Newsletter of the Section on Classification and Indexing

No. 20
December 1999

To all members of the Section

Ia McIlwaine
Chairperson
IFLA Section on Classification and Indexing

Standing Committee of the Section on Classification and Indexing

The Standing Committee has one corresponding member: Mandana Sadigh-Behzadi (National Library of Iran, Tehran, Iran) and one observer: Marie-France Plassard (IFLA UBCIM Programme, Die Deutsche Bibliothek, Frankfurt am Main, Germany).

Committee members mailing list

The following members have signed up for the Subject Access of Web sites and Digital Libraries working group:

Lois-Mai Chan
Friedrich Geisselmann
Gerhard Riesthuis
Irina Tsvetkova
Pia Leth
Jon Anjer
Marcia-Lei Zeng (Chair)
Ia McIlwaine
Edward Swanson
Dorothy McGarry
Julianne Beall

The following members have signed up for the Multilingual Thesaurus Guidelines working group:

Gerhard Riesthuis, chair
Lois-Mai Chan
Marcia-Lei Zeng
Friedrich Geisselmann
Julio Vaz dos Santos Rodrigues
Marje Aasmets
P Vidanapathirana
Magda Heiner-Freiling
Dorothy McGarry
Julianne Beall
Ia McIlwaine
Edward Swanson

If you wish your name be added to or deleted from either of the above two lists, please send an email to mzeng@kent.edu at any time.

Developments in classification and indexing in Estonia, 1998/1999

080 792.2Ix(474.2)Ix(091)I2MRF (history of the drama theatre in Estonia)

The short UDC number for the national bibliography is located in the 072 field. The abridged scheme (129 subdivisions) of the UDC was developed for this purpose in Nov.-Dec. 1998.

The 600-69X fields are used for subject headings. Topical terms are located in the 650 field: 651 - geographic(al) names; 653 - chronological terms; 655 - form/genre terms; 695 - special topical terms of the Estonian Academic Library (EAL); etc. All the terms/subject headings which are taken from the Estonian Universal Subject Headings List have the indicator 9.

The Estonian Universal Subject Headings List (Eesti üldine märksõnastik (EÜM)) was published by the National Library in June 1999. It contains 21,891 terms (on 59 themes). In addition to it, a separate list of form/genre terms has been compiled. In EAL a primary list (about 150 terms) had been made up on the basis of the library's alphabetical subject catalogue. The final list contains about 220 terms. It was worked out by the working group and disseminated in libraries.

The Working Group (WG) on classification and subject indexing of the Estonian Librarians' Association has the function of co-ordinating and promoting the classification and indexing activities in Estonian libraries.
An article about the WG in 1992-98 has recently been published by the WG's chairperson: Kirsel, A.-M. Liigitamise ja märksõnastamise toimkond [Committee on classification and indexing] // Eesti Raamatukoguhoidjate Ühingu aastaraamat, 10 (1998), Viljandi, 1999, pp. 54-58.

The greatest efforts of the WG in the 1990s were aimed at preparing the translation of the UDC into Estonian. The major task up to now has been to create the Estonian version of the UDC on the basis of the UDC MRF. The work was completed but publication of the tables was suspended for more than a year. Now the hopes are high again that the printing may be accomplished by the end of 1999.

The existence of UDC tables in itself does not guarantee the unity of classification decisions by different libraries. Unfortunately, INNOPAC does not permit the creation of a UDC authority file at present, so the WG started to unify the classification methodology in 1997/98. During 1998/99 classes 004 and 65.01 have been under discussion, finally resulting in definite decisions.

With regard to class 004 - Computer science, problems have been met on account of the newness of the class. Subclass 004.9 could have a more detailed structure. Compound indexes for classifying several themes , e.g. learning networks (004.78:378), teleworking (004.78:331), teleshopping (004.78:658.8), etc. have been recommended. Recommendations concerning 11 items have been recommended and circulated among libraries.

Part of the classification problems concerning 65.01 (theory and practice of organization) is caused by its relationship with 658 (business management, administration). Later on the interdependence between classes 65 and 33 (esp. 331, 334.7 etc.) will be discussed. The WG is continuing in this direction.

UDC-news has been treated in lectures and seminars. The EAL organized a second UDC day on October 28 1998. The themes of the day were: the development of UDC in 1997/98; similarities and differences of classes 33 and 65; the structure of class 39. Thirty-six librarians from all over Estonia, mainly from public libraries, participated in the event.

