descriptors � . ... . - suppression of certam relations (e.g. ImplIcit relattons) - sequence of relations in print 4.2 Retrieval -- user-definable reference codes for output ,;, Orientation: (e.g. in accordance with ISO/DIN) - display forms of thesaurus on screen (cf. also 3.1.): - layout: - alphabetical display pagination - systematic display line pitch - other forms of display - caption . - search mode for terms - typographic differentiation of descriptors/non-descnptors - navigation through semantic st ucture - other features : * Mode of Input: 3.3 Output to a File - entering of terms Formats of Output: - direct selection of terms from screen thesaurus - ASCII file * Control of Input: - Special format required by other syst�m (i.e. retrieval software, thesaurus maintenance program) - rejection of unknown terms - replacement of thesaurus terms not admitted for the 4. Indexing and Retrieval representation of concepts: 4.1 Indexing - replacement of compound terms by semantic factors . - replacement of non-descriptors by descriptors (in thesaUrI * Orientation: - display forms of thesaurus on screen (cf. also 3.1): with preferred terms) - alphabetical display - replacement of terms from secondary language by terms - systematic display from dominant language in mutilingual thesauri - other forms of display - automatic inclusion of all synonyms (in case of thesauri - search mode for terms without preferred terms) - navigation through semantic structure ,;, Formulation of Search Strategies: * Mode of Input: - automatic generic search option - entering of terms - automatic search for related terms - direct selection of terms from screen thesaurus .- automatic inclusion of search term predecessors * Control of Input: * Updating: - rejection of unknown terms - statistics on the use of search terms

ASTM Symposium, Cleveland 1991 Terminology for Knowledge Transfer, The ASTM (Amer .Soc.for Testing Materials) Sym­ Vienna, 1991 posium on Standardizing Te rminology fo r Better Com­ A second announcement on the 3rd Infoterm Sympo­ munication: Practice, Applied Theory, and Results sium to be held under this topic in Vienna, October will take place from June 13-14, 1991 at Cleveland, OH. 1991 has been released. Its sUbtopics cover: Terminology It is sponsored by the ASTM Committee on Terminology work by subject specialists; Preparation of high-quality in cooperation 'with the American Translators Associa­ multifunctional terminologies; and International Coopera­ tion (ATA), the Terminology and Linguistic Services tion. The following seven themes have been outlined: Directorate of the Government of Canada, and the General theory of terminology and special theories of Canadian Standards Council. Authors are invited to terminology - Terminology unification in theory and present papers on practical methods and techniques of practice - Terminologies and knowledge transfer - terminology work, selecting and defining terms - analyz­ Terminology standardization - Tools to support termi­ ing and classifying concepts, applying terminological nology unification - Applications (in different countries) principles and theories to terminology problems, measur­ - Dataflow management (bibliographical, terminological, ing and enhancing the quality of terminology products, factual data; availability of terminological publications). computer assistance in terminology work, structured Forfurther information turn tothe Conference Secretariat documents as terminology database, experience with at Infoterm, P.O.Box 130, A-1021 Vienna, Austria. online or CD-ROM terminology products, the nature and use of vocabulary control - thesauri, indexing, etc., standardizing terminology in ASTM - ·terminology policies in technical committee, and national and inter­ national activities in terminology. For further informa­ tion contact: Mr. Richard A. Strehlow, Martin Marietta Energy System, Oak Ridge, TN 37831-6088, USA.

Int. Classif. 17 (1990) No. 3/4 Ganzmann: Evaluation of thesaurus software 157 Roy Rada Computer Science Department, University of Liverpool, England

Maintaining Thesauri and Metathesauri

Rada, R.: Maintaining thesauri and metathesauri. An institution which indexes many documents and Int. Classif. 17(1990) No. 3/4, p. 158 -164,30 refs. maintains a large thesaurus may have a special staff for Maintaining a thesaurus is a time-consuming task which should thesaurus maintenance. This thesaurus group consists of go hand-in-hand with the indexing of information and should be experts in the domain of the literature to be processed. supported by software. To connect different document data­ bases their respective thesauri should be related. The most These experts have a grasp of the terminology and straightforward way to support this by computer is to map the semantic subtleties of the subject field. The thesaurus terllls of one thesaurus to those of another. Such a mapping group is responsible for collecting index terms, and creates one kind of metathcsaurus. As citation systems are ex­ tended to include full-text online, a new thesaurus may be used making the thesaurus as up-to-date as possible for the to index individual paragraphs of a document, and a meta­ index group. The index group indexes the documents thesaurus may apply to a universe of paragraphs. To illustrate according to the latest version of the thesaurus and sug­ these principlcs several computer systems are described which help people maintain thesauri and metathesauri. Particular gests new index terms to the thesaurus group. success has been had hy the National Library of Medicine with its Medical Subject Headings and its_ Unified Medical Language System. (Author) 2.2 MeSH Maintenance At the National Library of Medicine, which indexes about 300,000 documents per year, about 100 people are devoted to indexing. About a doten index terms are as­ signed to each document, but an indexer is expected to 1. Introduction spend only 15 minutes on each document. The indexing The advent of computerized information systems has staff is complemented by a thesaurus staff of about 5 increased the importance and practicality of large-scale, people. The thesaurus, the Medical Subject Headings well-devcloped thesauri. Furthermore, the communica­ (MeSH), which they maintain includes about 15,000 tion networks in the world make connecting one main headings in an 8-level hierarchy. thesaurus to another important. How can these While the instructions to indexers emphasize that they thesaurus systems be maintained in a way that is sup­ must base their indexing work on the whole document ported by the computer? This paper addresses the soft­ rather than on just the title and abstract, protocol analy­ ware side of the thesaurus maintenance problem. sis has revealed that in fact the indexers place a heavy emphasis on terms available in the title and abstract. T.'1e 2. Thesauri thesaurus staff also give great heed to the terms in the The effectiveness of a retrieval system is largely de­ bibliographic citation (which contains the title and ab­ pendent upon the document classes in it. These classes stract but not the full article). Accordingly, indexing by may be based on terms from an indexing language (12). mapping terms from the citation against the thesaurus One of the most popular forms for an indexing language can often produce correct index terms for the document is a thesaurus. A thesaurus is a set of concepts in which (18). each concept is represented with at least synonymous The strategy for maintaining MeSH at the National terms, broader concepts, narrower concepts, and related Library of Medicine proceeds along three parallel paths: concepts (5). * If an indexer does not find a needed concept in the thesaurus, then the indexer proposes a new term. The 2.1 Principles indexer uses a special paper form from the thesaurus In the development of a bibliographic citation system, group to indicate the details of the request and then one dilemma is whether to build a thesaurus first and a thesaurus specialist investigates the feasibility of then use it for indexing the documents or to index the this proposed change. documents by free terms and then construct the The thesaurus specialists annually review the usage thesaurus after accumulating free terms. As is often true files for MEDLINE (the database for articles in­ in life , the middle ground is particularly attractive, and dexed with MeSH). If a MeSH concept is not used that means in this case building a thesaurus and indexing in indexing, then the concept is a candidate for re­ documents hand-in-hand. moval from the thesaurus.

Int. Classif. 17 (1990) No. 3/4 158 Rada : Maintaining thesauri and mctathe�;aurj , The thesaurus group each year reviews one or more possible to circulate various requests for changes to areas of biomedicine. For instance, recently the bi� MeSH on the computer and to record on the computer otechnology sections and medical informatics sec� comments. The voting of individuals could be recorded tions of MeSH were reviewed and updated. as well so that decisions would become automatic. The largest number of changes which occur to the Critiquing strategies about the structure of MeSH thesaurus traditionally originate from indexer sugges­ could be automatically invoked. At the moment MeSH tions. But whatever the route to a proposed change, a has at the top-level over a dozen sections and some con­ thesaurus specialist responds to the proposal and the re­ tain thousands of terms while others contain only a few sponse is reviewed both by the entire thesaurus group hundred terms. When a very large thesaurus is accessed and by a Library-wide committee. via a computer screen, the user may need to cliange the As recently as two years ago, the computer tools for contents of the screen many times to find the terms for MeSH maintenance at NLM were negligible. Every spe� a query (27). By balancing the thesaurus the number of cialist had a computer terminal on his or her desk and levels required to get to any given concept is less. The would routinely access MEDLlNE to discover the pat­ , computer could suggest ways of improving balance. terns of usage of a term in the literature, but the com­ Finally, additional relations might be added to the puter system did not directly cater to decisions about thesaurus. The MeSH system has been refined so that thesaurus updating. In order that the yearly update to instances of hierarchical relations are distinguished. But MeSH occurs, a very difficult set of programs had to be a variety of non-hierarchical relations are still not part invoked. Two members of staff would spend several of the thesaurus. Theories of analogical inheritance months meticulously re.cording the approved changes to could be applied by the computer to the thesaurus, if MeSH into the database and then preparing a publish­ these non-hierarchical relations were present (15,22). able record of those changes for the publication of MeSH in paper form. 2.3 Otller Thesaurus Strategies About 4 years ago a requirements document was The strategies for maintenance of thesauri at other in� developed for an improved thesaurus maintenance sys­ stitutions do not take wide advantage of the computer. tem. This document specified that the system should At Elsevier Scientific Publishers, the Excerpta Medica * support the recording of information not unlike that indexers submit proposals to a senior editor who may on the set of forms already used, then order .thesaurus changes. At Derwent Scientific * be either easily usable by all the people who normally Publishers, procedures for changing the thesaurus are were involved with MeSH forms or there should be not formalized nor are there professional staff whose a parallel paper system, principal task is to maintain the thesaurus. , The market place for electronic directories has wit­ produce data that made its way into the thesaurus nessed a wide variety of products over the past few years files of the MEDLlNE system, and (25). The Yellow Pages provides product and service * produce the publications of MeSH. information to an enormous audience. Each advertise­ A prototype system was first developed on dBASE ment may be viewed as a small document. In a typical III + because it gave flexibilityin database and interface scenario, a salesperson has a booklet which lists the index options. To' commun�cate with MEDLINE was, how­ terms available for classify ing Yellow Pages entries. The ever, a problem and conversion routines were needed thesaurus only has two levels. When a salesperson needs to get data to and from MEDLlNE. a new thesaurus term to characterize a product or serv­ The Library's computer division was planning exten­ ice, the salesperson sends a request to the editors of the sive use of the Model 204 database management system Yellow Pages. An Electronic Yellow Pages allows access from Computer Corporation of America. The thesaurus to information from the Yellow Pages in a more flexible group had hired temporary programmers to implement way (4). To take adequate advantage of such an inter­ the thesaurus maintenance prototype, and the computer face, a stronger thesaurus and thesaurus maintenance division was not in a position to support a dBASE III + tools are needed. development. To reduce compatibility problems the When small, academic thesauri are made for particular decision was made to develop the final thesaurus main­ projects, the methods are often less indexing-based and tenance system in Model 204. more committee-oriented (21). A small committee The 'Model 204' thesaurus maintenance system has chooses terms based on a variety of unstated principles facilitated reco'rd keeping and has removed the months and produces a thesaurus by hand. If some principles of of difficult labor at the end of the MeSH year. The in­ good thesaurus structure were more widely recognized, troductory matertal to the publications - those parts software tools to guide people in this kind of structuring which summarize the changes to MeSH of the past year might be more popular. Principles of balance and inheri­ - are relatively easily produced now, whereas' they were tance are not appreciated, and more work is required to difficult to produce before the new system was used, But demonstrate the value of such principles. the new system has not replaced the use of paper forms The major document database systems provide basic in proposing changes nor in circulating information that thesaurus update commands (2) but do not guide main­ is to be discussed at meetings. tenance. BASIS, BRS/Search, and STATUS use the The requirements document noted possibilities which ANSI standard structure for a thesaurus. In STATUS have not been feasible to implement. It would be one cannot relate one narrower term to two broader

Int. Classif. 17 (1990) No. 3/4 Rada: Maintaining thesauri and meta thesauri 159 terms without repeating the narrower term. Main­ TYPE VENN DIAGRAM EXAMPLE tenance of the BASIS thesaurus is done via the Basis No Common Cis empty English and German Menu Language as are other manipulations of the 0 database. BRS/Search provides menu-driven software for routine thesaurus maintenance, and STATUS pro­ Identical Group @8 Dewey Decimal vides a thesaurus handler for adding or deleting terms. With all three systems it is possibie to load a thesaurus F.. sperant{} in batch mode. BRS/Search has a verification procedure External Common @ 0 to validate recipro eating relationships of terms in the thesaurus. Figure 1. Group versus Common Language: G is one of the Several new systems have been recently developed group languages. C is the common language. which include sophisticated thesaurus maintenance pro­ cedures. The THSRS system can present thesauri 3.1. Mapping graphically and advices users about changes to the create an internal common language is by thesaurus so that the number of criss-crossing lines in One way to the graphical depiction is reduced (6). The PRESS sys­ mapping or matching. A variety of rearrangements of terms can be performed to determine whether a kind of tem is intended to support the reuse of software docu­ syntactic match exists between two terms (17). For ex­ ments and generates thesauri from a collection of index­ ing forms (1). Some systems avoid the traditional ample, a computer program might determine that 'short­ ness of breath' is equivalent to 'breath shortness' based thesaurus but use linguistic and word-frequency tools to on a rule that removes the preposition and inverts the dynamically generate synonym classes that are com­ order (11). parable to those which one would want in a thesaurus . (26). Mapping is easily done by computer, and through this method a metathesaurus can be maintained. In the first instance a set of thesauri are connected at a central site by an automatic mapping. On an annual basis changes 3. Melalhesauri to the thesauri can be collected, and the mapping redone. A meta thesaurus transcends a set of thesauri. If people While such automated mapping is clearly inferior in the want to query several bibliographic citation systems quality of its product to what could be achieved with simultaneously, a metathesaurus can help. Historically, more knowledge-based methods. it is perhaps the only there are many examples of metathesauri. In the 1960's, practical method to regularly updating a metathesaurus. a connection between the Universal Decimal Classifica­ The International Classification of Diseases (ICD) is tion and other classification systems was established (7). the World Health Organization's official classification The Armed Services Technical Information Agency and scheme for world-wide morbidity and mortality. The the Atomic Energy Commission have linked their index­ Systematized Nomenclature of Medicine (SNOMED) is ing languages (28). The National Library of Medicine a thesaurus for patient records (30). EMTREE is the has developed the Unified Medical Language System thesaurus used for indexing Excerpta Medica. In a pro­ (UMLS) which connects many of the indexing languages ject of the European Economic Community to connect of biomedicine. The problem is to formulate translation patient records with the medical literature ICD and schemes that preserve the original meaning as much as SNOMED have been mapped into EMTREE (23). possible, while allowing different groups to create or There are fifteen different facets (A through N and Q) change their languages with as much autonomy as in EMTREE and 1 � different classification areas in ICD possible (13). which are numbered 1 through 17. A randomly selected In order for all groups to be able to communicate using set ofICD terms camefrom thirteen different ICD classi­ a common language, each group must either use the com­ fication areas. They mapped into three EMTREE facets: mon language itself or be able to translate between the C. D, and G. Figure 2 shows the mapping ofICD terms common language and the group language (see Figure to EMTREE facets. It is not surprising that an index 1). When no common language is shared by groups that matching of ICD terms to EMTREE would result in the wish to communicate (for example, English and Ger­ vast majority of EMTREE terms being from the C facet, man), each pair of groups must be able to have pairwise since the C facet includes Physical Diseases, Disorders translations made into and out of each other's languages. and Abnormalities and the entire ICD tree relates to the When the common language is the same as all group classification of diseases. languages (for example, Dewey Decimal Classification Over 50% of SNOMED terms have been directly Scheme), the need for consensus is large. Since all groups mapped to equivalent EMTREE terms with simple have to agree on changes, the only changes that are easy matching rules. First exact matching between SNOMED for a group to make will be those that do not affect other and EMTREE terms yielded EMTREE equivalents for groups or those that are valuable to all groups. With an 32% of the SNOMED terms. Replacing SNOMED external common language (for example, Esperanto), a terms with preferred term equivalents from EMTREE group communicates with another by first translating its as defined in the online EMTREE synonym list yielded language into the common language, and then having another 16% of the SNOMED terms. Reversing word the receiving group transrate the common language into order of unmatched SNOMED terms yielded another its own group language. 4% ofthe SNOMED terms. IfSNOMED and EMTREE

Int. Classif. 17 (1990) No. 3/4 160 Rada: Maintaining thesauri and metathesauri terms were morphosemantically analyzed further * The Information Sources Directory contains infor­ matches would be found (19). mation about publicly available biomedical informa­ tion resources. For each source the Directory con­ tains the source's scope, vocabulary, syntax, rules, and access conditions. * The Metathesaurus c::ontains concepts from a variety of biomedical vocabularies. Each concept has a canonical representation that includes a semantic type. Mappings between different vocabularies are handled by this canonical representation. * The Semantic Net represents relationsh!ps among ABC D E F G H J K L M N Q the semantic types. For instance, 'Virus' and 'Or­ ganism' are semantic types, and they are related by the 'is-a' relationship. The semantic type 'Disease' is connected to 'Organism' via the 'caused-by' relation. In the early stages of the Metathesaurus development generic frames were tried as a knowledge representation scheme (3) (8). The fundamentals of medical knowledge \\-ereto be manually encoded by experts into this Meta­ thesaurus. After some trials, however, the inability to 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 create a substantial, useful knowledge base meant another strategy was needed. Lexical matching techniques offered a viable ap­ proach to an early prototype for the Metathesaurus. In January 1988, a UMLS research group at the University of California, San Francisco proposed a step-by-step approach to building a Metathesaurus by beginning with mapping (29). This first proposal became the plan for the first UMLS Metathesaurus prototype and is called Figure 2: The dotted lines show the mapping of leD to Meta-I. EMTRFP (It the top levels of the two tree structures, The purpose of Meta-l is to a_ssist computer informa­ tion retrieval applications and to provide a browsable, machine-readable reference tool. Meta-l contains terms An extension to mapping is merging. One type of from one or more of MeSH, DSM-III (a thesaurus of merging is illustrated by the Vocabulary Switching Sys­ mental disorders), SNOMED, ICD. and CPT (a tem (VSS). The VSS contains the subject descriptors thesaurus of medical procedures). Where possible, the from 15 indexing languages in the areas of physical official name of a concept is the name used in MeSH. science, life science, social science, and business. With Meta-l contains all MeSH and DSM-III terms and any it, a user is able to automatically generate document re­ from SNOMED, lCD, or CPT that lexically map to quests in 15 indexing languages based on a request in MeSH or DSM·III. The link between any two terms just one (17). VSS utilizes five file types: term file, word from different thesauri is classified as either an identical, file, stem file, stem phrase file, and concept file. The a synonym, or a related link. Human intervention is first four are inverted files, and these provide access to required to properly determine whether a link is of the the fifth file. VSS has been shown to reduce search pre­ type 'related' or 'synonym'. After lexically based map­ paration time, improve search strategies and retrieval, pings were ascertained, subject experts reviewed and and increase usage of existing databases. revised the mapping results.

", Each Meta-l record contains three types of informa­ tion: 1) basic facts, 2) relationships, and 3) usage data. 3.2. UMLS The basic facts include the name of the concept, the The Unified Medical Language System. (UMLS) was names of the thesauri in which the concept is located, formally initiated by Dr. Donald Lindberg of the the semantic type (such as Organism), and syntactic National Library of Medicine in 1986. The first two years category (such as noun). The relationships which will of the effort w€re devoted to development of a require­ be represented in a record include broader than, nar­ ments document for the UMLS. The basic assumption rower than, synonym to, and lexical variant. For each is that (10) if the .UMLS is to be improved. it must be term, the hierarchy from the root of a thesaurus to that used. Accordingly, the UMLS will be widely available. term is eXIJicitly recorded in the record for that term. To increase availability, UMLS develop ments do not For instance, the term 'coronary vessels' has a record depend on any particular hardware or software systems. whose relationship fields include two paths from anat­ UMLS has three fundamental components: an Infor­ omy in MeSH (see Figure 3). The explicit encoding of mation Sources Directory, a Metathesaurus, and a this information in each record allows for rapid retrieval Semantic Net: and display of this hierarchical information.

Int. Class;!. 17 (1990) No. 3/4 Rada: Maintainingthesauri and metathesauri 161 semantic primitives for the CMIT diseases in each Two Paths root-to-leaf path. This intersection set describes the Anatomy Anatomy node inheritance properties of each root-to-,leaf path. Use this set to decide what CMIT diseases to add to Cardiovascular System Cardiovascular System the MeSH hierarchy. 4. Repeat step 3 till no more diseases from CMIT are Arteries Veins added to MeSH. Coronary Vessels Cpronary Vessels This algorithm is able to iteratively suggest a small set of diseases which may be connected into the meta­ thesaurus via a narrower-than relation, based on the in­ Figure 3: Position of coronary vessels in MeSH hierarchies. heritance properties of nodes in a root-to-Ieaf path (20). The usage data for MeSH terms is extensive and in­ cludes a list of the frequency with which a qualifier is 4. Individual Documents used to modify the term in MEDLINE and a list of MeSH Document reuse involves classifying existing docu­ terms for other concepts that frequently occur in the in­ ment components and combining these components into dexing of the same articles, For instance, if the term a new document. A system called Many Using and Creat­ 'coronary vessel' occurs 200 times with 'arteriosclerosis' ing Hypertext (MUCH) has been developed to support and 300 times with 'aortocoronary bypass' in MED­ this document reuse via a metathesaurus. Each LINE, then this information would be available in the generated document has an outline which has important 'coronary vessel' record. relations to a thesaurus. The system supports collabora­ Meta-l is a massive information source that occupies tive work and indexing and gives guidance on the struc­ about 100 megabytes. Much of this information occurs ture of the thesaurus. This approach to paragraph index­ in the form of the explicit hierarchies recorded with each ing and document reuse illustrates the superset approach concept and the frequencies of Occurrence that are re­ to a common language. From a global common language corded with each concept. An application of Meta-I users can acquire local languages. could represent the hierarchical information more suc­ The thesaurus model may be seen as a set of link ob­ cinctly but then require more computation time to dis­ jects. Each link object specifies a source node, link type, play the hierarchy around any given concept. Due to its target node, and pointers to paragraphs. A group of link large size Meta-l is distributed on a CD-ROM. objects with the same source node form a frame (see A HyperCard browser for Meta-l has been developed Figure 4) (16). A node-link-node triple is a richer index and is distributed by the National Library of Medicine. for a paragraph than just a node. By placing paragraphs To use that interface one must have a MacIntosh com­ on the links, one can say that a paragraph is about a puter, but one can use Meta-I without the HyperCard certain relation between two nodes - not just about a interface. The interface is not intended as the tool to node. The range of allowed link types goes beyond the connect Meta-I to other applications but rather as a way ANSI standard types. to better see the information in Meta-I. The Meta-I in­ formation can be accessed from the CD-ROM and util­ hypertext ized without the HyperCard browser. With the Hyper­ � The history of link objectl _ _ __> hypertext can Card browser one can point to a concept's relation to h� link: has be .... target: history another concept and then see the new concept on the hypertext screen. Selecting a button labeled 'co-occurring terms' paragraph: 77 link object2 causes another window to open which displays the cooc­ Links are the link: use curring terms in MEDLINE indexing. cornerstone ... target: links CM,IT is a knowledge base of medical diseases with paragraph: 13 several attributes for each disease (9). MeSH, SNOMED, and CMIT were mapped into a relational Figure 4: Graph,ical and frame representation arc given side-by­ data model in one UMLS-related experiment. The over­ side to show the frame representation. all algorithm for merging MeSH, CMIT, and SNOMED is: Programs to help authors in dynamically generating 1. Attempt lexical matches between SNOMED, MeSH, various outlines from the thesaurus have proven invalu­ and CMIT disease names, including all combinations able in guiding the reuse of text and the generating of a of synonyms and eponyms for a specific disease name. cohesive, new document ..The program traverses the 2. Extract a dictionary of disease attributes, called thesaurus and prints a document from the paragraphs semantic primitives, for each disease in CMIT using encountered during the traversal. The encountered various parsing algorithms and use lexical matches be­ terms become the outline of the document. A textbook tween CMIT semantic primitives and MeSH and has been automatically generated in the aforementioned SNOMED synonyms to add synonyms for the seman­ way and been published. tic primitives to the dictionary. A theory of outlines is being developed which relates 3. Find all root-to-Ieaf paths in MeSH relation which to thesauri (14). While an outline is intended to reflect contain two or more MeSH diseases mapped to CMIT some model of the world that makes an impact on read­ diseases, then compute the intersection of the set of ers, a thesaurus and an outline serve the same purpose

In!. Classif. 17 (1990) No. 314 162 Rada: Maintaining thesauri and metathesauri of organizing concepts into a hierarchy. An outline has tation databases but also the full text of the documents an ordering among siblings which is not normally impor­ which are cited (24). After the user has browsed a tant in thesauri. Sometimes the character of the hierar­ thesaurus and through it found relevant citations, he or chical relations in an outline manifest inheritance prop­ she can elect to browse the relevant documents. The role erties and thus reflect what should be true of thesaurus of the thesaurus in browsing facilities for a single docu­ relations. At other times, the outline is like the high level ment is only partially understood. of a thesaurus in that it identifies the facets of a topic rather than the hierarchical relations. Acknowledgements: The author worked at the National Library of Medicine several years ago but is indebted to colleagues there The ongoing experiments with the MUCH system for an update which allowed this paper to be written. Doug occur on a network of Unix workstations with a proto­ Johnston and Peri Schuyler designed the Model 204 MeSH type document reuse tool that has a relational database maintenance system. Peri Schuyler is in charge of the MeSH Section. Joe Hutchins W

Int. Classif. 17 (1990) No. 3/4 Rada: Maintaining thesauri and metathesauri 163 (16) Minsky, M.: A Framework for Representing I

INDEKS - VERLAG

INDEKS-VERLAG. Woogstr. 36a , 0-6000 frankfurt 50

TKE '90: TERMINOLOGY and KNOWLEDGE ENGINEERING

Proceedings of the 2nd International Congress on Tenninology and Knowledge Engineering. 2-4 Oct, 1990. University of Trier. fRG. organized by the Asso­ ciation for Tenninology and Knowledge Transfer. and the Intern·ational Infor­ mation Centre for Terminology (infoterm). Edited by Hans Czap and Wolfgang Nedobity .2 Vols,. together 696p .• ISBN 3-88672-204-X. - 205-8; together OM 12 5,-

Vol.l contains the seven keynote lectures and the 34 papers of sessions I-III. Vo1.2 the 39 papers of sessions IV-VIII and a list of the authors with �heir addresses and page numbers. The k�.>:notes are by A.Ozhincharadze; A.M. TJoa and R,Wagner; W.Rauchi A.Melby; S.Mllke, S.Pinano, H.Uchida and T. Yokol ; M.Schaar; G.Engel and H.Picht. The session topiCS are: Terminology, knowledge theory, and knowledge engineering: New applications; Knowledge­ based systems; Natural language processing and knowledge engineeringi Docu­ mentation languages and ordering of knowledge; Electronic dictionaries; Information management in organizations; Computer. support in technical com- . munication; Terminology and knowledge transfer tools.

Th�re are still some left-overs of the2 proceedings volumes of the first TKE-Congress in Trier in 1987. INDEKS Verlag offers both vol umes, amounting originally to OM 12 5,- now for only OM 75.-

Int. Classif. 17 (1990) No. 3/4 164 Rada: Maintaining thesauri and meta thesauri 14th International Online Information Meeting, London Reports The provisional program of ONLINE 90, to take place from 11-1 3 Dec.1990 at Olympia 2, London, GB, and Communications offers - among many other sessions - one on Document Indexing and Markup with four papers altogether of which only one is actually dealing with indexing, viz.: Klaus KOTZIAS, Siemens AG: How to respond to Linguaggi Documentari e Basi Dati - different language particularities by indexing texts using automatic text analysis. - Among the socalled Satellite Documentary Languages and Databases Events a seminar on Perspectives in Indexing and Search­ a The Consiglio Nazionale delle Ricerche with its ing is held by Ev BRENNER, USA, on Frid y, Dec.14, Istituto di Studi sulla Ricerca e Documentazione Scien­ 1990, for Brit.pounds 201.25. According to the an­ tifiea invites for a meeting in Rome, Dec.3-4, 1990 and nouncement he will extend on 1) A review of pertinent offers the following program consisting of three sections: developments in the past 35 years necessary to better (1) Languages: which are they and how arc they con­ understand today's technology, 2) The Relationship structed. (2) Classification use in a database. (3) Free between indexing and searching - the interface problems, languages and automatic indexing. The speakers and and 3) A review of practical developments reasonably to their papers will be: (ad 1) Georges van SLYPE: The be expected in the coming years that will affect indexing e construction of a thesaurus. - Paola COSTANZA CAPI­ and search effici ncy and budgets. * It is said of Mr. TAN!: The ISO International Standardization: an Brenner, that in 1990 he will have attended the ASIS instrument and an objective. - Fiora IMBERCIADORI: mid*year meeting in Fort Lauderdale, FL, the 2nd � The new thesaurus for education in Europe. - Renata International Chemical Information Co ference in BORRUSO, Linguistic aids in the ITALGIURE-find Montreux, Switzerland, in September, the Annual system of juridical documentation research. - Roberto EUSIDIC meeting in Helsinki, Finland, in October, and POTENZA: Thesaurus of earth sciences. - Costantino the Annual ASIS meeting in Toronto in November. As a CIAMPI: Experiences of a semi-automatic thesaurus special feature of the seminar he will review these construction. - Bruno FELLUGA, Sandra LUCKE, Mario meetings, highlight developments, and analyze important PALMERA: From a descriptor system to a thesaurus in trends. In addition, he will present a special report on the environment field. - Antonio CAMELLI: THES Executive Information Systems (EIS), an area in which GIUR: A thesaurus to support ajuridical database and as he has recently been making some special studies. For a verification tool of the law lexicon. - (ad 2) Ingetraut further information please turn to Learned Information DAHLBERG: Philosophical bases of conceptual ordering Ltd., Woodside, Hinksey Hill, Oxford OX1 5AU, England. systems. - Jacques MANIEZ: Actuality of documentary classification systems. - Giliola NEGRIN!: Objectives Meeting on Data and Concepts, and Courses in and multi-disciplinarity of a "Finalized Project" - Paola Darmstadt 1991 VENEROSI: The use of a representation language in a user-system interface. - Emilia CAMPOCHIARO, Fran­ At the Technical University of Darmstadt, Depart­ cesco MARCELLI: Thesaurus and classification: an ment of Mathematics, Professor Rudolf WILLE will integrated system of indexing and retrieval of online again organize a meeting of his Special Interest Group on parliamentary infor�ation. - Alberto MARINI: Support Concept Analysis of the German Society for Classifica­ procedures for the use of classification schemes. - (ad 3) tion, as already in the pre�ious years since 1986. Interested Fulvia NALDI: Automatic indexing and retrieval of colleagues are invited to hand in proposals for papers on archives. - Paola CARRARA : Organization and classifica­ the relationships between actually existing data and tion of dates. - Haria VANNINI PARENTI: Information conceptual thinking. The meeting, starting o� Febr.21, retrieval in the EXI-RPTS project. - Antonio ZAMPOLLI: 1991 at 4 PM and ending on Febr.2 3 at 1 PM, is preced* The automatiC treatment of natural language. - Marh::ne ed by 3 courses on Formal Concept Analysis which start CHOUKROUN, The contribution of VIDEOTEX on Febr.20 at 9A5 AM until Febr.21, 3 PM, given by ergonomics in the multicriteria retrieval of research data the Research Group on Concept Analysis. The three banks: TELELAB and TELETHESES, two research data courses run in parallel and cover (1) Introduction, (2) banks on MINITEL. - Maristella AGOSTI: Hypertexts Conceptual Scaling, (3) Conceptual Knowledge Systems. potentiality for information retrieval. -Ciulio GROPPI: The courses and the meeting are in German. Participa­ Retrieval techniques of a single subject in different tion is free of charge except for DM 20.- needed for databases. mater,ials given to the participants. - Registrations for Chairmen of the sessions will be: Vittorio NOVELLI, all events should be sent to: Prof.Dr.Rudolf Wille, FB Antonio A. MARTINO, Alfredo SERRAI, and Giulio Mathematik, Technische Hochschule, schloBgartenstr.7, - OCCHINI. D-6100 Darmstadt. Except for four papers in English, the rest isin Italian. A proceedings volume is foreseen. , Cataloging Heresy: Challenging the Standard ' For further information please turn to the Secretary: Bibliographic Product Signora Tamara FARNES I) Istituto di Studi sulla Ricerca e Documentazione Scientifica, Via C. de Lollis 12, This will be the theme of a Congress for Librarians 1-00185 Roma. Tel., 06/4452351, Telefax, 06/4463836. scheduled to be held at St.John's University in Jamaica,

Int. Classif. 17 (1990) No.3/4 Reports and Communications 165 New York, on Feb.18, 1991. With support from a grant participating. It is directed by Professor Andre Lich­ by the H.W.Wilson Foundation, nationally known nerowicz of the Academy of Sciences of Paris. Chairman librarians, thesaurus designers, and representatives of of the Program Committee is ProLJ .Arsac, University bibliographic utilities have been invited to address the of Paris VI. theme. The papers scheduled for presentation are: The purpose of the conference is to present the state Bella H .WEIN BERG: A theory of relativity for catalogers. of the art in storage, retrieval and diffusion of non­ Norman ANDERSON: The non�neutrality of descriptive structured information found in text, image and sound. cataloging. - Kathy WHITEHEAD: The Art&Architecture The previous conference, RIAO 88 had been held at Thesaurus as an alternative to LCSI:1. - Sally SINN: The MIT, Cambridge, USA in March 1988 and was a resound­ development of classification and subject heading systems ing success. The call for papers is addressed to re­ for medicine. - Sanford BERMAN: Cataloging tools and searchers from all countries in academic or industrial Copy: The myth of acceptability - A public librarian's research. Specific themes for the conference include: point of view. - Mary PARR: Standard cataloging- data A - Lingustic analysis for automatic text treatment and the academic library: The technical services manager's (Automatic indexing, Automatic abstracts, Natural point of view. - Liz BISHOFF: DCLC and the master language interrogation, Multilingual interfaces). record concept. - Ed GLAZIER: The display and index­ B - Construction and utilisation of large linguistic ing of customized catalog records in RLIN. - Lucia knowledge bases (electronic dictionaries, thesaurus, RATHER: Standard cataloging data: The view from the bilingual dictionaries). Library of Congress. - Sheila INTNER: Rejecting standard C - Confidentiality in information retrieval systems. cataloging copy: Implications for the education of D - Multilingual interrogation and computer assisted catalogers. translation. Because the acceptability of standard cataloging data E - Automatic extraction of factual information from concerns all types of libraries and all subject specialities, full text. this call is for additional papers for inclusion in the F - User interfaces and ergonomy of information re­ volume to be published by Learned Information. Papers search systems. of varying length are acceptable. Manuscripts should be G - Artificial intelligence for user aid, and for personali­ double spaced, with ample margins, and should use the zing systems. author-date system of citation. The deadline for sub­ H - Intelligent navigational aids and automatic data mission is Dec. 1S, 1990. All contributed papers will be structuring in hypertexts and hypermedia. refereed by a committee of experts in the theory and I - Neural nets for computer aided information research. practice of cataloging. Contributors will be notified of J - Data entry systems (OCR, automatic structure the referees' decision by Jan.1 S, 1991. recognition, document preparation norms). Authors of accepted papers will receive complimen­ K - New applications (Software engineering and informa­ tary registration for the full-day Congress (which in­ tion research systems program and documentation cludes lunch) and a copy of the book upon publication. retrieval for re-use, production of intelligent tutorial A list of the authors and titles of accepted papers will and documentation systems in software. development. be distributed to attendees at the Congress. - For further Automatic image indexation via pattern recognition. information contact: Dr. Bella Hass Weinberg, Division Optical memories, such as Videodisk, CD-ROM, CD-I, of Library and Information Science, St.John's University, Optical numerical disks. Multimedia systems managing Jamaica, NY, 11439, USA. Tel., (718)990-6200; Fax, text, sound and images. Content addressable electronic (718)380-035 3. mail systems. Voice entry and speech recognition). The papers - to be submitted in 4 copies, 20 pages Making Connections. ACH/ ALLC 91 maximum, and to arrive before Oct.30, 1990 - must be validated by a prototyp-e or a working model to be The 1991 International Joint Conference of the demonstrated during the conference at a time separate Association for Computers and the Humanities (ACH) from their oral presentation. Such demonstrations and the Association for Literary and Linguistic Com­ should run on standard materiaL - Addresses: CID. puting (ALLC) will be held on March 17 -2 1, 1991, at 36 bis rue Ballu, F-75009 Paris (for Europe) and Ms.M.-T. Arizona State University in -Tempe, AZ, USA. Sub­ Maurice, 220 East 72nd Street 10F, New York, NY mission of abstracts on computer-aided topics in literature 10021, USA (for North America). and linguistics, and in humanities disciplines are invited Together with. this call for papers, a Call- for Product until Oct.1S, 1990 to Professor Daniel Brink, Depart­ Demonstrations has been sent out with an invitation ment of English, Arizona State University, Tempe, AZ, to companies and organizations which market innovative 85287-0302, USA. and competitive hardware or software related to the conference themes to exhibit these at RIAO 91. The specific themes are listed on a leaflet which can be RIAO 91. Intelligent Text and Image Handling obtained from: CID, 36 bis rue Ballu, F-75009 Paris. A call for papers has been released for this meeting which is to take place on ApriI Z-S , 1991 at the Universi­ tat Autonoma de Barcelona, Spain under the sponsor� ISCCR 91. Call for Papers ship of the European Economic Community and several other institutions. The conference is organized by the This call of the FID Committee on Classification Centre de Hautes Etudes Internationales d'Informatique Research for the 5 th International Study Conference on Documentatire (CoI.D.) with several other institutions Classification Research in Toronto, June 24-28, 1991,

Int. Classif. 17 (1990) No.3/4 166 Reports and Communications on Classification Research for Knowledge Represen�a­ CDMARC Names complements the complete Library tion and Organization has been published already in of Congress Subject Authority File which is also available Int.Classif. 90-2 on the inside back cover page. It is on CD-ROM (for USD 300/370) and includes the same mentioned here only for reasons of reminding that components as those included in a CDMARC Names abstracts between 500-1000 words are due by Oct.15, subscription. CDMARC Names contains 2.3 million 1990 to"be sent to Prof.Nancy J. Williamson, FIDICR personal, corporate, series, and title authority records Secretariat, Faculty of Library and Information Science, on 3 discs. Three display options are always available: University of Toronto, 140 St. George Street, Toronto, thesaurus image (heading and any variants), brief card Canada M5S lAl. image, or the USMARC tagged-record format. The hardware needed? Any CD-ROM drive on the market that supports the High Sierra standard and International Indexers Conference , Yugoslavia uses drivers compatible with Microsoft DOS extensions A preliminary announcement has been released for will do. The CDMARC system works on IBM PC/XT/AT this conference to be held in Hvar, Yugoslavia from or compatibles having at least 6401( RAM (with at least 22-25 Sept.1991. For more details contact: Dr.John 530K RAM free for use with CDMARC Names), 5 mega� Gibson, Landors, Wichenford, Worcester WR6 6XY, bytes of free hard disk space, and running DOS version England. Tel., 0886-888209. 3.1 or higher. For further information write to Library of Congress, CDS, Washington, DC 20541 USA TEL. (202)707 6100; 2nd Albany Colloquium 1991 FAX (202) 707 1334.

