Earlier this afternoon, Lauren C., Leslie, Patty, Chris, Jean-Paul and I attended (watched? listened to? whatever) a NISO Webinar called Bibliographic Control Alphabet Soup: AACR to RDA and the Evolution of MARC. The program consisted of three presentations related to RDA and the future of the MARC format.
The first speaker was Barbara Tillett, the Chief of the Cataloging Policy and Support Office at the Library of Congress. She discussed the history of bibliographic control up to RDA (Resource Description and Access), which is intended to be a new cataloging code to supersede AACR2. RDA grew out of an attempt to develop an AACR3. RDA attempts to incorporate FRBR principles (which has been discussed in a number of other entries. If you have any questions about it, just ask me), and tries to be more universal than AACR2, which is tied to the English-speaking world. Furthermore, RDA reflects changes in technology (both in terms of the content it describes and how content is described), changes in focus (bibliographic description is not just for a local library, but for an international audience), and a change of view (moving from describing items to, in FRBR-terms, describing entities).
So, what does all that mean in practical terms? Well, the RDA code has two major areas it describes, elements of records (in database talk, entities and their atrributes) and relationships (between elements of a record, and between various records). RDA simplifies a number of the descriptive rules for cataloging, using a “take what you see” approach. Rather than qualifying information with parenthetical statements and re-ordering data, as AACR2 requires, RDA would note information as it appears on an item, which will make it far easier to harvest data automatically. RDA also makes the rule of three (the rule that no more than three authors can be listed) optional, gets rid of Latin phrases in notes, dispenses with GMDs (general material designations), gets rid of the “polyglot” designation, allows for more complete data in authority records, etc. All of these changes are made with an eye toward allowing data to be harvested and generated in automated ways more easily from the records, making the records more intelligible to users, as well as strengthening the relationships between records for related and derivative works.
The second presenter was Diane Hillmann, the Director of Metadata Management Services at the Information Institute of Syracuse. Hillmann was very knowledgeable about her topic, but moved very quickly and assumed a lot of familiarity among her audience with the topic she was discussing. It was fairly confusing, but we were able to identify her main point, which was that the exclusive use of MARC by libraries limits us in exchanging data outside the library silo. Nobody else uses MARC, nor are they likely to. Descriptive metadata use outside the library world is exploding and we’re not in on it. To get libraries into the general metadata game, part of the project of the RDA developers is to develop a vocabulary with defined data elements that can be used to create cataloging records, but that are also searchable and intelligible to the Web in general.
The third and final presenter was William Moen, the Director of Research from the University of North Texas’s School of Library and Information Sciences. He discussed a research project he and a team conducted from 2005 to 2007, in which they studied how many of the fields and subfields available in the MARC format were used and/or indexed by libraries in their bibliographic records. They did frequency counts and analyses of more than 56 million MARC21 bib records from the OCLC database. 211 fields and 1,596 subfields were used at least once. Looking at records in the Books, Pamphlets and Printed Sheet format, Moen and his team found that 7 fields appeared in all of the records, while 15 fields occurred in more than 50% of the records. Many, many fields had very few occurrences. The 656 field had only one occurrence. About 60% of all fields and subfields are used in less than 1% of the records. This led Moen and his team to consider the idea of developing core bib records in the MARC format that use a limited number of the currently available fields. By identifying the fields that are used in all bib records, combined with the most commonly used fields, Moen and his team developed proposed core bib records. However, Moen does not advocate simply leaving the decision up to statistical analysis. If we are to move to a more streamlined core MARC record, he suggests that catalogers think long and hard about what is actually needed in the bib record, and that the MARC format be revised with an eye toward supporting the FRBR-defined user tasks (he also asks if we really know which content designations are needed to support a given user task).
As the broadcast part of the webinar wound down, Lauren, Leslie, Patty, Chris, Jean-Paul and I engaged in a lively and interesting conversation about the issues raised in the presentations that last well-past our scheduled end time. That struck me as a very good sign that this webinar was quite worthwhile.