Finally, I would like to point out that a lot of work has been done to get INNOPAC going.

The methodology of the classification is rapidly changing in the period of transition from card catalogues to electronic ones.

As Estonian librarians are less experienced in subject indexing, more attention has been paid to this sphere. The unification of subject headings among the libraries of the ELNET Consortium is supervised by the National Library.

The WG and EAL strive for the unification of UDC classification decisions. There are still problems in connection with the depth of the classification of various library materials, e.g. monographs, periodicals, dissertations, conference materials, collective works and fiction.

Marje Aasmets
Estonian Academic Library

Subject indexing and classification in France, new developments 1998-1999.

The planned developments have been implemented, especially with the total revision of a new edition of the Indexing manual ("Guide d'indexation"). This will be published in November 1999 and will contain a more significant theoretical part, presenting the principles of the indexing language, as well as all the free-floating subdivisions and indexing rules, distributed through large subject fields. (For further details about RAMEAU, see its presentation in the IFLA document on "Principles underlying Subject heading languages (SHLs)")

The new developments have also been concerned with a better access to RAMEAU. Authority records on CD-ROM ("Les notices d'autorité de BN-OPALE") can from now on be consulted in three different formats: in "lite" (thesaurus format), INTERMARC and UNIMARC. They are also accessible on the Web site of the Bibliothèque nationale de France (http://www.bnf.fr) through the new catalogue BN-OPALE PLUS.

A new complete edition of the RAMEAU products will come out in next November. With regard to these products, a large survey was carried out in June among all the institutions of the network. On the basis of results obtained, with more than 300 replies, the proposed range of products will be reconsidered and supplemented soon (Internet access, etc.)

In 1999, international co-operation also continued with our traditional partners and was more intensified. It will be particularly strengthened with Belgium as part of the cultural co-operation agreement signed in March between France and the French community of Belgium.

And, moreover, it has been extended to French-speaking Arab countries. The national library of Tunisia and some other important Tunisian libraries joined the RAMEAU network as active partners at the end of 1998. This co-operation should lead in the next years to an adaptation of RAMEAU in Arabic.

Lastly, it is necessary, of course, to emphasize the results obtained by the working group on multilingual subject access which enabled a very successful co-operation between the British Library, the BnF, the Deutsche Bibliothek and the Swiss National Library.

The final report of our working group (March 1999) shows the feasibility of linking our three SHLs (LCSH, SWD, RAMEAU) to offer multilingual subject access to bibliographic files. On the basis of this report, the MACS project (Multilingual access to subjects) was approved and a prototype will be developed.

This collective report was presented during this IFLA Conference (Cataloguing workshop, meeting 155, Thursday, 26 August) by its reporter, G. Clavel (SNL), under the title : "The need for co-operation in creating and maintaining multilingual subject authority files".

With regard to classification, issues related to the translation of the 21st edition of the Dewey Decimal Classification into French were exposed last year during a workshop of our Section held at IFLA Amsterdam. The proceedings of this workshop have been just published in a bilingual version by Forest Press ("Classification décimale Dewey : perspectives francophones / Dewey Decimal Classification: Francophone Perspectives").

This translation of DDC 21 has been followed by continued work with the Dewey editorial staff. This has enabled the proposal of some important changes in DDC concerning specific points for France , which had never been made before.

Thus, new numbers have been developed for division 944 (French History), as well as a complete revision of the area Table for France (-44), which from now on will be better suited to the reality of the administrative organisation of the territory. The French edition of the DDC in French will be available on CD-ROM next autumn.

Max Naudi
Bibliothèque nationale de France

Indexing and Classification in Germany 1998/99

The rules are now in use in four countries: Germany, Austria, Switzerland (since 1998 by the Swiss National Library) and the northern part of Italy.

A new edition of the collection of examples was published in June 1999 (Beispielsammlung zu den Regeln für den Schlagwortkatalog (RSWK) /Bearb. von Gerhard Stumpf unter Mitarbeit von Hans-Jürgen Schubert. 2.vollständig neu bearbeitete Auflage nach der 3. Auflage des Regelwerks. - Berlin: Deutsches Bibliotheksinstitut, 1999. ISBN 3-87068-615-4).