A Call for Papers has been sent out from the School of Information Science and Policy, State University of New York and the Albany Classification Research Organization for a conference in Albany, NY, Dewey Decimal Classification Nov.14-16, 1991 to continue the series started with Even though the 20th edition of Dewey Decimal "Classification in the Computer Ag e", held there from Classification was just published in January 1989, work Nov.18-19, 1988. has already begun on planning for the 21st edition. The The four conference themes are: Language and DDC Editorial Policy Committee will meet in retreat in Classification, Learning about Classification, and Legal Ocotober 1990 to discuss the 21st edition. Discussion and Economic Impacts of Classification. Abstracts items include proposed schedule changes, editorial rules, of papers - not to exceed 250 words -should be submitted frequency and number of editions,international concerns, as soon as possible and must be received before July I, and the nature and impact of electronic products. 1991. All papers delivered at the Conference will be Suggestions and comments from users may be forwarded published in a proceedings volume. Abstract headings to Peter J. PAULSON, Executive Director, Forest Press should include titles, authors, and institutions or agencies. OCLC, 85 Watervliet Ave. Albany, NY 12206-2082, Authors should classify their submissions according to USA. the four conference themes mentioned above. An ad hoc subcommittee of the American Library sponsors will include - next to the two Conference Association Subject Analysis Committee is soliciting organizations named above - the Council on Library comments from users on priorities for the 21st edition. Resources, State University of New York Research The committee is chaired by Jan DeSIREY, Hennepin Foundation, and various governmental groups, public County (Minnesota) Library. and private organizations concerned with the advance­ (From ASISISIGICR News, August 1990) ment of classification theory and applications. Information on registration will be sent to all members of ACRO ISKO, and CSNA. ALlSE members who have selected "Classification" as an area of scholarlyl teaching interest will also be notified. Abstracts and requests for New Abridged Dewey Published in June 1990 information on the conference should be sent to Dean DCLC Forest Press Division has published the 12th Richard S.Halsey, School of Information Science and abridged edition of the DDC. The one-volume abridge­ Policy, Nelson A.Rockefeller College of Public Affairs ment includes all of the user-convenient features of the and Policy, University at Albany, SUNY, 135 Western four-volume DDC 20 published in 1989. With about Ave., Albany, NY 12222 USA. Tel., (518) 442-5232. one-sixth of the schedule entries it is designed for BITNET RH611 ALBNYVMS. libraries of fewer than 20,000 titles, particularly for school libraries and small public libraries. In its 857 pages it contains also a manual and is available in America CDMARC Names and the AsialPacific region for USD 65 until Dec.31, The Library of Congress Cataloging Distribution 1990 and USD 75 thereafter from OCLC Forest Press, Service has recently announced the availability of its 6565 Frantz Road, MC 685 Dublin, OH 43017-0702. authority file on names on CD-ROM. A one-year sub­ In Europe and other areas of the world it is available scription can be had for USD400 (U.S.) or USD475 for Brit.pds 43 until Dec.31, 1990 and for Brit.pds 50 (international). A subscription includes fully cumulated thereafter. Orders from these areas should bc sent to quarterly updates, a reference manual in a 3-ring binder, DCLC International, 7th Floor, Tricon House, 51�5 3 keyboard template, and retrieval software. Hagley Road, Birmingham, B16 8TP, U.K.

Int. Class;f. 17 (1990) No.3/4 Reports and Communications 167 Indian National Terminology Bank Et hnicity Glossaries

The Commission for Scientific and Technical Termi­ A new project is going to be launched soon by IITF ' nology (CSTT) has launched a special scheme of compil­ focusing on terminology work in the social sciences. ing a national register of authors and translators who arc Fred Riggs from the University of Hawaii has already committed to the cause of enrichment of scientific prepared a first version of this terminology collection in and technical literature in various disciplines in India. English. The structure of this glossary is rather unusual The objective for compiling a national register is to in traditional lexicography, but is perfectly suited to create a data bank of manpower existing in the field for descriptive terminology work, since the entries start undertaking original writing or translation of various with a definition, various contexts from scientific scientific and technical subjects, an official release said . literature and then provide various terms usually denot� The CSTT was set up in 1961 in pursuance of a ing the concept in question. This concept�oriented lay� Presidential Order, The functions assigned to it were: out is very promising in comparison to traditional word� a) determination of the principles of terminology oriented dictionaries, particularly in the social sciences evolution with their pluralistically and multi�paradigmatically b) co-ordination of the existing terminologies available structured subject fields. in Hindi and other Indian languages A French version of this Glossary on Ethnicity has c) preparation and publication of approved terminological just been completed by Eric de Grolier, a Russian glossaries version has also been prepared . The first stage of our d) preparation and publication of university level books project will now be to make a comparative evaluation and other reference literature. of the classification schemes designed for these three In a nutshell the task assigned to it was to take all versions and to propose a unified classification, if this measures to promote Hindi and other Indian languages is viable. Then we are planning to process all the data to replace English as a medium of instructions at all by suitable microcomputer terminology management levels of higher education. software and design a user�friendly layout for publica.; It is functioning under the Ministry of Human Re� tion. Christer Lauren and Heribert Picht will investigate source Development, Goyernment of India, New De1hi� whether a joint Nordic version of the Glossary can be 110001. It has so far evolved and published more than organized. In addition we will try to find ways for the five lakh (500,000) terms belonging to all major university preparation of a German version. disciplines in the natural sciences, social sciences, Since ethnic and/or linguistic conflicts are unfortu� humanities, engineering and medicine. nately increasing in number and intensity all over the It is now engaged in evolving terminology for the world, terminologies pertaining to ethnicity research newly developed disciplines of minor engineering, train� (political science, sociology, sociolinguistics, anthropol� ing technology and management sciences. ogy etc.) in many languages are becoming more and Besides publishing work of terminology of space more important. The transparency of concepts used by science, computer science, steel and non�ferrous metal� researchers is vital to the promotion of projects investigat� lurgy, the CSTT brought out a consolidated glossary of ing into the state�of�the�art and into ways of improving administrative terms comprising 12,000 entries belong� the situation of minorities. ing to general administration and designations to promote You are invited to participate actively in this practice� the use of Hindi in official transactions pertaining to oriented project, which will certainly yield positive defence, posts and telegraphs, and railways. results not only in the subject field concerned, but also Over 15 ,000 Pan�Indian terms belonging to various in terminology research, particularly in descriptive disciplines have been identified and published in the terminology work and multilingual and interdisciplinary form of glossaries for free distribution among scholars, tcrminography. For further information please contact writers, journalists and translators. Dr.Gerhard Budin, IITF, Sensengasse 8, A�1090 Wien. The CSTT has undertaken an innovative scheme in (From, IITF News (1990) 112, p.3-4) deVeloping a computer based national terminology bank for storing, coordinating and analyzing the entire scien� Terminology Workshop, Kent, OH, 1991 tific and technical terminology in Hindi and other Indian Description of the curriculum and registration materials languages. for a Terminology Workshop to be held at Kent State The organization has also been coordinating with University, Ohio, from June 7�10, 1991 can be obtained Hindi book academies and state textbook boards, from Dr. Sue Ellen Wright, Institute of Applied Linguis­ engaged in the production of quality books on various tics, Kent State University, Kent, OH, 44242 , USA. The university subjects. So far 2200 books in Hindi and workshop will focus on fundamental concepts of termi� 6200 in other Indian ' languages had been published on nology theory and will provide intensive daily practice various university subjects either directly or through at the computer for each participant. The fee is usn 350 state language academies and text book boards. and includes instructional materials, use of computers, M.P. Satija and university housing (not meals),

Int. Classif. 17 (1990) No.3/4 168 Reports and Communications Rafael CAPURRO, Stuttgart; Mary DYKSTRA, Halifax, NS; Robert FUGMANN, Idstein; Charles HILDRETH, Springbrook Drive, IL; Erich MATER, Berlin; and ISKO News 4 Brian VICKERY, London. The animated discussions centered on the nine postulated of Don Swanson in his article "Information retrieval and the future of an illusion" (published in J.Amer.Soc.Inform.Sci. 39(1988) No.2, p.92-98). A brief summary of the nine postulates Tools for Knowledge Organization and the had been distribpted to the participants beforehand. The discussions on the possibilities of computers vis-a-vis Human Interface the capabilities of man heralded the very theme of the Report of the First ISKO Conference conference. It is recognized that there has been a spec­ tacular progress in computer-aided information retrieval, 1. Introduction yet the machine will always have its built-in limitations. The human being must always remain in command of The first biennial conference of ISKO was held at the machine. In serving the reader and analyzing their Darmstadt, Technical University, Aug,14�17, 1990. queries it is impossible for a machine to have the re­ Consistent with its aims the topic was aptly chosen to be quired human judgement. The postulates truly warned "Tools for Knowledge Organization and the HUman against the total algorithmic text analysis and search. A Interface". The conference was co�sponsored by the machine cannot recognize meaning, judgement is always German ISKO Chapter, the Committee on Classification elusive and thinking has no rules. Computers cannot and Thesaurus Research of the German Documentation think. Society (DGD/KTF) and by FIDICR, the Committee on The second plenary session held on the morning of Classification Research of the Federation Internationale Aug.15 at 9 AM with Winfried GODERT, Hamburg, in d'Information et de Documentation. the chair �aw Charles HILDRETH speaking on "End h re as n impressive gathering of about 200 experts � � � � users and structured searching of online bibliographic partlclpatmg In the conference and its preconferences. databases; recent research findings" . He pointed out that They represented 23 countries of North America (USA the system designers have not investigated well enough and Cana a), Asia (including participants from China, � the several major causes of subject search fai lures. Japan, IndIa, and Israel), and Europe, with participation Describing the methods of structured searching of online also from the East-European countries, such as the catalogues he examined the limited role of classification USSR, Hungary, Czechoslovakia and the GDR. in improving subject searches. He went on to describe The conference was divided into 12 plenary sessions two online catalogue research projects involving classifi­ and 12 workshops, with three workshops running in cation-based retrieval and browsing. The first project parallel conducted between the plenary sessions each evaluated the retrieval efficiency of the system combin­ day. In all about 60 papers were presented and discussed ing classification with probabilistic retrieval methods. according to the program as published in advance in The second project examined the effect of two classifi­ International Classification 90-1 . Most of the papers cation�based search aids on retrieval performance and were already there in a finely edited and published user satisfaction. pre-conference volume and distributed free of charge to Immediately followed the third plenary session the participants (1). with Winfried G(jDERT still in the chair. Masanobu T�e remaining papers, the synthesis by Eric de FUJIKAWA from Tokyo spoke on "Concept theory Groller and the summary of the panel discussions will be and facet analysis of knowledge units with emphasis included in the second volume soon to be published by on artificial intelligence research". Explaining the the Indeks Verlag. process of thinking he semantically defined the term In the opening session on Aug.14, at 6 PM, Dr.Inge­ "concept" and related it to image, pattern structure, traut Dahlberg, director of the conference and president structural relations and representations. He maintained of ISKO, welcomed the participants and referred to the that only concepts lead us to knowledge and intelligence. symbolism in the Darmstadt wedding tower on the The value of research in knowledge engineering and Mathildenh6he (depicted on the conference program) artificial intelligence lies in explaining the nature of and he mosaique in the museum as possible guides � knowledge. From his practical experience he stated to faIthfulness and love to be kept in mind during the that at the Japan Information Center for Science and days of the conference. There was also a note of welcome Technology he has designed a programme to select by the representative of the Technical University, abstracts which contain sentence structures in which Prof. Dr. Gerhard Knorz, speaking on the present state of certain terms have particular relations through the word this University. matching process. In the following we will only report on the 12 plenary . The fourth plenary session was chaired by Tim seSSIons The acc unts of the 12 workshops as given by : � C.CRAVEN, London, ant. Alan GILCHRIST Brighton the chaIrpersons III the 11th plenary session will follow UK, spoke on " Knowledge Organization and he Huma in a separate report. ; � Interface". He described his lecture as his personal and in comprehensive views on the situation in the U.K. He 2. The Plenary Sessions pointed out that despite the massive research only a few The first plenary session immediately following the tangible products have resulted. At present it is not . . m�uguratiOn was a panel discussion chaired by Dr.Win­ feasible to build an expert system that is intended to fned SCHMITZ-ESSER, Hamburg. The panelists included carry out intelligent retrieval. In knowledge organization

Int. Classif. 17 (1990) No.3/4 ISKONews4 169 there is no substitute for hard intellectual work. "The problems of intellectual and computerized contents The fifth plenary session was chaired by Winfried analysis" and stated that human intellectual capacity SCHMITZ-ESSER. Otto SECHSER, Zurich, spoke on and language ability are only partly known. Contents "Classification issues in databases from machine-readable analysis of documents by computer can only be done text data". He described the system and components of with selected strings of terms according to pre-formu­ TELEDATA which is a factographic database in Zurich. lated characteristics. Such algorithmic processes are to The database represents an alternative category of a be distinguished from the mental capability for abstrac­ retrieval system comprising both business oriented tion as well as for evaluation. DBMS and text data. Its classification problems are The 10th Plenary Session (which in a sense was the important not only in connection with indexing and last session with an individual technical paper) had Otto end-user retrieval but also in the process of generating SECHSER, in the chair with Hanne ALBRECHTSEN, and updating the database. He asserted that we should Birkeroed, Denmark, presenting her "Software concepts" develop a hybrid system of human and machine indexing focussing on knowledge organization and retrieval of that automatically processes the easy part of the job and re-usable software components and concepts>. Experience may assist the human experts to do the rest. with a Prototype Re-use Support System (PRESS) The sixth plenary session was held with Alan developed at Computer Resources International (CRI) in GILCHRIST in the chair. Brian VICKERY spoke on the framework of the ESPRIT project PRACTITIONER "Classificatory principles in intelligent interfaces". He was described. It was conc�uded that investigation in· expressed that intelligent interfaces to databases need to data processing terminology should be intensified for incorporate several kinds of knowledge, such as knowl­ approaching a domain analysis on an empirical bases. edge of information retrieval techniques, of subject The future work may include development of strategies structure and terminology of search statements. Organ­ for automatic indexing and semi-automatic thesaurus ization of knowledge involves the use of various methods compilation as well as considerations for a faceted of representation, including structures common in thesaurus approach. bibliographic classification: hierarchy, facets, semantic In the eleventh plenary session chaired by Ingetraut categories. He illustrated with examples the coming DAHLBERG, Frankfurt, the chairpersons of the twelve together of the techniques of artificial intelligence and workshops presented the summaries of the contents and information science. Knowledge structures are being issues raised of their respective sessions (given below). combined with those of AI. So the variety of knowledge The concluding 12th session was chaired by Douglas needed in intelligent interfaces is being effectively FOSKETT, Gerards Cross, UK much to his delight and captured and represented. happy memory of a similar occasion at the 1957 classifi­ The seventh plenary session was chaired by Norbert cation conference in Dorking when it also happened that MEDER, K61n. Roland HJERPPE, Link6ping, Sweden, Eric de GROLIER, Paris, presented - as in Darmstadt _ spoke on "A framework for characterizing systems for his Synthesis of the entire conference. This time it was knowledge organization: A first basis for comparison listed under ten points. Among the issues he emphasized and evaluation" . He stated that comparison and evalua­ the need for investigations into the basis of a universal tion of different systems of knowledge organization is scheme and urged liberal financial support for such a loaded with problems as most of the tests have been project. He also drew attention to the need for user focussed on IR and the results achieved in searching. The studies in the use of classification and more investigation manner of comparing classification systems takes into into the process of navigation. consideration only a small, albeit important part of the In the panel discussion to follow, entitled "Recom­ sequence of processes and products of which a classifica­ mendations for the 90ies" panelists were Ken BAKE­ tion system is made. However, some questions regarding WELL, Liverpool; Karen Markey DARBENSTOTT; the level of hierarchies and distribution of items in each Douglas FOSKETT (chair); Winfried GODERT; M.A. class, remain still unsolved. GOPINATH, Bangalore; Eric de GROLIER; and Peter The eighth plenary session was chaired by Loll J AENECKE, Pforzheim. Their recommendations ranged ROLLING, Luxembourg. Karen Markey DRABEN­ from understanding the nature of thinking to the pattern STOTT, Ann Arbor, spoke on "Experiences with online of knowledge production and its assimilation in society, catalogues in the USA using a classification system as a semantic analyses, users interfaces, and questions of subject searching tool" . She traced the development of terminology. Other questions concerned standardization, online catalogues in the USA that have facilities for structure of universal systems, use of natural languages, subject searching and browsing through machine-read­ deep structure, impact of hardware, technology in able text of library classification. She dwelled upon her our work, user studies and their needs, user-oriented experiences from 198 3 to 1986 developing and testing organization of knowledge in libraries, the importance of subject searching and browsing capabilities in an experi­ book indexes, and interdisciplinary cooperation. The mental online catalogue using the machine readable panelists agreed that knowledge organization research DDC-19 schedules and relative indeX'. The results of the needs inter-disciplinary cooperation. It is interesting to experiments demonstrated that the DDC provided new note that ISKO is already striving for an integration of and fruitful subj ect searching capabilities that were not the conceptual approaches of classification and indexing possible through the alphabetical and keyword search� research with linguistics and AI research. A major result ing. In the face of the new hardware and MARC format of the discussions: what we expect is a system where developments she suggested some improvements in the man should not consider himself the servant but the original experiments in her new and continued research. master in controlling computer work. The human mind The ninth plenary session was held with Maria DO­ is indispensable in information retrieval and knowledge MOKOS, Budapest in the chair. Erich MATER spoke on organization. It is therefore unrealistic to strive towards

Int. Classif. 17 (1990) No. 3/4 170 ISKO News 4 fully automated retrieval systems but instead towards Workshop Reports at ISKO Conference 1990 Computer-Assisted Information Retrieval Systems The twelve workshops of the First International ISKO (CAIRS). This finding justifies and strengthens at the �onference, Aug. 14-17, 1990 in Darmstadt, were embedded same time the theme of the conference. into the ge eral program of altogether twelve plenary sessions . . � With ten invited papers, They covered on the first day fundamen­ 3. Other Events and Conclusion ta� probl�ms and systems for knowledge organization and dealt With retneval aspects and applied classification on the second The deliberations were unreserved and extensive. The day. The workshops, with four papers each, were held in parallel. . conference not only drew scholars of outstanding merits t�ree at a time. In order to offer to the participants an overall and of notable names, but also the new breed of practi­ picture of the contents of all workshops the chairpersons were . sked for a report of their respective sessions. What follows tIOners of hard- and software, as well as library admin� � IS the contents of the 11th Plenary Conference Session with its istrators, teachers and researchers, information consultants workshop reports. and free-lancers in the information profession. We do not need to repeat the eminent names of colleagues Report on Workshop A: "General Issues" mentioned already above who as early as in the thirties, Chair: Ken BAKEWELL, Liverpool. Rapporteur: Feli� fifties and sixties of this century had written the major zitas B£LKE, Bonn. publications in the field of knowledge organization and In the first paper by J .A. BOON, Pretoria, S.A. ("The are now still working at the forefront of research in this integration of technology in the organization and field and reported at this conference about it. For many dissemination of information") man as well as machine a participant it was a rewarding and satisfying experience are both taken as information processors. Boon argues to see and talk to these our pillars' personally, One is for their sophisticated interaction. Nevertheless, he sees reminded of the words in the Preamble of ISKO: The information systems a social systems using information Society "provides personal contacts and opportunities technology. He assumes that some basic information of cooper tion to the world-wide community of col­ � problems are not yet identified, so the knowledge leagues .... . worker - as an information processor - has remained, On August 15, a general rneeting ofthe ISKO member­ until now, a somehow unknown being. How does he, ship took place. The Scientific Advisory Council met on in fact. collect, organize, process and disseminate infor­ August 17; suggestions had been invited from the mation? What SOrts of interrelationships are there in man's members present for the future programme of ISKO. It mind? I-Iuman cognition and learning has to be inves­ was decided to hold the 2nd International Conference in tigated and considered when designing information India in 1992 to coincide with the Ranganathan Birth systems that will do more than organize and manipulate Centenary celebrations in that year. bibliographic records. Until now, system designers have he conference was preceded by two pre-conference � met with individual functions of information processing semmars on August 14: one on "Intelligent Interfaces only. To get integrated knowledgelinformation systems, for Subject Access in Libraries" conducted by Charles there is needed the information function and related HILDRETH and Karen Markey DRABENSTOTT and software taking care of the various correlated human chaired by Heiner SCHNELLING, Giessen; and one on activities in this field, viz. administrative functions (for ' Thes urus oftware" with a number of papers pub­ : � . � example, word processing, diary keeping, scheduling); ltshed 10 thlS Issue of International Classification (90-3/4). transfer functions (accessing information from dis� A third event on that 14th of August was the precon­ tributed files, electronic communication); real informa� ference of the SCCAC (Society for Conceptual and tion functions (create information, collect, store, analyze, Contents Analysis by Computer) with some outstanding organize it). papers; Eric de Grolier in his Synthesis mentioned Robert FUGMANN, Idstein, (in "Unused possibilities especially the one by Rudolf WILLE, Darmstadt on in indexing and classification"), argued for the superiority "Conceptual knowledge systems" as of parti ular � of intellectual techniques in information processing and interest. denied the equivalence of the algorithmic approach. The participants had an equal treat of excursions and The intellectual techniques still include unexploited get-together events to lighten the burden of the day's capabilities; for instance by vocabulary categorization. work. During the meal in the "Knights Hall" of the Beyond this, it should be considered that algorithmic Auerbach castle on Aug.I6, Ken BAKEWELL "served techniques still remain in an experimental stage to a well" with a hilarious talk on "Humor in Indexing". good amount. Also, an unlimited variety of paraphrases On Aug.I7 in the afternoon, many took the chances for a single concept is a barrier to adequate algorithmic to make use of organized visits to a number of interest­ processing. i�g places, such as the Deutsche Bibliothek, the GMD's The contribution by Pat MOLHOLT, Troy. NY, lIbrary and documentation center on information ("Standardizing and codifying related term links for science, the Hoppenstedt Publishing House, and also the improved information retrieval") as a replacement for famous Mathildenh6hc of Darmstadt with its artistic the paper by Vesa Suominen from Finland, previously gardens and museums in Nouveau Art, a great experience included in the program, seemed rather specific within for all indeed. "General issues", however, semantic relationships M.P ,Satija belong to the core of information tools like thesauri and classification systems. Mrs.Molholt reported that for the (1) Fugmann, R.(Ed.): Tools for Knowledge Organization and thesaurus work she had done to design the Art and the Human Interface. Proceedings .... Frankfurt: INDEKS Verlag Architecture Thesaurus (AAT) of the Getty Foundation 1990. 280p. ISBN 3-88762-020-9. Advances in Knowledge more than 100 relationships have been identified which Organization, Vol.1 (VoL2 in preparation) can be reduced to 20-30 link types. Besides the taxonomic lnt. Classif. 17 (1990) NO.3/4 ISKO News4 171 relationship, known since Aristotle, the nonhierarchical mentary set"). They argued that, in a text, noun phrases relations became interesting since the 60s and they were refer to objects in the world, and thus.to factual knowl­ examined for applicability in knowledge structuring and edge. In order to accomplish knowledge retrieval from information retrieval. The main problem of related terms document sets, we have to switch from the intensional is, that we do not know what kind of relations are meaning of noun phrases in the text to their extensional involved here. A further problem arises when statistical reference to classes of objects in the world. analysis-puts out term associations, because co-occurence In the concluding paper, Marek CIGANIK tried to of terms does not provide a semantic characteristic. Mrs. show, how "Key faceted structures in a text" can Molholt will meet these difficulties by her research on function as a background for the understanding of the related term structure for the AA T. The relation­ a text. ships will be defined and applied to a subset of the AAT and will be tested in an information retrieval Workshop C: "Terminology" system. That system will be compared with another Chair and Rapporteur: Yukio NAKAMURA, Tokyo system using Boolean logic, but without relations. She Workshop C was on terminological problems in the assumes that the result will prove the superiority of the original sense of terminology. It included four presenta­ use of relationships over singletermsortermcombinations. tions. The first one by Miss Widad MUSTAFA ELDADI Gerd BAUER, Schleswig, (in "Promoting creative from Villeurbanne, France described the contribution of processes by a thesaurus-like representation of knowl­ terminology to the theoretica] conception of classifica­ edge structures") demonstrated his ambitious goal in tion languages and thesauri. She asks for better collab­ supporting scientific problem solving. He has chosen oration between linguists and information specialists. multidisciplinary examples when transforming textually Also the need for a terminological thesaurus was men­ formulated questions in a structured form. His TGW­ tioned. Close cooperation ' among terminologists and Method, the Structural Diagram-like Knowledge Rep­ documentalists, for instance, in the field of indexing resentation (thesaurusartige graphische Wissensdarstel­ languages for the analysis of certain linguistic problems lung) visualizes the network of concepts. By this, creative are explained. Some examples in telecommunication in a thinking is stimulated to change aspects and to find new vocabulary in English and French were studied with the combinations of concepts. Analogies are discovered, corresponding Arabic terms in addition. Ms.Mustafa and unknown connections can be derived from known Elhadi presented some new ideas and gave special relations. That, of course, is a creative art! attention to multi-referential terms - which I think is a new and important idea. Workshop B: "Algorithmic Text Analysis" The second paper presented came from INFOTERM, Chair and Rapporteur: B.ENDRES-NIGGEMEYER, Vienna. Christian GALINSKI and Gerhard BUDIN SaarbrUcken presented a new stream in the Vienna School of Termi­ This workshop had certainly the advantage of con­ nology. The paper entitled "Specialized encyclopedias fronting us with topics currently on the top of the heads and lexicons processed on the basis of the methods of in natural language processing: connectionists models, terminological engineering" uses knowledge engineering knowledge representation and retrieval, and semantic in the terminological approach. It is to be expected relations. We heard four papers, the speakers being that much interesting work will develop in this respect in Renate DEFFNER from UnterschleiBheim, Gerda the forthcoming years. RUGE from MUnchen. Richard BOUCHE from Villeur­ The fourth paper was presented by Miss Elsemiek ten banne, France, and Marek CIGANIK from Bratislava, PAS from the Free University of Amsterdam, entitled Czechoslovakia. "On the Role of definitions in terminology". This is an It was Renate DEFFNER who acquainted us in her extension of the principles of terminology and displays paper with Hans GEIGER with a connectionist system different types of definitions, i.e. representations of architecture. ("Associative word recognition with definitions of certain terms. She distinguishes three connectionist architectures"). She explained that a types of definitions, the generic, partitional and the connectionist model follows the information processing functional one. Often two or more types of definitions principles of neural networks, what means in particular are used mixedly in the single definition of a term which that retrieval happens by spreading activation. She causes difficulties. At the end the author concludes that had made good experiences with this model in a word types of information necessarily contained in the defini­ recognition task. Some 5000 streetnameswererecognized, tion of a given concept should be studied further. I think and the system coped well with spelling errors. this is a also a new approach for the terminological From Gerda RUGE (together with Christoph methodology . SCHWARZ) we learned (in "Linguistically based term The third paper was given by Wu GUANG-WEI from associations. A new semantic component for a hyper­ Shanghai, China. He made a detailed study of term term system") about new developments in the REALIST equivalence between English, Chinese and Japanese. It system that helps to find good search terms for free text showed the necessity of adding new criteria for establish­ retrieval. The main innovation in REALIST is a semantic ing multi-lingual thesauri. He gave very detailed explana­ relation among terms which describes their semantic tions on these problems. However, the original work is relatedness. It is based not on statistical cooccurrence, so extensive that we cannot go into details here. He sent but on linguistic assumptions about thesemantitsimilarity the full paper to the ISKO General Secretariat from of concepts in head/modifier and conj unction relations. where a copy can be requested. One of the important R. BOUCHE, Sylvia LAINE and Jean-Paul METZ­ remarks made by him is that the usual standard for GER developed a new perspective to the topic of noun thesauri, especially the one for multilingual thesauri phrase retrieval (in °Knowledge retrieval from a docu- prepared by Europeans are not adequate for oriental

Int. Classif. 17 (1990) NO.3/4 172 ISKO News 4 languages, in this case Chinese and Japanese. And I Within this single framework each paper contributed myself, have the same feelings in this respect, especially to a different aspect of universal systems. My only regret when working on the Japanziation of the ISO standard is that we did not have enough time for discussion and for monolingual thesauri. Also we are going to establish questions and a synthesization at the end. I think we another ISO standard on multi�lingual thesauri. We could have worked on some interrelationships between found that there are general items that cannot be applied these presentations which would have produced some to the oriental languages. This was also noticed by rather interesting results. Gerhard BUDIN in his question. These are the general streams of the terminological Workshop E: "Knowledge Organization in problems which have been dealt with in this session. Special Systems "

Chair and Rapporteur: PavIa STANCIKOVA( Bratislava The four papers presented and discussed in this Wor kshop D: "Knowledge Organization by workshop focussed on four different subject fields: Universal Systems" biology, soil protection, anaethesiology and Indology. All of these fields were looked at from the point of view Chair and Rapporteur: Nancy WILLIAMSON, Toronto of terminology, concept an'alysis, classification and It was a very successful session partly because of the indexing schemes. coordination of the topics � they fitted together very Gerhard BUDIN, Vienna, (in "Terminological prob­ well � in the session, partly because of the expertise of lems in the organization and presentation of biological the speakers. It began with a presentation "On the prob� knowledge") discussed the following main topics: lerns � in general � of unified classification systems" . Eric Philosophy of science of biology and construction of a de GROLIER, Paris, with his usual vigourous self and knowledge base for zoology; the systematic character of with stimulating ideas set an excellent tone for the rest biological knowledge; the concept of "species" as one of of the session. He presented an analysis of various views the key concepts in biology; the organization of biological of categorization with examples of the use of these knowledge, mainly its problems of their maintenance categorizations in each case. Included in these cate� from the point of view of genetics or newly discovered gorizations were (1) Consensus, (2) Philosophical points species. A new trend in this field was foreseen - "evolu­ of view, and (3) Semantic categories. He' concluded his tionary classification" and its main advantage: flexibility. presentation by looking at practical problems which he The presentation of Erich WEIHS. Munich, took our had identified as 1) Fundamental to categorization and attention to the Bavarian soil protection system with to a unified classification system. These were gene� references to special systems as geology, water ecology. ral categories, 2) Descriptions, and 3) Notation. He feels References arc managed over one thesaurus and data that all of these need further investigation and studies in over data banks, maps and documents. The queries can terms of contemporary organization of knowledge. This be formulated as space related with coordinates and paper provided an excellent framework for the other topographic information, as e.g. village names, rivers, three papers. names of maps and administration related names and The first of these was given by Edouard SUKIASYAN borders. Discussed were the topic of the thesaurus of the of the Lenin State Library, Moscow. He described the soil protection system. It is not a hierarchical thesaurus background of the BBK, the Library Bibliographical but one structured as a network . Also the problem of Classification of the USSR - with its structure and its im­ inverted terms was discussed. portance in science and technology. Mr.Sukiasyan The position and organization of subject contents of concluded by saying that while this was a classification anaesthesiology waS presented and discussed by Pad mini used in the USSR it could be used in other political RAJ, DUsseldorf, mainly from the point of view of contexts as well. various library classification schemes (Universal and His paper was followed by the one of Gerhard special ones), It was pointed out that not all of the RIESTHUlS, Amsterdam, and Steffi BLIEDUNG, systems used are cognitive, bibliographical, bibliothecal Berlin, on a new approach to the Universal Decimal and self-perpetuating which I consider an essential Classification. They described an experiment using the requirement in any classification scheme. UDC in the development of a thesaurus: a very successful The fourth presentation of A.A.N. RAJU, Hydera· application, based on Class 314. It demonstrates that it bad, lead us to the problems of Indology classification could be done and showed the importance of the rela� and organization. It was pointed out that Indology tionship between a classification system and a thesaurus. has an important and prominent place in Indian libraries, I think it presented an intriguing possibility for new especially in academic libraries. Most of them use the approaches to the UDC and for new uses of the UDC in DDC, the UDC and the CC, but various branches of the online environment. Indology as, e.g., philosophy, religion, language, literature, The final paper looked at a different aspect of universal and culture are not satisfactorily elaborated in the DDC systems. Werner BIES, from the Free University of or the UDC. It was stressed that CC is in the better Berlin discussed the controversy over the alphabetic position with respect to various branches of Indology. subject cataloguing rules in the German library context. The discussion focussed on a necessary standardization He demonstrated to my satisfaction the reasons for difficulty in achieving an acceptable consensus on codes of the problems of Indology classification. of this kind. It is of particular interest to North Americans Workshop F: "Thesaurus issues" where there are no such cataloguing rules at the moment. There is a good deal of discussion about whether or not Chair and Rapporteur: Robert FUGMANN, Idstein we need a code and what that code should be. The first of the four papers presented was read by

