WGC Minutes - April 11, 2001
Present: David Banush (recorder), Yumin Jiang, Marty Kurth, Linda Miller, Margaret Nichols (chair), Jean Pajerek, Lois Purcell, Cecilia Sercan, Pam Stansbury, Tom Turner, Iris Wolley, Sarah Young.
1. Metadata recommendations for digital image collections
Tom led a discussion of the recommended metadata requirements for the preservation of digital image collections. These recommendations are included as part of the report "Establishing a Central Depository for Preserving Digital Image Collections", which builds on work completed with IMLS grant funding and is available at http://www.library.cornell.edu/imls. The Digital Preservation Policy Working Group, charged with taking the IMLS grant findings and creating a set of Cornell policies, was co-chaired by Anne Kenney and Oya Rieger. The other members of the group were David Block, Erla Heyns, Peter Hirtle, George Kozak, Joy Paulson, and Tom.
As Tom noted, millions of digital image files have been created at CUL over the past decade. As a result, we have a great deal of legacy digital data. Some this data had been in aging formats or on older media, and much of it was lacking information about the files themselves (the software/hardware used to create them, etc.). Thanks to work done with the IMLS grant, these problems have been resolved. The IMLS grant was instrumental in developing a policy to address the issues of long-term preservation and access to these images. It will be especially important to ensure that necessary administrative metadata is in place to manage the files as they age and technologies advance.
The guidelines would ideally be used for images stored in a proposed central repository, but could also apply to any images created around campus, regardless of where they are stored. The online report discusses many issues, including selection, legal and technical matters.
The descriptive metadata guidelines are general and are not highly detailed. Some of the primary recommendations include:
- The use of the 899 field in all CUL MARC records for collections of images.
- The creation of PURLs where appropriate.
- The use of the values recently added to the 007 field (image bit depth, file format, quality assurance targets, antecedent/source, level of compression, and reformatting quality).
- Cross-walking MARC data to populate fields in other metadata structure formats.
- Following the guidelines laid out in the CUL Cataloging Procedures for Networked Resources, as well as such documentation as Nancy Olson's Cataloging Internet Resources, LC's Draft Interim Guidelines for Cataloging Electronic Resources, and LCRI 1.11A, governing reproductions.
In response to a question from Jean about how catalogers would know some of the information needed for the 007 field, Tom stressed that staff creating bib records for these collections would have to be provided with the needed details from the onset; we will not be able to guess the values later. Cecilia noted that microforms have for years given this sort of information on the film or fiche itself and wondered if digital images would be similar. Tom expressed hope that they would be.
Some lingering questions Tom pointed to included the use of 007 by other institutions, as well as file migrations: if files are moved from format to format, how will we keep track of this and update the records as needed? He also noted that while processes for keeping track of this data might be cumbersome for technical services, they might also be inadequate; preservation metadata needs are far more detailed than what MARC 007 can provide. Finally, questions remain regarding the maintenance of synchronization when there are multiple collections of metadata about the same resources, as well as administrative metadata generally, which MARC is ill-equipped to handle and for which there are not yet standards (though a NISO group chaired by Oya Rieger is working to develop standards).
Tom welcomes feedback on the proposal from any interested parties.
2. CONSER Module 31 regarding cataloging of remote access computer file serials
Yumin led a discussion of CONSER's interim policies on multiple distributors of electronic text and the changes to Module 31 of the CONSER manual dealing with electronic serials (available at: http://www.loc.gov/acq/conser/mod31pt1.html.) As she noted in documents distributed prior to the meeting, a PCC task force has been charged with recommending long-term policies on cataloging multiple manifestations of electronic resources. In the meanwhile, CONSER has recommended a policy for dealing with records offered by multiple aggregators. The policy permits a single record approach when both print and multiple electronic aggregator versions are available, and also permits separate records for each distributor or aggregator. However, the policy forbids a single record for multiple aggregator versions.
Other guideline highlights include the use of uniform titles with multiple qualifiers, with aggregator names added to e-versions of print titles, and a policy of linking print versions through 776 fields to their (several) electronic counterparts, but not linking various aggregator versions to one another. The policy also instructs catalogers not to separate electronic multi-version records if they are encountered.
Yumin pointed out that CUL has used multiple electronic versions on one record, with one serv,remo location in the OPAC for all holdings and a Gateway record with qualifiers and the holdings specific to the particular aggregation. She also noted inconsistencies in our practices with aggregators: Elsevier titles have separate records, while others like JSTOR and ProQuest are done as multi-version single records. The new CONSER guidelines may further add to the confusion for both users and catalogers.
After some discussion of the guidelines and their implications for CUL, we agreed to ask Jim LeBlanc to update the current cataloging procedures for networked resources so that we would know how to handle these situations in the future. We also agreed to leave our current multiple electronic version records as is, even though there may be some confusion about them. Sarah pointed out that users may not be as confused about these things as we believe, as they are often satisfied with finding an electronic version regardless of its aggregator.
3. Proposal for abbreviated-level, machine-generated records for e-journals in aggregators
Marty led a discussion of a proposal for creating, in a largely automated fashion, brief or "sleek" records for individual serial titles in aggregator sets. He noted that there is a need to have some level of access to the titles as soon as the aggregation is made available, but that currently, our means of doing so are rather limited. One approach would be to use a static list on a Web site to alert users to the availability of certain titles, but that method represents a kind of separate catalog for searching. A more optimal solution would be the possibility of generating brief Voyager records from title lists provided by vendors when the aggregator is activated and then replacing those records with full-level records in cataloging projects conducted as we do now.
The NERL Wiley project allowed a test of this method. Marty noted that a working group consisting of Adam Chandler, Lois Purcell, and Cecilia Sercan had followed the recommendations of the PCC SCA Task Group on Journals in Aggregator Databases, as well as the CONSER minimal record elements, in constructing the templates for machine-generated brief records.
There were a number of questions about the encoding level used. Cecilia said that encoding level 3 (abbreviated) had been chosen over encoding level 7 (minimal level) because AACR2 capitalization rules will not be followed in these records. Given that we anticipate full cataloging for these titles, some members of the group wondered if encoding level 5, for preliminary records, would be more appropriate.
Tom noted that the MyUpdates component of MyLibrary would want to include these records in its data capture. He asked that the 948 codes be constructed so that VgerSelect would be sure to pick them up.
Sarah asked about how holdings would be maintained, if at all. If the aggregator adds retrospective holdings at a later date, would we provide only the holdings data available initially? If not, how would we propose to update the holdings? David wondered about selector requests for full cataloging if no copy were available. Marty responded that holdings maintenance overall would continue to operate as it does currently. With regard to brief records, if the coverage of a given aggregator changed dramatically before the brief records for it were replaced by full records, it would be relatively straightforward to replace the original set of brief records with an updated set. Selector requests for full-level cataloging of individual titles in aggregators would move through the e-cataloging stream as they do now.
Marty noted that IRPC will be discussing the Wiley project at an upcoming meeting and asked what other groups might also be consulted. Tom suggested that General Selectors or CD Exec also provide their feedback.
Minutes: David Banush