Ellen C. Rappaport
Albany Law School
Serials Meetings Galore!
What a treat was this summer's ALA conference for a serialist: I attended serials meetings for four days! I began with the LITA preconference: "The Future of Serials Control: Implementation of the MARC21 Holdings Format." It introduced the audience of librarians and system vendors to the Holdings Format, and to its application in serials control systems. The CONSER Publication Pattern Initiative was discussed, and use of its records was demonstrated. System vendors played an important role in the program, as they do in the Publication Pattern Initiative, because the Publication Pattern database will be useful only if our serials control systems can load, export and use its data, and if their systems facilitate contributions to that database. System vendors spoke frankly about what their systems can and cannot do today with this data.
This program was followed up the next day by a discussion of "What does a system vendor mean when he states that his system is 'in compliance with the MARC Holdings Format?'" at the MARC Formats Interest Group (MFIG) meeting. Out of this meeting and the Publication Pattern Task Force meeting came a draft paper to help us evaluate systems' use of the MARC Holdings Format. Just as we had to learn to evaluate what a system vendor meant 20 years about by "fully MARC compliant," referring then to bibliographic records, we now must develop similar questions — benchmarks — for the use of the MARC Holdings Format. That paper by now should be posted on the CONSER web site, for broader discussion.
* * * * *
The Publication Pattern Initiative database is growing. By June it had reached its stated goal of more than 1,000 pattern records input by participants, in addition to the 40,000 records loaded from Harvard data in February 2001. Input is continuing. In addition, members of the Publication Pattern Task Force are assessing the pilot project, recommending changes to the Holdings Format, and considering efficient workflow patterns for this data.
The Publication Patterns Task Force meeting considered the concept and use of Universal Holdings Data. This refers to a pattern-and-holdings record (in the MARC Holdings Format, of course) which represents the entire run of the publication and the publication patterns during the run of the title. The pattern fields (853/854/855) would express all the pattern changes of the title — showing changes in enumeration, chronology and frequency. The holdings fields (863/864/865) would express all the parts of the title, similar to a bibliographic 362 field, but expressed in the MARC Holdings Format, probably in Z39.71 style. The concept of a Universal Holdings Data Record is also a draft paper which should by now be posted on the CONSER web site, for broader discussion.
News just reached me at this writing of a program about the Publication Pattern Initiative for AALL 2002: "Publication Patterns: Creating Connections in the Serials World." More information about this next time.
* * * * *
Jean Pajerek's article in the September 2001 issue of TSLL reported on two MARBI papers which were passed in June 2001, dealing with changes to the MARC21 Bibliographic Format which will allow implementation of the proposed changes to AACR2, expected in June 2002. We need to begin to discuss these changes with our system vendors, because all will need to be implemented by some time next year; some are significant changes.
Proposal 2001-04 made field 260 (imprint) repeatable for resources that change over time, so that earliest, current and intervening publishing information can be included in the record. The earliest place/publisher are important to identify the title; later changes are important for checkin (Is this the right record?) and claiming (Who publishes this now?). In the past, we've put this information into notes, which were enough for 3 x 5 cards, but usually can't be retrieved in an OPAC.
Now we must think about — and discuss with our system vendors — how we want this additional place and publisher information displayed in our OPACs, and in our serials control systems. The dates for the earlier and intervening titles will be stored in $3; how do we want to display them? The proposal, available at lcweb.loc.gov/marc, suggests some possible displays.
Proposal 2001-05 added the new code "i" for "integrating resource" to "s" (serial) and "m" (monograph). An integrating resource is one that is changed by means of updates that do not remain discrete and are integrated into the whole. They may be finite or may be continuing. Examples are updating loose-leafs and updating web sites.
Proposal 2001-05 added a new frequency code to the bibliographic and holdings formats: "k" for continuously updated — more frequently than daily. An example is the OCLC database. Also added were the codes "l" (updating loose-leaf), "d" (database) and w (updating web site), added to existing codes of "m" (monographic series), "n" (newspaper) and "p" (periodical). Lastly the code "2" (integrated entry) was added to the Successive/Latest entry code. Code "2"will be used for integrating resources according to revised AACR2 Chapter 12, and for electronic serials that do not retain their earlier title. They will be given their most recent title, with earlier titles in 247 fields — yes, the old "247" is being revived, with a new application.
All of these changes will affect our OPACs and our serials control systems. We must think about what fields should display, field labels in the OPAC, which categories are retrieved if we ask for periodical titles or specify serials, how we will index and display 247 fields. Will we re-catalog all our loose-leafs with code "i" and if not, how will newly cataloged loose-leafs display with old ones?