The Deutsche Bibliothek (DDB), Frankfurt, Schweizerische Landesbibliothek, Bern, the Bibliothèque Nationale, Paris and the British Library, London work together in a project MACS (formerly MUSE) for crosswalks between the German, French, and American subject authority files (SWD, RAMEAU, LCSH). In the first step the subject headings in the fields of theatre and sports were examined. Geneviève Clavel-Merrin reported at the IFLA conference at Bangkok.

Classification
A new Working Group on Classification was established, which deals with the problems of a German edition of DDC. Members are Magda Heiner-Freiling (DDB, Chair), representatives from different German union catalogues, from Switzerland (Schweizerische Landesbibliothek, Bern) and Austria.

The task of the working group is a feasibility study. Up to now it has not been decided whether the DDB will use DDC. If we use DDC in Germany we will make as few changes as possible. We will store decomposed Dewey numbers in the records. The German edition should be bilingual. Concordances to German classification, e.g. RVK, should be made.

The use of UDC is declining - it is now used in about four or five big libraries in German- speaking countries.

The Regensburger Verbundklassifikation (RVK) is spreading wider. About 60 university libraries in Germany, Switzerland and Northern Italy use it.

Friedrich Geisselmann
Universtät Regensburg

Subject Cataloguing in the Netherlands 1998/'99

1. The UKB (the consortium of the Royal Library and the University Libraries) and PICA decided to continue the Common Subject System GOO used in their catalogues and in the National Catalogue for the next five years. A third edition of the Classification for this system - the Dutch Basic Classification - has been published. This classification is also available on the Internet at the site of the Royal Library (http://www.konbib.nl). An English translation is in preparation. English and German translations of the second edition are available at the same Web site. The Thesaurus of the GOO system (see the paper of G. Riesthuis and R. Storm at IFLA Amsterdam 1998) is still growing with about 200 descriptors per month, mainly proper names. An English version of the Thesaurus will be available in the near future.

2. Several new thesauri were published in the last few years, among them a new edition of the Thesaurus for the Public Libraries, the Biblion-Trefwoorden Thesaurus. Further I would like to mention the English edition of the Vrouwenthesaurus [European Women's Thesaurus. A structured list of descriptors for indexing and retrieving information in the field of the position of women and women's studies. Amsterdam : IIAV, 1998. ISBN 90-6981-012-3. - Not available on the Internet]

3. Several updates of the Dutch edition of UDC have been published, based on the Extensions and Corrections to the UDC published in 1997 and 1998.

4. The first edition of "Woordsystemen : theorie en praktijk van thesauri en trefwoordsystemen / onder red. van H. Magrijn … [et al.]. - Den Haag : NBLC, 1997" was sold out within two years. A revised edition will be published at the beginning of 2000.

5. The Royal Library in collaboration with other libraries maintains a reviewed, classified system for access to Internet sources, called DutchESS (Dutch Electronic Subject Sources). The classification used is the Basisclassification. It is also possible to search with keywords. The URL for library and information science is www.konbib.nl/dutchess/06.

Gerhard Riesthuis
University of Amsterdam

Report from Norway 1999

1. The most important development in Norway is that the translation and editing of the new Norwegian abridged Dewey Decimal Classification, based on DDC21, finally can start. The contract between the National Library of Norway and OCLC has been signed. The Norwegian Directorate for Public Libraries is co-financing the project. The project period is estimated to be one year with the possibility of some extension. The editor will start her work at the beginning of September this year. A steering committee and a consultative committee - the latter based mostly of members from the Norwegian Committee on Classification and Indexing - have been established. Subject specialists from libraries using Dewey will be consulted for terminology. The Web pages of the project (in Norwegian only) can be found at: http://www.nb.no/html/dewey.html

2. From 1999 The Norwegian Committee on Classification and Indexing has been subordinate to the National Library of Norway. New terms of reference have been made. The newly appointed committee will elect its chair at its September meeting. The web pages of the committee (in Norwegian) are located at: http://www.nb.no/nkki/

3. The seminar "Days of Knowledge Organisation" is a yearly event at Oslo College. This spring the theme was "Collection development in a new age". The aim was to discuss integrating internal and external resources in the online catalogue, with retrieval through classification and indexing as a central point. Some presentations can be found at: http://www.hioslo.no/bibin/kurs/korg99/ - but only a few are in the English language.