Int. Classif. 17 (1990) No.3/4 ISKONews4 173 Alice STERN and Norbert RICHETTE from Luxem� Workshop G: "Online Retrieval" bourg: "On the construction of a superthesaurus based Chair: Ingeborg STOLTZENBURG, Frankfurt, Rappor­ on existing thesauri". The aim of this project is to teur: Anatol VASILJEV, Delft compile several thesauri from various fields, in different There were speakers from four different countries, languages, and in different conceptual structures and only one paper of these has already been published in formats and of different degrees of specificity into one VoLl of the proceedings. single master thesaurus, in such a way that the origin of Winfried GODERT, Cologne, stressed the necessity of each descriptor remained visible. Such a thesaurus can an interactive design of different subcomponents of any serve several purposes, among them are the purposes of subject access facility. The four subcomponents are indexing, giving aid to and providing the raw material enumerated in his abstract (p.161 of the proceedings). for the construction of new thesauri, assisting in auto� He also presented a typology of users' queries, and gave matic translation by virtue of its multilingual character, examples of a number of design problems. and linguistic research. Hemalata IYER, Albany, NY, presented a comparative user study showing the power of Ranganathan's facets, The paper of Tim c. CRAVEN from London, Ontario, on the verbal plane, in an online environment. The bond was entitled "Automatic structure modification in the strength of facets was utilized and the use of Boolean graphic display of thesauri". Two-dimensional thesauri operators for the different facets was standardized. The can display their relational network in greater conceptual two control groups in the experiment were: (1) an transparency than the conventional, linearly arranged unstructured quorum search, (2) a conventional Boolean thesauri. But there are always geometrically different search. The results indicated that the searches based on possibilities of arranging the mutually interconnected the faceted model performed much better. Precision was descriptors. These possibilities can vary widely with higher than 30%, recall was also better (ca 20%). respect to their perspicuity. The closer the geometrical , Anatol VASILJEV, Delft, NL, presented the human proximity into which closely related concepts are side of the human interface. He recommends a broad brought and the less the connecting lines between the approach to the design or redesign of the IR system, descriptors cross over, the more does such a display serve envolving brain-mind research, computer information the intended purpose. To achieve this goal purely science and artificial intelligence, neural networks, manually would require much time. Therefore a computer cognitive science, etc. The ways of human thinking program was developed which automatically maximizes should be taken into consideration. The Law of Requisite the transparency of the relational network of the de­ Variety (LRV) is recommended as a guiding principle. In scriptors in a thesaurus. Sometimes, it even proves the second part of his paper he showed the differences in necessary to introduce dummy descriptors to achieve the syndetic structure between the conventional subject satisfactory transparency. headings card catalogue and a subject heading access "Dynamic thesauri: The cognitive function" was the point file of an online catalogue being presented as a title of the paper presented by Lorna REES-POTTER, KWIC-index. He recommends the introduction of more MontreaL The idea pursued in this project is to semi­ "see" references to match the variety in users' queries automatically generate a thesaurus from the full text of better (LRV). documents. This approach is based on the observation Mary DYKSTRA, Halifax, N.S., presented a study that cited papers seem to influence the terminology as concerning the possibility of using PRECIS as a model in chosen by the authors of the citing papers. The citing the design of an expert system for text analysis and papers have much in common with respect to the retrieval. The analytico-synthetic rules for the formation terminology in which they discuss the findings of the of both the syntactic and thesaural relationships in cited paper. The corresponding noun phrases were PRECIS have potential relevance. The four parts of her manually extracted from the citing papers and they were project are given in her abstract on p.171. The realiza� recognized as meaningful candidate terms, which were tion of the project will be carried out with the Group then organized into subgroups based on conceptual Quatro, Montreal (see Abstract). similarities. Ongoing research aims at building individual Workshop H: "Knowledge Organization in concept records for each of the candidate terms, and at Special Systems II" noting also the related terms in these records. The relations to be taken into consideration are hierarchy, Chair and Rapporteur: Horst KORNER, Munich equivalence, and association, but also categorial relations We had one change in our program: The first paper such as facet relations. by Pnina WENTZ was not given, instead, Annelise Mark PEJTERSEN, Roskilde, DK, came in and thus we did In the fourth paper, Michael EISNER, Darmstadt, have four papers in our workshop. Altogether I can say reported on ongoing developments of the socalled that all four papers had not only theoretical approaches ARBOR system. It comprises a thesaurus with its to their problems but they also went into practical typical concept relations and is applied to the recording applications. You might have seen the demonstration of and storing of information on objects of classical ar� Ms.Pejtersen outside with Icons of the Book House. Her cheology or history of art. The progress recently achieved paper was mainly on the subject analysis for setting up a in the system is that the specific thesaurus not only system like that. There was a very detailed investigation serves for setting up a database but now serves also for of user behaviour and so on before they started out in retrieval functions. For example, several kinds of Greek drawing the icons for the computer and making the temple were classified. New archeological objects can system work. then be equated with one of the temple types in the The second paper was by Hans CZAP of the University store by a retrieval procedure. of Trier. He spoke about the increasing complexity of

Int. Classi!. 17 (1990) No.3/4 174 ISKO News4 economic and industrial documentation. For instance if Thereafter we had two papers presented by Lisbeth ' a new drug is applied for commercial distribution in BJORKLUND, of the LIBLAB Department of Computer Germany in one case it took 200,000 pages of documen­ and Information Science in Linkoping University, tation of text. before the product could come on the Sweden. The first paper she presented was on Hyper­ market. In another case it required 70,000 pages for a class. She showed four hypertext applications for a new telephone system by Siemens. The problem is that classification scheme, Storyspace, Guide, HyperCards the increasing complexity of these industrial products and NoteCards. From this paper it became clear, what cannot be handled by conventional systems. He went hypertexts can do for the creation, presentation, struc­ into some details of the theory of information systems, turing, and interconnection of classification systems and especially of industrial use and came up with some first for access to them. theoretical and implementational steps. He is very much In her second paper she talked of special projects interested in making contacts with whoever would like within the LIB LAB Department, the " Archival Informa­ to work in the same direction. tion 2000" project. It will probably become a part of The 3rd paper was by Josef HOLZL of the university the national archival data base in Sweden. The paper of Economics in Vienna on "Expert systems in product showed clearly that hypertext is a very useful tool to and commodity economics". Here again there was a bring users far more easily nearer to archival structure theory which went back to Wittgenstein and others, and which they are confronted with when looking for the practical side was whether in product and commodity information. information systems it is useful to have human experts The fourth paper was presented by Roland H]ERPPE, or to use expert systems; or whether a combination also from LIBLAB Department of Computer and Info­ of asking the experts to spend more of their time in rmation Science. He presented, as a special example of developing the expert systems so that also not so expert an application of hypertext a small section ofP .M.Roget's salesmen or so could make use of expertise of the real thesaurus. Very good pictures showed not only that this experts. One example was that you could have an thesaurus which was originally meant_as a classification, expert system on a portable computer. The salesman can once presented as hypertext was not only a possibility to take it to the customer. Even the very complicated get the users in, but was also a device to discern con­ products like radio transmitters with various frequency textual structures and also some standard meanings' in a variables and quartz and what not the salesman can word and in the clusters. come up with a precise offer and give the prices and so All in all it can be stated that in this workshop it was on suiting thus the individual customer needs. well presented how hypertext is about to open new ways The last paper was by Jutta VISCHER of Hamburg. for users to come to the information sources as well 'as a She reported on her thesis work at the Hamburg Poly­ new means of discerning meaning from existing struc­ technic for Library and Information Science on the tures which may have been produced for quite other "Harmonized System for the naming and encoding of purposes. commodities of the international trade". This harmonized From a technical viewpoint it might be said, one and system took some ten years to develop and came out for a half hour for four of such high-calibre papers, all a variety of countries trading with each other, it was related to practical experience and therefore very adopted in 1988 by a lot of countries. It gives the first valuable, was just too short. I had to press my speakers four hierarchical levels fixed for everyone using the quite a bit. And although the papers were very well system. But in the lower levels the countries can differ. presented in written form, I wish that we had had more She analysed this with the methods of classification time to have a closer look at the charts, and ask more theory. And in her work that she docs at the University questions. So, next time it would be my wish to have of the Deutsche Bundeswehr, Hamburg, she can also about 2 hours or so for such a session. make use of her experiences. Wo rkshop K: "Retrieval from Un iversal Systems " Wor ks hop J: "Hypermedia" Chair: Barbara KELM, Frankfurt; Rapporteur: Nancy Chair and Rapporteur: Winfried SCHMITZ-ESSER WILLIAMSON, Toronto The first of our three speakers and four papers was There were four speakers and three presentations Arek LESCH from Pforzheim, who, as a member of the each dealing with one of the major classification schemes. SEL Akatel, is active for the European RACE project. The first two speakers, Hannes HUG and Meta WALSER In his paper - together with Peter SZABO - on "Hyper­ from the ETH ZUrich combined their presentation. They media approaches" he gave the basic definition of spoke on the system ETHICS developed for the library hypermedia and hypertexts. From this definition it of the Swiss Federal Institute of Technology. It is an became clear that hypermedia cannot only be used online catalogue plus a circulation system used by as a tool for knowledge organization but also that a network of nine participating libraries. The interesting hypermedia must be regarded, and, in fact are, the set of thing about this presentation was that it was demonstrated all available media now and in the future. Both, the tool as a live system with the two people participating. They and the thing, are being brought together to form a new were actually doing retrieval for us connected with their type of system which is capable of navigating through library in Zurich by telephone and bringing the results the world of science and make it easier and much more upon a screen. It demonstrated very well the system, it attractive to the users to get to the sources of knowl� provided for retrieval by keyword/subject terms and by edge. In a most impressive way he showed his charts, and the Universal Decimal Classification. This operating how this can be done. And in fact, a new world of access system has a parallel with the research work which to knowledge and control of knowledge was visible in his Karen Markey has been doing in the USA on the Dewey speech. Decimal Classification.

Int. Classif. 17 (1990) No.3/4 ISKONews4 175 The second paper was given by myself on the Library that such systems should be built by skilled knowledge of Congress Classification and its move toward com� engineers. puterization. The description of the MARC format has Michael SCHOPEN's paper was on "Cross file search­ now been approved in principle for testing by the ing of biomedical databases at DIMDI". DIMDI, the Library of Congress and the Dewey Decimal Classifica� German Institute for Medical Documentation and tion. There is a hope that this format will be flexible Information in Cologne, offers sequential search of an enough to be used for other classification systems unlimited number of files. The programming language such as the UDC, Bliss and whatever. If this does come GRIPS-L and the EXTRACT COMMAND were developed about, such common use of the same format would as tools for cross file searching. It was proposed that encourage the international exchange of information specific menu-driven systems would be designed for each based on classification. data base which would prove to be extremely useful. The final paper was given by M.A.GOPINATH from It was pointed out that a future approach would include Bangalore. He discussed information retrieval using the the transfer of user-defined concepts from one database Colon Classification. He described the system using an to the other followd by translation of these concepts inference engine and a user interface as a filtering into a specific data base language using intelligent user process and showed the importance of analytical classifi� interfaces. cation in online retrieval. It is a good sign that n.ew applications are being The significance of this workshop was that we worked reported in the area of retrieval technologies involving together in the discussion of the three major systems studies in information science and artificial intelligence. which are important in information and online retrieval User friendliness is being given due consideration. -and while many of us might like a new classification for Many of the systems being deVeloped are dynamic and information retrieval most of us know that this is not designed to respond to a changing environment. going to happen soon, So we must look seriously at making the best use of those systems we have. Workshop M: "Indexing" Workshop L: "Retrieval Technologies" Chair and Rapporteur: Gerhard RlESTHUIS, Amsterdam Claus POULSEN from the National Library of Chair and Rapporteur: Krishan KUMAR, New Delhi Education, Copenhagen, in "An indexing concept As the first of four papers Alina VICKERY, London supporting subject access for innovation and creativity" presented hers, entitled "Knowledge organization in reported on his proj ect. He is experimenting with intelligent tutoring systems". She described EUROSAGE, information retrieval not using classification systems, an intelligent tutoring system, being developed by thesauri or free text. Instead, his hypothesis is that the TOME Associates to help managers to gain better subject indexing is already done by the author. He uses understanding of business opportunities in Europe. She two means: (1) General review articles, (2) Tables of discussed the ways of organizing and representing contents, indexes and the like. Based on this hypothesis knowledge. It was mentioned that dictionary with he built an information retrieval system called PARA­ thesaurus relations, classification, semantic categories, DOKS. frames on semantic analyses, hypertext, etc. were highly R.G,PRASHER from the University of Saugar, India, suitable tools for this purpose. Use of artificial intelligent gave a report on his research in the field of indexing and hypertext techniques for building moduls of the journals. His research is based on the work done in the system appear to have interesting possibilities when field of back-of-the-book indexing, PRECIS and POPS\. completed. M.PARAMESWARAN from the University of Calicut, Veronika OECHTERING, Dortmund, presented the India, spoke on the problem connected with the making paper entitled "On the problem of transparency of of a chain index for a systematic catalogue in which the individual computer technologies in subject-oriented Dewey Decimal Classification is used. Most problems online retrieval". She put forward a model of human­ are the result of shortcomings in this system. computer interaction (HCI). She had the view that transparency is a negative criterion in evaluation and Mirj a IlVONEN from the University of Tampere, manifests itself as intransparency. It was concluded that Finland, reported on a study of interindexer consistency. all themes showed big transparency problems. She As in most studies of this kind she found that consistency thought that intermediaries will never get exact knowl­ was very low. Important factors for this were: (1) edge of the client problems. During the discussion, the central aspect of the work was described in different R. FUGMANN pointed out that an ideal combination ways, (2) some indexers left the central aspect out, (3) would be that there should be a joint effort of the end some indexers indexed peripheral aspects, and (4) users and the intermediaries. Mention was also made of central aspects were sometimes expressed in more than the need for more research on interdisciplinary hurnan� one way. computer interaction. This ends the two reports on the First International ISKO Marlene ROCKMORE, Maynard, MA, presented a Conference, We are very grateful for an unexpected rich result, for a real harvest of valuable contributions with new ideas and paper which dealt with "Facet analysis and thesauri for helpful methods, of new aspects and applications. They are a corporate information retrieval". She described TIMS, a treasure for thorough examination, further investigation, research single thesaurus system to support retrieval �f corporate and development in our field of knowledge organization. The information. The special feature of TIMS being a set of reader is invited to look into the proceedings volumes (reference above) and gain a more comprehensive survey and a deeper tools for machine-assisted indexing called the indexer's insight into the subject presented and discussed. May this then workbench. She claimed that a faceted thesaurus struc­ bear fruits in any further professional work and for the benefit ture is very effective in its environment and the system of the future of the field of ISKO's concern! has proved to be simple and easy to use. She emphasized I.Dahlberg

Int. Classif. 17 (1990) NO.3/4 176 ISKONews4 to allow for mutual discussions. The terms of reference under which it met have already been published in FIDI CR News and all of these were duly fulfilled. The final FIDleR News 28 recommendations were unanimous and were put before the Management Board by the Chairman at a- meeting held in March 1990 and before the FID Council at its meeting in Yugoslavia the following month. The general recommendations ""ere accepted, though the manner in Congratulations to Ingetraut DAHLBERG on the which they may be implemented will undoubtedly have success of the ISKO conference in Darmstadt, Federal to be limited by financial constraints. The President has Republic of Germany, in August. 1 already indicated thi� in print. FIDICR matters The individual memb-ers did a great deal of investigat­ ing between meetings in addition to reading past papers Unfortunately, the Chairperson of FIDleR was and recommendations from FID. In order to discover unable to attend the 45th Conference and Congress of more about the users of the scheme and their needs FID in Havana, Cuba, in September because of complica­ each member undertook a mini-survey, attempting to tions in travel arrangements. This was regrettable and gain as international and diverse response as possible. It apologies were expressed to the delegates through Ben was conducted informally, usually in person, and fol­ Goedegcbuure, Executive Director of FID and an lowing a series of points that was agreed beforehand, "Annual Report" was prepared and sent. A summary but without too structured an approach. This was both of that report follows: revealing and heartening and it was felt that this was a In 1989190, FIDleR's major project was to aid in very successful part of the Group's work, as half a day's efforts to strengthen the future position of UDC as a informal discussion with a user brought out much more viable classification system for libraries and information information than a survey or questionnaire by post centres world-wide. Three current members of FID/CR, could possibly have done. It revealed considerable Dr. Ia MCILWAINE, United Kingdom, Professor Ger­ ignorance about the classification and how it works on hard RIESTHUIS, Netherlands and Professor Nancy the part of the users and it was particularly encouraging WILLIAMSON, Canada served on the Task Force on in that it demonstrated a great enthusiasm for the UDC System Development, chaired by Dr. MCILWAINE. scheme, far beyond any expectations. A more detailed report of the activities of the Task The Task Force agreed that if the UDC was to con­ Force is given below. tinue to occupy a place in the forefront of bibliographical Activities planned for 1990191 include the 5th Inter­ classifications there was a need for extreme urgency. national Study Conference on Classification Research More money was needed to make it more secure and (ISCCR '91) to be held in Toronto, Canada, June 24-28, outdated sections needed revision without delay if users 1991. A call for papers was issued in June 1990, and the were not to be diverted to its competitors. Computer conference will be co-sponsored by the Faculty of science is just one example of a discipline where great Library and Information Science, University of Toronto, dissatisfaction is felt at present and users are working the Canadian Classification Research Group, the IFLA out their own versions which they will implement if Section on Classification and Indexing, the International FID does not improve the situation soon. Publicity for Society for Knowledge Organization (ISKO) and the the scheme was also seen to be lacking and it was felt American Society for Information Science SIGleR. that if there were a better flow of communication there Theoreticians, researchers and practitioners in the fields might be greater interest. of classification and organization of knowledge are As a priority, the Task Force recommended the invited to participate. The hope is that some important creation of a machine readable master version, from directions for research will result from this conference, which different language editions and editions .of vary­ and that FIDICR will be in a position to identify some ing size could be developed. It recommended that this specific research proj ects which it might foster and help should be a version of about 60,000 terms (i.e. about to shape its future. three times the size of the Dewey Decimal Classification, At least one funded project would be desirable if and roughly the size of the German and English Medium this could be achieved. One approach to funding might editions). It was proposed that this should be in English. be a project which might attract corporate sponsorship. Until there is such a tool, the work of revision is handi­ Members of FIDleR and others who have suggestions capped since at present no authorized text exists in an for possible projects are invited to send their ideas for easily manageable form. The report anticipated that research proj ects to the Chairperson of FIDICR. funds would be released to enable this to be done and that suitably qualified personnel would be employed on TaskForce for UDC System Development a two year contract to accomplish the task. It recommend­ This Task Force, which has now completed its work ed that a small panel (a maximum of three members) and submitted its final report, held three meetings should be set up to plan the implementation of this between August 1989 and February 1990. The first, proposal. This panel's first task would be to revise which was during the 1989 IFLA General Conference in the former publication. It was proposed that future Paris, was followed up by an informal discussion later revision should be much more consistently faceted than that same week and the final meeting lasted two days. had been the practice in the past and that the ultimate The intermediate meeting lasted one day but was organiz­ goal should be a fully faceted scheme supported by a ed to coincide with the October meetings of the UDe semantic network. Management Board and chairmen of the eRe's meeting, The Task Force was very concerned over the manner

In!. Classif. 17 (1990) No.314 FIDleR News 28 177 in which revision work was organized and untertaken. It IFLA Division of Bibliographical Control, felt that the CRCs were not yet all functioning in the Section on Classification and Indexing manner that had been anticipated when they were Grig­ The primary activity of the IFLA Section on Classifi­ inally set up. Lack of funds and insufficient voluntary cation and Indexing this year was the on-going work of expertise make their task a difficult one. A good starting the Section's Working Group on "Guidelines for Subject point for increased participation might be the involve­ Authority Files." The North American members of the ment of those organizations who had expressed interest Working Group met twice in Chicago, USA in 1990 in assisting, in their answers to the questionnaire recently during meetings of the American Library Association circulated by FID. It was also proposed that the small and the European members met for two days in Frank­ working group responsible for planning the creation of furt, Federal Republic of Germany. Drafts of the "Guide­ the online version and for producing revisions to the lines" were discussed and revised at these meetings and Principles .. . and the "Code of practice" should act as further revision took place when all of the f(lembers of an intermediary between the CRCs and the Management the Working Group met together during the IFLA Con­ Board, so that there was expertise available to assist ference in Stockholm in August. The current goal is when points of conflicting interest occured and to co­ to have a draft of the "Guidelines" ready for world­ ordinate the work of the five committees. There was wide review in February 1991, with the expectation general agreement that in the interest of speeding up that comments and reactions 'received will be discussed revisions the rule forbidding re-use of a number for a during the IFLA Conference in Moscow in August, 1991. changed concept for at least ten years should be recon­ We rner STEPHAN, Federal Republic of Germany, is sidered, the current Chairperson of the Working Group, having Particular attention was paid to ways in which the taken over from Barbara KELM, Federal Republic of scheme might be better publicized, More frequent information on recent editions and revisions in progress Germany, at the Stockholm conference. The Section's programme for its open session during was needed and various proposals were made, including the Stockholm conference included two papers, "Classifi­ the publication of an annual list issued by FID of what cation and Indexing in the Nordic Countries", by Tor was available, greater use of the professional press and HENDRIKSEN, Director of the Norwegian School of the possiblitiy of setting up user's panels to improve Library and Information Science, Oslo. He discussed communication and stimulate interest, classification and indexing in Denmark, Finland, Iceland, Following the Council's approval of the Task Force's Norway and Sweden stressing Nordic co-operation in report and its statement that it intends to continue its classification and indexing and Nordic involvement in support of the UDC a small group has already begun international library organizations. A variety of classifi­ redrafting the Principles .. , and plans are being made cation schemes are used and the handling of the verbal for the creation of a machine readable master version. subject approach differs from one country to another. The Task Force hoped that this would be completed In the second paper, Arj a-Riita HAARALA, Director of within a two year period and that more radical revisions the Tampere University of Technology Library, Tampere, would be implemented before the end of the century. Finland, discussed, "The Role of UDC in Finnish Classifi­ . There seems no reason why this should not happen if cation Policy". UDC is very widely used in Finland to sufficient funding can be found to carry out the pro­ classify monograph and serial collections, for shelf posals. I.e. McIlwaine, arrangement, for bulletins of new acquisitions, for Chairman bibliographies and for reference databases. Included in 1 President's column, FJD News Bulletin, 40 (4), 1990, p.SS. the paper was information on recent activities related to the UDC at the management level within FID. Current New Publication revision work and the re-structuring of the UDC revision process were discussed. A recent publication on UDC is The UDC: Essays fo r a One of the founding members of the Section, E.R. New Decade, edited by Alan GILCHRIST and David SUKIASIAN. has written a paper on the "IFLA Section STRACHAN, This collection of ten papers describes on Classification and Indexing, 1977-1988. This paper re cent managerial activities and technical applications was available at the Stockholm conference. Also the of UDC and is published by Aslib in London. Section continues to publish an annual Newsletter which provides a means of sharing differing national experiences MARC Format for Classification and of fostering an international perspective on classifi­ At the 1990 Annual Conference in Chicago, USA, in cation and inde:x.ing. June, the ALA's Committee on Machine Readable Dorothy McGarry Bibliographic Information (MARBI) gave "provisional" approval to a US MARC format for classfication. The ASIS SIGICR Library of Congress and the Dewey Decimal Classifica­ The 1990 ASIS Annual Meeting in Toronto, Canada, tion Editorial Office will now begin to test the format November 4-8, will include a number of papers and for eventual implementation, It is expected that a version presentatirms of particular interest to FIDICR members. of the format will be published by the Library of Con­ ASIS SIGICR will be sponsoring two programmes. A gress Cataloging Distribution Service (CDS) in the spring programme on "Applications of Classification" will focus of 1991. It is hoped that the format will be sufficiently on the use of existing classification systems, methods flexible for adaptation for use for such classification and problems in the development of classification schemes as the National Library of Medicine Classifica­ systems for document collections, proposals for the use tion, UDC, Bliss and other systems. of classification in automated text analysis and classifi-

Int. Classif. 17 (1990) No.3/4 178 FIDleR News 28 cation for technology and problem solving. Papers wpl RYMPLE (University of Illinois, Urbana-Champaigne) be presented by Clare BEGHTOL, Ernest BRETON, and Jennifer A. YOUNGER (University of Wisconsin­ Mary DYKSTRA and Nancy WILLIAMSON. The second Madison): An Investigation into the Relationship be­ programme on "Indexing Software; the Theoretical and tween Subject Indexing Principles and Online Subject the Possible" will explore the capabilities of stand­ Searching Strategies. - Marta J. FERNANDEZ (Guate­ alone indexing software and discuss the issue of maintain­ mala City, Guatemala) and Caroline M. EASTMAN ing indexing standards. Presentations will be made by (University of South Carolina): A Comparison of Tax� Linda FETTERS and Cecilia WITTMANN. onomic Knowledge Structures, - Janice GLASGOW Additionally SIGleR will co-sponsor three sessions: (Queen's University, Kingston, Ont., Canada); Imagery a practice oriented session on "Thesaurus Software and Classification. - Stephen C. HIRTLE (University Specifications"; a session on "Navigating Through Large of Pittsburgh): Ordered Trees: A Structure for the Data Spaces" which will explore different approaches Mental Representation of Information. � Robert KRO­ to this problem kind including associative searching using VETZ (University of Massachusetts, Amherst': Informa­ hypermedia, current and projected tools for the University tion Retrieval and Lexical Ambiguity. � Elizabeth D. of California MELVYL system, graphic interfaces and LIDDY, Carol A. HERT, and Philipp DOTY (Syracuse thesauri; and a session on "Using Machine Readable University): Roget's International Thesaurus: Con­ Dictionaries" will examine the use of such dictionaries ceptual Issues and Potential. Applications, - Mark T, in natural language processing. MAYBURY (Rome Air Development Center): Generat­ An AS IS contributed paper session on "Information ing Natural Language Definitions from Classification Handling Programs and Systems" will include a paper Hierarchies. - Mary MICCO (Indiana University of by Jessica MILSTEAD on "Thesaurus Software Packages" Pennsylvania): Report on Linking Subject Headings to and one on "Design for a Subject Search Interface and LC Classification Numbers and Suggestion for Automat­ Online Thesaurus" by Marcia BATES. ing the Classification Schedules for the Explicit Purpose Following the ASIS conference, on Thursday No­ of Improving Subject Access in Online Public Access vember 8, at the Toronto Conference of Data Base Catalogs. - Amadeo NAPOLI, Roland DUCOURNAU, Users, SIGICR and SIG/HCI (Human Computer Inter­ and Claude LAURENCO (Universite Louis Pasteur, face) will co-sponsor a program on "Thesauri as Knowl­ Strasbourg, France): An Object-Oriented Approach to edge Bases for Information Retrieval." Papers will be Classification. - James D, Palmer, Yiqing LIANG, and presented by Alina VICKERY, Deidre STAM, Marlene Lilian WANG (George Mason University): Classification ROCKMORE and Trish YANCEY. as an Approach to Requirements Analysis. - Eric RAMES and Bernard DURIN (MATRA Espace, Toulouse, France): A Classification Model for Reusable Software 1st ASIS SIGICR Classification Research Components. - Marlene ROCKMORE (Digital Equipment Workshop Corporation): Computer-Aided Knowledge Engineering Twenty papers were accepted and are scheduled for Corporate Information. - Dolly SAMSON (Weber to be presented at the CR Workshop to be held at State College): A Comparison of Classification Schemes ASIS'90 on Sunday, November 4, 1990. These papers in Support of Knowledge-Based System Test Planning. - represent a wide range of interests and disciplines, but Philip J. SMITH and Rebecca DENNING (Ohio State of course they express a common belief that classifica� University), Steven J. SHUTE (AT&T Bell Labs.), and tion schemes are a key element in information systems Lorraine F. NORMORE (Chemical Abstracts Service): and information science projects. We are very pleased Toward the Development of Semantically-Based- Search with the reports of the important research or work on Systems. - Larry M. STEPHENS (University of South information systems that will be presented at the Work­ Carolina): The Classification of Semantic Relations shop. Based on Primitive Properties. Although initial proceedings will be distributed only (From ASISISIGICR Ne ws, August 1990, p.2-3) to participants, they might be published for wider U.K.: CRG 276 and 277 distribution next year. Announcements would be made as soon as details are known. The 27 6th meeting of the British Classification Co-chairs of the workshop are Susanne M. HUMPHREY Research Group was held on April 26th, 1990 at the (National Library of Medicine) and Barbara H. KWAS­ University College London with 8 members present. NIK (Syracuse University). Authors and titles of papers The members expressed their deep regret at the recent are as follows (titles may still change): death of Mr. SANDISON who had been a faithful James D. ANDERSON (Rutgers, The State University member of the Group for about 25 years and had made of New Jersey); Ad Hoc, User-Determined Classifica­ many useful contributions to its discussions and delibera­ tion Based on Faceted Indexing. - David BEARMAN tions. In recent years he had also served as a most reliable (Archives and Museum Informatics) and Toni PETER­ Treasurer. SEN (The Getty Art History Information Program); The topic of this meeting were the practical and Retrieval Requirements of Faceted Thesauri in Inter­ theoretical implications for subject indexing in large active Information Systems. - Ernest J .BRETON (U­ scale bibliographic databases in the online environment. niversity of Minnesota, Deluth); Functional-Representa­ The paper by Joyce Butcher and Ross Trotter, Building tion of Technology. - Cheryl Ventura CONWAY (Allied­ on PRECIS, presented at the 1989 IFLA Conference in Signal Aerospace Co.): Semantic Classification of Paris, formed the basis of the discussions. Members Maintenance Information. - Bruce B. COX (U.S. Patent were reminded that the British Library was developing and Trademark Office): Improving Intellectual Access to a new system for use in indexing the humanities and the U.S. Patent Classification. - Prudence W. DAL- social science collections. Miss MEDLIN explained the

Int. Classil. 17 (1990) No.3/4 FIDleR News 28 179 intentions of the British Library that controlled supject ties of working with the computer�hidden relationships data, pitched at a fairly broad level were needed to included in a PRECIS string Of with the thesaurus which secure recall, but that precision could be achieved had been built up together with the PRECIS indexing without control. It became obvious in the discussion operation but still inaccessible, Mr. FOSKETT reverted that the changes planned would result in numerous to the matter of specificity and emphasized the impor- problems. The CRG members pointed to the dangers tance of providing different levels to suit differing if PRECIS indexing was abandoned, especially if the circumstances. He thought that the Group should indexing would not remain enough specific. consider how differing approaches could be incorporated At the further meeting on July 5th, the topic of the into one usable system. Mr. MILLS proposed to provide April meeting was continued, viz. levels of specificity the detail and then to truncate. Mr. NEILSON referred in indexing. Miss Medlin circulated a set of imaginary to the use of notation for that purpose. The discussions titles as examples for the new approach to be taken by were to be continued at the October 25 meeting. the British Library. It was agreed that specificity relates At this July meeting, it was also related that FID to the hierarchical level at which a concept appears in had now relinquished any responsibility for the Broad the indexing language, whereas exhaustivity implies System of Ordering (BSO) and that the three panel all subject index descriptions based on the summariza- members who had been in charge, were permitted to tion principle; one can be exhaustive without being do what they wished with it! (Summary from unconfirm- specific. After a longer discussion, involving possibili- ed minutes.) �------�

Thesaurus Software on Offer in Europe Babylone Documentation Anthesis Abbreviations: 413, Belles Partes Relations: E = Equivalence, H = Hierarchy, P = Proximity F -142 00 Herouville-Sain t-Clair SSTFI = Search Systems with Thesaurus Features Integrated (SSTFIj Relations: H, E, P, and other typical features; TFI = Thesaurus Features Integrated TMS = Thesaurus Management System Mac-OS)

(1) ENGLAND Documenta Platinum TMS - Thesaurus Management System 90, rue de Miromesnil F-75008 Paris Pyramid Computer Systems, Ltd. (Relations; E; SSTFI; Mac-OS) 59 Vastern Road H, Reading, Berkshire, RGl 8BY, UK (Relations: H, E, P, as well as a facet facilitYj MS-DOS. JLB Doc Comes as a monolingual and a multilingual version. A JBL Informatique standalone system) 39, Boulevard Victor Hugo F-92110 Clichy . CAIRS (SSTFI; Relations; H, E, P; MS-DOS and UNIX) Leatherhead Food Research Association Randalls Road Alexandrie Leatherhead, Surrey KT22 7RY, UK Qisk International (SSTFI) 2, rue de Bourets F*92 150 Suresnes CALMS (SSTFI; Relations; H, E, P; Mac-OS) Pyramid Computer Systems, Ltd. 59 Vastern Road Biblio-Tech Reading, Berkshire, RGI BBY, UK Platinum (Computer Aided Library Management System with 90, rue de Miromesnil TMS integrated , MS-DOS) F-75008 Paris (A library system with TFIj Relations: H, Ej Mac-OS) Micro-Status Textbase Management System Harwell Computer Power Colibris Curie Avenue Platinum Harwell, Oxfordshire, OXll OQW, UK 90, rue de Miromesnil (SSTFI; Relations; H, E, P, and others. MS-DOS, UNIX, F-75008 Paris TSO, and others) (SSTFI; Relations; H, E, P; Mac-OS) (2) FRANCE (3) GERMANY THESAUPLUS Progiplus CICADE Tour Neptune Display Softwaresystems GmbH&Co. Vertriebs-KG Cedex 20 Westenhellweg 72 F-920B6 Paris-La Defense D-4770 So est (Relations: 1-1, E, Pi running under MS-DOS, UNIX, (Full Text SSTF; Relations; H, E. Under MS-DOS, VMS and in other TEXTO environments) UNIX, VMS, and others)

Int. Classif. 17 (1990) No.3/4 180 FIDleR News 28 PROTERM CDS-ISIS Progris GmbH Free to developing countries from UNESCO Auguste-Viktoria-Stralk 64 (A thesaurus component coming as mainframe, mini D-1000 Berlin 33 and micro versions by p, Vasarhelyi, Budapest) (Standalone thesaurus system; Relations; H, E, P, and other useful features; MS-DOS) MultiMate Advantage 2 by Logic Soft INDEX (SSTFL Relations, E, P. MS-DOS) (by Ernst Lukas, vended by,) agi - Information Management Consultants TIN LIB Georgenstr. 120 by AGICOM AG D-8000 Munchen 40 Holbeinstrage 46 (Standalone thesaurus system; as many different rclations CH-4051 Basel benveen terms in one or different languages as needed (a library system with TFI) can be implemented in DATAFLEX kernel; MS-DOS, UNIX, VMS) (This compilation was made by Dr. Winfried Schmitz­ Esser, Oderfeldstr. 13, D-2000 Hamburg 13; any helpful DOMESTIC .additional information on these and further programs KTS Informations-Systeme GmbH or changes of addresses is welcome) Leopoldstr, 87 0-8000 Munchen 40 (SSTI; Relations, H, E, P; IBM-MVS and CICS, as well as online)

LIDOS TermNet Seeks Pa rtner to Sell Land Software Entwicklung 9-Language Thesaurus Software Postfach 1126 The International Network forTerminology (TermNet) D-8507 Oberasbach b. Ntirnberg has developed a 7�language Micro ISIS-based thesaurus (A library housekeeping system with some TFI ;Relations: software (in English, French, German, Italian, Spanish H, E, P, and Partitive; MS-DOS) and Greek) and is about to enlarge it by two further languages. It looks for 2-3 partners interested in sharing and distributing this software which is based on CDS! (4) NETHERLANDS ISIS, a menu-driven generalized information storage and retrieval system designed specifically for the computeriz­ BIDOC ed management of structured non-numerical data bases. PTT-Borg To this a thesaurus module has been designed in order BIDATA Gebouw 3, 2e Etage to allow multilingual searching on Micro ISIS data bases. Stationsplein 7 The system is used for maintenance purposes and NL-9726 AE Groningen supports the following actions: (A standalone thesaurus module serving the H, E, P creating descriptors in seven languages, three of them relations, under MS�DOS) are compulsory others are optional; creating non-descriptors, in one language at a time Turbo-Thesaurus (no translation supported); D, van Bremen Informatic Advies B.V. adding and deleting relations, BT/NT, R T, USE/USED NL-Den Haag FOR addition is performed in one chosen language, (Standalone thesaurus construction and maintenance other languages and reciprocal is generated; program; Relations: H, E, P; MS-DOS) changing keywords spelling: one or more languages at a time; adding and updating scope notes and facet numbers (5) MARKETED IN VARIOUS COUNTRIES (i.e. classification codes). While processing, this program keeps track of two differ­ ALEPH ent "data stamps" for term matlipulation and relation­ Ex Libris Ltd. ship manipulation. P.O.Box 10568 The thesaurus database is a set of ISIS records. Tel Aviv 69085, Israel A record may contain: (A library system with TFI) Descriptors in 7 languages (Relations BT, NT, RT, USED FOR), Scope Notes, Facet numbers (or classifi­ BASIS plus cation codes); Date stamps. by Battelle No n-Descriptors in one language, USE relations for (SSTF!) that language, Scope note, Facet numbers (or classifi­ cation code), Date stamps. BRS-Search For further details, please contact: TermNet, Mr. C. Ga­ by BRS, US, also UK, D linski, Heinestr. 38, POB 130, A-l02l Vienna, Austria. (SSTFI) Phone (0222)2675 35-308; FAX 216 32 72.

