Sophia McMillen distributed the Subcommittee's draft proposed shared database authority work guidelines. She noted that the draft was essentially the same as one distributed previously. Sophia said the general principle behind the guidelines is to establish headings according to existing cataloging standards.

Sophia said one question still under discussion within the Subcommittee is headings without authority records. Some people feel every heading needs an authority record. Other people feel such a policy is too impractical. Sophia said the Subcommittee might leave the decision to each cataloger's discretion. She noted that the two situations where all are in agreement that an authority record needs to be created for a heading are (1) when the heading needs a cross-reference, (2) when, for a series, you need to record series treatment decisions. With respect to cross-references, Sophia noted that one question to be resolved is whether locally-created cross-references can be protected from overwriting during import/merge procedures in the authority database. This is an especially important question for subject heading cross-references, since new LC subject authority records will be loaded into the shared database via bulk import on a regular basis. Sophia observed that some testing indicates there may be no way to protect local cross-references from being overwritten during batch loads.

With respect to series treatment decisions, Sophia directed members to page 5 of the handout, which showed a sample shared series authority record with MARC data from different libraries describing their series treatments. Different tags hold data for different aspects of series treatment (e.g. tag 644 Series Analysis Practice, tag 646 Series Classification Practice). In each tag, subfield a holds the code for the treatment practice, while subfield 5 holds codes for institutions to which the field applies. The Authority Subcommittee proposed that in subfield 5, libraries use their MARC Organization code (formerly known as the NUC code) to conform with national practice. The SCCC agreed to the subcommittee proposal to use MARC organization codes for series treatment decision data, rather than any locally-defined codes.

The SCCC also agreed with the Authority Subcommittee proposal that the core tags for an authority record would be the 1XX tag containing the heading and the 670 tag citing the work cataloged that prompted the heading. If a 4XX tag is added for an alternative form of entry, it should have a 670 citing the work cataloged that prompted inclusion of the 4XX.

The SCCC worked out a timeline for initiating work in the authority database (based on the assumption that there will be at least one more meeting before summer):

1. Review of the Endeavor default authority tag tables would be completed by the next meeting. Michelle suggested to K.T. Yao that if her Tag Table Subcommittee needed assistance with the review, they might tap members of the Authority Subcommittee.

2. Development of training materials for local creation of authority records in Voyager would be completed by the next meeting. It was hoped that the Authority Subcommittee might be able to do training as part of the meeting and/or do training via a distance education format such as e-mail or the Web.

3. Importing authority records from the LCNAF into the local authority database might be implemented earlier than the next meeting, once the Authority Subcommittee finished doing functionality testing and finalized procedures. Libraries with access to the LCNAF via OCLC or RLIN such that they can make copies of authority records for import could contact Sophia if they wished to start importing records.

   Michelle asked if any SCCC members would like to help test authority functionality. Diane Johnson volunteered to help with testing.

   With respect to operator profiles, it was proposed that authority work only be open to those who already have add bib capability. It was also proposed that there be three levels of access:
   

   • Edit only (for those only interested in adding cross-references or series treatment data),
     
   • Add+Edit (for those interested in importing LCNAF records and/or locally creating authority records),
     
   • Add+Edit+Delete (for those interested in assisting with authority database clean-up projects that are likely to be an outgrowth of the authority record bulk import activity).
     

K.T. Yao showed members the two virtual hosts being used to test the two different WebVoyage interfaces for CJK display. The glyph server treats CJK characters as images. Users do not need recent versions of web browsers or special font files to see the characters. But the characters cannot be saved as text for copying into another document the way Western characters can, and because only someone who can see the images can know what the characters are, the interface is not ADA compliant. The Unicode server is supposed to treat CJK as Unicode-compatible text. It therefore only works with browsers that support Unicode and have the necessary fonts loaded. K.T. represents UH on Endeavor's Unicode Display Task Force.

