66th IFLA Council and General
Jerusalem, Israel, 13-18 August
Code Number: 093-123-E
Division Number: IV
Professional Group: Bibliography
Joint Meeting with: -
Meeting Number: 123
Simultaneous Interpretation: No
The national bibliography of a small country in international context
With a part on subject cataloging by
National Library of the Czech Republic
Prague, Czech Republic
The Czech Republic belongs to the so-called "post-communist" countries. The communist governments had no desire to support library efficiency, including contact with developed countries or enhancing automation. As a consequence, most libraries were ignorant of international standards. After the 1989 revolution, the Czech Republic initiated many dramatic changes, and the Czech national bibliography, too, benefited from this course.
The paper describes the changes implemented in the compilation of the Czech National bibliography, specifically the introduction of UNIMARC, AACR2R, LCSH in both Czech and English versions, some of which link to UDC-MRF notation. Thanks to the retrospective conversion project, the majority of printed monographs published in the Czech Republic during the 20th century is now available via Internet (WorldCat, CASLIN - Union catalog of the Czech Republic, and the database of the National Library) and also on CD-ROM (Czech and English versions).
The challenges faced by the introduction of international standards in a relatively short time-span, the question of "loss of national identity" versus joining the international community, and the importance of national authority files for cooperation on both national and international levels are discussed. A short demonstration of the bibliographic and authority records from our online catalog, based on the Israel system ALEPH500 and also the CD-ROM version of the Czech national bibliography will be included in the presentation.
The Czech national bibliography represents the publications of a small country in the heart of Europe with an annual production of about fifteen thousand printed monographs and three thousand serials. Non-book materials, cartographic documents and electronic resources (both local and remote) registered in the Czech national bibliography total some five thousand titles annually. Most of these documents are published in Czech, a language which only a small percentage of the world population can speak and understand. However, Bohemistics is a subject represented at many universities throughout the world. The libraries of these universities and also national libraries claiming universal scope and/or university functions have large collections of so-called "Bohemica".
When we first started to use foreign databases and travel abroad after the revolution in 1989, we started seeing American libraries' records for Czech imprints in OCLC. We did not like the quality of some of the records. Mostly, the records were quite brief and contained spelling errors, not only in the descriptive parts of records but also in the headings (both names and subjects). When we mentioned this fact during our visit to the Library of Congress, we were told that cataloging Czech imprints is not a priority in U. S. libraries and it was suggested that in order to alleviate this problem we should contribute records representing Czech imprints ourselves. This was a wonderful example of applying the UBC principle that each country should be responsible for the cataloging of its own publications, which has been known in theory rather than practice in our country. How easy to say, but how difficult to bring this wonderful and logical idea to real life, especially because the cataloging should adhere to international standards. When no international standards have ever been applied in a particular country, it can take many years to achieve what for some countries has been quite easy and an obvious goal; we needed ten years to catch up.
The following pages describe the ten years we needed to be able to produce our national bibliography on CD-ROM (also available on the Internet) and of high enough quality that foreign libraries are able to reuse our records without too much post-processing. I will describe not only our successful results, but I will let you have a look into blind alleys and mistakes that made our way longer. I do not expect that our example can prevent other countries from making the same mistakes. Each country has to go its own way and make its own experience -- including its own mistakes. That is life.
A couple of years ago we were unable to accept advice and learn from the experience our colleagues from abroad had acquired. Only now, after acquiring our own experience (and evaluating our own mistakes), we know that they were right and that our way could have been shorter, easier, and cheaper . And even now, we are sometimes unable to give up some practices we are used to, and I am afraid that this will be recognized as a mistake years from now. But let us look at the ten years as they truly were, not as they could or should have been.
The first significant step in the direction of automation and acceptance of international standards was the project "Automatizovany system ceske narodni knizni bibliografie: ASNB-K" (1) [Automated System of Czech National Book Production: ASNB-K]. Since 1983 data were created using a regular word processor: the goal was not the creation of a searchable database but only the ability to prepare print copy of the "Czech National Bibliography" for publication each month. The project staff was charged with analyzing the records thus created and studying international cataloging standards. Luckily, since all the records used ISBD punctuation, there was a sort of unifying structure, and for this reason it was possible to convert them recently into the rest of our modern database. The employment of ISBD in the "Bibliography" was viewed by Czech librarians for years more as a curiosity than a means to standardization.