4. The National Library of Norway has developed a searchable subject index based on the Norwegian abridged Dewey Decimal Classification. The index, which is based on chain indexing, is compiled in connection with the classification of national bibliography entries from 1984. The index additionally allows searching to document descriptions in the databases Norart (Norwegian index to periodical articles), Norbok (Norwegian national bibliography), Norper (Norwegian list of serials) and Sambok (Norwegian union catalogue of monographs). The location is http://www.nb.no/baser/emner/

5. The database Nordic library information from the National Library of Norway gives detailed information about Norwegian libraries. The directory will be revised this autumn. Among the improvements is the provision of detailed information about classification and indexing systems used in the individual Norwegian libraries.

Jon Anjer
Oslo College

Report from Portugal

The most obvious alteration in the present edition is the addition of two new chapters, one concerning specific instructions for the analysis and indexing of special subjects, such as Literature, Culture, Civilization, History, Philosophy, Religion, etc., the second concerning the rules for indexing of special documents that, by their nature require additional indexing criteria. These include reference works, children's books, manuscripts, early printed monographs, textbooks, graphic material and serials.

Another important change is the inclusion of an alphabetical index which is a useful tool for quick and easy consultation.

Classification
With regard to classification in Portugal, we should highlight the updated, revised and enlarged edition of the UDC Manual used by the National Library and by the libraries integrating the National Bibliographic Database - Porbase (CDU: tabela de autoridade, 2nd ed., updated in accordance with Extensions and corrections to the UDC Vol. 18, 1996). This edition comprises changes to Class 9 - History as well as the inclusion of a new subclass in Class 0 - 004 - Ciência e Tecnologia de computadores, replacing subdivision 681.3., This UDC Manual is made up of loose-leaves so as to permit the inclusion of future updates according to UDCC directives and the National Library's classification policies.

Julio Vaz
Biblioteca de Arte, Fundação Calouste Gulbenkian, Lisbon

Romania's developments in classification and indexing, 1998-99

As a result of the publication of the UDC systematic tables and the efforts of the National Commission for the UDC, beginning in March 1998, all Romanian libraries (including not only academic libraries and the National Library of Romania, but also the public libraries) moved to use the same UDC version, including class 004, the changes in class 9 and the new version of class 8.

Under the joint co-ordination of the Bucharest Central University Library and ABIR (Romanian Academic Libraries Association) the creation of the Romanian Thesaurus (based on UDC) is continuing. The National Library is assisting those university libraries which are already involved in the process.

Not unconnected with classification, and as an means of assistance with classification, indexing and cataloguing activities, the National Centre of Cataloguing-in-Publication improved and enlarged its activities by publishing and circulating regular (monthly) booklets which include bibliographical records for the books in the process of publication.

So as to make the processing of publications more efficient and in order to create an online catalogue in more libraries, on account of the developments in automation for cataloguing, classification and information retrieval, the National Library of Romania initiated a programme of co-operation and resource sharing in the field of cataloguing and classification. Questionnaires were sent out to all types of library, to find out their availability and ability to participate in implementing this catalogue with their own bibliographic records, particularly for current Romanian publications. In October a meeting of the cataloguers and classifiers who are interested in participating in this programme will be held, in the course of which the principles, responsibilities and terms relating to the project will be set up.

Adriana Kiraly
"Octavian Goga" County Library, Cluj

Report from Sri Lanka

The University of Sri Jayawardanepura library follows the Dewey Decimal Classification scheme for Buddhism, using a phoenix schedule with a short notation, as follows:

2B0 General Buddhist works
2B1 Religion and philosophy
2B2 Tripitaka
2B3 Life of Buddha
2B4 Popular Buddhism
2B5 Monastic life
2B6 Buddhist sects
2B7 Buddhist arts
2B8 Comparative Buddhism
2B9 History of Buddhism

These practices indicate a demand for a faceted structure for the religion classes of enumerative schemes.

Since there are faceted features in this basic classification it could also be applied with a suitable notation for a modern classification scheme. For instance, the University of Sri Jayawardenepura Library has adopted the same notation as the Dewey Decimal classification by using 4S0 for 491.48 and 8S0 for 891.48 to reduce the length of notation for Sinhalese language and literature the sake of economy and convenience.