K.T. also reported that UH Hamilton had begun efforts to convert all of its Chinese records with Wade-Giles romanization to pinyin romanization, (something OCLC, RLIN and LC have already done). Hamilton purchased pinyin versions of its RLIN Chinese records, which would be loaded into the shared database to overwrite their Wade-Giles counterparts. Fred reported that pinyin loads into the production database would begin the week of Jan. 28-Feb. 1, 2002.

In addition to the Chinese records, Hamilton decided to reload all of its RLIN records, since CJK records suffered EACC code corruption in the migration from CARL to Voyager. (Certain EACC code elements ended up being treated as subfield delimiters, resulting in characters not displaying.) Fred said the loads would be done 5,000 records at a time, first Chinese, then Japanese, then Korean. At an estimated load rate of 400 records/hour, the RLIN reload was expected to take about a month.

1. New Librarian
   Fred Allen informed the group that Carol Kellet had accepted the vacant librarian position in the Systems Office. Once Carol has settled in, she will take over Fred's duties as the Systems Office catalog module support person, and as the Systems liaison to the SCCC.

2. Load schedules and keyword index regeneration
   Fred reported that the RLIN record loads would be going on at the same time as Marcive record loads. The load schedule called for Marcive loads to be done mornings and nights, RLIN loads to be done afternoons. Fred explained that adding such a large volume of records to the database would necessitate periodic regeneration of the keyword indexes to maintain optimal system response time. While the first regen had yet to be done, plans were to do them on Sundays. During an index regen, the production database is frozen. Fred asked if anybody customarily did cataloging work on Sundays. Catherine Thomas at Law said she did. Since the regen was expected to take 9 hours, Fred asked if it would be workable if the database were frozen from 6 am to 2:30 p.m. on Sundays. Catherine said such a schedule would be workable.

3. BatchCat
   Fred said he had been looking at the Endeavor BatchCat product, a sort of programming toolkit that allows you to create a front end for batch cataloging operations. Fred said he had been able to do limited batch deletions. He hoped to do some testing with batch location changes and batch item type changes.

4. Cataloger's Toolkit
   Hamilton has been investigating Cataloger's Toolkit, a product developed by Gary Strawn and used by catalogers at Northwestern University. Sophia said it is an alternative front end for Voyager cataloging. Fred said he hoped to follow-up with Gary Strawn on how the produce works. Lisa Sepa asked if individual sites could decide to use Cataloger's Toolkit on their own. Sophia said she would try to find out.

5. Upgrade to Voyager version 2001.1
   Fred informed the group that Systems was proposing to upgrade to Voyager 2001.1 after finals, (tentatively on May 20 & 21). He reminded the group that Voyager is totally down during an upgrade. Michelle observed that the upgrade would be the first test of Systems ability to take a snapshot of the shared test database and restore it after the upgrade resets it to a generic Endeavor training database.

6. Marcive records
   Fred reported that the list of tags Hamilton Gov Docs had given as needing protecting during Marcive loads was: 024, 035, 050, 090, 590, 591, 898, 971. Fred asked other libraries that catalog government documents to let him know if they would like any additions to the list of protected tags.

7. LCSH loads
   Fred reported that he had done a test load of LCSH records that morning using a file of 37 records, some new, some with existing counterparts in the database. Fred said he would have a report to the Authorities Subcommittee by Tuesday or Wednesday. Noting that the Subcommittee had problems importing records manually, Sophia asked Fred what setting he used. Fred said in duplicate detection, he had the system check on the 010 and he checked the Replace box. Fred said he would check to see if the import triggered anything in the global headings change queue. Sophia observed that there are a lot of outstanding questions involving the global headings change queue, such as who oversees it, and how do server jobs get run.

At the last Hawaii Library Association meeting, an OCLC representative gave a presentation to Sophia McMillen and Thelma Diercks on possible ways that libraries not already using OCLC could become users via various group access options. Michelle said she had a copy of the documentation from the presentation and could provide copies to anyone else who was interested. Nadine Leong-Kurio and Lenore Maruyama said they would like to have copies.