Int. Classif. 17 (1990) No.3/4 FIDleR News 28 181 (3) Library Classification: Fundamental and Procedure, with 1008 graded examples and exercises. Madras: Madras Library Association 1944. (This is the only book not as yet reprinted by UBSPD in Book Reviews 1988-1990 under their programme mentioned.)

(4) Elements of Library Classification. Paona: NK publishing Ranganathan's Reprints in Library Classification House 1945. (Now in its 3rd edition 1967, published by Asia Publishing House, Bombay). Dr.S.R.Ranganathan (1892-1972), the father of the Indian Library Movement, has been the most prolific (5) Classification and Communication. Delhi: University of library science author of all times. His writings, especially Delhi 1951. on classification, won him the acclaims and admiration (6) Philosophy of Library Classification. Copenhagen: Ejnar of peers all over the world. His testament on classifica� Munksgaard 1951. tion forms the very rock of the current theory of library classification. His findings communicated through the (7) A descriptive account of the Colon Classification. New medium of books, journals, reports, and lectures have Brunswick, NJ, USA: Graduate School of Library Science. Rutgers State University 1965. (Published again in 1967 in pervaded and constitute our current knowledge of the India by the Asia Publishing House, Bombay). subject. His books arc the librarian's lore. The Sarada Ranganathan Endowment in Library Science, established in 1962, isaregistered trust and a legal In the Colon Classification (CC) first published in and intellectual legacy of Ranganathan to promote 1933 after eight years of thinking and practical applica­ research in library and information studies. Inevitably, it tion, Ranganathan gave to the bibliographic world a new has now shouldered the responsibility to develop and species of library classification, now called analytico­ perpetuate Ranganathan's thoughts. One of its aims is to synthetic classification. It is essentially a faceted system. create and publish research literature. So far it has been It treats the entire universe of know ledge as a single publishing a quarterly journal Library Science with a whole. Nevcrtheless, like other classification systems, Slant to Documentation and Information (1964�) and a it divides the universe of knowledge, at the first instance, few research booklets such as A purview of the CC-7 into manageable main classes called Basic Subjects. The (1970). Later, it took upon itself the publication of number and scope of basic subjects in the CC is always annual lectures delivered under its auspices. The first changing. The CC distinguishes itself from other systems major project published by it was the 7th edition of the in the further division of a main class. Each basic subject Colon Classification in 1987. Thereafter, the Endowment is further divided into fundamental and indivisible embarked on a new programme of reprinting Ranga­ concepts called Isolates. According to his postulates, all nathan's books. That is in direct fulfillment of Ranga­ the isolates belonging to any basic subject can be grouped nathan's wish as he was anxious to see his books updated at the most into five categories called the Five Funda­ and always available. In Section AE9 of his Prolegomena mental Categories (FFC), the least number of categories of 1967 he wroW for any bibliographic classification. These categories 'It is my policy to take the assistance of my younger colleagues exist independent of basic subjects. (Ranganathan in preparing new editions of books. This is done in the hope that wanted to abolish the oncept of main classes but was it will help them to develop into competent authors, and that ¢' there will be some competent young colleague of mine to take left with no time for research in this direction). In a care of the future editions of such of my books as may continue given basic subject within each category the isolates are to be of value and in demand.' subgrouped into facets. A facet is the totality of isolates His books certainly continue to be in demand. More belonging to one category and obtained on the basis of a than a dozen books on different subjects have been single characteristic of division. Within a facet, the published so far. All these books have been offset isolates are arranged into arrays and chains by applying printed with the imprint of the Sarada Ranganathan the logically formulated Principles of Helpful Sequence. Endowment with the sole distributor's rights remaining The order of concepts and classes was of fundamental with MIS UBS Publishers' Distributors (UBSPD) (5, importance to Ranganathan as was their infinite com­ Ansari Road, Delhi-110 002), India's largest book binability. The isolates can be combined in the citation exporters and distributors. (In Europe, these books are order of the five Fundamental Categories, called Person­ being distributed in an arrangement with UB5PD by the ality, Matter, Energy, Space, and Time (PMEST) to Indeks Verlag, Woogstr.36a, D-6000 Frankfurt 50, construct an infinity of class numbers from a relatively FRG). It is a common knowledge that it is the UBSPD brief schedule. CC formulates the nccessary grammar to who have financed and looked after the production of construct class numbers for different subjects. these books with professional skills. It ensures therefore Since its publication the CC has been accepted as the a good production and a wide distribution network. most useful method to cope with the turbulently advanc­ Ranganathan's maj or books on classification are: ing knowledge. Its methods have been adopted by all the

(1) Colon Classification. Madras: Madras Library Association major living classification systems. 1933. (Now in its 7th edition (1987). The last edition The Colon Classification of 1933 was a result of produced by Ranganathan was the 6th ed. (1959), again intuition without any objectively guiding principles at issued with amendments in 1963, published by Asia Publish­ hand. The theory behind the CC and the principles of ing House, Bombay). library classification were only published four years later

(2) Prolegomena to library classification. Madras: Madras in the Prolegomena of 1937. This book is veritably Library Association 1937. (Now in its 3rd edition of 1967, known as his magnum opus. The second edition was published by Asia Publishing House). revised by his British colleagues and admirers and was

Int. Classi!. 17 (1990) No.3/4 182 Book Reviews published by the Library Association in 1957. Now, this Verbal Plane, and the Notational Plane gave the para� book is in its 3rd edition published in 1967. Its method mount position to the Idea Plane. Ddinition, scope, and is essentially comparative. It provides norms to evaluate the order of the main classes/basic subjects should be any classification system, also outlines a methodology to based on the modes of growth of subjects in the Universe design library classific:Jtion schedules. It is a weighty of Knowledge. Knowledge is regarded as of a social work indeed. character. The scope of a basic subject is always chang­ Library Classifi cation,' Fundamental and Pr ocedure ing and depends upon literary warrant. Ranganathan (1944) is a book on practical classification by CC and had the considered view that library classificationists the DDC. In fact, it is a book of seminal ideas which must quickly respond to the changing social and intel­ later sprouted into very fundamental concepts. Efforts lectual world and their needs; indeed , it must hold a are on the way to revise it and adapt it to the current mirror to the changing world of knowledge and its editions of the CC and the DOC. various sources and embodiments. Altpough Ranga­ The Elements (1945) are a published form of a series nathan could not be credited to be the sale inventor of lectures on fundamentals of library classification first of faceted classification, it can be stated that without delivered in 1944 at the University of Bombay, and then him there would have been hardly any species of this in the UK in 1956. The second edition was edited by his kind of classification. He gave it the name and the intimate British friend Bernard I. Palmer (1910-1979) concrete substance. He applied the idea of fundamental and was published in the UK by the Association of categories to library classification and focussed his Assistant Librarians in 1960. The third edition was thoughts intensively upon it. He was also a great apostle published in India in 1962. In this little book, Ranga­ of technical terminology. He asserted time and again nathan - with the help of 108 subjects taken as examples that no discipline has ever grown to maturity without - illustrates the Principles of Helpful Sequence (order) its standard terminology. Many new concepts and their and other basics of library classification. In the epilogue, terms were thus created by him which even started to he exhorts fellow librarians to learn and use technical enrich the English language at large. Notation was for terminology for effective communication. Although the . him simply a tool to implement the findings and desicions book aims at the uninitiated, it is full of profound of the Idea Plane, relegating notation thus to the menial thoughts, and still forms an indispensable class room level. However, his notation comprises, at present, teaching and learning aid. 74 digits of different species, indeed a high water mark Classification and Communication (195 1) was the of a notational system. Also he equipped his notation fruit of a post-graduate degree course in Library Science with many devices to make it versatile and infinitely at the University of Delhi as well as of a visit to several hospitable to new subjects. No doubt, in the pursuit industrial libraries in USA in the summer of 1950 by of a large capacity and hospitality, his notation became invitation of the Rockefeller Foundation. The 38 rather complicated to the extent of appearing frighten­ chapters examine the possibilities of classification ing to the user. Ranganathan never realized the irony principles in communication and where they can play a that it was his notation that has turned many a poten­ useful part or where not. The book ends with an appeal tial user away from his classification. to the library profession to give up any attitude of For Ranganathan, classification was a tool of subject resistance to progress and re-thinking in the sphere of analysis and individualization rather than a marking and classification and to Unesco and other international parking device. For him, to classify meant to translate organizations and foundations to provide for research in co-extensively (faithfully) the subj ect of the book (in classification. the natural language) into an artificial language of In his Ph ilosophy of library classification (1950) ordinal numbers. And it was Ranganathan who raised written on the invitation of Dr.j ean Anker, the librarian the study of library classification to its soaring intel­ of the National Science Library of Denmark, he describes lectual heights and who invented many abstract and the history, nature and limitations of library classifica­ sometimes even inarticulate concepts in the bOJlndaries tion. Some of the chapters deal with the relation of of its theory. From a sort of an art he developed it classification with cataloguing and give guidelines for into a science. Now it is easy to teach the principles the organization of research in library classification at of library classification. When it is acquiring the in­ every level. This book laid the seed for a dynamic theory gradients of a full discipline, it is to Ranganathan, by of library classification. The latter two books are ad­ all consensus, that the major credit ought to be given. dressed to the general audience amongst librarians, All the Ranganathan reprints listed above had been linguists and philosophers. out of print since long. The new generation of Indian The Descriptive account (1967) is an outcome of library scholars and Ranganathan admirers had great explaining the salient features of the CC in detail in a difficulty in acquiring them. Most of them had to seminar on the "Intellectual organization of information" content with reading a second-hand Ranganathan. held at Rutgers University in 1964. The book is still Since he still pervades the Indian library schools curric­ important for its "Seminar panel discussion" (p.243-285) ula, the value of his books remains intact. The substance where Ranganathan answers many expert questions still lies in the Prolegomena and in the Elements. Their regarding the CC and library classification raised by availability through these reprints will keep Ranga­ many well-known librarians. nathan available and alive; and they will go a long way The philosophy enshrined in these books is in brief: in perpetuating and advancing his thoughts. Library classification should be based on dynamic Notwithstanding the utility of these bare reprints, postulates and principles. His own division of classifica­ the revised and updated editions (as Ranganathan had tion work into the three planes: the Idea Plane, the hoped and wished) were most desirable. Even these

Int. Classif. 17 (1990) No.3/4 Book Reviews 183 reprints would have multiplied their value and utility technical language for the given subject, which of had these been printed with some editor's introduction necessity was expressed at the time concerned by explaining their relevance, limitations and the purpose paraphrasing, definition-type linguistic constructions. of reprinting his books without revising them. Even the This precluded its proper indexing at the time con­ opportunity has not been availed of to correct the cerned. In such cases one may attempt, through full­ earlier misprints, something mentioned as advanced text searching in the texts stored , to round up the information in context of those times still remains relevant documents by using as search criteria the as such thus confusing the readers. Even the Colon Class expressions presumably occurring in the paraphrasings. Numbers printed on the verso of the title pages are The texts thus found may then, in a single step, be according to old editions. The Prolegomena has been re-indexed using a suitable new descriptor, mentioned as Volume 1; what the other volume(s) Nor is one obliged, in indexing with the aid of an inte­ will comprise of is not told. The second edition (1960) grated thesaurus, to use one simple specific word as a of the Elements has been reprinted, while in 1962 its descriptor. Other words with identical meaning occurring 3rd revisions edition was brought out and used widely in linguistic usage may likewise be used after having in India. Ignoring these venial omissions, the Ranga­ been admitted to the vocabulary in a suitable function. nathan scholars will gratefully appreciate the bonanza Machine-internally they will then be converted into the of his books. Thanks to the Endowment for such a corresponding descriptors. Moreover, an interactively thoughtful initiative which is a part of Ranganathan's . maintained thesaurus - unlike a printed one - is always birth centenary celebration on a large scale with a fully up to date, an advantage of great importance variety of programmes throughout the year from August particularly in joint indexing operations, such as are 1991 to August 1992 . possible when working in a network . It is not in the last M.P,Satija place due to these advances that, as the author correctly notes, the thesaurus is currently experiencing a renais­ Dr.M.P.Satija, Guru Nanak Dev University, Department of sance as a tool of information science and practice. Library and Information Science, Amritsar-143 005, India No major progress docs the author expect for the increased employment of integrated thesaurus in com­ mercial online databanks and for improved indexing. Where she does perceive possible advances owing to MEISS, Brigitte: Thesaurus, integriert, online, dyna­ integrated thesauri is the field of the designing and miseh - Neues Arbeiten mit DOMESTIC, (Thesaurus, operation of a user's own, internal documentation. integrated, online, dynamic - A new way of working (This is not surprising, for in the online data banks the with DOMESTIC). Munchen: KTS Informationssysteme procedure shun the costs of any careful indexing. The 1989. 103p. costs for such omissions are then left to the user, who An "integrated thesaurus" as described in this book initially, however, usually underestimates these costs is installed into the literature documentation program as well as the magnitude of the information loss inherent in such a way that one may have interactive access to it in working without a thesaurus. In a self-managed at all times without having to leave the program. Even project, however, both types of costs make themselves while literature analysis is going on, such a thesaurus felt, and one is responsible oneself for the information may be expanded, reorganized, reduced as well as loss occurring). amended, and after any such operation it is immedi­ Access to the contents of the book is provided by ately available again in its modified form for continued an index, albeit rather a succinct one; however, it is work. All this is shown by meanS of examples and with supported by a well�arranged table of contents. The the aid of DOMESTIC, a PC-assisted literature documen­ printing is outstandingly clear and overseeable and free tation program. of errors. Some flaws noted are the facts that at many For any subject from the field covered by the doc� points in the book the distinction between "concept>! umentation concerned the indexer or qucrier can display and "descriptor" is blurred and that among the few the relevant parts of the thesaurus on his or her screen existing rival products of DOMESTIC named in the and thus readily make an optimal selection, which is book the possibly most serious competitor, at least facilitated furthermore by the alphabetic and systematic in Germany, is not mentioned. access provided to the thesaurus. These improvements Robert Fugmann thus have the effect of increasing indexing and retrieval quality. Dr.R. Fugmann, Alte Poststr. 13, D"6270 Idstein They positively affect indexing also insofar as the integrated thesaurus offers improved possibilities of MEINL, Franz: Sachmerkmale, SchHissel zur technischen revising the indexing of documents. This is desirable Gestaltung, Beschreibung und Information, (Subject when the framework of interests of the system users characteristics. Key to their technical formation, descrip­ has changed or when a special field needs to be indexed tion and information). Ehningen : Expert Verlag 1990. more specifically. A further important, but rarely 374p. 1SBN 3-8169-0051-8 discussed problem can likewise be tackled in this connec­ tion: it is often particularly important, and not only Although written by a technician mainly for techni� in patent documentation, to retrieve the earliest doc­ cians, this book also outlines the incorporation of the uments on a given subject. However, in such an early "Subject Characteristics Classification System" into the developmental stage of a subject, no keyword suitable economic process and the technical-organizational for thesaurus use has usually been developed yet in the functions of a given enterprise. In his former function in

Int. Class;f. 17 (1990) No.3/4 184 Book Reviews German industry the author supervised the elaboration GOTTSTEIN, Klaus (Ed.); BEHRMANN, J., BOCK­ of an EDP-assisted subject information system designed WERTHMANN, W., MENACHE, G.(Comps.), Directory for a versatile assortment of goods. The experience of International Cooperation in Science and Technology. gathered and insights gained by him in this task as well Index of Institutions in the Federal Republic of Germany as in his cooperation on the DIN standard committee Cooperating with DeVeloping Countri.es. 2nd, rev.& HSachmerkmale" (Subject Characteristics) (NSM) are enlarged ed. MUnchen, etc., K.G.Saur 1989. 256 p., exploited as well as deepened and expanded in this ISBN 3-598-10626-2. A Publication of the Research book. Though mainly conceived for self-study, it can Unit Gottstein in the Max Planck Society. also be used as a teaching aid. This directory lists institutions in the Federal Republic Its central topic is the analysis of the interactive of Germany including West Berlin which cooperate with relationship between a group of objects of basically institutions in developing countries in science, technology arbitrary nature and the associated structure of char­ or letters, or which work on problems related to develop­ acteristics. This central topic is treated in an inductive ing countries. For each institution, a brief description of fashion and illustrated by significant examples whose its respective activities is given. Altogether 595 institu­ similaritics and differences �re lucidly worked out. tions are listed (p.25-200) with their addresses, descrip­ From this analysis the exact conceptual clarification tions, partner countries and partner institutions. The of the thematic field is derived. In an almost natural book is written in English to facilitate the transfer of way the author arrives at an application of Boolean information to scientists, science administrators, students algebra which he terms "Merkmalalgebra" (algebra of and government agencies in developing countries, characteristics). As his unique scientific achievement interested to establish contacts with institutions in one may regard his having brought the structuro-Iogical Germany. Of special importance are the indexes to this interrelationships to the "mathematical point". In the directory (p.201-256), numbering seven, partly estab­ chapter "Parallelen" (Parallels) he finds an elegant lished from the fields of an entry, viz.: 1) Subject index, connection to the Formal Concept Analysis according 2) Name index, 3) Index of cities in which German to R.Wille, which he expects to furnish a broad scientific institutions are located, 4) - of partner countries, S) - of foundation for the subject characteristics classification. partner institutions, 6) Name index of German institu­ This connection is then used for a comparison of product tions in English translation, and 7) Index of abbrevia­ and literature information. Both in the one and the tions and acronyms. other, however, the formalizability of the interrelation­ The Preface (p.1�24) includes also listings of the ship and thus also the applicability of EDP depends on references and a selection of additional literature, of . the limits of model formation: a rigid descriptive scheme 29 databases supplying electronically stored information, cannot encompass "reality in its entirety". Among the of 15 Federal Ministries involved in cooperation with non-quantifiable properties this is true already of the developing countries, of 16 Ministries of Regional object's own, inherent characteristics and even more so Governments involved in cooperation with developing of its relationships with the technical-economic environ­ countries, and finally of a summary of the topics of the ment as described by the relation characteristics. Also projects described in the volume: in the fields of energy, deserving of mention here is the purpose-specific inter­ health, nutrition, raw materials, communication/doc� linkage of various groups of objects as required for umentation, history/culture, and environment. expert systems. Awareness of existing limits and of the All in all, a very useful book indeed, especially for all problems still to be solved is not repressed by any means, colleagues interested in international cooperation, for it is just as indispensable for the sovereign deVelop­ communication and understanding. In 1979 it has had ment and handling of subject information systems as its first edition, however, as the editor stated, "far from the structuro-Iogical weaponry. complete". Work towards this second edition started in Although mathematical formulations make up a 1980. At the end of 1990, facing the enlargement of sizable portion of the work, the reader is not expected Germany by its Eastern part, a new situation is given, to have advance knowledge surpassing 'lower algebra'. calling also for a necessary new updating of the directory. What is expected, however, is his concentrated participa­ May it not take another decade for the collection and tion and effort to acquire an overall view. The author presentation of its data! endeavors to make reading easier by means of com­ I. Dahlberg parisons and thrown-in quotations, while it is also quite noticeably supported by three Annexes. Of these, Annex A alphabetically and systematically lists and explains the technical terms used, Annex B presents the aforementioned subject information system from the author's industrial practice, while in Annex C a realistic economy calculation is performed to show conclusively that the not inconsiderable expense of establishing a subject information system is certainly worth while if such a system is professionally and energetically develop­ ed and utilized. Josef H6lz1

Prof.Dr.}. Holzi, Institut ftir Technologie und Warenwirtschafts­ lehre, Wirtschaftsuniversitat, Augasse 2-6, A-1090 Wien, Austria

Int. Classif. 17 (1990) No.3/4 Book Reviews 185 o FORM DIVISIONS 5 ON SPECIAL OBJECTS CS (TAXONOMIES) ***

01 Bibliographies 51 On Taxonomies in the Form and Structure Area 1 199 02 Li te ra tu re Rev i ews 52 On Taxonomi es in the Energy and �la tter Area 2 03 Dictionaries, Termi nologies 187 53 On Taxonomies in the Cosmos and Earth Area 3 04 Classification Sy stems & Thesauri (CS & T) 187 54 On Taxonomies in the Bio Area 4 05 Periodicals and Serials 188 55 On Taxonomies in the' Human Area 5 06 Conference Reports, Proceedings 188 06 Un Taxonomies in the Soc io Area 6 07 Textbooks (whole field)* 57 On Taxonomies in the Econ. & &Production Area 7 08 Other t,lonographs (whole field)* 190 513 On Taxonomies in the Science Inform. Area 8 09 Standards. Guideli nes 191 59 On Taxonomies in the Humanities Area 9

1 THEORETICAL FOUNDATIONS 6 ON SPECIAL SUBJECTS CS & T

11 Order and Classification 61 On CS «& T in the FOffil and Structure Area 1 12 Conceptol ogy in Classification 191 62 On CS T in the Energy and t·latter Area 2 198 13 Hathematics in Classification 191 63 On CS & T in the Cosmos and Earth Area 3 198 14 Sy stems Theory in Classification 64 On CS & T in the Bio Area 4 199 15 Psychology and Classification 191 65 Un CS & Tin the Human Area 5 199 16 Developillent of Science and Cl assif. 192 60 On CS & T in the Socio Area 6 199 17 Problems in Classification 192 67 On CS & T in the Econ.& Production Area 7 199 18 Classification ResearchCd 192 60 On CS & T in the Science & Infoffil. Area 8 199 19 Hi story of C1 ass ifi ti on 192 69 On CS & T in the Humanities Area 9 199

2 STRUCTURE AND CONSTRUCTION OF CS & l' 7 CLASSIFICATION AND -LANGUAGE **

21 General questions of CS & T 192 71 General Problems of Natural language 199 22 E1 ements of CS & T 192 72 Semantics 200 23 Construction of CS & T 192 73 Automatic Languaye Processing 200 24 Relationships 192 74 Grammar Problems 200 25 Nume rical Taxonomy 193 75 Question-Ansltering Sy stems, Online Techniques 200 26 Notation, Codes 76 Lexicon/Dictionary Problems 201 27 Revi sion, Updating, Storage & t·la intenan(C 194 77 General Problems of Termi nology 201 28 Compati bility and Concordance bebleen 194 78 Subject-Oriented Temlinology Work 202 29 Evaluation of CS & T 194 79 Problems of �lultilingual Systems 202

3 CLASSING AND INDEXING (METHODLOGY) 8 APPLIED CLASSING AND INDEXING (C & I)

31 Theory of Classing and Indexing 194 8182 General Problems of Applied C&I 202 32 Subject Analysis 194 Data C&I 203 33 Classing and Indexi ng Techniques 195 83 Title C&1 34 Automatic Classing and Indexing 195 84 C&1 of Primary Literature (Except 85) 203 35 Manual and Automatic Ordering 195 80 l:look Index; ng 203 36 Coding 86 C&1 of Secondary Literature 203 37 Reclassification 87 C&1 of Non-Book Materi als 203 38 Index Generation and Programs 195 88 C&I in Subject Fields 203 39 Evaluation of Classing and Indexing 196 89 free

4 ON UNIVERSAL SYSTEMS 9 CLASSIFICATION ENVIRONMENT

41 On Universal CS & T in General 196 91 Organization Problems in general 42 On the Universal Decimal Classification (UDC) 196 92 Persons and Institutions in Classif.& Indexing 203 43 On the Dewey Decimal Classification (DOC) 197 93 Organisation on a National & Internat1 . Level 204 44 On the Libr.of Congr.Classif. (LCC)& LCSH 197 94 free 45 On the Bl iss Classification (BBC) go95 Education and Tra ining in Classif.& Indexing 204 46 On the Colon Classification (RCC) 197 free 47 On the Library Bibliographical Classif. (LBC/BBK) 197 97 Economic Aspects in Classif.& Indexing 48 On Other Universal Systems 198 98 User Studies 204 49 free 198 99 Standardization in Classif.& Indexing 204

'Personal Author Index 204 * Monographs on special topics at special subdivisions Divisions of 04, 5. 6, 78, 82, 88 acc. to Infonnation Cod)Jll;I ** CS & T stands for Classification Systems and Thesauri Classification (ICC) described and displayed in Int.Classif.1. C & I stands for Classing and Indexing 9(1982)2, p.87-93 and in Int.C1assif.& Indexing Bibliogr. Language abbreviations: 69 11 A further division of this scheme (by one and sometimes two Bulgarian fi Finnish p1 Po sh levels) was publ i shed in Int.Classif.12(1985)No.3, p.147-151 os Czech fr French pt Portuguese ch Chlnese hu Hungarian ro Romani an da Danish it Italian ru Russian de Gennan ja Japanese sh Serbocroa ti on There are abstracts in our holdings'to almost all of the ref­ en Eng1 i sh 1 i Lithuanian sl Slovakian erences cited. We offer to send copies of these abstracts to es Spanish n1 Dutch uk Ukrainian anybody interested. Please write to the Editorial Office.

186 Int.Classif. 17(1990)No.3/4 Classification Literature ., FORM DiVISIONS

03 Dictionaries, Terminologies

3/4-001 036.75 3/4-013 042.3-da Yakubajtis, E.A.(Ed.): Architecture, protocol') and testing of Dansk Central for Dokumentation. Danish UOC Committee: Danish open informa tion networks: An expl anatory dlctJOnary of 600 Abrid ed UDC Edition. Class 62 621 : En ineerin Sciences� terms . Orig. en-ru. Moskva. SU: Fin.; stat. 1989. 191p. c anlca an ectnca ng1neenng. ng. a. yng y, D : Dan­ -----rhe

3/4-002 036.757; 715 3/4-014 043.3 Herden. W. ; Hein. H.-W.(Eds.): Kurzlexikon Wissensbasierte Sy­ Comaromi. J.P. ; Beal l, J., Natthews, W.E.Jr., New, G.R.(Eds.): sterne. (Lexicon of knowledge-based systems) Orig.de-en. r'lUn­ Abridged Dewe� Decimal Classification and Relative Index. De­ .• •• Ny: chen,DE: Oldenbourg 1990. 96p 54 refs ISBN 3-486-21718-6 v1sed by Me lV1 1 Dewey . EdltlOn 12. Urlg.en. Alba ny, Forest Gives short and longer explanations to some 55 terms and con­ Press OClC 1990. XIII,857p. , ISBN 0-910608-42-3 cepts in systematic order. Contains a survey on the new features of Abridged Ed.12, an Introduction to the DOC, a Glossary and an Index to intro­ 3/4-003 038-68 duction and glossary. There is also a Manual introducing into Huber,R. ; Rieth, R.{Eds.): Glossarium Arti '> 7: Festun en, For­ the classing of the difference Tables and Schedula Classes, a teresses. Fortifications. Worterbuch zur unst de. r en . survey of the Policies and Procedures of the Library of Con­ �lUnchen, DE: K.G.Saur Verlag 1990. 406p. gress Decimal Class if.Div., Relocations and Reductions, and a The volume deals with the vocabulary of European mi litary en­ comparative and an equivalence table for 780 Music. Thereafter gineering after the invention of gunpo�lder demonstrated in fol low ,the tables, the Schedules and the Relative Index. relation to castles and fortresses. These are shown in pict­ ures with the vocabulary in three languages and a defi nition 3/4-015 043.6-99 in Genna n. Comaromi,J.P. et al . (Eds.): Dewey Decimal Classification: 200 Religion Class. Reprinted from ed.20 of the Dewey Decimal Clas­ 3/4-004 . 038-824 slflcatlOn wlth a revised and expanded index and manual notes Neuheuser, H.P. : Book review of Watne. P. (Ed.): Dictionary of from ed.20. Orig.en. Albany, NY: Forest Press 1989. 191p. archival tenninol ogy. Dictionnaire de termi nologie arch ivi­ stlgue. (£ngI1sh and FrenCh wlth equlvalents 1n DutCh, German. Congress Italian. Russian, and Spanish. 2nd ed. MUnchen, etc. K.G.SaurI, 044 library of Classification and Verl .19BB. 212 p.) Bibli othek. Forsch.& Praxi'> Vol 14. No Library of Congress Subject Headings 1990. p.94-96 3/4-016 044.6-667 DE 3/4-005 038-83 Runchock,R. ; Droste,K. (Eds.): Class K. Subclasses KK-KKC. Law Mikhalevich, V.S. (Ed.): A dictionary of cybernetics. 2nd ed. of Gennanh. library of Congress ClaSSlflcat10n Schedules. Com­ Orig.ru. Kiev, SU: Gl .red.Ukr.Sov.Entsiki. 1989. 751p. bined W1t additions and changes through 1988. Orlg. en. Contains 2400 terms and the theory and practice of information Detroit, MI: Gale Research Ind,. 19B9. ca 600 p. processing and management. A reprint of a document of 1982 prepared in the Subject Cata­ lqging Div. of the Library of Congress by J.E.Goldberg. 3/4-006 038-83

Dictionary of computing. Orig. ru, transl .from en. 3/4-017 .• 044.6-667 FR Moskva , SU: Mashinostroenie 1989. 567p. Runchock, R Droste, K. (Eds. ): Class K. Subclasses KJV-KVW. The handbook contains over 4000 terms used in computer techno­ Law of France. Library of con1� 88.ress Classification Schedules. logy, electronics, mathematics and logic. It was compiled by Combl ned wlthm: changes through Urlg. en. more than 50 professionals from the US and the UK. (In Eng­ Detroit. Gale Research Inc. 1989. ca 300 p. lish printed by Market House Books, Ltd.) Created by J.E.Goldberg. based on a prel iminary draft by J.Wawrzkow. 3/4-007 038-87 Frieling, W.-R. : Worterbuch der Verla�ssprache. (Dictionary of 3/4-018 044.6-668 publishers terminology) Dng.de. Berlin: .Fneling 1989. 96p. library of Congress. Subject Cataloging Oiv.: Classification. Class KJ-KKZ. law of Europe. Orig. en. 04 . Classification Systems & Thesauri Washington, DC: library of Congress 1989. XXXI,599 leaves , Prepared in the Subject Cataloging Div. by J.E.Goldberg.

041 General Classification Systems and Thesauri

3/4-008 1 041-de-63; -66 Deutsche Bib iothek; 8i b liotheksverbund Bayern (Ed.): Schlag- wortnonndatei SWD . Teilaus abe Recht und Verwaltun . Stan : Okt.1989. Standard Subject File. Section on Law and Admini­ stratl on) Orig.de. Augsburg, DE: Arb.gem.f.jurist.Bibliotheks­ u.Dokumentationswesen, Universitatsbibliothek 1990. XVII,626p. 047 Library Bibliographical Cla�sification (aBK/LBC)

3/4-009 041 3/4-020 047.6-81; -56; -58; -85; -86; -82 Atkins, Th .V.(Ed.) ; Ostrow, R. : Cross-ReferenceNy: Index. A �989uide. Bibliothekarisch-Biblio ra hische Klassifikation:1 BBK A. Ada - to search tenns. 2nd ed. Orig.en. New York. Bowker t1erte assung; systematlsc er e1 , rary 1 10- 970p., ISBN 0-8352-1918-6 graphical Classification, adapted version, systematic part) A total of 42,000 terms have been drawn from the fol lowing Orig.de. Berl in, DE: Nethod. Zentrum f.Wiss.Bibliotheken u.ln­ sources: LCSH. Sears, Reader's Guide, NY Times Index. Public formations- u. Dokumentationseinrichtungen 1989. 96p. Affairs Infonn.Serv., ERIC Thesaurus, Psychological Index The­ Vo1 .5.1 contains classes: 81 Culture. 82 Science of Science, saurus, and Subject Guide to the Information Access Databases. 84 Sports, 85 Mass Media, Journalism. 86 Publishing, Book The Index is selective and lacks cross-referencing. Trade , 87 Book and library Science, Information and Documenta­ tion, 88 Museums. Archives.

042 Universal Decimal Classification 048 .7 Patent Classification Systems 3/4-010 042.1-cs-43; -44 3/4-021 048.7 Ustredi Vedeckych Technickych a Ekonomickych Informaci:zoor: Cze­ Deutsches Patentamt; Eidgenoss.Amt f.Geistiges Eigentum; �ster­ Q9Y.choslovak Full UDC Edition. Class 57: Biology. Class 58: On g.sk. Nitra, CSFR: Ustredna podohospodarska kniznica reichi.• sches Patentamt: Internationale Patentklassifikation. 5. T9"S"9. 296p. Aufl Bde.I-9. (Internat10nal Patent Classlflcat1on, 5th ed., vahd from Ja n.I, 1990. 9 voTs.) Ori-g.de. Koln, DE: Heymanns 3/4-011 042. 1-cs-63; '-65; -16 Verla� 1989. 1516p. Ustredi Vedeckych Technickych a Ekonomickych Informaci: Cze­ The nlne vol umes cover: Daily Suppl ies (141p. ); Working proce­ choslovak Full UDC Edition, Class 35 Public AdministratiOn; dures (338p.); Chemistry, Metal lurgy (263p. ); Textiles, Paper Class I,36: CSfR: Social Welfare; Class 389:5) 1989. Metrology. Orig. cs. (67p. ); Building, Minin,9 (Hp.); Mechanical Enjlg., lights, Praha UvlEI (Konvlktska 2U3p ., Heating, Weapons, Exploslves {184p. }; PhYsics (157p. ); Electrl­ ISBN 80-212-0053-7 cal Engg.(l46p.); Introduction & Classification System (306p.)