Exchange Formats, Bibliographic Description and Names
"Serious" library automation that considered broader international standards and their implementation in Czech libraries started later -- only in late '80s. There was almost no money available for library automation; and since most of the libraries (including the National Library) could not afford any professional library system, the only possible way was to use the general database system CDS/ISIS offered by UNESCO free of charge. It was necessary to develop a Czech library application for the CDS/ISIS system. The National Library produced the "Modularni automatizovany knihovnicky system: MAKS" (2) [Modular Automated Library System]. In 1989 MAKS formed the basis not only for the Czech national bibliography but also for automation in hundreds of Czech libraries which were using CDS/ISIS in the '80s.
The first exchange format was prepared and published as the core module of MAKS in 1989. In developing the conventions for an exchange format, we, albeit aware of UNIMARC, contemplated the question: UNIMARC or national format? We decided to develop our own national exchange format. It was the "Vymenny format pro bibliograficky (dokumentacni) a katalogizacni zaznam" (4) [Exchange Format for Bibliographic and Cataloging Records]. Fortunately, our exchange format employed considerable granularity and thus conversion into UNIMARC later (during the mid '90s) was not very difficult. Providentially, it used the principles of ISBD, and in this oblique manner ISBD was accepted by many Czech libraries.
While MAKS was being tested, it was becoming more and more obvious that our cataloging rules were obsolete, especially as far as headings were concerned, something ISBD does not cover. Our old rules had outlived their usefulness, existing foreign models were viewed with skepticism, and to prepare new rules expeditiously was not manageable. Furthermore, our professionals demanded rules for special materials. During the discussions about preparing new cataloging rules, the suggestion to accept AACR2 was introduced, but it was rejected. We had always been strongly influenced by foreign cataloging rules (Prussian instructions; Soviet cataloging rules) and we did not want to be influenced by yet another set. However, we accepted that we are not an island, so a compromise was proposed: develop Czech cataloging rules based on AACR2, or rather on their positive principles only, and supplement them with other foreign rules. A working group of 40 librarians presented a 597-page opus; their new "Pravidla jmenneho popisu" (5) [The Rules of Descriptive Cataloging] was to have been published in 1993. But the new Rules were never published.
The combination of the positive principles of different sets of cataloging rules produced a very good theoretical result, but it was inconsistent and its application in "real life" was far from easy. When our librarians started to travel abroad after 1989, they brought back the knowledge that cooperative cataloging on both national and international levels is necessary, standardization is essential, and locally developed rules can hardly compete with international standards. Obviously, a small country with a not too copious book production, in a minor language, has no chance of displacing already existing cataloging rules, under which millions of records already exist. If we wanted to become part of this world, we had to accept its rules. At last we realized that specific national cataloging rules, no matter how perfect, unavoidably lead to isolation. The 40-member team was discharged as discussion about the option of translating AACR2 was gaining ground. We decided to translate AACR2 and ISBD and to analyze all their details and implications. We wanted to examine with an unbiased, critical eye especially those rules that may be unacceptable to Czech cataloging tradition, and to document our opinions with specific examples. The same approach was taken in regard to UNIMARC.
In the '90s, Czech libraries began to implement integrated library systems. In 1992, the Andrew W. Mellon Foundation was presented with the CASLIN Project (Czech and Slovak Library Information Network). In their Statement of Intent, the directors of four main Czech and Slovak Libraries (including the national libraries of both countries, responsible for their national bibliographies) agreed to lay a solid foundation for a nation-wide library network to provide both domestic and foreign users with easy, fast, and unrestricted access to the information stored in, or mediated by, libraries and information centers. Since the beginning of CASLIN, international standards for cataloging were considered very seriously in connection with its union catalog. In 1994 the CASLIN Board of Directors approved and announced the following requirements for cataloging for the CASLIN union catalogue:
- All records have to be created in UNIMARC
- The guiding rules are AACR2R
- CASLIN union catalog records must follow ISBD.
