As of 22 April 2009 this website is 'frozen' in time — see the current IFLA websites
This old website and all of its content will stay on as archive – http://archive.ifla.org
As of 22 April 2009 this website is 'frozen' in time — see the current IFLA websites
This old website and all of its content will stay on as archive – http://archive.ifla.org
and
Bohdana Stoklasova
As our libraries started using integrated systems, it is becoming clear that international standards must be followed. It is evident that UNIMARC is a well-thought through theory but it needs many improvements and adjustments to be equally successful as an internal format. We see linking fields as one of the main obstacles. There is a need to create a UNIMARC USERS GROUP to discuss all the problems and issues. Libraries ought to strive for more uniformity in their use of UNIMARC, regardless of whether they use it as their internal format or only as a conversion tool.
For a truly effective cooperation among libraries, both the use of a common format and adherence to one set of cataloging rules are a prerequisite
Perhaps the main stumbling block in our decision-making process was our decades-long isolation from the West and our stubborn adherence to stifling rules from the past, even after we learned about new and better ways. In the decades prior to 1989, our lack of knowledge of what Western libraries were doing led to the development of the CDS/ISIS-based Czech (and Slovak) Exchange Format. We were vaguely aware of UNIMARC but, frankly, did not fully understand it. Besides, in those years it was not yet clear whether UNIMARC would remain just an attractive theory or whether it would be put to practical use. So we built into our own Format some "improvements". Although the CDS/ISIS-based Czech Exchange Format is, unintentionally, essentially a simplified version of UNIMARC, it is not quite compatible with it and our "improvements" even increased the incompatibility. For instance, we had to conform to CDS/ISIS limitations such as the availability of only 99 fields or the inability to utilize indicators and repeatable subfields. We removed fixed fields, we eliminated the "duplicate" recording of authorship (found in 200$f,$g and 7XX fields), and so forth. While we considered our recording of authorship only once in each record a definite improvement, we later learned that this feature was a major flaw of our system, making conversion of these records into UNIMARC rather difficult. Let us warn others who may be contemplating following the same route that modification of standards is always dangerous since valid objections and obstacles become easily supplemented by subjective ones. The result is something far removed from the initial international standard. Despite our isolation from the West, we did have some contacts; although we did not pay much heed to our foreign colleagues' advice to accept UNIMARC as is, we have, unwittingly, built into the Czech Exchange Format a high degree of "granulity" which helped us to overcome the other defects which we discovered while converting 500,000 records during the last 3 years.
After many oscillations, our post-1989 re-joining the international library community contributed to our belief that accepting international standards would improve our daily work and also render our records more acceptable to the rest of the library community. But it took ten years before UNIMARC became the norm in Czech libraries. In the mid-1990's, the National Library of the Czech Republic (NL) and several other institutions in the Czech Republic ceased to use CDS/ISIS and started using various integrated systems, with the same old questions: which standards to choose and how to implement them. After several years of using automation, newly gained contacts with foreign institutions and at last getting exposed to international standards, we opted for accepting UNIMARC, AACR2, LCSH, and UDC. We decided to implement all changes at once: a new automated system as well as bibliographic description and subject analysis. After years filled with arduous efforts, we achieved satisfying results, vindicating our choice. The National Library even participated in OCLC's testing of UNIMARC-USMARC conversion program. The result is a fully working conversion system allowing us to contribute into WorldCat both current and retrospective editions of the Czech National Bibliography in UNIMARC. And we are able to use UNIMARC records from WorldCat for cataloging our foreign acquisitions.
We had to choose from several automated library systems, be it those available from foreign vendors, e.g. ALEPH, TinLib as well as some home-grown versions, e.g. Lanius, KP-sys, Rapid Library. We had to consider how to combine these systems with UNIMARC, AACR2 and LCSH, both in Czech and English versions. Of course, there are many problems resulting from this coexistence. We will here elaborate only on two of them: (a) the general ones (they should be carefully observed by the Permanent UNIMARC Committee (PUC)) and (b) the local ones (they can be interesting and inspiring to other countries)
In 1996, we have completed the Czech translation of the UNIMARC Manual and we published it in a loose-leaf binder allowing easily for updates. UNIMARC is now the main format used in Czech libraries, with the Czech Exchange Format becoming a secondary one (mainly in libraries using CDS/ISIS before their migration to some form of an integrated system). We have also translated the Authority manual and we now create our authority records in adherence to it.
All the records contained in NL databases (monographs, serials, various non-books special formats as well as periodical articles) are available in UNIMARC. Most of them are also available on the CD-ROM version of the Czech National Bibliography. UNIMARC also became the main format for retrospective conversion: the number of records generated from current cataloging is annually enriched with more than 100,000 high quality records resulting from retrospective conversion of the printed version of the Czech National Bibliography.
Because we selected AACR2, we had only limited sources where to find instructions for and answers to UNIMARC-related problems. Very useful were the UNIMARC-USMARC Conversion Tables prepared by the Library of Congress. In many cases though we had to rely on written PUC consultations, assistance from colleagues from abroad familiar with USMARC, and of course also slowly on our own growing expertise. Currently, we are preparing a Czech edition of AACR2/UNIMARC interpretations and examples. All these tools are essential for consistent application of the rules, especially those that are a novelty for Czech catalogers. We do not encounter major problems in the area of bibliographic description since we have been using ISBD for close to two decades. Much more frequently we have problems with choice and form of access points, specifically corporate bodies and uniform titles, less frequently the type and form of subject heading based on LCSH.