Int.Classif. 17(1990)No.3/4 Classification Literature 187 048- Special Classification Systems and Thesauri access points in computer-based retrieval . - 3/4-033 McAl l is­ ter-Harper, D. : Dewey Decimal Classification 1n th e online 3/4-022 048-56 environment: A study of libraries in North Carolina. - 3/4-034 Hill, J.S.: Things are taking a little longer than that: A re­ ThesaurusAZ: of ERIC Descri �990tors,. 12th Edition. Orig. en. Pnoen1x, uryx Pr ess 640p . ISBN 0-89774-561-2 sponse to De�ley Decimal Classification in the online environ­ Covers over 9,500 terms in four descriptor displays. ment. - 3/4-035 Bland,R.N.: Toward the catalog as a tutorial guide to the 11terature. - 3/4-036 Summervi lle,L.: Cataloging, 3/4-023 048-627 classification, and pedagogy: Toward the catalog as a guide to Fachinformationsverbund Internationale Beziehungen: Thesaurus the literature: reaction. - 3/4-037 Wilkinson,C.L. : Intellec­ Internationale Beziehun en und Landerkunde . (Thesaurus for in­ tual level as a search enhancement in the online environment: ternational relations, and countries Orig.de. Ebenhausen ' DE: Summation and implications. - 3/4-038 Comaromi ,J.P.: Summation Stiftung Wissenschaft und Pol itik 1989. VIII,340p., of Classification as an Enhancement of Intellectual Access to ISBN 3-927674-00-1 Information in an Online Environment. Book review by C.Hafner in Nachr.Dok.41(1990)No.3, p.204 3/4-039 06.88-05-13 3/4-024 048-657 Book revi ews of Wei nberg, B.H. : Indexing: The State of Our Deutsche Sti ftung fUr Internationale Entwicklung (DSE): The­ Knowledge and the State of Our Ignorance. Proceedings of the saurus fUr wirtschaftliche und soziale Entwicklung. 1: Alpha­ 20th Annual �jee ting of the Jlmer-lcan Society of Indexers, New b�tischer Tei1, 2: S�r�kturierter Teil, 3: Register Ten . Eng- York City, 13 Nay 1988. Medford, NJ: Learned Information 1989. 11sch-Oeutsch, Franzos lsch-Oeut<;ch, Spani sch-Oeut<;ch. 12. , rev. 134p. Auf1 . (Thesaurus for economic and <;ocial development, 12th by V.J.Nurcombe, in Library Review Vol 39, No 1, 1990. p.52 rev.ed.) Orig. en-de; fr-de; e<;-de. Bonn, DE: DSE, lentrale by D.Dunn, in Spec .Librarie<; 18(1990)No.l, p.85-86 Dokumentation 1989. 274+46+144p., ISBN 3-924441-45-6 by [.B.Christianson, in Libr.Quart. 60(1990)No.l, p.85-86 Covers 2,610 descriptors and 1,218 non-descriptors structured (in Part 2) into 19 fields and 102 groups. 3/4-040 06.88-06-13/15 Chiaramel la,Y.(Ed.): 11th International Conference on Research 19BB.and Devel opment FR:in Inf ormation Retrieval198&.. Grenoble, •• June 13-15 Grenoble, Presses Un lVerslte 680p ISBN 2-7061-0309-4 The papers have been arranged under the following headings: Special session, SIGIR Award, Natural language processing, 3/4-026 048-91 Cogn itive models, Paral lel distributed proce�s, applications, Schreiner, M. (t�ch� .ed.): Fachsystematik und Einzf:"lsprachen­ Quantitative models, Thesaural models, Interface, Data bases , s stematik der Logical models, Artificial intell-igence, Set oriented models, Blbl10thek des1SC Instltut<;1 fUr All erne ine S rach­ wlssenscha t er est a en e ms- nlversltat Unster. Implementation techniques. 3. Aufl . (Subject classification and classification of lan­ 3/4-041 Sparck Jones, K. : A look back and a look forward. - guages of the Library of the Institut for General Linguistics, 3/4-042 Smea ton, A.F. ; Rij<;bergen. C.J.van: Experience on in­ corpord ting <;yntactic proC!�ss ing of user queries into a docu­ University of MUnster/W, 3rd. ed) Orig.de.•• MUnster/W, DE: Bi­ bliothek de<; Inst.f.Allg.Sprach\�iss WWU 1989. 305p. ment ofretrievdl <;trategy. - 3/4-043 8onzi. S.; Liddy. E.: The usp anaphoric resolution for document description in infor­ 05 Periodicals and Serials mation retrieval. - 3/4-044fi Antoniadis, G. ; Lal lich-Boidin G .. Pol ity, Y., Rouault, J.: French text recognition model for infor�ation retrieval system<;. - 3/4-045 Jacobs, P.S.; Rau, L. 3/4-027 056 .156 F.: Natural lan�uage techniques for intel ligent information Habel, Ch. ; retrieval . - 3/4-046 Pevzner, B.: Precedental data bases : how KanngieBer1,,S. , Strube,G. (Ed�.): Kogn; tionswissen­ schaft, Band 1, Heft 1990. (Cognition Science, Vo1 .1, No .1, and why they are worked out and used. - 3/4-047 Case, D.O. : 1990) ong.de. Re lde1berg, DE: Springer Verl . 1990. 46p. Huw do the experts do it? The use of ethnographic methods as According to the Editorial , Cognition Science is understood to an aid to understanding the cognitive proce<;sing and retrieval supply th!,!oretical and methodological basic a<;<;umptions for of large bodies of text. - 3/4-048 Belkin, N.J. : On the nature function of explanation 1n intelligent information re­ research lnto natural and computenzed cognition. The editors. and hold �hat cognit)on �akes place in processes of infor'nation trieval . - 3/4-049 Salton. G. ; Buckley, Ch.: On the use of handhng, rRsultlng 1n the fact that cogn itive processes are spreading activatlOn methods in automatic infonnation retriev­ identical with computing processes. To do cognition <;c ience al. - 3/4-050 Brachman,R.J. ; r�cGuinnes<; .O.L. : Knowledge repre­ means therefore, to con<;-Ider mental achievements and the ir sentation, connect ionism.H.: and conceptual retr-ieval. - 3/4-051 processes from the point of view of information process ing and Yu, C.T. ; t�-jzuno, Two learning schemes in informatlon re­ to understand them as computing processes or the ir result<;. trieval . - 3/4-052 Wong, S.K.M. ; Yao, Y.Y., Bollmann, P. : li­ The first number contains the follmling article<;: �Ie nder, K.F. near structureH. ,1n lnformation retrieval . - 3/4-053 Hirabayashi. F.; t�atoba, Ka<;ahara Y.: Information retrieval using im­ Wagener, M.: lur Verarbeitung rauml1.:icher Informationen: Model­ le und Experimente. - Kaufmann, Semantische und kOflzeptu­ pression of documents as a clue. - 3/4-054 Fuhr, N. ; HUther, elle Aspekte der Weg-Praposition "durch". - Brewka, G.: Bevor­ H.: Optimum probability e<;timation based on expectations. - 3/4-055 Crouch, C.J.: A cluster-based approach to thesaurus zugte Tei l.theorien: WissensrevisionH., in einem Ansatz zum Oe­ fault-SchheBen. - Mohnhaupt. Rehkamper, K. : Gedanken zu construction.• .li±:Q§l - 4 05 Harman, 0.: Towards interactive query einer neuen Theorie der Kognition. Da<; Buch von T.Winograd( expansion tl i1e tri cher,P. ; Fuhr,N., Lustig,G. , Schwant­ F.Flore<;: Erkenntnis - Maschinen - Verstehen. lur Neuge<;tal­ ner, �1 ., Knorz, G.: The automatic indexing system AIR/PHYS. - tung von Computersy<;teillen. ;;4-058 Kok, A.J.; Rolman, A.M. : Retrieval based on user be­ aVlour. - 3/4-059 Simpson, P.: Query processing in a hetero­ 06 Conference Reports, Proceedings geneous retrieval network. - 3/4-060 Brajnik, G.; Guida, G., Tasso, C.: IR-NLI II: Applying man-machine interaction and ar­ tificial intell igence concepts to information retrieval . - 3/4-028 06.86-10-22/24; 751 3/4-061 Tague, J.; Schultz. R. : Some measures and procedures for evaluation of the user interface in an information re­ lunde,P. ; Agrawal,J.C. (Ed<;.):IV: Empi rical Foundations of Infor­ mation and Software Science Emp1 rlcal I� ethods of Evalua­ trieval sy<;tem. - 3/4-061 Teskey, F.N. : Intel ligent support tlon of Man-Mach1ne Interfaces: Proc.4th Symposlum, Atianta ,GA N.:for interface system<;. - 14-063 Ganzalez-Rubio. R. ; Couprie, Oct.22-24, 1986. New York-London: Plenum Press 1987. X,533p. , A parallel multiprocessor machine dedicated to relational ISBN 0-306-42817-2 and deductive databases. - 3/4-064 Bo<;c, P.; Gailbourg, M. : Flexible selection among objects: a frameword based on fuzzy YSTh�1S; papers . cove: : User-system in�erfaces: Key issues; - Anal­ - EvaluatlOnj User-system Interaction ; Information sys­ sets. - 3/4-065 Damier, Ch. ; Defude, 8.: The document manage­

tem deSign and development; Software engineering; Software and ment component of .• a multimedia data model . - 3/4-066 Furnas, system performance evaluation; Software tran<;ition tools and G.W. ; Oeerl'lester,S Dumais,S.T., Landauer.T.K.�man,R.A., techniques; Information, kno�lledge, and value. Streeter. L.A., Lochbaum, K.E. : Information retrieval using a singular value decomposition model of latent semantic struc­ 06.88-04-8/9; 756 ture. - 3/4-067 Croft,�I.B.; Lucia, T.J., Cohen, P.R. : Retriev­ ing documents by plausible inference: A prel iminary study. - 3/4-068 Nie, J.: An outline of a general model for information retrieval systems. - 3/4-069 Schauble, P.: An information The issue contains the papers as part of the 2nd structure dealing with term dependence and polysemy. - 3/4-070 Annette Lewis Phinazee Sympo<; ium, of Library and in- Barthes ,Ch. ; Glize,P. : Planning in an "expert system for autom­ formation Sciences, North Carolina Central University, April ated infonnation retrieval . - 3/4-071 Zarri , G.P. : Conceptual 8-9, 1988, Durham, N.C. The ten papers are arrange in 5 parts: representation for knowledge bases and "intel ligent" informa­ Spel ler, B.F.Jr. : Access to Information in an Online Environ­ tion retrieval syHems. - 3/4-072 Das-Gupta, P.: Rough sets ment: Editor's Introduction. - 3/4-030 Chan, L.H. : The Library and information retrieval . - 31:-07� Bookstein,A.: Set orient­ of Congress Classification Sy stem in an online environment. - ed retrieval . - 3/4-074 Choue a, .; Fraenkel , A.S., Klein, §/4 031 Saye,J.D. : "The Library, of Con�ress Classification Sh .1.: Compression of concordances in full-text retrieval sys­ ys tem in an online environment ' A reactlon. - 3/4-032 High, tems. - 3/4-075 Jimenez , G�C.: Access by content of documents W.M. : Library of Congress Classification Numbers as subject in an office information system.

188 Int.Classif. 17( 1990)No.3/4 Classification Literature 3/4-076 06.B9-03-13/17; 771 me�tation - an information management tool for both public- and International Conference on Terminolo� Standardization and prlVate-sector.• organizations. - itCE102 Hirtz, H. ; Nistrup �lad­ Unlflcatlon. Int. Inform.Communlc.Educ. 01 8, No 2 .• 1989. sen. B Engel . G.: OANTERM/OR application. The DANTERM p.222-223 Group. - 3/4-103 Vol lnhals, 0.: TERM-PC - .•the terminology sy­ stem. - 3/4-10� Kanemore, K.; Oshikawa, Y Nakahara, S.: CJE MUlTIBUILD. A inguistic data processing system for English 3/4-077 06.89-06-26 and Oriental language'). - 3/4-105 Westerberg. K.: Why term banks on CD-ROM? - 3/4-106 Henning, J.-M.: MicroCl!zeau IV. - �/4tlO) Del Bigio, G. : CDS /ISIS. Computerized Documentation en. ys em Integrated Set of Information Systems. - 3/4-108 Stoll. Short report on this Sheila S. C.-H.: Tenninology \'Iork on PC. - Intner and which took place during the Convention of the American library AS'iociation at Dallas. The fol lowin 4 06.89-12-18/21 speakers and their papers: 3/4-077a Chan, l.�1.: Importance9 and uses of classification schedules online. - 3/4-077b Paul­ 3/4-oJ/c son. P. : Dewey Decimal Classification online. - Wil­_ liamson. N.: ,LIbrary of Congre'>s Cla'isification onl'ine. 3/4-077d- Garrison. W. : lCSH/LC name authorities online. _ A concluding panel di'>cussion treated the future needs and Issue No.n contains the 18 papers of topical area Ion which uses of onl ine subject tools in local '>ystems, moderated by a report had been published in Int.Classif.90-1. p.28-29 by E. Michael Gorman. f1ater. Al l the 18 papers are in German. Issue No.78 conatins 3/4-O78 06.89 10-3/5; 124 the 21 papers of topical area 2 on "Complex information sup­ denkO'I.S. (Ed. ply", of which only 3 papers have som!! bearing on .our field ' ��������� and which have be!'!n added at the end of the following list. ��� ��� 3/4-110 Bonitz. H.: Wissen - Information - Informatik. - ��i����� 3/4-111 �Ieber. F.: Referatestruktur al s Grundmuster der Infor­ i;! matlOnsverdichtung. - 3/4-112 Dietze, J. : Oer semantische Ab­ stand in einem Wortfeld als Grundlage der Wissensreprasenta­ 3/4-079 06.89-10-3/5; 771 tion. - 3/4-113 Hater.E.: Ziele und r�ethoden automatischer In­ Current problems of terminology: Synopses of reports presented haltserschlieBung. - r:-ll\ Scheller, B.: Hissensstrukturen conference. 3-5 Oct.1989. auf Grundlage von a ten- nformationssystemen. - 3/4-115 at a sCIent,f,cV methodoloSUgicar 1989. l;lOP . Poetzsch, E.: Tendenzen der Entwicklung und Nutzung faktogra­ Ong. ru. ladwostok. , phi scher Datenbanken - dargestellt an Beispielen aus der Bio­ technologie. - 3/4-116 GUnther, L.: Aufbau und Nutzung eines datenbankgestUtzten faktografischen Informat ion'ssystems fUr Baustoffe/Bauelemente - FADABA. - 3/4-117 Lalu�ko. N.S.; Bon­ dar, V.V.: Aufbauprinzipien fUr die faktographische Datenbasis "Technologien chemischer Produktionsprozesse". - 3 4-118 Letz. sibilities of util ization. Papers of a Col B.: Transformierung von faktografischen Daten in elne lssens­ 5-6 Oct.1989) Orig.de. Berlin. DE: Deutsches basis fUr Nuclear Data. - 3/4clJJckchmidt., L.: Bauelement. e - tut 1990. 179p .• ISBN 3-87068-890-4. OBI Materialien 90. Datenspeicher als Basis eines -Systems. 3/4-120 Gasch. All of the following 17 papers are in the German language. U.: Zum Problem der rlissensreprasentation und Inferenzoptimie­ 3/4-�11 Kelm. B.: Die SWO als Arbeitsinstrument fUr die Sach­ rung im Expertensy'>tem - Entwicklungsumgebungen. - 3/4-121 ersc leBung der Deutschen Bibliothek und die CO-R�-Ausgabe Killenberg, H.: Entwurf wissensverarbeitender Systeme. - der Deutschen Bibliographie aTs Beispiel fUr einen OPAC mit 3/4-122 ZAvada. J.: Von faktographischen Informationssystemen RSWK-Ketten. - 3/4-082 Godert, W.: Aufbereitung und Recherche zu Expertens,ystemen. - 3(, 4-123 Sejful-Hulukov, R.R.: Wissens­ von nach RSWK gebildeten Daten in der CD-R(X.l Ausgabe der Deut­ umfonuuny bel der Entwic ung geologischer Expertensysteme. - schen Bibliographie: - rr4-�8� Stephan, W.: Kooperative Datei­ 3/4-124 SchUler, W.; Bockl isch. St.: Ein lern algorithmus zurn fUhrung un? redaktlOne e u gaben der Deutschen Bibliothek. An lernen und Aktualisieren von Wissensbasen in Expertensyste­ 3/4-084 GelSelmann. F.: Aus der Arbeit der zentralen Redaktion men. - 3/4- 5 �lahl, A.: Lernalgorithmen zur automatischen Re­ des bayerischen ¥erbundes fUr die SWD. - 3/�-08� Schubert. H.­ levanzbewerF ung von Referaten. - 3/4-126 Jarosch, H.; MUller, J.: Aus der Arbelt der zentralen Redaktion unc en des bayeri­ H.-D.: Flexible SteuerungsUbergabe in rechnergestUtzten Infor­ schen Verbundes fUr dit;�WD. - 3/d-OB6 Traiser,�I.: Derzeitiger mationstechnulogien. - 3/4-127 Meyer, H. : Die Hodellmethode - St�nd der �WD-Systematlslerung un erste Oberlegl.lngen zu ihrer erkenntnistheoretisches Instrument auf dem \-lege lOr kilnstli­ Welterentwlcklung. - 3/4-087 Loth, K.: Strukturierung der chen Intelligenz. - 3/4-128 �larkscheffel. B.: Konzeption einer Schlagworter der Schl agl',lortnormdatei. - 3/4-088 Rinn. R.: 13e­ flexiblen Anfragesprache fUr ein wissensbaS'jerte'> Ingenieur­ ziehu�gen der SWD zu anderen Normdateien. - 3/4-0�9 Auer, G. : lnformationssystem. - 3/4-129 Anan'eva, T.N.: Einige theoreti­ EDV-Elnsatz und SacherschlieBung an den asterrelc lschen �/is­ sche Fragen der .Ausarbeitung sprachlicher �Iittel fUr automati­ senschaftli.chen .�i�liotheken. - 3/4-090 Hahne, H. : Oberlegun­ s ierte lnformatlonssysteme. - 3/4-130 Osokina, N. : Orientie­ gen Uber elne moghche Anwendung der RSWK sowie Mitwirkung an rung der Informationssysteme auf das Informationsprofil. der SWD seitens der Deutschen Bilcherei. - 3/4'6091 Beaujean,M.: 3/4-131 06.90-05-21/25; 42.6-6 Nutzung der DB-Dienstleistungen bei den Stadt Uchereien Hanno­ 5 .T , Bud set, 21.-25. i ver. - ib4b�921 Galsterer, 8.: StadtbUchereien DUsseldorf und Arntz, H.: "re4*F'�u�ncr0 fn" "��fCc rff �*�'P n�����r.���o the ab­ Report on this 2nd �pertext Conference of the Association for stracts of the papers to be found in Vol .2. The volume starts Computing Machinery in Pittsburgh, 5-8 Nov. 19B9 with an intro­ with the plenary session papers; the workshop papers are ar­ duction into this field of interest in informati·on science, ranged according to the program. Workshop topics include: Ge­ informatics, writing. linguistics, psychology, etc. It out­ neral issues, Algorithmic text analysis. Terminology. Kno\�l­ lines also the factors of success and gives short descr-iptions edge organization by universal and by speical systems, Thesau­ of a number of panel discussions. ,>uch as "Interchanging hy­ rus issues. Online retrieval, Hypermedia. Retrieval technolo­ pertexts". "Indexing and hypertext", "Expert systems and hy­ gie�. and Indexing. pertext". "Cognitive aspects of Designing hypertext systems". 3/4-133 Vickery,B.C.: Classificatory principles in intelligent interfaces. - 3/�-131 Hjerppe,R.: A framework for characteriz­ 3/4-099 06.89-11-22/23; 77.213 ing systems fur now edge organization: a first basis for com­ Proceedings of the International Symposium "Terminology in Ad­ parisons and evaluation. - 3/4-lh3 5 Drabenstott, K.M.: Experi­ vanced Microcomputer Appl ications". Vienna. 22-23 November ences with online catalogs ,n 1989 .. Ong.en. �ermNet News,. No 26. 1989. 45 p. e USA using a classificationH. : �,ssue contalnes an openIng speech by Pierre Lerat and the system as a subject searching tool . - 3/4-136 Albrechtsen, following papers with a list of participants at the issue end: Software concepts: Knowledge organization and the hUman inter­ ¥14-1�0 Melby, A.: Terminology - an indispensable tool for in­ face. - 3q-13� Fugmann. R.: Ununsed opportunities in indexing onna lon manangement. - 3/4-101 Schaar.M.: Terminology docu- and class] lca ion. - 3/4-138 Deffner, R.; Geiger. H.: Associ- Int.Classif. 17(1990)No.3/4 Classification Literature 189 .• ative word recognition with connectionist architecture'i generator and optimization of Minitel consultation. - 3/4-191 3/4-139 Ruge, G. ; Schwarz. Ch.: Linguistically based term asso­ Klose,G. ; luck.K.v.: The representation of knowledge in LIlOG. ciat10ns. A new semantic component for a hyperterm system. - 3�4-1f2 Tarnowski . M.: Text-understanding -H.:extracting knowl­ 3/4-14 Mustafa EThadi. W.: The contribution of termi nology to e ge rom written texts. - 3/4-193 Gerlach, Terminological the t eoretical conception• of classificatory language'i and U • reasoning in the natural language consultation system WISBER. thesaurus indexation 3/4·141 Guang-Wei, W.: Survey of term 314-194 Nol ler. J. -U. : How to communicate expert knowledge. equivalence between English and German. Japanese and Chinese. 3/ -195 Sorensen, H.S. : The use of knowledge-based frames for 3/4-142 Sukiasjan.E.: Description and analysis of the Library· terms 1n EUROTRA. - 3/4-196 Serbryakova. I.N.: Non-neutral le­ Bibliographical Clas'ilfication (SSK/lSC). - 3/4·143 Bies. W. : xical units in scientific texts and the problem of their human Die Rezeption von Regelwerken zur SacherschlieBung: Da'i Bei­ and machine-assisted human translation. - 3/4-197 Schiller,A.; spiel RSWK. (Reception of subject cataloguing code'i.•: the ex­ Steffens. P. : A two-level environment for morphol09ical des­ ample of the RSWK - Rules for Subject Catalogu ing) 3/4-144 criptions and its appl ication to problems of German inflectio­ Raj . P.: Position and organization of the subject content of nal morphology. - 3/4-198 Veda. Y. ; Ahlen, S., Kogure, K. : De­ anaesthesiology in different library classification systems. clarative control of generation using typed feature structure. 3/4-145 Craven, T. : Automatic structure modification in the 3/4-199 Buchan,R.l. : Retrospective indexing (RI) - A computer­ graphic 1 display of thesauri. - �?4VI46 Eisner. M.: New thesau­ aided indexing technique. - 3/4-200 Genuardi, M.T.: Knowledge­ rus qua i ties of ARBOR. - l.L.!::l!Z. aS11 jev, A. : Enhancement of based machine indexing from natural language text: knowledge the subject access vocabulary in an onlinp. catalogue. - base1.; design, development and lIIaintenance. - 3/ 4-201 �Iars, N.J. 3/4-e04g Vischer, J. : Das Harmonisierte System ZlJr Beze ichnung Van der Vet, P.E. : A semi-automatically generated knowl­ und ierung der Waren des intp.rnationalen Handels. (The Har­ edge base for direct answers to user questions. - 3/4-202 monized System for the naming and encoding of international LuteS-Schaab, B.: An online thesaurus for fact retrieval .- trade commodities). - �i4.1t9 Lesch. A. ; Szabo, p. : Hypermedia 3/4-203 Rogozhn ikova, T.: Individual information thesaurus: approaches. - 3/4-15 Jork und,L. : HYPERClASS. Four hypertext associative fields and a model of semantic devel opment. - •• applications of a c Passification scheme. - 3/4-151 Bjorklund. i¥4-20j Sare, F.; GUntzer, U. , Mittermeier, J JUttner, G. : l.: Problems of knowledge organization in an archival environ­ n ana ys'is of drug interactions using descriptions in system ment - KAM . - Y4-152 Hjerppe, R.: The role of classification TEGEN with regard to learning behaviour and user acceptance. - in Hypertext. ssue<; in implementing Roget's the'iauru'i as a 3/4-205 Suneeti, R. : Thesaurus in law: some problems of con­ hypertext. - 374-153 Hug, H. ; Walser. M.: Retrieval in tlw £TH structlon. - 3/4-206 Seetharama, S.: Toward'i establishing con­ database using the UDC. - 3/4-151 Vickery, A. : Knowledge orga­ cordance among medical classification systems. - 3/4-207 Vel­ nization in an intelligent tutonng system. - 3/4-155 Schopen, ho-lopes, R. : Automated access to multilingual information. - M.: Cross fi le searching of biomedical databases at DHlD!. 3/4-208 Beliaeva,L.; Piotrowski , R .. Sokolova, S.: Principles

3/4-15•.\praSher, R.G. : Index and indexing. - 3/4-157 Parames­ of 11nguistic automata and their information bases design. - waran : Chain procedure and Dewey Decimal Classification. - 3/4-209 �laes. J.-M.: The•• lntell igent dictionary project. - 3/4-1�8 livonen. M.: The impact of the indexing environment on 3/4-210 Mayer,R. ; Geer.D Ha�ne.K.-H. : On how to bring Hy per­ lnter ndexer consistency. - f/4k159 Molholt. P.: Standardizing text to termbanks. - 3/4-211 Bouillon. P. ; Tovena, l. : Word and codifying related term in S for improved information re­ formation and computational dictionaries. - 3 4-212 Kim Deok­ trieval . Bon; Hyu n-Jae 1m, Chio. K. , Kim, G. : KOCP: an exten ed DCP for Korean and Engl ish. - 3/4-213 Lemnitzer, L.: A conceptual 3/4-160 06.90-10-2/4 schema for a mult"ifunctional lexical database: a proposal .- Czap, H.; NedobHy, W. (Eds.): TKE'90: Termi noloqy and Knowl ­ 3/4-214 Merl o. J.C.; litzer, deS.: Terminological electronic edge Engineering. Vols.l&2. Proceed1 ngs 2nd Internat!onal Co n­ dictionar-ies as tool s for access to kno�lledge. - 3/41215 Pol i­ gress of Ter!mnology and Knowledge Engineering. Applic(ltions, karpov, A.A.; Kolodyazhnaya. L.: A system for comp1 atlon and 2-4 Oct. 1990, Un-ivers ity of Trier, FRG. Orig. en. Frankfurt/ analys is of electronic philological dictionaries using a per­ Main; INDEKS Ver1 .1990. VI II,682p. ISBN 3-88672-204-X, -205-8 sonal computer. - 3/4-216 Ray, J.; Nadkarni, M.: Electronic The volumes cover onl.y those papers that had been available dictionaries - an IndIan experience. - 3/4-217 Tanaka, Y.; Yo­ early enough for printlng before the conference. Some more pa­ shida, S.: Preparation of a concept dictionary. - 3/4-218 Czap, per'> were presented and wi ll be published el sewhere. On the H. : f.ol1struction and representation of concept<; in enterprises. other hand. not all of the authors of \-Ih ich the papers are 3/4-219 Shacklett, M.E.: Integrating information <;y.stems into listed in the following could afford to part-icipate. corporate strategic planning and ••plan execution. - 3/4-220 3/4-16g Dzhincharadze, A. : The role of termi nology in creating Bertrand-Gastaldy, H.i Paquin, L.C Dupuy, L.: The need for knowle ge data bases. - 3/4-162 Tjoa, A.M.; Wagner, R.: Basic infonnation and knowledge management. - 3/4-221 Eisele, H. : conceptual elements of knowledge ba,>ed systems. - 3/4-163 Les frustations du terminologue. - 3/4-222�. ; Frei both. Rauch, W. : Information sc ience in an information society. - G.: Tenni nological and lexical knowledge for computer-aided 3/4-164 Melby, A.: Terminology and technical documents : the tran'>lation and technical writing. - 3/4-213 He-in.R.G.: Hy per­ importance and limitation of forma l systems. - 3/4-165 Miike, text: indispensable infonnation systems or communicators. - S.; Amano,S. , Uchida,H. , YOkOl,T.: The ,s tructure and function 3/4-224 luckhardt,H.-O. : Automat-ic terminology support -in com­ of the EDR concept dictionary. - 3/4-166 Schaar, �1 .: f�u ltil in­ puter-aided and machine translation. - V4-2a5 Shreve, G.M.: gual information and knowledge management. - 3/4-167 Engel ,G. , Requirements analysis, empirical researc an prototyping in Picht, H. : New professional profile'i in knowledge engineering the software engineering of workstations for computer-assisted and knowledge transfer. - 3/4-168 Haas,J.: Treatment of uncer­ translation. - 3/4-2�6 Wright, S.-W.; Larson, C. : Building the tainty in social -science expert systems. - 3/4-169 Budin, G. : universal engine: mplementation of transplatform MTX. - Scientific knowledge structures. - 3/4-170 Nizhny, S. V. ; Sto­ 3/4-227 Scherf, W.G. : Computer-assisted translation. the work­ chik, A.M. : Some clear and unclear questions on searching in­ station and the translator. -// 4-22B Diamessis, S.E. : Ongoing terrelationships between encyclopedics and knowledge engineer­ efforts in Greece. - 3/4-229 1rayeva,S. : Compound tenn engin­ ing. - 3/4-171 Averbukh,K.J. : Term as a subject of study and a eering. - 3/4-230 Dilger.W. : Transfering knmlledge bases among tool for knowledge fiXation and knowledge transfer. - 3/4-172 hybrid knowledge representation <;ystems. - 3/4-231 Kukulska­ Sorensen, B.: "Level structuring" of knowledge and its applic­ Hulme, A. : Speed understanding of an unfa� domain. ation to an expert system domain with a particular view to the J/ 4-232 Osadtschuk, M.: Util ization of semantic frames in stu­ perspectives for tenninology work in general . - 3/4-173 Selov, 3/4-233 Dieguez. M.I.; Bur­ Y1ng and•• teaching tenninology. - S.D.: Terms, termability and knowledge. - 3/4-174 Citkina, F. dach,A.M Cabrera ,P.I., Garbarini,B.C.G. , Hormann.V.P. , lazo. A. : Comparative tenninology theory: problems . goals, methods, R.; A terminological microbank for Chile and latin America. - applications. - ¥4-175 Gorodetsky, B.Y.: Cognitive aspects of 3 1 234 Gal inskiU'i, Ch.: Knowledge transfer to/from countries tenninological p enomena. - 3/4-176 Grinev, S.V. : Terminology 111Lf i 1anguages ing non-European scripts. - 3/4-235 Gaudin. and knowledge theory. - 3ft-all Peschl ,N.F.: Some critical re­ F. : Sociotermi no logy and expert di scourses. - 3/4-236 Guespi n, flections on symbolic know e ge representation. - 3/4-178 Pet­ L.: Sociotenninology facing problems in standardization. - rova, LA.: Gnostic aspects of the term. - 3/4-179 Gurevitch, y/ 4-2Y Hosbayar, Che; Hasbagan: A study on l<1ongolian termino­ B.l. : The descriptive approach to visual knowledge representa­ ogy or physical concepts. - 3/4-238 Nedobity, W.: Methodolo­ tion and visual processing. - 3/4-180 Bakarsic. K. : Science gical and practical aspects of computer-ass isted knowledge citation index for the field of archeo1 ogy: hermeneutical and transfer to developing countr-ies. - 3/4-P9 Sans. J.: EOQL: infornmtics aspects of SCI based systems of regional basis. Glossary of terms used in the management 0 qual ity. - 3/4-240 3/4-181 Benn, W.: Model ling multiple paradiglri support in a Su, Wubin: China National Technical Commi ttee on Terminology KBMS. - 3/46 182 Fujiwara, R.: Intel ligent design of control for Standardization (CNTCTS). - systems com '1n1ng knowledge-based systems with conventional design methods. - 3/4-183 linster, M.; Gaines. B.: Supporting acquisition and interpretation of knowledge in a hypermedia 07 Textbooks environment. - 3/4-184 Skuce,D. ; Meyer, L.: Computer-assisted conceptual analys1s: an essential component of a terminolo­ 3/4-241 07.32 gist's workstation. - 3/4-185 Niggemann, J.: Representation of Langridge, D.W. : Subject analysis: principles•• and procedures. neuroanatomical knowledge: description language ADL. - 3/4-1d6 London, GB: Bowker-Saur 1989. VI,146p ISBN 0-408-03031-3 Otman, G.: Semantic networks in terminology: the case o� 3/4-242 07.32 3/4-117 Wallmannsberger, J.: Hypertext approaches to terminol­ Foskett, D.J.: Book. review of 3/4-241. ogica information processing. - 3/4-1:8 Endres-Niggemeyer,B.: In: J.Doc. Vol 46. No 1, 1990. p.59-61 A procedural IOOdel of abstracting. an some ideas for its im­ 3/4-243 07.32 plementation. - 31-1[9 Defrise. Ch. ; Nirenburg,S.: A language Hannabuss. St. : Book. review of 3/4-241. for representing ex meaning. - 3/4-190 Bu10t, T. : Dialogue In: library Rev. Vol 39, No 1. 1990. p.53

190 Int.Classif. 17(1990)No.3/4 Classification Literature 08 Other Monographs 3/4-255 124 �lurashov. V.K. : Towards a quantitative assessment of knowledge

3/4-244 OB.l representation. Orig..• ru. Paa, Niranda Lee: Concepts of information retrieval. Englewood, Na uch.-iekhn .lnform Ser.l, 1989. 19p., 7 refs. CO: Libraries Unlimited 1989. XVI+285p. , ISBN 0-87287-405-2 Knowl edge-frames are represented by a family of sets. In such a system the fol lowing is possible: to compare incoming mes­ 3/4-245 08.1 sages by content; for one and the same elements to fonn dif­ Soergel . D. ; Book review of 3/4-244. ferent parts of the system, to form hierarchies by the degree In: library Quarterly Vol 60, No I, 1990. p.83-84 of generalisation. Input of messages into the system and ag­ gregation of new parts are expressed quantitatively by a change in estimate D for finite sets D=Nln Nn, where n is the THEORETICAL FOUNDATIONS number of elements in the system and N is the number of sets. 1. �(Author, abbr. ) 3/4-256 124 115 General Order of Objects, Actions, etc. 1,1.: See 3/4-072 UenlJ, K. ; Ishizuka, Knowledge representation 1989.and util isa­ tion. Ori9 .ru (transl .from Ja panese). MOSKva : M' r 22Up . 3/4-246 119; 167 Atranslation of the second vol ume of a lO-volume series on Davies. R.: The creation of new knowledge by information re­ knO\�ledge engineering written by prominent Japanese �pecial­ trieval and claSSlhcatlOn. On g.en. J.Doc. vol 46, No 4, 1989 ists in the field of AI. p.273-301. 90 refs. The paper revierls previous work on producing knowl!'!dge by in­ 128 conce\ts 3/4-270in certain subject fields formation retrieval or classification and describe� techniques See a so by which hidden kno�lledge may be retrieved, e.g. serendipity in browsing, use of appropriate search strategies and, possib­ 3/4-258 128-72 ly in the future. methods ba'ied on Farradane's relational in­ Solov'eva, E.A. ; r�atorin, S.L: Model ling of conceptual knowl­ dexing or artificial intell igence. edge: a systemic bionic.• approach. un g. rU. iranS1.1nio en. Nil uch.-tekhn .lnform Ser.2, Vo l 23, No 3, 1989. p.2-8, 36 refs. 12 Conceptology in Classification 3/4-259 128-72 Solov'eva. E.A. ; �latorin. S.I.: fllodelling methods and models See also 3/4-110, -162, -176 of conceptual knowledge. Orig. ru, iransl.lnto en. Nauch.-tekhn .lntorrri��-ser.2. Vol 23, No 4,1989. p.2-8. 38 refs. 3/4-247 122.173 Shaw, M.L.G. ; Gaines, B.R.: Comparing conceptual structures. 3/4-260 H.: 128-82 Consensus, confl ict, correspondenceI, and contrast.341-363. Orig. en. Marloth, Book review of Folbert. O.G. . Hackl, C.: Der In­ Knowledge Acqu,s,t,on vo l No 4, 1989. 21 refs. fonnationsbegriff in Wissenschatt und lechnik. Munchen-\�ien: The paper outlines a methodology for eliciting and recognizing R.Olden60urg 1986. 299p. (The information conceptI, in science the individual differences in the use of terms by different and technology) Orig.de. Nachr.Dok. 41(1990)No p.60-6l authors. This can be used to focus discuss ion between experts on those differences between them which require resolution, 13 Mathematics in Classification enabling them to classify them in terms of differing termi nol­ ogies, levels of abstraction. disagreements, and so on. The methodology promotes the full exploration of the conceptual See also 3/4-059, -064, -072, -253. -255 framework of a domain of expertise by encouraging experts to operate in a Irbra in-stormi ng" mode as a group, using differing 3/4-261 133 viewpoints to develop a rich framework. (Author, abbr.) Santdmarina,J.C.; Chameau. J.L. : Fuzz¥ wi ndows and classifica­ tion systems. Orig.en. Int.J.f.lan-Machlne Studles Vol 32, No 2, 1990. p.181-201, 8 refs. 123 Concept construction A fuzzy sets based structure for classification systems is See also 3/4-263, -451 proposed. The basic idea is to use "window�tI to represent the constraints on the possible value., variables may take. The 3/4-248 123; 154 formal ism is very simple, however, this simplicity makes it Vojshvil lo.E.K.! Concept as a form of thought: Logico-episiem: attractive in the development of kno�lledge based systems. The 0109ical analyses. Orig.ru. Moskva, SU: MGU 19B9. 239p. most salient features of this structure include the possibili­ ty of developin!:! composite solutions, searching for lacunae. 1232 On definitions and creaUng a case based representation of knowledge with an See also 3/4-048, ,-184 avenue for model ling learning. (Author)

3/4-249 1232 15 Psychology and Classification Petrov, Yu.A. : The scientific-practical importance of the ef- fectivisation of definitions. Orig. ru. Logico-filos.issled., No 1. 1989. p.38-46 See also 3/4-259. -457 124 Knowledge representation 3/4-262 153 See also 3/4-050. 07B, -112, -179, -191. -326 Neill. S.D.: Body Engl ish: The dilemma of the physical in the ob"ectification of sub"ectlve knowled e structures: Ihe role 3/4-250 124 of the body in thinking. J.Doc. 46 1990 No 1. p.1-15. 42 refs. Font, J.M.: =L �e,",S '�'9�r�a�phFie Sc -;dneT.c�o�nnc�e �p�t". (The concept graphs) Evidence is presented to show that the body plays a continu­ Orig.fr. Bul LCent.hautes� eiud. lnt.ln",torm.doc., No 35, 1989. ing and fundamental part in the process of thinking and the p.11l-1l6 development of the knowledge used to think. The conclusion is that, if this is true. then there is a crucial part of our 3/4-251 124 subjective knwoeldge base which cannot be represented in an Bobrov. A. I.: A knowledge representation languageI for develop­ objectified scheme, whether for artificial intell igence or ing intelligent systems. on g.ru. In: Metody sredsiva sozda­ information retrieval . and therefore we cannot expect to get niya sovrem.avtomatlzlr.inf.-poisk.sistem: Tez.dok1 .2 region. a computer to think as a human being thinks. (Author) nauch.prakt.konf. , Penn', 12 Okt. 1989. Perm.SU 1989. p. lO-ll 3/4-263 153; 123 3/4-252 124 Robertson.W.C.: Detection of cognitive structure with protocol data: Carlucci . A.L.; Lenzerini. M., Nardi, D.: Knowl edge represen­ predicting pe rformance of PhYS 1CS transfer problems..• tation in AI systems. Orig.it. Riv.iriform. Vol 19, No 2, 1989. Cognitive Science Vol 14. 1990. p.253-280. 20 refs Appendix p.93-110. 19 rets . The article presents a cognitive map proposed to be associated with understanding of the "system concept", one component of the physics principle of Newton's second law. A definition of 3/4-253 •• 124; 132 Castellanos, P.J.B. ; Gonzalo.M.R Patos,S.J., Rios,C.: � the concept is fol lowed by the results of a problem-solving ebraic structured model of formal knowled e representation experiment designed to investigate whether or not good problem frames and rules Int.J. omput. at . 0 0 solvers possess cognitive structures similar to the one pro­ p.I-9. 9 refs. posed. Think-aloud protocols were col lected as subjects solved a series of physiCS problems involving Newton's second law. 3/4-254 124 Coding schemes were used to analyze these protocols and to Laishchev. LV.; Shekhovtsov, O. I.: Data manipulation on the develop a quantitative index intended to reflect the extent to 5 which subjects possessed certa in components of the proposed semantic su erstructure1 of a factual ty e IR �tem. Orig. rU. n: eto y sredstva sozdaniya sovrem.avtomatizir. inform.­ structure. This index proved a highly significant predictor of poisk.sistem: Tez.dokl.2 region.nauch.-prakt.konf., Perm', 12 perfonnance on Ne�ltGn 's second law transfer problems contained Okt. 1989. Penn, SU 1989. p.29 in a written exam. (Author. abbr.)