This declaration was extremely important - it had to be followed and it did not allow us to change direction when we faced opposition introducing international standards - and we really did!
Despite of our declaration of principles, it was not easy to implement the agreed-upon standards, not least of all because the documents were not available in Czech right from the outset. It took much longer to translate all the necessary tools (7-9) than we anticipated, mainly because in many cases there was no Czech terminology for some of the new concepts. Without examples of application of the rules in actual practice, it was very difficult to grasp certain concepts fully and translate them meaningfully.
In regard to AACR2, we reached Catch 22: without rule interpretations it was almost impossible to work with the rules, but without using them, it was very difficult to create workable interpretations. Before any interpretation could be agreed upon, it was crucial to try the rules out, to understand their ramifications, and to "live" with them before we could distinguish between what we did not like because it was new and different or what was fundamentally alien to our tradition. The initial implementation of AACR2 was very demanding on our catalogers.
When our databases were small, we tended to pay much more attention to bibliographic records than to authority records. The need for authority control remained underestimated for a couple of years. Only as our databases began to grow and as we started up- and downloading records, did we start paying the attention to authority control that it deserved. A new department for authority control was established only recently. It is now responsible for cleaning up the access files accumulated during the long period when authority control was unappreciated, and for creating authority files for personal names and corporate bodies.
Subject Cataloging and Subjects
For providing our users (both local and remote) with full access to documents, implementation of international standards for bibliographic description alone would not be sufficient. It would also be necessary to provide data of high (standard) quality for subject access. The need for change (mainly for subject headings) became obvious at the beginning of the '90s, but we were ready to do it only by the late '90s. It had not been that difficult with classification since we had a long tradition of using UDC -- we only migrated to the UDC-MRF in 1999.
Because we wanted to have subject headings that would be acceptable not only at the national, but also at the international level, it was necessary to change from the specialized Czech subject headings based on our old national standard, which was very different from any international one. It was necessary to choose one of the international systems (we have chosen LCSH) and apply its principles to our national subject headings. Of course, the Czech and English versions do not correspond 100%, but it is much easier to add similar Czech terms in English to the LCSH in the records downloaded from OCLC for foreign acquisitions and vice versa, as records are contributed OCLC, to add LCSH equivalents to the Czech subject headings in Czech national bibliography records, than to create subject headings in both languages and that are based on completely different philosophies.
Since subject headings are expressed in national languages, it is difficult to find and apply any "international" recipe. After much debate, LCSH was finally chosen. However, it is considered useful to meet local needs and requirements as well, so some modifications of the LCSH scheme were formulated:
- Direct form of geographical subdivisions in subject headings of the NLSH is preferred: i.e., geographic names immediately follow the main heading or main heading/topical subdivision combination
- Form subdivisions have been taken out of subject heading strings and assigned to a specific field (608 in the UNIMARC format)
- Generic headings representing classes of persons or types of corporate bodies are used more often in our system in order to;
-Add more access points to bibliographic records
- Allow links by named entities (persons; corporate bodies) between subject headings and the corresponding UDC notations
- Take out of name subject heading strings any additional thematic, geographic, or chronological information and to put them in a "class of person and types of corporate bodies" subject heading string to support easier automatic subject authority control and maintenance of named entities
- Add a set of subject headings to a document by using terms based on concepts in order to make the "broad" subject headings narrower; this makes it easier to establish links between subject headings and descriptors (thesaurus or controlled subject language)
These changes (due to the differences between LCSH and NLSH) mean a shift in our system towards a postcoordinated system, which we find more effective in the online environment.
- Using the UDC classification system, it was decided to:
- Improve multidimensional searching for OPAC users by linking topical subject headings, main headings/topical subdivision combinations, and genre/form headings in the subject authority files to the appropriate UDC notations
- Observe the IFLA recommendation, "to consider possible relationships between subject authority records and classification" (Guidelines for Subject Authority and Reference Entries, p.1)
- Improve the quality of subject data to minimize individualistic approaches to the subject matter of a document
- Use the middle terms of UDC notations, since the UDC (top level) selected points currently used in our libraries do not entirely identify topics
- Use only those UDC notations that identify the subject and genre/form, and express other components (e.g., geographic, chronological) by codes or verbal expressions
In the process of adding English equivalents to subject headings in our national bibliography records, we try to respect the most important recommendations of LC subject cataloging policy, but it is really a demanding process for us. We hope the situation will improve after implementation of the new version of the ALEPH system, namely, when we have the possibility of using the Z39.50 protocol.