There was a substantial development of language and literature mostly associated with religion. Sinhala literature was divided into major form classes as follows:

Gadya      = Prose
Padya      = Verse/Poetry
Champa      = Prose and verse mixed

Padya (verse) was further subdivided into sections such as:

Asna        (a descriptive message)
Gie         (a song)
Kavi        (a rhymed verse)
Kiyaman     (a recital)
Prasasti    (panegyric)
Sandesa     (a message sent through a bird)
Sehella     (a narrative verse)
Virudu      (verse in the question and answer form)
Yadinna     (a prayer)

This classification was based on the literary warrant of collections developed over the years and followed by the libraries without definite notation.

...

The National Archives Department wanted to upgrade the quality of cataloguing and started the publication of the Ceylon National Bibliography in 1963. A.J. Wells, General Editor of the British National Bibliography, gave instructions and contributed in the compilation of this issue with British practice. The Anglo-American Cataloguing Rules of 1908, which were used in the English speaking countries of the world at that time, were the standard. Even when the Sri Lanka National Library and Documentation Services Board took over this responsibility in 1973, the cataloguing practice of following the Anglo-American rules has been established.

The courses conducted by the Sri Lanka Library Association, University of Kelaniya and University of Colombo are taking the necessary steps to keep the curriculum updated with modern practice, specially with revisions of classification schemes and cataloguing rules. The majority of university and research libraries use the latest edition of the Dewey Decimal Classification. Special libraries find the Universal Decimal Classification is more suitable for classification and documentation. The seminars and workshops conducted by the Sri Lanka Library Association and the Sri Lanka National Library and Documentation Services Board keep library personnel well informed about current developments and make them involved in continuing professional developments.

It is significant that links with the Internet and the close association with international library and information organizations will benefit library and information services in Sri Lanka and promote contact with the modern information society. Sharing of experience at national level as well as international level will be more useful to upgrade the organization of knowledge in the country.

P. Vidanapathirana
University of Sri Jayewardenepura

Reports from Sweden

It is therefore of great interest to investigate whether there are automatic alternatives. At the Library of the Swedish Parliament last year a project was started in order to investigate if there are programs that can deliver assistance to the indexing activity, giving the indexer another role (that of supervisor and administrator), or replacing the human indexers completely.

When I, in the first phase of this project, sought information about automatic indexing, I discussed with Mr Alcidio Pereira, documentalist at the European Parliament, in order to learn more about the pilot project in automatic indexing that went on in 1995 at the European Parliament. Mr Pereira recommended that I to contact Joint Research Centre in order to learn more about recent research in the field. In Ispra, Italy, Dr. Ralf Steinberger works in a project on automatic indexing. He was very helpful; giving me both much advice and processing some of our documents in order to get an idea of the outcome. I also read with interest whatever I could find about automatic indexing. At the IFLA Satellite Meeting in Lisbon 1993 some contributors dealt with automatic indexing. In his contribution Robert Fugmann express a pessimistic, but well-grounded view: "The process of phrasing natural language text is an inherently indeterminate process, i.e. a process whose progress and results are unpredictable. The reason for this is that an author has an entirely free hand in the choice of the wording of his text." The problem lies therefore not in insufficient hardware or software development, but in the very attempt to try to handle all possible situations in language, when only some can be anticipated. Fugmann claims with emphasis that the attempt to write a machine algorithm that simulates an indeterminate process is impossible. Dr. Ralf Steinberger - who with interest acquainted himself with this IFLA volume - reassured me that he none the less believed it was possible.

In early spring we employed a computer linguist to make a thorough investigation of the market and to take care of installation and tests.
So far it is clear that there are no "off-the-shelf products" on the market. There are tools with different functionalities that can be adapted and adjusted in a higher or lesser degree. Another conclusion is that for this purpose it is a great disadvantage to belong to a small speech area. Many linguistic products are developed for the English-speaking area. They could, of course, be translated, but the companies do not seem to be too interested in paying for a translation. There is an interesting lesson to be drawn here. In spite of the impression of the expanding possibilities of information technology that one gets through articles, exhibitions, reports, etc., when it comes to the concrete application the picture changes. Maybe you should live in another country, speak another language, work on another platform, be a part of a much greater organisation, etc., in order to opt for the news being presented. Another lesson to be learnt is that many products contain old things in new clothes. Behind "rules", "categories" or "topics" to be determined and maintained in different ways one can detect good, but old formulas.