Int.Classif. 17(1990)No.3/4 Classification literatUre 191 3/4-264 153 ment capture system are examined . Two examples of bar code use Samuels. A.R.: Information clustering and problem solving. are provided: trucking industry applications and accounts pay­ Libr.Trends Vol 38, No 2, 1989 . p.157-170. 39 refs. able applications. (Author. abbr.) People absorb information holistically rather than separately from different levels ranging from the smal lest to the highest 19 History of Classification visible, from the subconscious mind to the written record. A major problem in the transfer of information lies in recogniz­ See also 3/4-514 ing that matching information dissemination techniques at all organizational levels to the characteristic ways people absorb 3/4-276 196 information is critical in ensuring the success of providing Baudoin de Courtenay, LA. : On the artificial languaae (1905) . the right infol1l1a tion at the right time. (Author) Ori9 .ru. In: Probl .struktur.lingv., 1985-1987. Moskva, SU 1989. p.379-385. 13 refs. 3/4-265 153 Reprint of an article by the author from a journal called "Es­ Va�ilenko. L.r. (Ed.): Problems in ehil0S0�hiC methodology. peranto" in which he affirms the legitimacy of existence of Ong.ru. Moskva, SU; AN SSR. Inst.flios. 189. IJ3p. "constructed ". or "philosophic" languages, based on a logical Collection of papers. classification of human ideas or notions and on designating them by appropriate symbol s, next to natural and international 154 concert formation languages. Listed are several scientists who accepted the idea See a so 3/4-248 of an artificial lan9uage and suggested their projects of such languages. such as Descartes. Dalgarno. Wilkins, Leihniz, etc. 3/4-266 154 3/4-277 198-825; 43 Markman, E.M.: Constraints children place on word meanings. 1 Cognitive Science Vol 14. 1990. p.5/-II, 39 refS. Lamonde, Y. : Une classification universel le our 'etude des bibliothegues et e la 11 ralrle au XIXieme slec e. A general 156 Cognition and New Knowledge . Knowledge Acquisition classlflCatlOn for the study of 19th century libraries and See also 3/4-027. -047, -175 book sellers) Orig.fr. Doc.et bibl . 35(1989)No 2, p.53-58, 12 refs. 3/4-267 156 The author suggests the use of the 1876 edition of Dewey for �Shapiro, E.L.: Knowledge of the world and its information mo­ the classification of 19th century printed materials and pro­ Orig.ru. Na uch.-tekhn.lnform.,Ser.2, No 9, 1989. p.z-4, poses an outline of a general classification for historians 16 refs. studying pr-jnted materials during the 'course of a comparative study of libraries and vendors in the 19th century. 3/4-268 156; 872

Tufte. E.R. : Envisioning information. Or.•ig. · en. Cheshire, CT: Graphlcs Press 1990. 126p $ 48.- STRUCTURE AND The book is a sort of a "Huseum of Cognitive Art", containinQ 2 with its more than 400 illustradons examples for the "idea CONSTRUCTION OF CS T of visual thinking" and provides "practical advice about how &; to explain complex material by visual means". Its extraordi­ 21 General questions of CS & T nary examples illustrate the fundamental principles of 'infor­ mation display. See also 3/4-134, -150, -202. -203 157 Computerization of Thought Processes See also 3/4-127. -451 3/4-278 212 Jodasova.p. ; Samek.T.: The termi31(1nolo§ � of retrieval2, languages. 3/4-269 157 Orig.cs. Ceskoslovenska Inform. 9)No p.45-46. 3 refs. Ar�tificial intelligence vis-a-vis knowled�e and thought: Cogi- Orlg .ru. Tartu, sO: Onwersltet 1 89. z40p . 213 Computerization of Classification Sys tems Se e also 3/4-429 . -448. -474 16 Development of Science and Classification 3/4-279 213; 756 Pitternick, A.B.: VocabuA.(Ed.):laries for online subject sea hin ' See also 3/4-246. -437 Orig.en. In: Kent. Encyclopedla of Llbr.& In Vorm. Sci. Vol 45, 1990. p.339-420, 44 refs. 3/4-270 161. 128-82 "Precontrol" and "postcontrol" are discussed with relation to 8onHz. M.: Information - Kno�'l edle - Informatics. Orig.en. vocabularies, the first exami ning subject authority systems Int.Forum Inform.& . Doc. 15(1990 No 2, p. 3-7. 21 refs. and their use in online searching, the last analyzes the pro­ (English translation of 3/4-110) cess of control ling vocabulary online. The author discusses the view of informatics. proposed by Yu. A.Shreider, as the scientific discipline concerned with the 3/4-280 214 interaction between knowledge and information. (Abstr.abbr.) Bara. 1u .: Problems of modelling and tenninological control in documentary lnformatlon retrleval systems. A reexamination of 3/4-271 161 the "thesaurus" concept. Orig. ro, en. Budd, J.M. : Research in the two cultures: The nature of schol­ Probl .lnform.sl Doc. Vo t 23, No 3, 1989 arship in SClence and the humanities. Orig. en. Collect.Manag. Vol 11. No 3-4. 1989 . p.I-21, 54 refs. 3/4-281 214 t.!!.i.:.Pozhariskij ,LF. : On the lexical com�OSition of8.1989. the IR thesau­ 3/4-272 165 Orig .ru. Nauch.-tekhn .lnform; . er.2, No p.15-24. Avaliani. S.Sh.: The nature of knowledge and of value.- 26 refs. Orig.ru. Tbilis;. SU: Metsniereba 1989. 182p. An IR thesaurus rich in phrasal units is less capable of inte­ gration (compatibilHy) and in addition impedes external data 3/4-273 165 base searching. Hence the conclusion. that there -is no point Ginetsinskij.V.I.: Knowledge as a didactic category: An exper­ in resolving homonYmi' in thesaurus terms by means of scope ience of didactic cogltology � ur lg. ru. notes; this operation is effectively carried out by the joint [enlngrad, sO: IZd-VO LGU lY89. 144p. action of the subject heading list and the semantic environ­ ment of the query descriptors. The capabilities of al l the software packages on which the IR system 1ingware tools can 17 Problems in Classification interact today make it possible to set the goal of develop­ ing highly effective IR thesauri which consist mainly of one­ See also 3/4-247. -362. -431, -474, -475 word terms. The paper oU,tlines the capabilHies for compound word conversion which are opened up during the automatic in­ 3/4-274 179 dexing of geological abstracts. (Author) Kimura, T. ; Yoshida, S. : Present state of sUb�ect catalogu ing and trends in subject headlng research: from t e flndlngs of a 215 Characteristics and kinds of thesauri study group. Orig.ja. Toshokan Zasshi Vol 83. No 4, 1989. See also 3/4-145 P.197-199. 9 refs. 3/4-282 215 Dietze. J.: Thesaurusl exik zur Wissensrepriisentation. (Thesau­ 18 Classification Research rus vocabulary for knowledge representation) Orig.de. Wiss.Z. Univ.Halle-Wittenberg. G. Vol 38, No 4, 1989. p.99-103 3/4-275 183; 313 Noirjean,R. : Bar code indexing - clearing the inpu tting bottle­ 3/4-283 215 neck for improved document management. Orig.en. Int.J.I�icro­ Dietze. J.: Die semantische Struktur der Thesauruslexik. graphlcs and Optlcat technot. vot I, No 3, 1989. p. 103-106 (Semantic structure of a thesaurus vocabu lary) Orig.de. The use of a bar code to identify and track an item in a docu- Wi ss.Z.Univ.Hal1e-Wittenberg. G. Vol 37, No 6, 1988. p.120-123

192 Int.ClassH. 17(1990)No. 3/4 Classification Literature 3/4-284 218; 984 3/4-295 233; 251 ; 332 Berkovich, E.M. ; Ron 'kin. M.L: Two-'>tage models of efficient Komonov, B.P.: A model for the multi-aspect description of ob­ or anization of subscriber service,> at computerlzed lnforma- �ects with I,the 1989. help of26 classes.3I 8 Orig.ru. Nauch.-tekhn .lnform .. tion centers. ng.ru. trans .lnto en. Nauch.-tekhn. inform. , er.2, No p. - , refs. Ser.2 Vol 23, No 4, 1989. p. 18-22 The model is based on the idea that the concept of Iclas,> ' is wider that that of a 'set ' . Any set is a class. but a class 22 Elements of CS & T need not be a set. Special composite classes and sub-classes are introduced to represent imprecise (or fuzzy) sets. Opera­ tions with composite classes are subject to multi-valued logic See al'>o 3/4-274 working on the '>dOle principle a'> operations with standard sets under binary logic. The possibility exists of extending Bool­ 3/4-285 222 ean al gebra and the field of sets in operations with composite Ri �ko .A. : Linguistic interpretation of 3,subject headings. Orig. classes which could also be useful in description of complex sk. Kni2. p. vedecke Inform. 2I(1989)No p.99-103, 6 refs. or imprecise properties of objects. (Author)

3/4-287 224; 254 3/4-296 I.: 233; 153 Zenovich. S.V.: A tabular-graphic clas'>ification model and its Solov'eva, E.A. ; f�atorin. S. Nethods of model ling and con­ appl ication in computerized information systems. Ong.ru. �eptual knowledge mod38els. Or"ig.ru. Nauch. -tekhn .1nfonn. ,Ser.2, Nauch.-teknn .lntorm. ,Ser.2, No I. 1989. p.Y-I2, 4 refs. No 4. 1989. p.2-8. ref,>. 225 Facets and Faceted Classification 234 Linguistic Support of Descriptor Languages See also 3/4-293. -317. -444. -446, -456 See also 3/4- 140 225 3/4-297 234 Buchan. R.L. : Intertwining thesauri and dictionaries. Orig.en. Infonn.Serv.& Use Vol 9, No 3, 1989. p. 171-175. 17 refs. Suggestion to combine the functions of a dictionary and a the­ saurus in a single toul . 3/4-289 225; 382; 388-26 Sana. Hikomaro: Facet tabulatlOn of index terms. Inform. Pro­ 3/4-298 V. I.: 234; 6376 cess.& Management Vol 26, No 4. 1990. p.543-548, 15 refs. Koroleva. Lingu lst1c eroblems of computerised data re­ A tr ial was made to display index tenus il) a table form with trieval systems, with speclal reterenCf> to oil and gas pro­ the aid of a computer. The target data are articulated subject �pecting economy. Orig.ru. In: Razvitie otra,>l .sistenlY NTI v indexes in which tellllS are linked with prepositions and commas usioviyakh poln.khoz.rascheta i samofinansir. Moskva, SU to form sentence-like phrases. A simple information extraction 1989. p.69-77. 10 refs. algorithm was devised by utilizing the highly structured syn­ t�ctic structures; no word dictionary is required. Preposi­ 236 Construction of Thesauri for Special Purposes tIOns and commas Me used as delimiters for extracting index See also 3/4-083. -084. -085 terms pertinent to the essential facets of a subject. The ap­ proach was appl ied to index modifications in the general sub­ 3/4-299 236 ject index to Chemical Abstracts. The modifications were pro­ Sol 'nitseva ,O.I3.: On the development of a subject heading list cessed on a personal computer laden with LISP programs for in­ for reference/i infonnatlon--servlce of-adademlc lnsbtubon formation extraction and table format display. Facet tabula­ �pecia 1 sts. Orig. rlJ . In: Intensif. i spo 1 'z .bi b1 iogr. resursov. tion of index terms not only facilitates r'l.pid scanning but Leningrad. SU 1989. p. 123-132. 4 refs. provides also document arrangements convenient for later re­ ference or review. In addition. it may help a reader quickly 237 COIHPuter-su3/ �ported-055 construction of CS & T understand index phrases. (Author. abbr.) See also

3/4-290 227; 754 24 Relationships Kharin,N.P. :, An out ut rankin method which reco nise'> automa­

tically speci led a�socla.• bve re at IOns e ween erms._ ng.ru. Nauch.-tekhn.inform Ser.2, No 9, 1989. p.19-23, 20 refs. See also 3/4-159, -456 A techn'ique of ranking retrieved items 11as developed for docu­ ment retrieval systems us-ing 800lean or Boolean-contextual 246 'lie ; ghts search logic�. The technique recognizes. dutomatically e'>tab1- See also 3/4-049, -051. -052 ished associative relations between terms in documents. The 3/4-301 ranking technique has been shown experimentally to lead to 5. ; 246 significant improvements in retrieval performance. (Author) Al -HaHamdeh, �Ii l)ett, P. : ofComp arison of index10 tenn weight­ ing schemes for the ranking Llbrpara. �raphs ful1, I-text2. doc1989.u­ ments. Ur lg.en. Int.J. lnform.& es. Vo l No 23 Construction of CS & T p.116-130. 22 refS. The paper discusses we-ighting scheme� that can be used to dis­ See also 3/4-203, -259, -280, -440 criminate between indexing terms in an experimental retrieval system for the ranking of the paragraphs comprising a full­ 3/4-291 231 text document. Nine different we ighting schems are compared Oka. Chihorni ; Kobayashi ,Y., Tokuyama ,G.: Deduction and induct­ using 20 full-text documents for which there are 130 associat­ ion in classification. Orig.ja. ULIS: Res . Report of Univ.of ed queries and relevance judgements. (Author. abbr.) Llbr. & Inform.scl. VO) 8, No 1, 1989. p.65-79, 17 refs. Discusses class ification schemes as they relate to anmia1s. 3/4-302 246 plants and books. Distinguishes analytic/deductive classifica­ TanigUChi, Sh.: Extensionby of the decision-theoretic model of tion from synthetic/indexing classification and examines information retneval fuzzy probabl Itt,,-. Ong.Ja. OL IS: existing schemes in terms of these 2 approaches. (Author) Res.Report of Univ.of Libr.& Inform.Sci. Vol 8. No 1, 1989. p.31-4/. 14 refs. 3/4-292 231.02 Proposes extensions of the decision-theoretical model using Svenonius, E.: DeSi%n of control led vocabularies. Orig. en. probabil ity measures of fuzzy events. The extensions here are Encyclopedia at [1 r.& Inform.Scl. Vo l 45. Suppl .lO. 1990. incorporations of index tenn wp. ights. mul tilevel relevance p.82-109, 96 refs. judgements. and inter-relationships between index terms. The term wei9hts, indicating relative importance of the terms in 232 Desi n rinciples for Classification Systems individual documents, are useful to estimate more exactly a ee a so power of a term to d i scrim; nate relevant and non-relevant documents. (Author, abbr.) 3/4-293 232; 225 Nesterov. A.V.: A document-information classification: a cate­ 247 Rel ational Data Files go rial-facet approach - the bas1c verSlon. Or lg.ru. See also 3/4-063 Preprint AN SSSR. SO GPNT8. No 4, 1989. 22p. , 13 refs. 3/4-303 247 3/4-294 233 Worboys. M. : Relational databa,>es: A theoretical primer. In­ Di Battista. G.; Kangassalo, H. , Tamassia, R. : Definition li­ form.& SoftwareTechii-QT-:--JT[I9B9)No 3, p.1l5-122. 14 refs. braries for conceptual model lin!!.. Orig.en. As a result of their tight and simple formal underpinning. re­ Data & Knowledge Engg. Vol 4. No 3, 1989. p.245-260. 17 refs. lational techniques are suitable to application to a wide Building a conceptual schema is one of the mo'>t difficult range of systems. Within the database community. relational tasks in AI application'> and therefore it is proposed to use a databases have become the domi nant type. There are now several 1 ibrary of concepts to facil Hate this ta'>k. Such a concept comprehensive product'> on the market that meet many of the library is a database with descriptions of concepts in a com­ requirements of a truly relational system. and relatlOnal me­ mon language. thods appear to be ever more extensively used. Relational

Int.Classif. 17(1990)No.3/4 Classification Literature 193 techniques al so have the advantage that they interlock well 29 Evaluation of CS & T with standard design methodologies and to newer ideas, such as object model ling approaches. (Author, abbr.) See also 3/4-486

25 Numerical Ta xonomy 3/4-317 292; 225 Ramachandran, M.: Citation1, order1989. in CC, UDC, and DC. Orig.en. Int.Libr.�lov. Vol Tr,""" No p.46-48 See also 3/4-295, -359 3/4-318 294 3/4-304 251 Lancaster, F.W. : The pe rspective: natural language versus con­ Di Nol a, A.; Pedrycz, W., Sessa, S., Sanchez, E.: Designing of trolled lan ua e. A new exa mination. In: Oppenheim,C. , Citro­ classification procedures with the use of equal ity and differ­' en, C.L., rl hths, J. . Eds. : Perspectives in information ence operators. Pattern Recognition Vol 23, No 8, 1990, management; Vol .l. London, G8: Butter�lOrths 1989. p.1-23 p.911-917, II refs. Tries to determine when natural language is likely to be pre­ ferred over the use of a controlled vocabulary, and vice 3/4-305 253 versa. Gupta, L. ; Sayeh, M.R., Tammana, R.: A neural network approach to robust shape classification. Orig. en. 3/4-319 295 Pattern Recognition Vol 23, No 6, 1990. p.563-568, 11 refs. Kralova, A .. Preparation of coordinatlon tables of information retrieval lanquaQes us.ng the aALEX data base. Ong. sk. 3/4-306 253 Kni2nice a Vedecke Inform. 21{1989)No 3, p.104-108, 6 refs. Mueni k istics for the dis- The Centre for Retrieval Languages of the Institute for Scien­ tific, Technical and Economic Information carried out an eva­ luation of infonnation retrieval languages using the SEMAN me­ 10 thod and the BALEX data base �Ihich contains about 190,000 le­ xical unHs. It is classed according to the UDC, the Union In­ 3/4-307 253 dustrial Classification system and the International Patent Li Wang; Dong-Chen He: Texture classification using texture Classification system. The SEMAN method has shown to be a spectrum. Pattern Recognlbon 23(I990lNo S, p.90S-9IO, IIrets . suitable method for the comparison of such systems. 3/4-308 254 Rybina, LA.: Partition systems qenerated by a hlerarchy. CLASSING AND INDEXING Orig.ru, transl.lnto en. Nauchn .-tekhn . Inform. ,Ser.2 Vol 23, 3 No 2, 1989. p.27-32, 13 refs. 31 Theory of Classing and Indexing 3/4-309 254 Shlien,S.: Multiele binar2� decisionI, 1990. tree classifiers. Orig.en. See also 3/4-137, -156, -275 Pattern Recogfl"l tlon Vol , No p.151-163, 32 refs. 3/4-320 311 3/4-310 258-51/4 l�uel ler,J. : Indexing and abstracting. In: North Atlantic Trea­ Brown, M.L. : Fitting and interpreting loqlinear interactions ty Organ izat'ion. Advisory Group for Aerospace Research and De­ in cross-classifications from health POl iCi and medicine. velopment: The organi<>ation and functions of documentation and un g.en. Meth. Inform.Med . Vol 29, 1990. p.I04- 12, 23 refs. information centers in defence and aerospace environments. Neuilly sur Seine, FR 1989. p.6.1-6.8, 9 refs. 3/4-311 258-835 Introductory article on the topic for the Technical Informa­ Teory,T.J.; Wei Guangping, Bolton,D.L., Koenig, J.A. : ER model tion Panel wh ich met in Athens, GR, 19-20 Oct. 1988. clusterin as an aid for user communication and documenratlOi1 1n ata ase deslqn. Commun.ACM 32 1989 No 8, p.975, 21 re s. 3/4-321 313; 934 Varfolomeeva, M.V. (Ed.): Scientific catalogu ing and catalogue 3/4-312 258-849 manage1nent: Collected papers. Orig.ru. Moskva, SU: Bibl . sov.,po Amma r, M. ; Yoshida, Yu ., Fukulliura, T. : Structural description estestv . naukam pri Prez idiume AN SSSR 1988. 196p. and classHication of signature images .lrrT9. en. Different aspects of central ized cataloging are considered Pattern Recognition Vol 23, No 7. 1990. p.697-71O such as: Centralized description, classification, termi nologi­ cal problem s of document title translation for the sY'itematic 3/4-313 258-91 file of the national union catalogues. Delcourt,Ch.: Where have al l the key-words gone? Computers & Humanities Vol 23, No 4-5, 1989. p.285-291, 14 refs. 3/4-322 313 It is argued that statistical model ling seems as val id in sty­ Taya ,H. ; t·larsushima,F., Sairaiji, H. , Takag i, H., Nakamura ,N. : listics as in any other field of the humanities and social J-BISC system as a tool for 'information retrieval of Japanese sciences. books. Indexinq, Part 1. Ori 9 .ja. Kagaku gijutsu bunken sab'i­ Su-Scl.(Technol .lnform.Serv.), No 9, 1989. p.4-13 27 Revision, Updating, & Maintenance 32 Subject A.nalysis See 3/4-357 See also 3/4-043, -082, -111, -242, -243 28 Compatibility and Concordance 3/4-324 3231: Riesthuis,G.J.A. ; Stuurman, P.: Trends in subject indexing, Contents 21,anal�sis.0 6, Orig. nl . See also 3/4-074. -206. -418 Open vol 1989. p.214-219, 21 refs. Much attention has been paid to translating the contents of 286 documents into indexing language terminology, however, less thought has been given to identify the meaning of documents. Al so, too little is known about the process by which a reader absorbs information. The Kyoto Universi Department Library decided in 1987 to change from own class ification sy.�tem to the 3/4-325 323 Nippon Decimal Classification (Nihon Jushin Bunruihyo), which Dykstra, N.: Subject analB, ysiS and thesauri : a background. is used in 80% of the Japanese university libraries. Pre-1986 Orig.en. Art Doc., No 989. p.173-4 books were excluded except for those from continlling series. Books were labelled in different colours. 3/4-315 287 Kristal 'nyj, B.V.: A unifonn system of rubr-icators lingustic basis for the network wlthin GASNTI. Orlg. ru. Nauch.-tekhn.infonn.,Ser.1, No 7, 1989. p.13-15 3/4-316 287 Riesthuis,G.J.A. ; Stuunnan,P.: Trends in subject analysis (3): coordinate indexing languages. Orig. nl. Open vol 22, No I, 1990 . p. 11- 15, 24 refs. 3/4-327 324.213 The authors conclude that the ideal indexing system is based Nohler, P.P. ; ZUl l, C., Geis, A.: Die Zukunft der computerun­ on a coordinate vocabulary which allows in retrieval both, the terstUtzten Inhaltsanal se cui . (The future of computer-aid- utilization of index terms and the notations of classification ed contents ana YS1S r1g. e. systems. ZUMA-Nachr., No 25, 1989. p.39-46, 15 refs.

194 Int.Classif. 17(199D)No.3/4 Classification Literature 33 Classing and Indexing Te chniques Hill National Center for Biomedical Communications {LHNCBC}, a research division of the National Library of Medicine (NlM). See also 3/4-295 The system uses an experimental frame language featuring in­ heritance and procedural attachments, and is designed to in­ 3/4-328 331 teract with trained MEOLINE (R) indexers by prompting them to Yu Tao: The general appl icability and specificity of documen­ enter MeSH (R) (Medical Subject Headings) terms as slot fil­ tation rules for indexing. Orig. ch. lers in completing document specific indexing frames derived Tu shuguanxue longx�n. No L 1987. p.61-63, 1 ref. from knowledge base frames corresponding to MeSH concepts. (Author, abbr.) 3/4-329 331.02 Hill. J.S.: Cataloging and classification. Orig. en. 3/4-340 348 In: ALA Yearbook of library and Information Services, Vol 14. Knowled�e-based indexin& g <;ystem. Orig. en. Chicago, ll: Amer.libr.As'ioc. 1989. p.86-7 Inform. etnevar [,br.Autom. Vol 26, No 1, 1990. p.8 Infol11lation on the indexing system employed at Reuters for the 3/4-330 334 Reuter' Country Reports, Reuter Textline and other texts from Khanshin. A.G. : A theory for coordinate indexing: an essaa, its Historical Information Products Division. Orig.ru. Nauch.-tekhn.1nfonn" Ser,z. No la, 1989. p. 19-2 , 10 refs. 3/4-341 348 Ostapenko, V.E.: Term identification and classification by means of elementary quantitative models. Ong. ru. 34 Automatic Classing and Indexing Nauch.-tekh.inform.,Ser.2, No 11, 1989. p ..24-28, 17 refs. Based on an analysis of the lexical/statistical structure of See a1<;o 3/4-041. -054, -113 three elementary models, the procedure of their integration into a system allowing one to identify and classify terms au­ 3/4-331 341 tomatically has been developed. (Author, amended) White Head, J.: Computerized indexing and retr-ieval sy'>tems. Orig.en. Inform.media technol . Vol 22. No 2, 1989. p.64-68 35 Manual and Automatic Ordering 3/4-332 342; 753 Chang, J.W. et a1 : t�u lti key access methods based on term dis­ 3/4-342 351 Boulanger,D.; �larch, S. T. : An apporach to analyzing the infor­ ACMjSIGIRcrimination and signature clu<;tering. Orig.en. In: 12th Int. Conf.on Re<;earch and Dev.elopment in Information Re­ mation content of exi sting databases. Orlg. en. trieval. Cambridge. MA: ACM Press 1989. p.1�6-185 Data Base Vol 20, tio 2. 1989. p.1-8, 33 refs. 355 File ordering/organization 3/4-333 •• 342 374-456 HUffman.G.D.; Vital .D.A Bivins,R.G. ,Jr. : Generating indices See also with lexical association methods : term0 uni uenes<;. Orig. �n. norm. roces<;. anagement 26 990 ,p. 9- 58, 6 refs. 3/4-343 353; 355 A software system has been developed which orders citations Ito, T. ; Nakashima, r� .: Modification of a multilevel indexed retrieved from an online database -in terms of relevancy. The descriptor file. Orig.en. system resulted from an effort generated by NASA's Technology Infonn.Systems Vol 14, No 4, 1989. p.317-326, 16 refs. Util ization Program to create new advanced software tools to The paper discusses a method of modifying the existing file largely automate the process of determining relevancy of data­ structure so as to cope .f'!.\'lith the problem that a file search base citations retrieved to support large technology transfer may need backtracking. i the examination of more than one studies. The ranking is based on the generation of an enriched block at some level , even for retrieving exact-matches. vocabulary using lexical association methods. a llser assess­ (Author, abbr.) ment of the vocabulary and a combination of the llser assess­ ment and the lexical metric. One of the key elements in rele­ 356 Hypermedia vancy ranking is the enriched vocabulary - the terms must be See also 3/4-098, -149, -150, -152, -183, -187, -210, -223 both unique and descriptive. This pap�r examines term unique­ ness. (Author, abbr.) 3/4-344 356 343 Gen., Lingu-istic and Statistical Nethods See also 3/4-046, -57, -066, -138, -139. -200, 3/4-334 3432 Panevov&, M.: Retrieval language PASSAT. Orig. cs. 3/4-345 356 Ceskoslovenska Inform. Vo l 31. No 4. 1989. p.l02-105. 6 refs. Fluhr, C.: Construction automati ue de liens h ertextes. (Au- tomatic constructIon of hypertext li••nks Orig.fr. 3/4-335 3433; 345 Bul l.Cent.hautes etud.in."inform.doc No 35. 1989. p.89-98 Jones,L.P.; Gassie, E.W.,Jr. : INDEX: The statistical basis for 356 an automatic conceptual phrase-lndexlng sY'item. un g. en. 3/4-346 pc: J.Airier.Soc .Inform.Scl. Vol 41, No 2, 1990. p.87-97 Kinnel l, S.K.: Hypertext on the Guide, version 2.0. Discussion of a program called INDEX, a statistical ly based Orig.en. Database Vol 12, No 4, 1989. p.62-66 phrase-indexing program and an extended vers ion, INDEXD which incorporates a dictionary which gives additional capacity for 3/4-347 356 statistical analysis and also certain capacity for syntactic Levin, M.Sh. :.Hierarchic hypertext6, 19 s�9.stem. Orig. ru. analysis. Nauch.-tekhn .lnform. ,ser.Z; No p.10-13, 13 refs. 3/4-348 356 3/4-336 .• 3439 Abramova. N.N Bevzenko, E.A., Belonogov, G.G. , Globus, E. I. , Shneiderrnan. B. ; Kearsley, G. : Hypertext Hands-On!: An intro­ Zelenkov, Yu .G., Novoselov, A.P. : Automated indexing of docu- duction and a new \'lay of organizing and+ acce ssing information. ments by ke�words and word combinations. Orig. ru . Reading: Addison-Wesley 1989. XI+165p. 2 diskettes Nauchn.-tek n.lnform.,Ser.2, No 4, 1989 . p.29-31, 4 refs. ISBN 0-201-15171-5 Book review by K.S.Catiin in: Inform.Processing & Management 3/4-337 3439 26(1990)No.4, p.559-571, 1 ref. Cawkel l. A.E. : Automatic indexing in the Science and Social Included with each copy of the book is a set of two IBM PC Science Citation Index CD-ROM. Orig. en. diskettes that contain a hypertext version of the book. Al l Electronic Library Vol 7, No 6, 1989. p.345-350, 10 refs. the software needed to browse this electronic version is in­ cluded on the diskettes. To run it one needs an IBM PC or com­ 344 Computer-aided Indexing patible with at least 384 K of memory, a color or monochrome See 374- 199 monitor (CGA, EGA, or VGA graphic adapter required to view graphics). and two 5.25 floppy disk drives or one 5.25 floppy 3/4-338 345 disk drive with a hard disk. Hancox, P. : A recursive algorithm for generating SLIC index entries. Program Vol 23, No 3. 1989. p.311-317. 7 refs. 357 Document Structuring See 3/4-065

3/4-339 347; 88-51/4 38 Index Generation and Programs Humphrey, S. ; Chien, De-Chih: The MedIndEx Sy stem: Research on interactive knowledge-based indexing and knowledge base man­ See also 3/4-009, -156. -2B9, -333, -334, -335, -474 ag en . an g.en. Bethesda, MD: National L,brary of r�edl clne 90 • � 3/4-350 382 19 53p., Rep.No.NLM-LHC-90-03 �.• • H itchc oc k. L.A. : The Pl aY ' s t�h e Thi n�9. , ,12fJ Cy o�u!...'c�a �n --'.f l�· n�d' -2i �t �! The MedlndEx (Medical Indexing Expert) System is an interac­ ��':f+���8 ���p � - tive. knowledge-based prototype for computer-assisted -indexing An assessment of play indexes. arlg. en. of the biomedical literature being developed at the lister RQ Vol 29, No 2, 1989. p. 248-249, 19 refs.