As should be evident from this discussion, we did not exactly dodge international standards, though we were slow at accepting them. Luckily, we do not feel that accepting them has in any way invalidated or lessened our national identity. On the contrary: when we first started seeing American libraries' records for Czech imprints in OCLC, we did not like the quality of some of the records. Now we contribute our own records representing Czech imprints to the WorldCat database. Obviously, this cooperation brings fruits to both sides, as we welcome any income we can get. Our contribution is an asset and our being able to import records representing foreign production, too, is, economically speaking, most welcome: we do not have to catalog these foreign documents and we can pay more attention to cataloging our Czech materials. What is equally beneficial is that in this way we see more proper application of international cataloging standards in these records, thus increasing our awareness of them. We do not consider this to be a loss of national identity. Rather, we consider this an argument for abandoning "national" stereotypes. But are those really national stereotypes, or are they "international" ones, in the sense that catalogers throughout the world do not like change? Whenever we fought for acceptance of international standards, there was always an underlying desire to preserve as much of the status quo as possible.
Another, more serious, problem that complicates the acceptance of international standards lies in the lack of truly international standards. It would be much easier for everyone at this juncture to be able to accept international standards as they are, were they truly international (and not mainly for the Anglo-American part of the world). That would guarantee that neither conversions nor the comparison of differences between different so-called "international" standards would be necessary.
This must become the case in the 21st century when catalogers, under economic pressure like most other segments of our societies, will have to compromise and develop and accept truly international standards. No one will have any incentive to develop and maintain local rules; no one will need to look for solutions in one "international" format (MARC 21) and try to incorporate them into the other "international" format (UNIMARC). This is now often the case with sharing electronic resources; it generally requires conversions followed by more or less demanding post-processing. No one will have the resources to devote to such undertakings. How long it will take before this vision becomes reality will depend on economic conditions, the level of international cooperation, and on users' demands on their libraries.
The current Czech national bibliography since 1994 is available on CD-ROM and via the Internet from the database of the Czech National Library and CASLIN - the Czech Union Catalog. The records are also available in WorldCat. All the records available in machine-readable form (since 1983) have been converted to UNIMARC and are available in the same ways. And what about the older records?
By the end of the year 2000, an important project, "Making the Czech Production of the 20th Century Widely Available", will have been completed. Its results are already accessible via the Internet as well as on CD-ROM. Most of the approximately half million records that cover Czech imprints from as far back as 1922 are included from the printed Bibliographic Catalog. However, there are records in a few other bibliographies from the early part of the 20th century that will also be converted. Most entries in the Bibliographic Catalog are, in fact, detailed citations that contain added entries, subject headings, UDC symbols, and also frequently, annotations. From an international viewpoint, this will be a great enrichment of the UBCIM and UAP projects; and the records are also available in the OCLC WorldCat database. And, of course, on the national level, the online availability of these records allows faster and cheaper retrospective conversion at many institutions in the Czech Republic.
In fact, nine large Czech research libraries are participating in this collaborative effort: each cooperating institution is converting specific sections of the paper version of the Bibliographic Catalog. The result of this project then decreases the number of cards they each have to convert from their own catalogs. Thanks to this project, by the end of the 20th century, the majority of Czech imprints (i.e., those delivered to the NL as the legal deposit copies) published in this century will be available in UNIMARC (and in WorldCat in MARC 21). I would like to mention that this collaborative effort had the blessing and financial support of the Andrew W. Mellon Foundation at the very beginning and later of our own Ministry of Culture, without whom this phase would not have been possible.
A special technology was developed for this project, a process consisting of three phases:
Step I -- Scanning:
Pages from the printed bibliographies are scanned by state-of-the-art equipment, which guarantees high quality images, essential for subsequent processing.