Most promising, so far in this project is a contact with advanced researchers in datalogy. Maybe they will process something with a good result, adapted for the different kinds of documents. In the Swedish Parliament there are a great many texts being produced each session due to a tradition of rather lengthy argumentation, both from the part of the MPs and also from the Standing Committees' side. To complicate matters further, the documents vary greatly due to the situation and the purpose of the document. To discuss in the Chamber, to write a question on a current, controversial subject, to present the party's programme in a whole field (education policy for instance), or to present a new legislative proposal constitute different issues. Nonetheless, they all end up as printed documents, as words and sentences, and for a programme they all look alike. For an indexer the rhetoric and the subjectivity of the political language and the strictly formalised presentation of the plain facts of a law proposal are kept aside, treated in different ways. If there is an automatic solution in sight, maybe it should be composed of specially adapted programs seeking for the most significant concepts in the text, but avoiding the pitfalls due to the document's nature. Thus a program should work as a trained indexer, i.e. a person who does not only read fast, extracting with ease the meaning of the document, but who takes into consideration the very document, its weight and structure.

To conclude, language is ever flexible and unpredictable, generally speaking. More specifically though, the language of authors of belles lettres, scientific works or philosophical works can never be anticipated by a program, neural or not. But the kinds of documents that constitute the basis for decision-making, or aim at influencing the decision-making process are in a high degree predictable and even repetitive. In these cases programs ought to be able at least to assist the human brain.

***

At the end of October I attended the Eurovoc seminar arranged by the Spanish Senate within the framework of the European Centre for Parliamentary Research and Documentation (ECPRD). ECPRD is a body administered jointly by the European Parliament and the Parliamentary Assembly of the Council of Europe (although Canada, Israel and the United States have an observer status). At this meeting users of the multilingual thesaurus Eurovoc, future users and those interested in becoming users were gathered. Before the seminar took place a questionnaire had been sent out to all parliaments within Europe and to the US Congress; 27 responses were received. There are as many as 15 users of Eurovoc; 11 parliaments use it for their parliamentary documents but it is also used for bibliographic references in as many as ten parliaments. The current status is that Eurovoc exists in 11 official language versions. For some users, for example the Belgian Senate, the multilingualism is an important reason for choosing Eurovoc as an indexing tool, since the country has two official languages. For others, it is its broad scope adapted to socio-economic matters that makes it so useful. There are a couple of unofficial versions and some ongoing translations in Eastern Europe. Within a not so distant future Eurovoc may exist in at least 15 versions and thus become the biggest multilingual thesaurus.

At the seminar, users presented their way of using Eurovoc, their technical solutions and their problems. As everyone can imagine it is a huge problem to maintain a common linguistic tool among so many languages and on top of it with so many different users with differing needs. We had debates and informal discussions about updating, re-indexing problems, the need of tracing changes in all the different versions, if it is possible to use the non-descriptor as a way of creating greater usefulness for each language version without loading the descriptor with meaning that goes beyond its original meaning, and so on.

We were informed about the new maintenance system, the distribution over the Internet and the planned fourth version. The working principle for the fourth version will be to perform a systematic work in different languages; taking care to avoid one language becoming a kind of first language and the others just translations of it.

Elisabet Lindqvist Michailaki
Swedish Parliament, Stockholm

The Division of Bibliographic Development and Co-ordination (BUS)
The Division of Bibliographic Development and Co-ordination (BUS) is concerned with the bibliographic quality of the LIBRIS database. As LIBRIS is a union catalogue where all participating libraries contribute to the quality of the database it is very important for us, at the National Library, to co-operate with other libraries. Therefore we have several reference groups within the bibliographic field. One important task of BUS is to produce guidelines for subject headings, classification and authority control. Our aim is to make all bibliographic information as standardized and easily convertible as possible both within a Swedish framework as well as internationally.

Classification
The classification system used by the Royal Library, most university libraries and all public libraries is called Klassifikationssystem för svenska bibliotek, SAB. It is an entirely domestic system used nowhere but in Sweden. The responsibility for the updating of this Swedish system is handled by a committee within Sveriges allmänna biblioteksförening, SAB, (= General Swedish Library Association).
Earlier this year the Committee presented an updated concordance between the SAB classification system (7th ed.) and Dewey (DDC21). BUS has been highly involved in this project. The concordance will be published by Btj (Bibliotekstjänst AB) as a printed booklet and negotiations are going on with OCLC for an online version as well. If the negotiations turn out successfully KB's intention is to provide the LIBRIS database with this online concordance - a good opportunity to bring the Swedish classification system into a more international framework.