Int.Classif. 17(1990)No.3/4 Classification Literature 195 The study attempts to devise a strategy for play index use 3/4-363 397 that will provide some assurance of efficiency and effective­ Feng. 5. : A comparative study of indexing languages in single ness. It clarifies the indexes' coverage characteristics and and multidataba'ie searching. Orig. en. tests their citation productivity with carefully selected play Can.J.Inform.Sci. vo l 14, No 2, 1989. p.26-46. 22 refs. titles. 3/4-364 398-546 3/4-351 383 Brenner, S.H.; rkKinn, E.J.: CINAHL and MEOLINE: compa rison of Kucianova, A.: Com ilin title and cor orate sub -ed entries. indexinq ractices. Orig. en. Orig.sk. Kni2.a ved.infom. 21 2989 No 5, p.214-220, 3 refs. au i Uled. lbr.Assoc. Vol 77, No 4, 1989. p.366-371, 19 ref'i. 3/4-352 383 Serebrjannikova, T. 0. : Compu ter-sup�orted estab 1 i shment of sub- ON UNIVERSAL SYSTEMS ject indexes to bibliograph 1p.s. Ong. ru. 4 Nauch.-tekhn.bibl .SSSR, No 2, 1989. p.28-32, 19 refs. 41 On Universal CS & T in General 3/4-353 384 Fetters, l.K. : A guide to indexinq19890 software. 3rd ed. Wash1ng­ See a1'io3/4-144 ton, DC: Amer.Soc.of Indexers cbp., ISBN 0-936547-06-5 Description of nine indexing programs and mentioning of 3 fur­ 3/4-365 418-432 ther ones which work with printed page proofs from a publisher "larkham, J.�'.: LCC, DOC, and Algae. Libr.Resources & Techn. or for creating a magazine index. The programs dg the alpha­ Servo Vol 34, No 1, 1990. p.54-61, 20 refs . betizing, formatting, and printing. The nine program) exist A comparison of the 'iystematization of algae -in the Li brary for the IBM PC family. The introduction covers help in choos­ of Congress Classification which put'i al l famil ies of plants lng a program and a survey on the organization of the reviews alphabetically together and in the Dewey Decimal Classifica­ on the nine indexing programs. It also provides a Comparison tion which arranges hierarchically in correspondence to bota­ Chart for the basic features, the edHing and displaying en­ nical classification. tries, the fomatting/printing features, and the sorting. The program features are explained in detail subsequently. 42 On the Universal Decimal Classificatilm 3/4-354 386 VeneZiano, V.: The "new" NOnS/LUIS indexes. Orig. en. See also 3/4-131, -153 Infom.Technol.& Libr. Vol 8, No 1, 1989. p.5-19 3/4-366 A. ; 42 .08 3/4-355 387 Gilchrist, Strachan. D. (Eds.): The UOC: E,>says for a new CDS/ISIS - IBM mainframe vers ion. Orig. en. decade. London, GB: Aslib 1990. 97p., ISBN 0-85142-265-9 Int.lnform.Commun.& Educ. Vol 8, No 1, 1989. p. 1l4-116 Thl-; essays include four "managerial" and s'ix "technical" pa­ pers eHher col lected from somewhere else or written especial­ 3/4-356 387 ly for this publ ication. 3/4-367 Strachan, P.O. : UDC revision Inozemtsev, V.A. : Some aspects of developme nt and classifica­ ,�ork in FlO. - .lii-3f!!! Scibor, E. : Shcherbina-Samojlova, 1.5. : tion of kno�/l edge representation languages. Orig. ru. A strategic approacn to revising the UDC. - 3/4-369 McIlwaine, Logiko-filos.issled., No 1, 1989. p.47-55 I.: The work of the UDC System Development Task Force. 3/4-370 Wil liamson,N. : The UDC: its future. - 3/4-371 Haarala� 3/4-357 387. 272 FlO pre-conference 'iemi nar on computer applications of the Sick, D. : Aufbau und Pfle e kom lexer naturl lchsprach19 ba­ UDC. - 3/4-0� 72 Loth, K. : Funk. H. : Subject search on ETHICS on sierter Dokumentatlonss rac en esaun . ue Ie tendenzen the baS IS the UDC . - 3/4-373 Karhu1a, P. : Use and usability und kritische Analyse e-iner ausgewahlten autonomen Thesaurus­ of the UDC in classification pra.ctice and online retrieval. - Software fUr Personal-Computer (PC) (Diplom-Arbe it). (Con­ 3/4-3�1 RinnY.:e, B.: HElECON system: economics databa'ies. 3/4-375 struction- and maintenance of complex documentary languages Nakamura, Ishikawa, T,: Expert sY'item for automatic UDC (the'iauri). Trends and critical analysi'i of a selected 'itand­ number assignment. - 3/4-376 Riesthui'i, G.J.A.; B1iedung,S. : alon� _� hesaurus software for PCs) Orig. de. SaarbrUcken: Uni­ The'iaurification of UDC. vers ltat des Saar1andes. 1989. 189p. (MS-Thesis work), refs. In chapters 2-5 the author presents the foundations, 'itate-of­ 3/4-377 P,: 42. 174 the-art and development trends of the thesaurus method in the Canisius, Entscheidunq Uber die UOC 1991. FID-Rat stel lt 1 ight of new technologies. Chapter 5 con.tains the draft of an We�chen fUr kunftige Entw-i cklungen. (Dec i'iion about the UDC evaluation sheet for thesaurus 'ioftware and in chapten 6-7 in 1991. FID Council prf'pares1. the way for future developments) the newest PC-Software for the'iaurus-management is evaluated Orig.de. DGD-Hi tt. Vol No 2, 1990. p.6 in an application test. (Author, abbr.& trans1.) The Counc il decided to cont-inue the support of the UDC until Dec.1991 and to find in the meantime an institution who would 3/4-358 387; 754 serve as its future publ isher. If no institution can be found, Zarri, G.P.: A knowledye rerresentation language for large the UDC \1i 11 have to be di'icontinued from the part of the FID. knowTed e bases and "intel 1 ent" information• retr'ieval s s- tems. nom. rocess. anagemen 0 p. 3/4-378 42.213 Neicu, S.D.: Computer-supported handling of UOC-numbers and 3/4-359 388-26 auxiliaries wlth regard to retnevai. Or lg. ro, en. Lawson, R.G. ; Jurs, P.C.: New index for clustering tendency Bib1ioteca, Bucuresti, No 3, 1989. p. 16-18. 2 ref'i. and its appl ication to chemical problems . Ong. en. Work on a UDC edition in Romanian i'i done by the Intitutul Na­ J.Chem.lnfonn.Comput.Sc1. Vol 30, No 1,1990. p.36-41, 17 refs. tional de Informare si Documentare Sti i ntifica si Tehnica in Kukarest since 1979. Ever 'iince 1987 thi'i work ha'i been compu­ terized for editing and index preparation. 39 Evaluation of Classing and Indexing 3/4-379 42.225 See also 3/4-067, -073, -125, -158, -538 Mandryka, T.N. ; Shchedrina-Samoj lova, 1.5.: Using a facet for­ mula for �erfecting the UOC 'icheme.10, Orig. ru. 3/4-360 392 1ra'i:iCll.-te hn .lnfonn.,ser.l, No 1989. p.13-I5, 20, 8 refs, Iivonen,M. : Interindexer consistency and the indexing environ­ ment. Int.Forum Inform.& Doc. 15(1990)No 2, p.16-21, 13 refs. 3/4-380 42.6-63; -72 New 'ichedu1e for UDC 65. Orig. 'ik. Sprav.Sloven.techn.knizn., 3/4-361 396 No 1, 1989. p.lO-ll Allen, B.: Recall cues in known-item retrieval. Orig. en. The Slovak Technical Library pub1 ished in late 1988 the sche­ J.Amer.Soc.lnfom.Sci. Vol 40, No 4, 1989. p.246-252, 44 refs. dules for Administration, Organization of production, commerce Examines the effect of different kind of cues on the recall of and transport. a text. 3/4-381 42.756 3/4-362 396; 173 Godert, W.: Die Dezima1klassifikation im Online-Retrieval. Blair,D.L ; Maron, M.E.: Full-text information retr-ieval : Fur­ (The UDC in the onhne-retneval) ung.de. Nachr.Dok. Vo l 41, ther anal sis and clarification. No 3, 1990. p.155-158 In orm. rocess.& anagement 0 26, No 3, 1990. p.437-447 In addition to a paper by H.J.Henne'i and 11.Bischoff (in Nachr. In 1985, an article by Blair and Maron described a detailed Dok.40(1989)p.355-360) in thi'i note some structural features evaluation of the effectiveness of an operational full text of the UDC are cons idered which may be of importance for an retrieval system used to support the defense of a large corpo­ online retrieval with synthesized UDC notations. rate lawsuit. The fol lowing year Salton published an 'article which cal led into que'ition the conclusions of the 1985 study. 3/4-382 42.88; 42.99 The fol lowing article reviews the initial study, replies to Rozenshil 'd, L.K.: Toward the uniform usare• of classification the objections raised by the second article and clarifies se­ tems. Orig. ru. In: Nauch.katalog1Z.il 1 org.kataiogov, veral confusions and misunder'itandings about the 1985 'itudy. Moskva, SU 1988. p.58-69. 5 refs.

196 Int.C1assif. 17(1990)No.3/4 Classification Literature 3/4-383 The four chapters descr-ibe: Introduction. The history of the ,J. , 42.8-817 Bauer. I.; Zajec ,J., Kritovac ,D. , Svarc Ni las-Bracovi c, N., Library of Congress law Classification (early years, struct­ Dolezal , V.: An approach to UDC appl ication to the classifica­ ure and scope of the anticipated Class K, the German law sche­ tion and indexing of research materials. Orig. sh. dule, method of development of law schedules), Structure of Infonnatol . Yugosl . Vol 20, No 3-4, 2988. p. 279-142, 22 refs. subclasses KK-KKC within Class K, and the subject arrangement of schedules KK and KKA. 43 On the Dewey Decimal Classification (DDC) 3/4-396 P. ; 448-92 Wilson, Robinson, N.: Form subdivisionsI, and enre. See also 3/4-033, -034, -277. -555, -556 libr.R�sources & Techn.Serv. 34 1990 No p.36-43, 17 refs. LC's scheme of genre description is incomplete and could be 3/4-384 43.07 extended usefu lly. Sifton, P. ; Dragani, N. : Workbook for DOC 20; Dewey Decimal Classification, Edition 20. orig.en. Ottawa, On t.: Canadian 3/4-397 448.172 Llbrary As soclabon 1989. 48p., ISBN 0-88802-252-2 Calderon, F.: library of Congres') Subject Headings: Vested in­ terest versu') the real needs of the lnformatlon socIety. Cat.& Classif.Quart. Vol 11, No 2, 1990. p.85-94, 19 refs. 3/4-385 .• 43.08 Friis-HanSAn, J.: Book revie�1 of Comaroml, J.P Satua, r� .p.: Amer ica's higher educat'ion system does not have a modern tool DfMey Decimal ClasslficatlOn. Hlstory and current status. New of infor�nation retrieval by subject. The failure to fu lfill Delhi: Sterl ing Publ . 1989. 178p. Orig. en. that 'llis.;; ion belongs to the library of Congress Subject Head­ J.Llbrarlanshl p Vo l 21, No 3, 1989. p.210-211 ings. Al though the consequences of such failure are awesome. no effart is 90ing on to adopt a different subject indexing 3/4-386 43.4 vocabulary. If that substitution does not happen soon enough. Merwe.M.S.van der: Comments on Oe�/ey 20. �lousaion Vol 8, No 1, harm \�i l1 be infl icted on the mi llions of students attending 1990. p.22-31 col leges and universities, as well as the legion of profes­ Critical review of the new edition from the point of view of sionals I"lho use research libraries al l over North America. The a teacher of classification in South Africa. crisis has to be brou9ht into the open and given the wide dis­ cussion it deserves , especially in the United States and 3/4-387 43.4 Canada. (Author) Berman, S. : 00C20 : the scam continues. Orig. en.

Libr.J. Vol 114, No 15, 1989. p.45-48, 5 refs. 3/4-398 • 448.213 . The interview with Sanford Berman p' inpoints his criticism con­ Joy, A.H.. ; Keane N.J.: tD"='JWl.CDl1ARCf. Subjects: The Li brary of Con­ cerning user-unfriendliness, lacking real ism, lack of informa­ gres.;; _�ubJect HeadIngs on un g. en. tion on changes and non-practicability in libraries, sho�m CD-ROM libt-arian Vol 4, No 9, 1989. p.36-45 by numerous examples. 3/4-399 448.756 •• 0. : 3/4-388 43.4 Lucas ,D. : Book review of t-la rkey. K Vizine-Goetz, Charac­ teristic� of .;;ubject authority records in the machine-readable Srinath.�1.: 00C20 : the striking differences from 19th edition. ..Qf ILA Bul l. Vol 25, No 2, 1989. p. 100-l03, ref. XXXIILibra�,J::l_?:'!i�. con ress Subject Headings. Dubl in, OH: OClC 1988. Cla sslt.Quart. vol 10, No 4, 1990. p.1l6-IIB 3/4-389 E. : 43.4 Zanobi, A. ; Grignani, uaderno Dewe . esercitazioni sulla 3/4-400 44B-6 Edizione 20 de�la Classificazione DecimaTe Dewey. Exercises Studwel l, II/ .E. · Subject �s tions 7: Some concerns relating on the 20th edltlon of the Dewey Decimal Classification) to thf' SOCldl �ciences.Orlg. en. Cat.& CTiadinys for the social sciences: clarification of the use of Schroeder,C.D.: Digital warrant: revised DOC 004-006 in a pu­ disciplMines versus to�ics; clarification of the meaning and blic library. libr.Resources & Techn.Serv. Vol 32, No 4, 1988. uSilgl' the subd :vislon "Civil ization"; use of "law and le­ p.367-377, 17 refs. yisl.3.t ;on" a"i d free-floating subdivision; and some suggest­ It is repor�ed that the revised schedules for data �rocessing, ions for dedl"iny with controvers;al soc ial topics. (Author) computer SClence and computer engineering (004-006) \'Ierf� easy to learn and apply. and browsability was improved by recl �s�i­ 46 On the Colon Classification (RCC) fication in the Boston PUblic library.

43, 6-967 3/4-401 0.: 14 .p. : 46.07 der: Andrews. Book review of Satija, Colon Classification llil!.. 1 1. __e_� A prac't rc-a introduction. New Del hi: [5S E5S Pub 7 refs. 1989. ��, 236p. J.librarianship 22(1990)No.1, p.59-60 for Dewey for local requirments is recommended as wel as use of a num- 3/4-402 M.P. : 46.172 ber of specific existing adaptations and internal options. Satija, Colon Classification, 7th edition (1987) : A cri­ Guidelines are given to ensure justifiable and �uccessful tical _i.ntrodl!£tion. Outlook on Research librs. Aug.1990. p.4-9 adaptations in the future, 3/4-403 46.293 3/4-392 43.6-967; 44.6-967; 6967 Husa :n,Sh.; /·l'.Jstdfa Khan, M.T. : Conceptual changes in the fun­ Iwuj i. H.O.: Africana in lC and DO classification schemes : a dalllE'nt_,!l categories•• in CC7. An evaluatlon. Ong. en. need for an Afncana scheme? Herald of libr.Sc Vol 29, No 1-2, 1990. p.45-50. 10 refs. J.Llbrananshlp vo l 21, No 1, 1989, p. 1-18 f1akes an evaluation of the many conceptual changes with re- 9ard to Fundamental Categories in Ed.7 of the CC. Traces the 3/4-393 43.6-99; 9843 developillent of the fundamental categories PMEST, explaining Woods, J.: Usin Dewe in a mission librar . Orig.en. Bull. each of the concepts with examples. Discusses the Hatter Mate­ Assoc.Brit.Theol .& Phllosph.libraries 2 1989)No 5. p. 11-13 rial Category. Matter Property Category and Matter Method Ca­ Presentation of notes on the adaptations made to class 200 in tegory introduced in the new edition of the scheme. (Author) Parternship House litJrary. a modern mission library formed in 1988 by the amalgamation of the post-1945 �tock of the 1 ibra­ 3/4-404 46.6-39 ries of 2 Anglican missionary societies. Kumar,B.K. : Geo ra hical divisions of East Asia in Colon Clas- sification Edition a : An a ralsa an a reconstructlOn. On the Library of Congress Classification rlg.en. u. 25, No ,1989. p. 44 The article present" an elaborate critique of the 'new sche­ and on the library of Congress Subject Headings dule' of China and its peripheral territories included in the chapter on 'Space Isola.tes' in the 7th ed. of CC taking ·into See also 3/4-030, -031, -032. -392 account the author's earl ier attempt on this subject and S.R. Ranganathan's "tentatively revised Schedule of Geographical 3/4-394 44.07 Divis ions of China". A reconstruction of a detailed schedule Dasgupta. A. : Book review of library of Con9ress subject cata­ of East Asian countries is given. (Author, amended) loging manual - SubJect Headl ngs. 3rded. On g. en. IASLIC Bul l. Vol 34. No 3, 1989. p.141-143 47 On the Library Bibliographical ClassiJ. 44.6-667 DE (LBC/BBK) 3/4-395 J.E. : Goldberg. librarf of Congress Class KK: law of GermanY. See also 3/4-142 A historical review of aw classlflcatlon. Ong. de. In: Hahn, G. , Kirchner, H. ; Parlament und Bibliothek. Int. Festschrift f. 3/4-405 47.6-39 Wol fgang Dietz zum 65.Geburtstag. MUnchen: K.G.Saur Verl .1986. Sukiasjan. E.R. : The area studies2. section1988. of the collection. p.328-344. 40 refs. Orig.ru. Bibliotekar' (OSSR) . No p.46-7

Int.Classif. 17(1990)No.3/4 Classification literature 197 3/4-406 47.6-56 3/4-418 487.626; 287 Kaback, S.M.: '�PI and APIPAT: the sum of the parts. Database Papava, A.: HardBBK cases. (Pedagogical literature in the 'lecond edit-ion of the tor publ ic libraries). Orig. ru. Vol 13, No 2, 1990. p. 22-26, 3 refs. Bibliotekar' (USSR), No 10. 1987. P.60-61 Description of the lIlerger of the database API PAT into WPI. that makes the Derwent abstracts in WPI available for most of 3/4-407 47.6-7; 47.173 the earlier references in APIPAT. Goltvinskaja, T. : Difficult cases: BSK classification of tech­ 487.753 nical literature55 for . re lonal libraries, Orig. ru. 3/4-419 1 8, lote ar No 2, 1988. p.6 -63 lepee. W.: EPOS : a ingu-istic tool for full text search. World Patent InformaUon Vol 11, No 4. 1989. p. 149-199 3/4-408 47. 6-962 EPOS (EPO Synonyms) i<; a system being devised by the European Makhotkina,L.N. : Ethnic common divisions in the BBK tables for Patent Office as a fu l1 text search tool to assist the exami­ regional studies catalogs . un g.ru. ner in querying databases. If an examiner knows one word ex­ SoV.Biht.. No 6, 1987. p.44-47 pressing his field of interest, EPOS del ivers to him all the ECLA (EPa classification) groups where this basic tenn is con- 3/4-409 47.812; 47.6-46; -71/2; -62 sidered relevant. (Author, abbr.) Babka, A.N. ; Tyrys hkina. Z.l.: Methods of reeeated coverage of literature by clas'ied catal09ues 1n agr,culture. economy , and 489 On CS for Other Spec ial Purposes See also 3/4-529 po litical sciences in the LBe tables. with spec ial reference 489 to the classedI catalogue of GPNTB SO AN SSSR. Orig.ru. In: 3/4-420 Nauch. katalog z.li t. i org. kata logov. f�oskva 1988. p.25-204 Superintendent of documents classification system. Orig. en. In: Pokorny, E.J., Miller, S.�1.; U.S.Government Documents. 48 On Other Universal Systems Littleton, CO: Libraries Unl imited 1989. p. 19-24 ON SPECIAL OBJECTS CS 3/4-410 483 JA 5 �Initial aims in the revision of Kihon Kenmei Hyomokuhyo. 3rd Orig.ja. Toshokan Zasshi Vol 83. No 5, 1989. p.264-265 3/4-421 519; 7578 Kihon Kenmei Hy omokuhyo (Basic Subject Headings), a 1 ist of Gawron.V.J.; Drury,C.G., Czaja,S.J., Wilkins,D.M.: A taxonomy only 4,270 headings, is meant for publ ic and educational li­ of independent variables affecting human perfonnance.21 InLJ. braries in Japan. The article discusses questions of oJni s�;ion �lan-Machlne Studies Vol 31, No 6. 1989. p.643-671, refs. of proper names, simple and compound wordsf synonyms, direct As part of an ongoing program to develop a Computer Aided En­ references to subordinate terms, entry of explanatory details. gineering (CAE) systelll for human factors engineers. a Human Performance Expert System, Human, was designed. The sy<;tem 485.213 contains a taxorionlY of independent variables which affect hu­ man perfonnance. This taxonomy was developed in a three-step process: review existing taxonomies; add independent vari­ ables used in Human; and remove redundancy and ambiguity. This Classifica­ process and the resultant taxonomy are described in this pa­ tion is now ava data base with per. (Author) all the advantages which this fonn wi ll updating and revision. retrieval aid. basis of version. ON SPECIAL SUBJECTS CS T translation into other languages. 6 & 3/4-412 485.27 62 On CS & T in the Energy and Matter Area 2 Outline of Nihon Jushin Sunruiho Ni on Decimal Classi fica­ tion10 draft 9th1 edition 0 , 2:0, Lan ua es. Or�g. ja. os an ass p. 143-146 See also 3/4-300 Revision proposals of the 8th ed. include linguistic�, lan­ guage in society, classification of languages. and Japanese. 3/4-422 6265; 242 t4 ishchenko,G.L.; Il' ina-Sidorova.A. Ya., Uskov,A.A. : The feasi­ 3/4-413 487 bility_ of automatic spec ification of paradigmatic relations in n 0hganic reaction thesaurus. Drig. ru. Jenkins. G. ; Function and a lication0 - is there • a difference? World Patent n ormatlOn ,p. refs. Nauc .-tekhn ,lnform. , Ser.2, No 7, 1989. p.20-22, 9 refs. In the 1 ight of recently agreed changes to Chapter I r I of the The possibility of an automated establishment of paradigmatic Guide to the IPC the meanings of the terms 'function' and 'ap­ relations in an organic reaction thesaurus using a formalized plication' are examined. It is suggested that technical dis­ IR language is considered. Three languages of structural in­ closures col lectively comprise a continuous spectrum of infor­ fonnation on organ-ic reactions were compared, the language of mation and that technical subjects are clusters of disclosures changing types of bonds vias selected. Principles for establ­ on particular 'planes' on this spectrum. The ideas of 'funct­ ishing paradigmatic relations in the thesaurus are formulated ion' and 'application' are explained as relativities between based on an analYSis of 700 organic reaction descriptions. these planes. (Author) 3/4-423 626 3/4-414 487 Attias, R. ; Dubois. J.-E. : Substructure systems : Concepts and classifications. Orig. en. Jenkins, G.: The use of technical planes in structur-ing the I, classification. J.Chem.lnfonn.CompuLSci. Vol 30. No 1990. p.2-7, 44 refs. wo rld Patent Information Vol 11, No 4, 1989. p. 187-192, 1 ref. The main approaches to substructure systems are reviewed in The article demonstrates; with examples, some of the practical the light of various classifications. allowing representation effects of the idea of planes on the design and developmEmt of of the same entity from different standpoints. A generic mo­ the International Patent Classification. (Author, abbr.) del , an extens ion of the graph-specific <;tructure. provides a basic lIIodel which unifies structure and substructure and 487 yields a fonnal tool for I'epresentation and evaluation. 3/4-424 629 refs. Timotin, A: ; Ghetaru. S. : Thp rational thesaurus of the tenns The article traces development of hybrid systems, includ­ included 1n the International Electrotechnical Comnnssion's ing their adoption -into the IPC. It explains their basic prin­ pUbjications.• Orig.ro. en. Probl . lnfonn.Doc. , Sucuresti Vol 23. ciples and reviews the offices and commerc ial organizations No �p .106-124 which utilize them. (Author) In the rational thesaurus of tenns contained in IEC publica­ tions, developed jointly by the Electrotechnical Research In­ 3/4-416 487 stitute and the Bucharest Polytechnic, each tenn entry is ac­ Snow. D.C.: Why 'hybrid' information systems? World Patent companied by synonymous, associated and subordinated terms. Information Vol 11, No 4, 1989. p.200-207. 10 refs. As the thesaurus specifies a variety of hierarchic relat-ions. Discussion of the origins of hybrid classification systems. it contains lIIore information than a dictionary and facilitates their characteristics, and their advantages and disadvantages. the incorporation of a concept into a structured body of The structure of technological information in patents and the knowledge. (Author, abbr.and amended) basic pattern of its description is reviewed briefly and the document classification and indexing component of hybrid <;ys­ 63 On CS & T in the Cosmos and Earth Area 3 terns is analyzed. (Author, abbr. ) 3/4-417 487.1 3/4-425 6392.232 Vries. S.de: Points of interest concerning the new IPC. Vera. J.A. : Design criteria for the Decimal Classification of World Patent Information Vol 11. No 3, 1989. p.115-120. 1 ref. revers of the hydro database at the Centro de Estudios Hidro­ Useful adaptations to new techniques and to keep the size of lraf{ wos. .• un g. es. the search files of the new IPC manageable are presented. ng. No 70, 1989. p.67-70, 73-74, 3 refs.

198 Int.Classif. 17(1990)No.3/4 Classification Literature (;4 On CS & T in the Bio Area 4 3/4-438 6835 Popowska,H. : Databases - classifications and typologies, Pt.2. 3/4-426 649 Orig.pl. Aktual .probl .infonn.i dok. 34)l989)No 3, p. 16-22 Lavieter. L.de: Towa rds a general thesaurus for envi ronment. A typo logy of databases established and used by the DIALOG Orig.n1 . Open Vol 22. No 2, 1990. p.49-51, 4 refs. service is presented. The need is suggested, however, for a deeper, multidimensional typology or classification of data­

3/4-427 •. 649 bases which would discriminate between content plane and ex­ Andreeva ,E.V. ; Ry zhikh E.P.: A classification of enviromental pression plane. (Author, abbr.)

situation knowledge:.• a des ign approach and structure. Orig.ru. Nauch.-tekhn .lnfo nn Ser.l. No 10. 1989. p. 16-20, 7 refs. 3/4-439 686 Nilanovic. �I .: linguistic problems of natural-to-information 3/4-428 649 language translation, with special reference to the mass com­ Nadezhina. M.E.: Spec ific features of a descriptor dictionarr munication thesaurus devel oped by the Yugoslav Institute of on environmental protection. Orig.ru. Journalism. Orig.sh. Praksa 24(1989)No 4, p.42-44, 5 refs. Nauch.-tekhn .lntonn. ,ser.2 Vol 2, No 5, 1989. p.18-19, 1 ref. 69 On CS & T in the Humanities Area 9 65 On CS & T in the Human Area 5

3/4-440 H. : 691.231; 692 3/4-429 651/4.213 Ovens. C. S. Nethodo logy for constructing a thesaurus in the Loginov,B.R.; V'yugin, V.V. : Automation technology for the in­ linguistics and literature fields. oril ' en. troduction of a bilingual thesaurus In med, CIne and health S.African J.llbr.& Inform.Sci. 57(1989 No 1, p.45-50, 8 refs.

care uSing a microcompu.• ter. On g. ru. Nauch.-tekhn.inform Ser.2. No 4. 1989. p.14-16. 3 refs. 3/4-441 693 Gyulane, G.: Comments on a prorosal for I,compl ement FlO numbers 3/4-430 Orig.hu. Flgyelo p.69-77. 9 refs. 1 656 Konyvtari 35 1989)tlo El lzarov.A.A; Brusnichkin,V.�I. : Educational databank inqware. Reflections on a work by Frigyesne Erdely-i entitled 'Flexible Orig.ru. In: Ped. informat. : bank ped.dannykh.Komp'yutenz.o­ Nus-ic Class ification System: suggestions for complementing the buch. Moskva. SU 1988. p.12-16 instrumentation area '. Cha llenges several aspects of the clas­ sification of popu lar and non-European instruments, and makes 3/4-431 656.172 suggestions for improvement. (Author) Thomas.A.R. : Terminological problellls : occurrence, causes, con­ se uences. and solutions 1in uidanceI":S. counsel lin and associa- 694 On CS 1 & T in the Arts e Ie s. n ..n. r. o ,po re s. See a so. 3-003

3/4-432 6598 3/4-442 694 Doucet,M. ; Filion, R., Garon, D. : Classific11": ation et analyse de Petersen, T.: The AAT Art and archi tecture thesaurus in the col ll":ctions d'objets de jeu selon sYsteme ESAR: rapport de MAilC format. Ong.en. Art Doc., 08, 1989. p.181-2 recherche. (Classification and analys is of col lections of games according to the ESAR system: research 3/4-443 694 report) Orig.fr. N.: Documentation et Bibliotheques Vol 35, No 4, 1989. p.173-185, Waddington, The AAT (Art and architecture thesaurus) appl i­ 23 refs. cation protocol . Ong.en. Art Doc., No 8. 1989. p.178-180 66 On CS & T in the Socio Area 6 3/4-444 694.225 SI":€ also 3/4-205 14h itehead. C.: Faceted classification in the Art and Architec­ ture �hesaurus. Orlg.en. Art Doc., No 8. 1989. p.175-7

3/4-445 6945 �likhajlova,V.G. ; Naumova ,I.M.: Perfecting an indexing language \�ith a view to reference-'informatlOn serVlce optlmisatlon. Orig.rlJ. In: Soversh.avtomatizir.inf.-poisk.i faktograf.sistem i sistem eletron.restavratsii arkh .. r10 skva, SU 1989. p.13-19

3/4-446 10 694.225 3/4-434 663 laltanne, H. : Facet classification the library of the Ap- Goch P.v. : plied Arts School . Orig.hu. Konyvtari Figyelo Vol 35, No 1, 1989. p.52-6"4;'"T2 refs. 9��_T 67 On CS & T in the Econ. & Production Area 7 696 in the Ethnic Sciences See also 3/4-392 3/4-447 6967 3/4-435 675; 233; 363 Roux. P.J.A. ; Rykheer, J.H. : Developins a South African master Zenovic,S.V. : A classification model in graphical-tabular form thesaurus for community informat1on. Ur lg.en. South Afr.J.f. and its appl icat)on In computer,zed lnformatlon systems. libr.& Infonn.Sci. Vol 58, No 2, 1990. p. 174-179. 13 refs. Ong.ru. Nauch.-techn.Inform. ,Ser.2, No 7, 1989. p.9-12 Description of rules for the graphical presentation of hierar­ 3/4-448 6988.213 chical taxonomic and classificatory structures. Introduction Col l ins, W.p. : Effects of computerization regu irements on the of normal forms and procedures for the coding of elements to 8aha 'i Classification Scheme in use at the Baha 'i World Centre be classified. Results -of the application of the proposed mo­ On g.en. Cat.& Classlf.Quart. Vol 11, No 2, 1990. p.69-83 del for the classification of projects in the building area. The Baha 'i Worl d Centre Library developed a classification for 8ahd 'i rnateria 1 s by expandi n9 li brary of Congress BP300-395 3/4-436 676; 268-76; 756 section. �nth the devel opment of a computerized cataloging Robinson, M.l. : A survey of SIC code searchable databases on system (called �IAGE) at this library, and particularly with Dialog. Online Rev. Vol 13, No 4, 1989. p. 313-319, 3 refs. the review of authorities for Baha 'i subject headings, this Standard Industrial Classification (SIC) codes are a type of classification has been further modified. The experience pro­ controlled vocabulary used by business searchers when search­ vides an interesting example of how the process of computeri­ ing for corporate and industrial data. Two four-digit vers-ions zation can affect classification schf'mf's. (Author) of the SIC coding scheme are used on business databases avail­ able through DIALOG. Some Dialog database producers have opted to create a variation of the codes for use in the ir products. CLASSIFICAT. AND LANGUAGE With the impending eight-digit SIC coding scheme from Dun and 7 8radstreet, it will become -increasingly difficult for the Dia­ 71 General Problems of Natnral Language log searcher to know which version of the codes to use when searching a particular database. A survey of the use of SIC codes in business databases on Dialog was undertaken. The pa­ 3/4-449 711 per presents the findings of this survey . (Author) Frumki na, R.H. (Ed.): Language and cogn itive activity. Orig.ru. Moskva, SU 1989. 144p. 676 See also 3/4-148 -154 3/4-450 711 68 On CS & T in the Science & Inform. Area 8 ph ilosophy Orig.ru. -lbl Galustyan, H.G. : The of linguistics. See also 3/4-136, Erevan. SU: Lujs 1989. 176p.

3/4-437 6811 3/4-451 714; 123; 157 Mirskij, E.M. : Knowledge or�anisation in the research20, front. Dahlback. N.: A symbol is not a s¥mbol. Orig.en. In: IJCAT-89: Orig.ru. Sistem.lssled.Aeto ol.probl .: Ezhegod. Vol 1988. Proc .llth Int.Jolnt Cont.Artlt.ln ell., Detroit, MI, Aug.20-5, p.225-247, 14 refs. 1989, Vo1 .l. San Mateo. CA 1989. p.8-14, 26 refs.