Step II -- OCR:
The scanned images are converted from a pictorial (TIFF) image into ASCII characters (unstructured text) using OCR.
Step III -- Structuring (Tagging):
In this final phase, the unstructured ASCII records are converted into UNIMARC records. Depending on the degree of clarity of the original record, tagging can be done automatically, whenever it is feasible to use a specific algorithm. Any records that do not pass automatic tagging are then processed manually, i.e., a staff member analyses the record and tags each field as needed. Our specialized software helps create high quality records in a short time.
This paper devoted to the Czech national bibliography in the international context discussed mainly the introduction of international standards for both bibliographic description and subject cataloging. Speaking about international standards, ten years ago we started from almost zero; and today, the Czech national bibliography for the whole 20th century is available via Internet in the database of the National Library (and through Z39.50 after the ALEPH500 implementation), in CASLIN - the Czech Union Catalog, and in WorldCat. The records are available in both UNIMARC and MARC 21. Since 1995 bibliographic description is prepared in adherence to AACR2R, and since 1999 the records are classified according to the UDC-MRF and contain LCSH in both Czech and English. Bibliographic records for printed monographs and serials (on both serial and analytical levels) are included together with non-book materials, cartographic materials and electronic resources (local and remote). Furthermore, we have started connecting bibliographic records to authority records in UNIMARC format.
Records from the Czech national bibliography prepared at the National Library have become heavily used by both Czech and foreign libraries. We do not have our own national standards any more, but we do not think we lost our national identity this way. On the contrary, we have become fully responsible for bibliographic control of Czech documents in all formats, not only theoretically but in reality - records created at the National Library of the Czech Republic have replaced original cataloging of Czech documents performed in many libraries abroad.
- ASNB-K : provadeci projekt automatizovaneho systemu ceske knizni narodni bibliografie. - Praha, Statni knihovna CSR, 1983
- Modularni automatizovany knihovnicky system. - Praha : Statni knihovna CSR, 1989. - 7 modules
- ISBD(G) : Vseobecny mezinarodni bibliograficky popis. - Praha : Narodni knihovna, 1993
- Vymenny format pro bibliograficky (dokumentacni) a katalogizacni zaznam. - Praha : Statni knihovna CSR, 1989
- Pravidla jmenneho popisu : verze 0.2 / Hana Vodickova ... [et al.]. - Praha : Narodni knihovna v Praze, 1992
- Zaznam pro souborny katalog. - Praha : Narodni knihovna v Praze, 1994
- Anglo-americka katalogizacni pravidla / American Library Association.- Praha : Narodni knihovna Ceske republiky, 1996
- UNIMARC manual : bibliograficky format / IFLA. - Praha : Narodni knihovna Ceske republiky, 1996
- UNIMARC/autority : univerzalni format pro autority / IFLA. - Praha : Narodni knihovna Ceske republiky, 1996.
- AACR2R/UNIMARC : schvalene ceske interpretace. Verze 2 (leden 1999)
- UNIMARC in Czech libraries / Edita Lichtenbergová, Bohdana Stoklasova [paper for the 64th IFLA General Conference, 1998],br> [http://archive.ifla.org/IV/ifla64/079-161e.htm]
- Guidelines for subject authority and reference entries / IFLA. - München : Saur, 1993
- Subject cataloging manual: subject headings / Library of Congress. - 4th ed. - Washigton, D.C. : Library of Congress, 1991
- Library of Congress subject headings : principles and application / Lois Mai Chan. - Englewood : Libraries Unlimited, 1995
- Retrospective conversion in Czech libraries / Bohdana Stoklasova, Miroslav Bares. -- Praha : Narodni knihovna, 1995
- Retrospective conversion in Czech libraries / Bohdana Stoklasova, Miroslav Bareš // International conference on library automation in Central and Eastern Europe. - Luxembourg : European Commission, 1997. -- s. 195-206
- Retrospective conversion at the National Library of the Czech Republic / Bohdana Stoklasova // 26. ABDOS-Tagung : Referate und Beiträge. -- Band 22 (1997), s. 134-141