Subject headings
BUS has created guidelines for subject headings. These are adapted for KB practice, but our intention is to make them useful for all Swedish libraries. We took the first step in that direction when making them available on our Web site this year. The guidelines follow the same principles as those stated by IFLA. We have also looked at the principles underlying LCSH trying to follow them as closely as we can. This is also in order to make our Swedish bibliographic records correspond better with international standards.

Subject heading authority list
The SAB classification system includes a separate alphabetic word index which is also used as a subject heading list. Alongside this index new subject headings are continuously presented on our Web site. These new subject headings are created by BUS together with a reference group with participants from other libraries. Proposals for new subject headings may come from any cataloguer at any library. In order to facilitate proposers' use of e-mail our Web site also includes a specific formula especially for that purpose. During this spring we have set up a database including the new subject headings from our Web site and a selection of headings from the printed SAB word index. This database will have two purposes. It will be used as an authority file integrated in the new LIBRIS system. And it will also function as a database available on the Internet for anyone who may have use for such a "word list". This is a project we are to finish this autumn.

Summing up - tasks ahead
Our main concern is to make our guidelines for subject headings more adapted to international standards. We will try to apply the Library of Congress guidelines to our Swedish context. To what extent we are not yet ready to say.

The other important task is the creation of the online subject heading list and to make this list user-friendly and available not only for librarians. We want this list to be a forceful tool for indexing and retrieval on the Internet as well as for printed material.

Pia Leth
Head of the Division of Bibliographic Development and Co-ordination (BUS) at the Royal Library, National Library of Sweden

Subject indexing and classification in the United States

Library of Congress Subject Headings (LCSH)
The implementation of form headings and form subdivisions continues. Previously, form headings/subdivisions were indistinguishable from topical headings/subdivisions in coding in the USMARC Formats. The distinction between topical and form data improves the faceting principles of LCSH.

The electronic version of LCSH now contains not only records for main headings and many heading/subdivision combinations, but also separate records for free-floating subdivisions. Records for more than 3,000 topical, form, and chronological free-floating subdivisions are in the process of being created and incorporated into the main file.

Medical Subject Headings
Medical Subject Headings (MeSH), published annually, continues to serve as both an indexing and a cataloguing tool for medical literature. In 1999, with the implementation of the Voyager Integrated Library System at the National Library of Medicine, the Library has moved to harmonize its practices for subject heading content and structure with the Library's indexing practices. The purpose is to facilitate cross-file searching and improve retrieval by assuring that the same terms produce like results for articles, books, non-print materials, electronic resources, etc.

In the Library's Web-based catalogue, called NLM LOCATORplus, "main headings or main heading and topical subheading combinations are no longer qualified by geographic, form, or language subheadings. Instead geographic descriptors and publication types (PTs), formerly known as form divisions, are carried in separate fields and are used in retrieval in co-ordination with main headings or main heading and topical subheading combinations." (Application of MeSH for Medical Catalogers, http://www.nlm.nih.gov/tsd/cataloging/catmesh.html). Nevertheless, for medical libraries that wish to retain the conventional subject heading strings, NLM will recombine subject headings, geographic descriptors, and PTs in NLM's established order to form the strings in MARC records to be distributed to these libraries. An example is shown below:

NLM LOCATORplus Structure
subj: Skin Diseases--diagnosis
subj: Skin Diseases--therapy
PT: Atlases

Distributed Subject Heading Structure
subj: Skin Diseases--diagnosis--Atlases
subj: Skin Diseases--therapy--Atlases

This change signifies a move of MeSH towards a more faceted, postcoordinate system.

Classification

Library of Congress Classification
With the availability of LC Classification data in electronic form, the Library of Congress is now able to edit the schedules more efficiently and publish them with greater frequency. Schedules published since August 1998 include Classes A, L, M, PR-PS, PZ, the P Tables, and the K Tables; and the following are ready for publication: K, KF, KJV-KJW, and PB-PH.
The K Tables, similar to the P Tables, contains a set of standardized form division tables, which eventually will replace all existing form tables in various schedules for Class K. The consolidation of tables previously scattered in different schedules represents another step towards standardization, uniformity, and greater consistency in the LC Classification.