Int. Cl ass if. 17( 1990)No. 3/4 Cl assification literature 199 3/4-452 714 3/4-467 p, : 733 Shingareva. LA.: Semiotic model<; in the sign typo logy and in Latocha. Ex loration von Aussa enraumen: Ein semantischer pwchol inguistics. Orig.ru. Ansatz. (Exploration of statement '5pace: A semantic approach In: Kul'tura. obshchenie, tekst. r�oskva. SU 1988. p. 174-202 1Jr"i"g.'"de . GMD-Studien, No 164, 1989. p.1-114. 42 refs. 74 Grammar Problems 3/4-453 714 Warner,J.: Semiotics. infonnation science, document" and com­ See also 3/4-197 pu ters. Orig.en. J.Doc. Vol 46. No 1. 1990. p.16-32. 93 refs, 75 Ouestion-Answering Systems, Online Techn. 715 Formalization of Natural Language See also 3/4-002. -121, -186 See also 3/4-001, -028. -060, -065, -068, -129, -133, -326 72 Semantics 3/4-468 751; 347 See also 3/4-069, -112. -141 le Loarer, P.: Djalo ue naturel our 1 'interro ation et index­ ation as'iiste. atura 1a ogue or compu er-aSS1ste re­ 3/4-454 722 trieval and indexing) Orig.fr. Rev. francaise de Bi bliometrie Bonzi. S. : Book review of Hir<;t. G. : Semantic interpretation Dec, 1989. p.63-66 and the263p. resolution of ambiguity, Cambridge University Press 1987. J.Amer.Soc.lnform.Scl. Vo l 40, No 6, 1989. p.438 3/4-469 751 Rornanov, A.A. ; Shemak0 in. Yu. I.:moae:-lJrT9 Infor�ation. retrieval systems 3/4-455 723 witl�i!.n interactive eration ru, transl.lnto en. Autom.Doc.& �lath .Linguistics 23 1989)No 2, p.15-20, 3 refs. Zdenek ,J.: Determi nation of the level of information31. relevance3, 1989. of a text. On g.cs. Ce skoslovenska fnform. Vol No p.65-67, 3 refs. 71:J'C. Dialoiue SY'items. Interactive Catalogs See a so 3/4-037, -038, -080, -081, -135, -190, -201 725 Semantic Networks and Associations See also 374-186 3/4-470 752 Akeroyd. J.: Information seeking in online catalogues. 3/4-456 725; 225; 243; 355 Orig.en. ,J .Doc. Vol 46. No I, 1990. p.33-52. 12 refs, Duncan, E.B.: Structurina knowled�e bases1, for designers of Three different interfaces for online publ ic access catalogues learning mater-ials. Hypennedla 1{l9B )No p.ZO-33. ref<;. werf-' evaluated u'iing transaction logging as a methodology. In two cases exercises were set dnd run against each interface so 3/4-457 725; 153 that result'i could be critically evaluated to indicate areas Shastri , L. : Default reasoning in s"ema ntic networks: a formal­ of wP'dkness. In all cases a random transaction log was derived ization of recognition and inheritance. Orig. en. and th is was used to diagnose system problem'i and identify Arbt.lntell. vol 39, No 3. 1989. p.283-355, 36 refs. patterns of behaviour. Results showed that searches were often difficult to <::ategorise, U'iP.\' ; employing different approaches 3/4-458 725 to the same end. Users also were prepared to undertake sus­ Romanov,A.I.: Use of semantic networks to describe some engin­ tained searching until something wa'i retrieved ; however they then frequently stopped. Search input problems are also iden­ eering objects. Ong..• ru. Na uch.-tekhn .Inform Ser.2, No 7, 1989. p.23-25, 2 refs. tified and possible areas for further re'>earch 'iuggested. (Author) 726 Semantics of Texts and Languages

See also 3/4-189. -192 3/4-471 • 752

Gre"nwootl,• D.: OPAC research in the UK. Orig. en 3/4-459 126 l.LibrarianshipVol-21, No 3, 1989. p.157-173, 18 refs. Andre,J. : Reconna issance de la structure 10 i;ue des documents {Logical structure recognition of a document Or-,g. fro 3/4-472 752 Bul 1.Cent.Hautes Etur:l .int.inform.doc., No 35. 1989. p.47-52 Book reviel�S of Hildreth,Ch.R.: The online catalogue••: develop­ menh_and directions. London: Libr.A'isoc.1989. 212p 3/4-460 726 Oriy.en. Dietze. J.: lexikalische Nakrostruktur und selilant� scher Ab­ by Ann O'Brien in: Libr.Assoc.Record 91(1989)No 12. p.727 stand, lexikallsche M1Krostruktur und semantlsche Ko nslsteill by Kathlp.en t'l .Hogan in: Ubl'.J. 47(1990)No.l, p.65 - Versuch ihrer computergestiltzten 8estimmulJ..[. (Lexical lilacro­ by H.-I,Hartley in: Educ.for Infonn. 8(1990)No.l, p.60-61 structure and semantic distance, lexicoll microstructure and semantic cons istency - attempt at their compllter'i!lpported de­ 3/4-473 752 tennination) Orig. de. Umstatter, W.: Book revie\� of Hildreth. Ch.fI. : Intel l igent in­ Wiss.Z.Univ.Ha1 1e, G. yo1 39. No 2. 1990. p.43-48 terfaces and retrieval inethods for subject searching in bi­ bhogrdpn lc retrIeval systems . Wa shinGton, DC: L1brary of Con­ 3/4-461 726 �Cataloging Distribution Service14. 1989.1. Orig. de. Go 1 ovi na, L. V .: i'S �e �m a� '�' t� i�cc Epe�r�c�e�p�t�i �o n� o�f- "c �o �he�. r�e�n �tc-co�,�' ","�o�n�i �c�a�t .,g o�ng ,s . HlbTfOfhek. Fonch.ll.Praxls Vol No 1990. p.90-91 Orig.ru. In: Kul1tura, obshchenie. tekst. Moskva, SU 1988. p.106-126 3/4-462 726 Heyer,G.; Ke se,R. : Intell igente Textverarbeitung. (Intel 1 igent text processing) Ong .de. PraXl s Infonn. verarb.u. Kommun i kat. , No 4, 1989. p.213-216. 6 refs. 753 On line Activities. Query 0atimizdtion See a Iso 3/4-042, -045, -0 9, -056, -058, -147, -332 3/4-463 i 726; 122 l!!.9...Norvig, P. : Marker pa ssing as a weak method for text nferenc- Cognitive Science Vol 13. 1989. p.569-620, 29 refs. 3/4-475 753; 174 Allen, B.: Knowled e or anization in an i nformat ion retrieval 3/4-464 726 task. Infonn. rocess. anagenrent o ,p. 5- Tazi . S. : Reconnaissance de la structure textuel le de') docu­ 20 refs. r�nts . (Textual structure recognition of documents) Orig. fro It Ita '> found that the Itay users organize their knowledge has 1fLiTT":Cent.Hautes EtuJ.int.infonn.doc. , No 35, 1989. p.37-45 an impact on their interaction with information sY'items. Be­ cause of this, such organizing structures may be candidates for inclusion in cognHive models of users. Other U'ier character­ 73 Automatic Language Processing, istics that may affect ;nfonnation retrieval can be investiga­ ted using th is type of experiment. (Author, abbr. ) See also 3/4-044 3/4-476 753 3/4-465 731; 196 Brakel , P.A.v.: Evaluating an intelligent ga teway: A rnethodo­ £. : ]Qgy. Mater, Zur automatischen Sprachverarbeitung und ko�nitiven 19"88 . Orig.en. South Afr.J.Libr.6 Infonn.Sci. Vol 56, No 4, lin,uistik. {Automatlc text proceS�'j l ng and c09nlbve Ilngulst­ p.277-290, 16 refs. ics Orig.de. Wiss.Z.Techn.Hochschl .Ilmenau Vol 35, No 5, 1989. p.27-33 753

new for the creation of subject bibliographles is proposed. It consists of two phases: first, the raw bibliogra­ phical material is downloaded from an online bibliographical

200 Int.Classif. 17(1990)No.3/4 Classification literature da�abase (e.g.DIAlOG) and then this material is processed 757 uSlng knowledge-based means. A meta-programming approach is applied in which the raw bibliographic material is viewed as a logic program upon which a second-order logic program is dev­ eloped. The second-order program creates the subject bibllo­ graphy by operating a rule base and the first-order logic pro­ 758 Online Activities in Subiect Fields gram. The entire system, named REX, was written in and See also 3/4-075, -155 used to create automatically a subject bibliographY on Applic­ ations of logic Programming in Decision and Control . (Authors) 3/4-489 758-26 Lynch, M.F. : The hiQh information technoloqy road to searching chemical structures. ur lg. 3/4-478 753 generic FG en. Palincsar, St. F. ; Garner, M.S. : Mapping the unmappable: Plumb­ Nitteilungsblatt Chemie-Information. No 16, 19B9. p.31-40 ing the depth of crossfile and cross-system navigation.

Orig.en. Online Vol 14. No 4. 1990. p.15-21, 4 refs. 3/4-490 .• 758-26 Techniques described provide possibilities for innovative Nilakantan. R. ; Bauman. N., Haraki. K.S Venkatarghavan. R.: cross file and cross-system searching. The procedures can be A ring-based chemical structural query system. J.Chem.Inform. replicated with most standard word processing software. & Comput.Sci. Vol 3D, No 1, 1990. p.65-68. 17 refs.

3/4-479 . 753 3/4-491 758-51/4 Whitcomb.l. : OCLC 's EPIC sys tem offers a new way to search the Van camp,A.J.: Subject code searching in biomedical databases. OClC database. Online Vol 14, No 1, 1990. p.45-50, 1 ref. Orig.en. Onl ine Vol 14, No 1, 1990. p. 90-92

754 Programs for Online Queries 76 Lexicon/Dictionary Problems See also 3/4-062, -071. -128, -290, -358

3/4-480 754 See also 3/4-297 Medvedev ,A.A.; Nikitin.A.I., Sushchanskaya. N.F.: A conversa­ tional system for maintaining a natural-conceptual d1ct10narv. 3/4-492 761 Or1g.ru. Na uch. -tekhn.inform. ,Ser.2, No 8, 1989. p.2-8, 14. Polikarpov, A.A.: Theoretical problems of appl ied lexicology. 5 refs. Orig.ru. Vestn.MGU. Ser.9. No 5, 1989. p.64-74, 29 refs.

3/4-481 754 762 Dictionary Structures Weye r, St.A.: uestin for the "DAO": Dow lIe<;t and intel li ent See 374-213 text retrieval . Online 01 13. No 5, 1989. p.39-48, 13 refs. 763 Construction and Updatin9 of Dictionaries 756 Classification S stems and Thesauri in Online Systems See also· 3/4-215. -217 ee a so a- , - , - , - -381 3/4-493 763; 768-446; -358 Berka, R.: pr pa ration of foreign terminological dictionaries. 3/4-482 756 Orig .cz. Zeme a. r:,"�f�o�n n;;C -_ �V�o�I-"'2 �8 �, �Nf,o:-i4�, �ln9, "8i l9r. �p �. 1"3H8E- �1 *4�2�, �2Gr �e ffs�.o · Clough.R. : ASSASSIN: The on oin develo ment of a text stora e and retrieval system. n orm.Serv. se go. p. 3/4-494 763; 768-825 3 refs. Pongoova, I.: Some aspects of lexicon modelling and its appl i­ cation in librarianship. Orig. sk. 756 Zemed .infonn. Vol 28. No 4, 1989. p. 131-137. 7 refs. Publikations- und Infor�ationssysteme. Ma tho u .Datenverarb. : 3/4-495 763 Sinha, R.M.K. : On artitionin a dictionar for visual text recognJtion. Pat ern ec09n1 lOn 0 p. 9 - on wood protection means 7 refs. No 1, 1990. p.34-35 764 Kinds of dictionaries See 3/4-152. -165

765/6 Monol ingual and Multi] ingual Automatic Dictionaries See also 3/4-208, -212, -214, -21 6

3/4-496 765 3/4-485 756.44 Chartron. G.: Lexicon management tools for large textual data­ Chan.l.M. : libra of Con ress class numbers in online catalo bases: The lexinet system. Orig.en. J. lnform.Scl, ,Principles searching. Or1g.en. RQ 28 1989 No.4, p.530-536. 6 refs. �a ctlce Vo l 15, No 6. 1989. p.339-344, 13 refs.

3/4-486 756; 298-51/4 3/4-497 .• 765 Wells. C.J. ; EVett.l.J., Whitby,P.E Whitrow.R.J.: Fast dict­ Rada. R. : Ranking documents 40,with a thesaurus. Orig. en. J.Amer.Soc.lnform.Scl. Vo l No 5.1989. p.304-31O, 16 refs. ionary look-up for contextual word recognition. Orig. en. Report on exploratory experiments in evaluating and improving Pattern Recognition Vol 23. No 5, 1990. p.501-508, 28 refs. a thesaurus through studYing its effect on retrieval. A formu­ la called DISTANCE was developed to measure the conceptual di­ 767 Data Bases in Dictionary Form stance between queries- and documents encoded as sets of the­ See also 3/4-209, -211 saurus terms. DISTANCE references MeSH (Med ical Subject Head­ 3/4-498 767 in9s) and assesses the degree of match between a MeSH-encoded 1. query and document. The performance of DISTANCE on t�eSH is EVens,N.: Book review of Boduraev. B. Briscoe. Eds. : Com­ compared to the performance of people in the assessment of utational fexlco ra h . orp na ura an ua e roceSSln . on­ conceptual distance between queries and documents, and is don: Longman 1989. XIV 310 . J.Ooc. 45 1989 No 4. p. 34-336, found to simulate with surprising accuracy the human perform­ 10 refs. ance. The powe r of the computer simulation stems both from the Collection of papers focussing on the development of diction­ tendency of people to rely heavily on broader-than (BT) rela­ ary databases for computer processing of Engl ish from the tions in making decisions about conceptual distance and from tapes of the Longman Dictionary of Contemporary English. the thousands of accurate BT relations in MeSH. One source for discrepancy between the algorithms ' measurement of closeness 3/4-499 768-82 SU between query and document is occasional inconsistency in the Makarenko,V.A. : TerminoloU SR.ical dictionaries on library science BT relatlons. Our experiments with adding non-BT relations to and bibliOgrap� 1n the S Dr lg.ru. leorlya 1 prakt.ob­ MeSH showed how these non-BT relations could improve document shchestv.-nauc .1nform., No 1, 1989. p.171-186, 5 refs. ranking. if DISTANCE were also appropriately revised to treat these relations differently from BT relations. (Author) 77 . General Problems of Terminology 757 Expert Systems See also 3/4-002, -070. -120, -122. -123, -124, -126. See also 3/4-235, -236, -506. -508, -509, -511 -168. -172. -194, -375 3/4-500 71.07; 79.1 3/4-487 757 Hohnhold, 1. : Obersetzun sonentierte Terminolo iearbeit. Eine Grundlegun9 fUr Praktiker. Translation-oriented terminology Endres-Niggemeyer, B.: Book review of Wormel l, I.: Knowledge •• work) Ori9. de. Stuttgart, DE: Intra 1990. 223p 80 refs. Engineering. Ex19p87.ert systems and information retrieval . london: laylor Granam mop . Url g. de. The book has four chapters: Subject translation and termi nol­ Nacnr.Dok. vo l 41, No 3, 1990. p. 196-197 ogical work; Contents-oriented and methodological aspects

Int.Classif. 17(1990)No.3/4 Classification Literature 201 of translation-oriented tenninology work; Organisation of 3/4-518 78-82 translation-oriented terminol ogy work. In Section 2 a chapter Tittenbrun, K. : Terminolo�ical tower of Babel. Orig.pl. Aktual . on Subject fields and subject field classification systems probl . infonn. i dok. Vol 3 , No 1-2, 1989 . p.22-26, 14 refs. can be found with a number of requi rements listed of such a To overcome the present situation in the information sciences system for tenninol ogy work. with the ambiguity of basic tenns, it would be desirable to create an international service for the registration of new 3/4-501 77.07 tenns and for an exchange of infonnation about them. Mohn,D. : Book review of Felber, H .. Budin, G.: Terminol09ie in Theorie und Praxl s. TUbl ngen 1989. (Tennlnology in theory and 3/4-519 78-82; 09.77 practice) ang.de. Fachsprache Vol 12, No 1-2, 1990. p.B2-83 Klodnicka,H. : Polish tenni noloqical standards in librarianshi� and I,bibliogr aphy. Orig.pl. Aktual . probl .inform.dok. 34(1989 3/4-502 77.179; 06.89-05-31/1 No p.38-39 Datta, J.: Workshop at Infoterm targets gu idelines for termin­ ology plannlng 1n the developlng countnp.s. On g. en. 3/4-520 78-82; 99 CH TermNet News, No 27, 1989. p.1 I, Zhou Zhiou: The erogress of termi nology of information and do­ The workshop took place on I·lay 31-June 1989 at Vienna and cumentdtion and Its standardI Zation 1n Ch ,na. Dn g. en. had participants from Africa, Europe, and North America. TermNet News, No 25, 1989. p.9-10

3/4-503 77.925 3/4-521 78-825 Reichling,A.: la creation de lexiterm. I,(The creation of lexi­ Kolesinskiene, N.: library terms. Orig. 1i. Biblioteku darbas, term) Orig.fr. Termi nol.& Traduct.. No 1989. p.lll, 172 No 6. 1989. p.21-22

3/4-504 77.99 3/4-522 78-83 Scientific-technical terminology: Collection of abstracts. Kuendig, A. : Rechnernetze: Begriffe, Konzepte, Anwendungen. Orig.ru. Nauch.-tekhn.terminologiya, No 10, 1989. 27p. (Computer networks: tellninology, concepts. applicatIons) Orig.de. In: Handbuch d.modernen Datenverarb.. Vol 26. No 148 3/4-505 771 iesbaden, DE: Forkel 1989. p.3-31, 32 refs. (8anque de Mots. Special number of the Centrp. de Terminologie et de Neologie (CIN) ). Orig.fr. 1989. p.5-129 W3/4-523 78-834; 778 3/4-506 Guespin, l.; laroussi, F.: Glottopolitique et standar­ VI) .; Dobrokhutova. I. Laufer, N. I.: �I !!.nf� o!.'n�n "a t" i�o, !!n_t�e�cO!!h!!.no'Cl�0' 9:tlly... !a�s disat10n tenninologique. - 3/4-507 Chuckwu , Y. ; Thoiron. P. : mirrored in a dictionary. Orig.ru. Refonnulation et reperage des termes. - 3/4-5 lerat, P. : les Nauch.--tekhn. Jnfonn. ,Ser.2, No 9, 1989. p.36-38, 7 refs. fondements thiwriques de la tenninologie. - -509 Otman, G. ; Terminologie et intel ligence artificiel le. - 3/4-510 Humbley. 79 Problems of Multilingual Systems J.: Terminologie et conscience linguistique.

771 See also 3/4-166. -195, -196, -207, -222, -224, -225, -227. -500 W : 3/4-524 791 ;1�����'��" ������10th F��icAO Congr.&p. Gen. Perschke, S.: EUROTRA. Orig.en. In: f-la chine translation summit AssenoblY.� 24-26 Oct. 1988. Hongkong 1988. E34-E38 2, Aug.16-18, 1989. Munich, DE. Frankfurt/Main: Deutsche Ges. fUr Dokumentation e.V. 1989. p.148-154 772 Fonn and Designation of Terms Overview of the mach ine translation project EUROTRA of the See 3/4-171, -173. -229, -507 Comm ission of the European Communities including its prin­ ciples (multilingual transfer-based system, sentence based 3/4-512 773 deep syntactic representation) and its stratificational ap­ Ootuka, H.: Rol e of tenninological knowledge in informa tion proach which subdivides linguistic descriptions in a number activit : a case stu in Ja an; an a lication to a knowled e of more or less autonomous dimens ions. Future plans foresee trans er centre. ng.en. n: ur on, ., ong, .: e use 0 the start-up of the industrial development, development of new information technologies in develop-ing countries. Proc. methods and tools for the reusability of lexical resources FID/CAO 10th Congress & Gen.Assembly, Beijing, 24-26 Oct.1988. in computerized applications, and the creation of standards Hongkong 1988. p.E27-E33 for lexical and terminological data.

776 Terminol ogical Data-210, Banks APPLIED CLASSING AND See also 3/4-099, 8 -226, -230 INDEXING (C I) 3/4-513 776 &: Stock,W. : Textwortmethode und Obersetzungsrelation. Eine �\etho-, 81 P�oblems of Applied C & I de zum Aufbau von komb,nlerten Llteraturnachwels- und Termino­ Ge.,er"-I logiedatenbanken. (Textword method and translation relation. A 812 Alph.& Classified Subject Catalogs method for the construction of combined literature-record and See also 3/4-035, -036, -080, -088 terminology databases) Orig.de. ABI-Techn. Vol 9, No 4, 1989. p.309-313, 14 refs. 3/4-525 d 812 Anisimova, G.M. : The use of main card index in the provi sion 777 Country and Language-oriented-234. Termi -237,nolo -5 I-Iork of information services to researchS, 1989. staff. Orig. ru. See also 3274-228, -233, H Na uch.-tekhn .lntorm. ,Ser.l, No p.20-21

3/4-514 777 FR 3/4-526 812 Groult, �l .: l'histoire du vocabulaire scientific en France. Olesova, E.S. : The conception of classed catalogue editing. (History of France's sC ient,f,c vocabulary) Orlg.fr. Orig.ru. In : Nauch.katalogiz.lit.i org.katalogov. Moskva, SU. TermNet News. No 27, 1989. p.2-4, 11 refs. 1988. p.105-112, 7 refs.

778 Special language Research 3/4-527 812 See 3/4-523 Rozenshi l'd, l.K. : Towards the problem of the classified cata­ logue and its place wlthln the tradlflonal Ilbrary structure: 78 Subject-Oriented Te rminology Work a conception. Orig.ru. In: Nauch.katalogiz.lit.i org.katalo­ gOY. Moskva , SU 1988. p.6-24. 39 refs.

3/4-516 78-775 3/4-528 812 Mironesko, E.M. : Semantic calquing as a \tay to fonn technical Rozenshil 'd, L.K. ; Usanova, LA. : The classed catalogue: its terms. A case study of the ra,lway terminology. Ong.ru. In: functional efficiency and place in the hbrary structure. lingv.issled.,1988. Problematika vzaimodejstviya yaz.urovnej . Orig.ru. In: Nauch.katalogiz.lit.i org.katalogov. Moskva, SU Leningrad, SU 1988. p.136-142, 14 refs. 1988. p.25-51, 4 refs .

3/4-517 78-814.09 3/4-529 812' 489 Augstkalns, J.A.: Toward the standardisation of classification Shapochnik, E._: The structure of the s stematic literatur�0 for tenninology. Orig.ru. In: Nauch.kataloglZ.lit.i org.katalogov. students, 8-9 ye ars o . ng.ru.- lote ar Moskva. SU 1988. p.152-157, 10 refs. 1988. p.60 Criticims concerning GOST 7.26-80 and GOST 7.39-82 on classi­ fication terminology as they do not seem to meet international 3/4-530 812.213 requirements. Ahlhelm, D.: Der Schlagwortkatalog fUr staatliche Al lgemeinbi­ bliotheken und Gewerkschaftsbibliotheken. (The subject catalog 78-82 Terminolo;& -239in the Information Sciences tor governmental general hbranes and union libraries). Orig. See also 3 de. Bibliothekar (leipzig) 44(1990)No 3, p. 139-143, 3 refs.

202 Int.Classif. 17(1990)No.3/4 Classification literature 814 Manuals. Rules. and Codes for Subject Cataloguing 3/4-543 863 See also 3/4�143 Nikitkina, N.V. : Toward the registration of subject indexes to Russian journals and serials 'in Russia and the USSR. Orig.ru. 3/4-531 814 In: Aktual .probl .bibliogr.bibliogr. leningrad,SU 1989. p.75-82, Unteregger�Vaerst. A.: D'ie "Reine Lehre" der RSWK: Besuch in 14 refs. den SWD�Redaktionen AUgsburg und MUnchen. (The "pure lesson" of the RSWK: visit to the SWO editorial departments in Augs­ 864 Citation Indexing See also 3/4-180 burg and Munich) Orig.de. Mitt. Vereinigg.Osterreichischer Bi­ bliothekare Vol 42. No 3. 1989. p.79-90. 8 refs. 3/4-544 864 Bawden, D. : Citation indexing. Orig.en. In: Manual of Online 3/4-532 814 Search Strategies. Boston. MA: G.K.Hal l 1988. p.44-83 Unteregger-Vaer5t, A.: RSWK:(RSWK: rette sich wer kann oder Revolu­ tion 1m Schlagwortkatalog. save yourself. or Re volut10n 864 in the subject catalogue) Orig.de. Mitt.Vereinig.tlsterreich. 8ibliothekare Vol 42_. No 3. 1989. p.104�108. 2 refs.

3/4-533 814 Simpkins. J.: How the pu bl i5hers want it to look. Orig. en. Indexer Vol 17. No 1. 1990. p.41-42 , 2 refs. 87 C 8< I onNon-Book Materials "It" is the index to a book commissioned to an indexer. A num­ ber of publishers have been asked for the specifications re­ garding an index to the ir book s. The result of this investiga­ 872 Picture Indexing, including PhotographS tion is presented here, di5tinguishing between 11 areas of ge­ See also 3/4-053 neral unanimity of requirement and 33 more specific instruct­ ions of 8 publishers. 3/4-546 872 Botha, W.�I. : A contribution towards indexin of hoto ra hs. 3/4-534 818 Orig.df. S.Afr.J.libr.& Inform.Sci. 56 1988 No 4. p.323-331, Aczel .E.: The national library's Ale1h subject heading catalog 6 refs. Orig.he. JSllC Bull. Vol 16, No 2. 988 . p. 46-49 An IR SY5tp.1l1 for a general collf>.ction of South African photo­ graphs -is described. Al though it was inspired by the principle 82 Data C8<1 appl ied in the Radio Times Hulton Picture Post Library. it has been developed as a completely new system. {Author}

See also 3/4-114. -115, -116. -117, -118 88 C 8< I in Subject Fields

3/4-535 •• .• 82-382 Libmann, F. ; Minasyan, N Duran J.-P. ; Commission of the Eu­ 3/4-547 88-39 ropean Communities, Dir.Gen.XIII, Telecommunications, Informa­ Conger. L.D. ; Shipley, R.M.: The megatenn challenge . {Negaterm tion. Indu5tries and Innovation: Demonstrationsprogramm Werk­I indexing espec ially for geographic names). Orlg. en. stoff-Datenbanken 1988-1989. (DemonstratlOn program matena s Database. No 12. 1989. p.120-2 databases 1988-1989) ong. de. Luxembourg. LU 1989. 41p. The program including the data about materials from eleven 3/4-548 88-397 US European data bases is meant to fonn a prototype for the pro� Binder,J. E.; Gustafson .N.M. , r'le rritt,M.: Geo logi cal/geographi­ motion of 5tandardization, integration and publ ic relations cal name subject access. Orig. en . about such data. Three development5 have been accompl ished: A Inform. Technol.& Uhr. Vol 8, No 4, 1989. p. 408-421 , 23 refs. User Guidance System, a Conunon Reference Vocabulary, and A 3/4-549 List of Infonnation Sources (D{)4IS). Tables show the different 0.: 88-51/4I, application fields, problems, materials properties and kinds Blake. Indexing med-ical - journals. Indexer Vol 17. No of materials. 1990. p.33-3q --

84 C 8< I of Primary Literature (Except 85) 3/4-550 88-634 SU f1en'shikh. V.A. : Computerised content analysis of documentary informa tion in pa rty bo dIes' computerlsed 1nformatlon systems. 3/4-536 841 Orlg.r-u . In: r1ezhotrasl.1nform.sluzhba. Moskva 1989. p.8-11 Engel , M. : SacherschlieBung in medizinischen und naturl'/issen� schaftlichen Dissertationen an der Universitatsbi'bliothek der 3/4-551 L.: 88-66 CA FU. (Subject analysis of med1cal natural-science disserta­ i3jarnason NacPherson, "I!!nd":e" ,x"i n9,-;�C: :\an� a"d"i !,-a nm l"ei!9a:clr '1 .:,i ti' e"r, ,a ;;t u", r""e . and * - tions at the univers�t.•y library of the Free ·University Berlin) Orig.en. Law Uhr.J. Vol 81, No 3, 1989. p.489-504, 74 refs. Orig.de. Bibl .-Infonn No 19, 1989. p.16-18, 2 refs. 3/4-552 88-934 3/4-537 842 'Holden, P.: IndexinG and retrieving chamber mus ic. Brio Vol Ichinowatari.K. , Miwa,M. , Motohashi .H. , Matsuyama ,S. , Idezoe, 25. No 1. 1988.P.21::Z4 M.: Analysis of patent information: Trends in housing �lan./, Description of the indexing system at the Guildhal l School of Orig.Ja. Joho hann (J.Inform.Process.& Manag .) vol 32, 0 f1usic and Drama l"ibrary which al lows the searching of virtual ­ 1989. p.583-595, 4 refs. ly all instrumentdl combinations when combined with the appro­ priate fac ilities of retrieval software. 3/4-538 842.39 Sluckin. A.A. ; .MarS-ak, N.A. , Borisova,I.N., Lukanin,V.A. : More 3/4-553 88�93 effective retrieval of atent letters accordln to biblio ra­ Kl emm, G.: Indexierung mit dem Adjektivzirkel . �1!jg lichkeiten p lC escnptlons.4 ng.ru. auc .-te n.Inform. , er. , 0 zur Erfassung musikalisch�expressiver Qualitaten in der Musik­ 1989. p.lI-D. refs. dokulllentation von Rundfunkarchiven. (Possibilities of indexing of musically-expressive qual ities in documenting the mus-ic of 3/4-539 844 broadcasting archives) Orig.de. Info 7, Baden-Baden Vol 4. Cramer.M.D.; Markland, M.J. : Newspaper indexing with Pro-Ci te. No 2, 1989. p.72-77. 13 refs. Orig.en. Coll.& Res.Libr.News , No g, 1989. p.807-8D8

3/4-540 844 NI CLASSIF, ENVIRONMENT L£. : 9 �.Okorafor, Indexin Ni erian news a• ers for effective re- Inform. eve opmen 0 p. re s. 92 Persons and Institutions 3/4-541 849 Li Caifing: Classification and cataloguing of multivolume pu ­ 3/4-554 924 Western languages. orig . blications in ch. Yates-�lercer, P.: AnJ. a 1preciation of JasonIS, Farradane, 29 se�06t. Tushu qingbao gonzuo (Ubr. Inform.Serv.), No 3, 1988. p. 37-38 1906-27 June 1989. ntorm .Sc1. Vol No 6,1989. p.30S- 85 Book Indexing (Guest Editorial). 3 refs. See 3/4-558 3/4-555 925; 43 Chan, M.L. : Decimal 1, Classification Editorial Policy Committee 86 C 8< I of Secondary Literature Annual Report July 1988-Ju ne 30. 1989.I, Orl g. en Libr.Resources & Techn.Serv. Vol 34, No 1990. p.l03-105

3/4-542 863 3/4-556 925; 43 Mamontov, M.A. : Bibliographic control of indexes to bibl iogra­ Intner, Sh.: Cataloging and Classification Section.34. 1988-89 �hic tools in the USSR: Present state and tasks . Ong.ru. Annual Report:-Llbr .Resources & lechn.Serv. Vol No 1.1989. n: Aktual.probl.blohogr.b1bhogr. Lenlngrad. SU 1989. p.67-74 p.100�101

Int.Classif. 17(1990)No.3/4 Classification Literature 203 98 User Studies 3/4-557 925 Ching Walu, J.H.: SCECSAL 7 and DOC 20 : a voice crving in the wil derness. MALA Bull. Vol 5, No I, 1988. p.61-63. 3 refs. See also 3/4-058, -061, -130

3/4-558 925 3/4-563 982 J.: Library Jargon: student comprehension Mulvany, N.C.: American8, 1990. indexers in13 the 19905. Orig. en. Naismith, R. ; Stein, Library Work. No p.16-19. refs. of technical lan ua e1 used b librarians. Orig. en. Text adapted from a speech del ivered to the Society of Index­ o ege e'iearc r. 1989 Sept. p.543-552, 14 refs. ers Conference at the University of Edinburgh, 7 April 1990. Jargon, or technical language, appears in both the writing and Three main topics were treated: The American publ ishing in­ speech of librdr-jans in the ir dealings with the publ ic. If the dustry, the markets for indexing, and the role of profession­ library patron cannot comprehend this language, the reference al indexing associations. transact ion is impeded. Thi s study measures students' compre­ hension of a selection of library terms that were derived from 926 International Societies and Groups actual reference interviews and library handouts. A mul tiple­ See also 3/4-503 choice test was administered to a group of freshmen. "Think­ ing-aloud" or protocol analyses were also run for infoY1Jla tion 3/4-559 926 as to the rea�;oning processes used by the subjects. Given the results of this testing which show that patrons misunderstand Hol ley, R.P. :_ Report from the Section on Classification18, and4, Indexing, 1988-89. lnt.Cat.! Blbilogr.ControJ Vol No library terms approximately half of the time, the researchers 1989. p.S9 offer librarians a continuum of solutions. (Authors)

3/4-560 926 986 Use of Subject Catalogues See 3/4-096 Ho' ley,R. : Classification and Indexing. Annual Report 1988/89. Orig.en. IFLA Annual 1989. p. 138-141 3/4-564 987 Jackson . K.M. : Load-ing Wilson indexing locallt - The Texas A&M 93 Organisation on a NatI. & Internatl. Level experience. Online Vol 14. No 4, 1990. p.42-4 To -improve user access especially for students the Sterling C. See 3/4-089. -095 Evan'i Library at Texas A&M loaded three Wilson databases as one file accessible via the "arne interface as the online cata­ log. Included were "Applied Science and Technology Index". 95 Education and Training "Biological and Agricultural Index", and "Business Periodicals Index". The Wilson databases are available on 80 NOTIS termi­ See also 3/4-167, -232 nals located in public services areas of the library.

3/4-561 953 3/4-565 988 J.�1.: How to use the major indexes to U.S.Government pu­ Edmunds, H.: Superclass6, 1989 consciousness.. Or-i9. en. Ross. Library Work, No p. 15-16 blications. Chicago, It: Aii1er.Libr.Assoc. 1989. 35 p. , Classification is not restri cted to library materi al<;. r1any ISBN 0-8389-0509-9 people need to be able to class ify produc ts, services, and a Introductory guide to help in looking up subjects in the six whole range of other information records which comprise sub­ "Inost important" indexes : the Monthly Catalog, CIS/Annual (in­ ject structured data bases. Describes a classification work­ dex dnd abstract volumes), the CIS Leg islative Histories of shop for a database of learning of portunities related to the U.S. Public Laws, the Congressional Record Index, the American Trainin9 Access Point Project (TAP . (Author) Statistics Index. and the index to U.S.Government Pub1 ica­ tions. For each index there is an explanation of its organi­ 3/4-562 959 zation, an example demonstrating how to find a spec ificd docu­ 8rundage ,C.A. : Teaching control led vocabulary and natural lan­ Illent, and an exercise in which users are asked to find simi­ guage to end-users of scientific online and Cu-R�1 databases. lar document by con'iulting the actual index. Orig.en. SC 1.& Techn .Libr. 1O(1989)No 1, p.3-13, 12 refs. J The paper reviews the similarities between scientific language 989 Use of CS & in Certain Institutio� and search language using controlled vocabulary or natural See 3/4-091 , -093, -097 language. It di scusses the nature of end-users in science li­ 99 Standardization braries, types of infoY1Jlation needs, and methods to teach the use of appropriate vocabulary. �ee j/4-159. -506, -510, -519. -520

Personal Au.thorIndex Berkovich.E.H. 3/4-284 Boulanger,D. 3/4-342 Cawkel l.A.£. 3/4-337 !Jerman, S. 3/4-387 Boyce. B.R. 3/4-484 Chan,L.�1. 3/4-030 077a 3ertrand-Gastaldy, H. Brachman,R.J. 3/4-050 -485 -555 3/4-220 Brajnik, G. 3/4-060 Chang ,J.W. 3/4-332 Abramova, N.N. 3/4-336 Auer. G. 3/4-089 Biebricher,P. 3/4-057 Brake-Gerlach. U. Chartron, G. 3/4-496 Aczel , Eo 3/4-534 Augstkalns,J.A. 3/4-517 Bies. W. 3/4-143 3/4-094 Chiaramella, Y. (Ed.) Ahlhe1m. D. 3/4-530 Avaliani.S.Sh. 3/4-272 Binder, J.E. 3/4-548 L. Brakel , P.A.v. 3/4-476 3/4-040 Averbukh, K.J. 3/4-171 Akeroyd, J. 3/4-470 Bjarnason r4acPherson, Brenner, S.H. 3/4-364 Ching Walu,J.H.Y. 3/4-557 Al-Hawamdeh ,S. 3/4-301 3/4-551L. Brown, M.L. 3/4-310 Choueka, 3/4-074 Albrechtsen,H. 3/4-136 Bakarsic, K. 3/4-180 Bjorklund, 3/4-150 Brundage,C.A. 3/4-562 Christianson, E.B. Allen,8. 3/4-361 -475 Bara, 1u. 3/4-280 3/4-151 Buchan.R.L. 3/4-199 -297 3/4-039 Amma r, M. 3/4-312 Barnhart, L. 3/4-019 Blair, D.C. 3/4-362 Budd, J.t4. 3/4-271 Chuckwu , Y. 3/4-507 Barthes, Ch. 3/4-070 Anan'eva,LN. 3/4-129 Blake. D. 3/4-549 Budin, G.L 3/4-169 Citkina. F.A. 3/4-174 Andre, J. 3/4-459 Baudoin de Courtenay, Bland, R.N. 3/4-035 Bulot. 3/4-190 Clough, R. 3/4-482 Andreeva ,E.V. 3/4-427 I.A. 3/4-276 Bobko, A.N. 3/4-409 Col lins. W.P. 3/4-448 Andrews, D. 3/4-401 Bauer, I. 3/4-383 Bobrov. A.1. 3/4-251 Calderon. F. 3/4-397 C6maromi ,J.P. 3/4-014 Anisimova .G.M. 3/4-525 Bawden, D. 3/4-544 Bonitz.M. 3/4-110 -270 Canisius. P. 3/4-377 -038 (Ed.) -015 Antoniadis,G. 3/4-044 Beaujean, L.M. 3/4-091 80nzi,S. 3/4-043 -454 Carlucci ,A.L. 3/4-252 Conger, L.D. 3/4-547 Arntz. H. 3/4-131 Bel iaeva, 3/4-208 Bookstein, A. 3/4-073 Case, D.O. 3/4-047 Cramer, M.D. 3/4-539 Atkins,Th.V. {Ed.) Belkin, N.J. 3/4-048 Bosc, P. 3/4-064 Castel lanos, P.J.B. Craven, T. 3/4-145 3/4-009 Benn. W. 3/4-181 Botha, W.N. 3/4-546 3/4-253 Croft. W.B. 3/4-067 Attias, R. 3/4-423 Berka . R. 3/4-493 Boui llon. P. 3/4-211 Catl in, C. 3/4-348 Crouch, C.J. 3/4-055

204 Int.Classif. 17(1990)No.3/4 Classification Literature