Back in the early '70s it took four or more years for the Library of Congress, the embryonic CONSER, and librarians to move from latest to successive entry cataloging. And this time we have a new player involved — our system vendors. We must be activists in our user groups, so that our systems are ready for the new fields and values in MARC bib records. We have some early warning this time; I hope we can make changes more quickly than last time.
The Publications Pattern Task Force has attempted to involve system vendors and the utilities in the development of the MARC Holdings Format and the pattern database. Most of the major system vendors have participated in meetings and surveys of the Task Force, and in the LITA Pre-conference in June. Perhaps this can be a model for similar discussions about changes to the MARC Bibliographic Format.
In addition to the two proposals, MARBI and its audience discussed five serials discussion papers, DP07 through DP11. MARBI discussion papers precede MARBI proposals. These grew out of the Publication Pattern Initiative's experience with using the Holdings Format. These five discussion papers are now on the MARBI web site www.loc.gov/marc. All but DP07 are expected to become formal proposals which will be on the MARBI web site before ALA's midwinter meeting in January 2002. The discussion papers recommend enhancements to the Holdings Format which will allow us more refined and accurate prediction patterns, some of which will be particularly useful to legal continuations.
* * * * *
One topic pervaded the meetings I attended at ALA Midwinter 2001 — the re-use of metadata. For several years now, the common plaint has been that we librarians can't make the best use of the Web because we can't keep up with its growth. And with the apparent shortage of catalogers, this problem may extend to our more traditional forms of material, too. Shared cataloging may no longer be enough.
Roughly speaking, metadata is identifying and descriptive information about any resource. Our traditional form of metadata is the catalog record. The term has been used recently for identifying information about electronic resources. Metadata usually pertains to information in machine-readable form, which can be copied, transmitted, edited and re-coded, for example from HTML to MARC.
AACR2 defined three levels of cataloging in 1978. MARC21 defines eight encoding levels (Leader, position 17), from Full level to Minimal level and also Prepublication level, a preliminary record. The Core Record definition, initially developed by the Program for Cooperative Cataloging (PCC) in 1996 identifies descriptive elements, access points and control fields which must be present in a core-level record. Although not all elements may be present, those which are present in a core record should be as good as those in a full record, following cataloging rules and with appropriate authority work. A study to investigate "the cataloging community's response to the concept of the core record" was conducted this year. (www.loc.gov/catdir/pcc/bibco/coretudeprop.html) In addition, a study of users' perspectives on the core record for books (www.loc.gov/catdir/pcc/modelcfinal.pdf) suggested that "in general, the set of required elements in the Core record meets users' needs quite well. All elements that users identified as most useful are required elements in the Core Record Standard." However, the users in this pilot study were less successful identifying a particular version of a title, with both Core and Full records. "This suggests that users need a record structure that is easier to use and to understand than the current structure, yet one that still supports sophisticated user tasks." This was a pilot study toward further development of standards for the Core Record. The report indicated many places where further research should be done. For serials, CONSER defines three levels of records: minimal, core and full.
We began to accept less-than-full bibliographic records because we had to. We've been sharing the workload since the Library of Congress first issued printed cards about 100 years ago. Machine-readable bibliographic records and OCLC allowed thousands of catalogers around the world to contribute to the shared cataloging endeavor. But even the most minimal-level record requires a human cataloger, and we're not keeping up. Here are some of the proposed solutions I heard at Midwinter:
Prof. Lois Mai Chan of the University of Kentucky School of Library and Information Science stated, during the Committee to Study Serials Cataloging review meeting at ALA Midwinter 2001, that in order to cope with the enormity of cataloging the Web, we will need to employ a "tiered approach to bibliographic control" in which some resources will get more bibliographic control than others. She went on to suggest new uses of LC Subject Headings and LC and Dewey classification schemes which may help. A related paper can be found at lcweb.loc.gov/catdir/bibcontrol/chan.html.
During the same CSSC meeting, Regina Reynolds of the National Serials Data Project made a convincing case for the re-use of bibliographic data. She too urged a "hierarchy of record levels" starting with "hand-crafted cataloging" for material of the highest research value. Material of medium research value should get automated description which is edited by a cataloger, and authority control. For material of the least research value, we should obtain metadata in automated ways.
She proposed that we seek partnerships with agencies outside the library world which create bibliographic data for their own purposes. For example, to apply for certain identifiers, publishers submit bibliographic data on forms, some of which are machine-readable. CIP, copyright registration, ISSN and ISBN, and DOI receive bibliographic data from publishers. Since their bibliographic data isn't exactly like ours, she suggested that we work with these agencies to facilitate their input of better bibliographic data which libraries can adopt or adapt. Librarians can develop forms which help shape the publisher's input, for example with pull-down menus of controlled terms and required fields.