Subject data in the metadata record
ALCTS/CCS/SAC/Subcommittee on Metadata and Subject Analysis
The Subcommittee on Metadata and Subject Analysis continues its study of the application of subject data, including classification and free-text and controlled vocabularies, in metadata records, particularly the Dublin Core records. Its charge is to recommend the most effective methods of supplying subject data for the representation of the contents of World Wide Web resources. Based on the deliberations and discussions during the 1999 ALA Midwinter Meeting and Annual Conference, a report has been drafted and will be presented to the Subject Analysis Committee, the Cataloging and Classification Section, and the Association of Library Collections & Technical Services for approval. For a copy of the full Working Draft of the report, please contact Subcommittee Chair Diane Dates Casey at d-casey@govst.edu

OCLC's CORC (Cooperative Online Resource Catalog) Project
In January 1999, OCLC launched a 12-18 month pilot project CORC (Cooperative Online Resource Catalog). Its purpose is to create a new, flexible system for cataloging Web resources. According to OCLC,

The new system "will offer both full USMARC cataloging and an enhanced cataloging mode for Dublin Core--a set of 15 metadata elements for describing electronic documents. For international access, CORC will support Unicode coding of characters. In addition, records can be imported into or exported from CORC using Resource Description Framework (RDF)-compliant XML (Extensible Markup Language), HTML, and MARC."

For subject data in the Dublin Core metadata record, OCLC has expressed an interest in implementing in CORC the ALA Subcommittee's recommendations put forth in its draft report mentioned above. OCLC will use LCSH in CORC but with a simplified syntax. It is considering placing topical data, spatial and temporal data, and form data in different elements in the Dublin Core instead of being combined in a subject heading string. This syntax is more compatible with the basic structure of the Dublin Core element set.

Lois Mai Chan
University of Kentucky

Classification activities in the UK, 1998-99

The proceedings of the UDC workshop held in Amsterdam were published in November 1998 in Extensions and corrections to the UDC no. 20. Copies may be obtained from the UDCC, PO Box 90407, 2509 LK The Hague, Netherlands.

Research continues to centre more upon ways of retrieving information via vocabulary alone, rather than structures and classification systems per se. The Library and Information Commission recently issued a call for research bids in the area of subject retrieval, which they are funding this year, but the successful applicants have not yet been announced. It is, however, a welcome sign that they have done this, as I cannot recall it happening before.

The Classification Research Group continues to meet on a regular basis at University College London, and its discussions continue to centre round the development of the revised edition of the Bliss Bibliographic Classification. There have also been discussions on the problems of "aspect classification". Development and publication of the BC2 continues, and there is a degree of collaboration with the UDC. Vanda Broughton who is now a full time Research Assistant employed by the UDCC at UCL still plays an editorial role in Bliss, and since both schemes are currently revising Chemistry there is a degree of similarity in the structures of the two classifications in this discipline. As far as publication of BC2 is concerned, Classes AY-B - Science in general and Physics are currently in the hands of the publisher and should appear before the end of this year. The Introduction to the scheme is also being reprinted - worth noting, since it contains one of the clearest expositions of the theory underlying faceted classification, and Chemistry should be ready for publication next year.

The British Library has completed its move into the new building at St Pancras. Concurrent with the move is the change from an in-house classification to DDC, and Joan Mitchell visited the Library in February to inform the staff of the former Science Reference Library about the classification, in view of the misgivings among some there over abandoning the Sandison scheme.

While she was in London she met with members of BSI and the UDC editorial team to discuss the feasibility of creating greater harmonization between the two classifications. The proposal for a joint area table was abandoned, and instead it was decided to create a linking device which would allow the user to translate the codes of one scheme into those of the other. A research proposal using Chemistry as the test discipline was worked out and submitted to the LIC under the call for bids mentioned above, but it was not successful. It has now been submitted to the Humanities Research Board, but the result will not be known until November.

A separate UDC update has been prepared for this conference, but revisions to appear in this year's Extensions and corrections include the new Table of Common Properties referred to at IFLA last year, expansions in the Area Table for North America, Class 2 - Religion, Tourism, Environmental Science and the customary range of new terms that need inclusion. It will be published in November. BSI have just completed a pocket edition of UDC which will go on sale in September, and there is a proposal to develop a multi-lingual CD-ROM - in English, French and German of Pocket edition size - at present under discussion. The UDC Web site may be visited for recent information at http://www.udcc.org. The software firm Muscat is marketing UDC under licence for a one year trial period as one of three retrieval devices available through their Empower knowledge retrieval software. It will be possible within the next few weeks to see a demo of the system at http://demo-suite.muscat.com/

I.C. McIlwaine
University College London