Our catalogers face three difficult areas: choose the correct access point(s), establish the form, and record it in the appropriate UNIMARC fields and subfields. Some Czech libraries implemented the change from original Czech cataloging rules to AACR2 and LCSH relatively quickly and without special obstacles, while in others the traditional processing lingered on. During the interim period, some records created by individual libraries, although all in UNIMARC, were quite different and incompatible. This experience led us to the conclusion that using the same format is not always the answer to a successful cooperation -- there is a need for commonly used rules and interpretations. Especially lately when we introduced cooperation also in the area of authority files, is this fact getting more prominent.
The UNIMARC Manual: Bibliographic Format can be evaluated positively in its relationship to AACR2. The only truly problematic aspect is the Linking Fields Block. PUC is aware of this problem and it suggested a solution: substitute strictly defined subfields with embedded fields. Due to this development, the NL abandoned its own project which would have helped our cataloging agencies.
At the present, we deal with this issue by sometime coding both 4-- and 975 fields, e.g.
At the same time, we expect that fixed subfields will be able to represent all types of data which can be
found in the linking fields, e.g. author/title added entry.
On the whole, we have only marginal criticism of the Manual, some of which we already forwarded to PUC in 1996. A few problematic and difficult aspects of the format remain:
When it comes to the creation of authority records, the UNIMARC format lacks coding to define series accurately, something that is possible to do in USMARC (see later).
UNIMARC still did not come forth with an adequate holdings format. Fortunately, we are able to code this information in ALEPH.
We view the more-or-less theoretical orientation of UNIMARC and its makers' limited cooperation with libraries which try to do their best to use it in their daily application, a definite weak point of the system as such. It is essential, for the further development of UNIMARC, to follow and study individual countries' additions to various indicators, fields and subfields which obviously reflect the institutions' requirements. We realize that it may not be possible to gratify all the national requirements since that could lead to a certain "softening" of this international standard. However, it is important to consider that constructive criticism is probably deserved and it should be dealt with positively.
In some situations, we code indicator and subfield with the value 9. In such fields we then record data which are not necessarily sanctioned by AACR2, but an institution may deem them important for its clientele. This way we also often deal with block 4. To illustrate this point, below are a few examples which we dealt with recently.
In the first record, AACR2 call for the use of a Collective uniform title, whereas indicator 9 in field 327 allows to record as access points also titles of individual works:
Or we use field 423 when we want to provide access also via the second work contained in the book, a common situation in AACR2:
This coding allows us to upload our records, created according to AACR2, in their correct version, without creating any problems for agencies in other countries. The appended data are identified as such so that an institution can evaluate and consider whether to accept the record as is, adjust it or not use it at all. When these extra data are of no use to an agency, they can be automatically stripped from all our records by the agency's loader rather than one by one as the records are being encountered. Next to the use of indicators with value 9 we already defined fields and subfields. Often, these are archival fields which we use for recording of older forms when they need to be included in our union catalog. We create additional new active fields in block 9 when AACR2 is insufficient for the description of certain types of documents such as incunabula and other old materials, where not even fields, added in the newest UNIMARC Additions are adequate for description of such documents, to the same extent that we are used to.
Only the National Library and research libraries use the above expansion of some of the UNIMARC fields. In general, for a monograph record to be included in our union catalog, it must contain
Our implementation of the UNIMARC/Authority guidelines is also progressing smoothly, although we have a limited experience with it so far. We prepared instructions for recording of personal and corporate names, uniform titles, series, and subject headings. We opted to enrich the series authority records with some USMARC fields since we consider them important, namely 050 Series call number, 642 Series numbering example, 643 Place and publisher, 644 Analysis practice, and 645 Tracing practice.
It became clear to us that for the creation of a quality record mere familiarity with international standards is not enough: we also have to create a mechanism which will govern the authority file, including its technical, organizational and financial security. This task would be greatly simplified if we could access good authority records created in UNIMARC by other national cataloging agencies.
The National Library is now also using international standards in the area of subject analysis. Coding UDC in field 675 is simple, but there is a marked difference in the subject headings which we now use. Subject Headings of the National Library (SHNL) are a variation of LCSH; starting in 1999, records included in the Czech National Bibliography will also contain regular LCSH, and of course records that we export from the WorldCat have English subject headings. The use of the LCSH model within UNIMARC renders subject headings more concise and easy to display, not to mention that all subdivisions are clearly delineated, thanks to the hierarchical arrangement. This too is a departure from our traditional treatment and presentation of subject headings.
We see one universal format allowing for the exchange of records without conversion to be the ideal solution. The BL/LC/NLC MARC Harmonization Initiative and UNIMARC improvements and changes will most likely result in different formats. The more similar the individual formats will be, the easier the conversion. For a truly effective cooperation among libraries, both the use of a common format and adherence to one set of cataloging rules are a prerequisite.