The metadata could be converted to MARC, loaded to OCLC and edited. To give the publishers a benefit from participating, it could then be returned in HTML, giving them better data to use for marketing on their Web pages.
Ms. Reynolds reported on a small study comparing NSDP bibliographic data supplied by publishers with the final AACR2 cataloging. She found that the publishers' data was about 75% correct, and could have been judged better than that. Two important fields — title (245) and numeric/chronological designation (362) were not counted as correct when the publishers' data didn't match AACR2 format. The information was correct, but the format wasn't. Ms. Reynolds suggested that we could use this data if our cataloging rules conformed more closely to common practice, such as more capitalization in the 245 field. A detailed version of this presentation can be found in Ms. Reynolds' paper at lcweb.loc.gov/catdir/bibcontrol/reynolds.html.
Records communicated in the ONIX format (Online Information Exchange), created by publishers for their use, are being used for marketing on such Websites as Barnes and Noble, Borders and Amazon. As an executive for Barnes and Noble pointed out during an ALA Midwinter 2001 program sponsored by NISO and BASIC (Book And Serial Industry Communications — formerly BISAC and SISAC), online book customers need more than the minimal bibliographic information in BIP. In the store, a sales person can help the customer to understand BIP listings. Furthermore, the customer can pull the book off the shelf and scan the summary on the jacket, the table of contents, the author's biography, and can even read a chapter or two. He knows how big the book is, and what kind of illustrations it has. ONIX has been developed to code this expanded description in XML, and has been mapped to MARC21. Like MARC, ONIX is a coding system which enables records from many sources to be transmitted and re-used. (www.editeur.org/onix.html, www.loc.gov/marc/onix2marc.html)
During the Midwinter 2001 NISO/BASIC program, Sally McCallum of the Library of Congress stated that while ONIX records are created by and for booksellers, libraries can benefit from them, too. In addition to the bibliographic data, ONIX records contain marketing information that library records don't now have: target audience, subject information, rights information, additional information about dimensions including weight, supplier and discount information. As Ms. Reynolds proposed, Ms. McCallum is working with the Association of American Publishers, Inc. to find ways to make these records more useful for libraries.
One LC project reported in LC Cataloging Newsline [v. 9, no. 13 (Nov. 2001)] has extracted table of contents data from ONIX records. A programmer at the Library of Congress has developed a program that scanned a file of ONIX records to produce over 10,000 records for tables of contents, now posted on the Internet. LC will add 856 fields to the related bib records, then will redistribute the records, with 856 links to this TOC file, through the Cataloging Distribution Service.
ONIX is now for books, but committees are being established to define coding for e-books and video. And first steps have been taken to include serials in ONIX. After surveying librarians, pubishers, serials agents, system vendors and standards bodies, its developers have proposed that serials information include "shipping data" which records when a specific issue has been sent out by the publisher. This is very useful data for claiming; the ONIX standard could make it more readily available.
ALA's ACLTS Committee on Cataloging: Description and Access (CC:DA) has formed a Task Force on ONIX International, charged to evaluate the relationship between library metadata such as AACR2 and MARC21 and the ONIX International standard. The Task Force will also identify the issues involved in using ONIX International metadata in AACR2 cataloging records. It will prepare discussion papers and rule revision proposals as needed, and will monitor projects that use ONIX International. It presented an interim report to CC:DA in June 2001, and will present a final report at ALA's Midwinter meeting in 2002. The interim report and other information are on the Task Force's Web page at www.ala.org/alcts/organization/ccs/ccda/tf-onix1.html.
* * * * *
The CONSER-at-Large meeting at ALA in June discussed possible new approaches to CONSER membership. One CONSER opportunity for law librarians would be Enhance Membership, the most minimal level of participation, in which one maintains existing records, dealing with title changes and cessations. CONSER is examining its training requirements and its quotas (minimum number of records you're required to improve, to remain a CONSER member). An Enhance Member must contribute a minimum of 75 transactions per year. A group of libraries can work together in a "funnel project" to achieve this minimum. For example, a funnel project is being organized among theological libraries. Is this something the AALL's OBS-SIS might want to explore?
* * * * *
In honor of its 200th anniversary last year, the Library of Congress sponsored the Bicentennial Conference on Bibliographic Control for the New Millennium. Its papers and resulting discussions are posted on the Web at lcweb.loc.gov/catdir/bibcontrol. Two interesting papers among many: "AACR2 and Its Place in the Digital World: Near-term Revisions and Long-term Direction" by Ann Huthwaite, who is chair of the Joint Steering Committee. And "Extending MARC for Bibliographic Control in the Web Environment: Challenges and Alternatives," by Sally McCallum, Chief of the Library of Congress Network Development and MARC Standards Office. The Bicentennial Conference Web site also contains the conference participants' recommendations for future action.