Authority Control Advisory Committee

Authority Control Advisory Committee Minutes


Date: 2005-10-04

Present: Jerry Anne Dickel, Ellen Ellickson, Eileen Smith, Karen Spicher, Manon Théroux (chair and recorder).

Absent: Eric Friede, Earl Roy, Rick Sarcia.


1. Membership

Rick Sarcia and Jerry Anne Dickel have renewed their membership on the committee for two more years. Integrated Systems & Programming (IS&P) has been invited to nominate a replacement for Marsha Garman, now in the SML Acquisitions Department. Susan Tucker has finished her term. Sadly, Earl Roy will be leaving Yale later this month. Manon will look for replacements for Susan and Earl.

2. MARS Processing of Current Cataloging

There has been no MARS processing of current cataloging since mid-February because of the committee's testing of the MARS service after its move from OCLC to Backstage Library Works. Testing had to be repeated several times because of errors on both sides. Our first test revealed that we had done the record extract incorrectly; subsequent tests revealed that MARS had done their setup incorrectly after the move from OCLC. Final test results were posted here:

http://www.library.yale.edu/cataloging/authorities/marsvoyagerQA2005.html

Most of the problems revealed by the testing were either known problems or problems not significant enough to affect our decision on whether to resume processing. However, the problems relating to our 880 fields (which contain nonroman characters) were significant enough that we thought it worthwhile to delay processing until a fix was in place. These problems had to do with second indicator values: 880s that correspond to 6xx fields are supposed to be coded 2nd indicator 4 but MARS was changing them to 0. Also, the non-filing indicators used for 880s will sometimes differ from the nonfiling indicators used for their associated parallel fields (e.g. the Hebrew article "ha-" requires a non-filing indicator of "3" in romanized fields but a nonfiling indicator of "1" in nonroman fields because the article is represented by a single character in the Hebrew script). MARS was making the nonfiling indicators in nonroman fields match those used in the romanized fields.

Manon has just sent off what should be the final test of the MARS fix for the 880 indicator problems. If all goes well, we could resume processing as early as the week of October 10. The plan is to send 1-2 month's worth of records every week until we are caught up with the backlog. All loads must be completed before the Unicode conversion of Orbis.

[Incidentally, it was just discovered during testing of Voyager 5.0 that the Hebrew character for the article "ha-" is a double byte in Unicode, which means that the nonfiling indicator needs to be "2" in order to retrieve the record. But this problem is separate from the MARS problem.]

3. MARS Processing of RLG Snapshots of Our Arabic, Persian, Hebrew, and Yiddish (APHY) Records

Cataloging staff in our nonroman units have been adding nonroman fields to their bib records in RLIN for quite some time. However, until we migrated to Voyager in 2002, these fields were stripped when the records were loaded into Orbis. As part of our Voyager implementation, we ordered a snapshot of all of our CJK records from RLG and loaded them into Orbis, overlaying the existing records with records that contained nonroman fields. However, because of various display problems, we postponed getting RLG snapshots of our APHY records until we migrated Orbis to the Unicode version of Voyager. This migration is currently planned for late December 2005/early January 2006. We have ordered a snapshot of our APHY records from RLG, will send the records through MARS processing, then load them into Orbis as part of our Unicode conversion. Catalogers of APHY materials will then start sending their daily cataloging to MARS using ExportQ like everyone else.

Some complications:

-- Although the RLG union catalog is in UTF-8 (Unicode), the RLG snapshot will be in MARC-8. RLG is not yet exporting records in UTF-8. The MARC-8 records will be converted back into UTF-8 as part of the Orbis conversion. Unfortunately, we will inherit some of the MARC-8 display problems with Hebrew records (especially with multiple and bracketed dates in 260 fields).

-- The snapshot records will contain obsolete area curator codes (e.g. jud98 05) in local uniform title fields. These local headings used to be added as a way of tracking new acquisitions but the practice was abandoned years ago. The headings, which will be tagged 699 or 630 (2nd indicator 4), will need to be deleted. Although the 699 field is not indexed in Voyager, it should be easy to find these headings using a keyword search and implement a special project to delete them.

-- The snapshot records will also contain special tracings that need to be retained. Unfortunately, these will be tagged using RLIN conventions, not our current Orbis conventions, and some of the RLIN tags (696, 697, 698, 699) are not currently indexed in Orbis. These fields will all need to be retagged. Unfortunately, the headings could contain almost anything (so it will be impossible to find the headings and convert the tags). We need to find out what our options are from IS&P. Some possibilities: Could Endeavor convert these for us? If not, could IS&P run a report to identify them even though the fields aren't indexed? Could Endeavor add the RLIN tags as new left-anchored indexes and also add them to the composite Local Subjects index (cleanup wouldn't be needed at all or could be done on an "as encountered" basis)?

Tag Comparison Chart for Local Headings (all fields have 2nd indicator=4):

Personal Names: RLIN=600 or 696; Orbis=692
Corporate Name: RLIN=610 or 697; Orbis=693
Conference Name: RLIN=611 or 698; Orbis=694
Uniform Title: RLIN=630 or 699; Orbis=695
Topical Subject: RLIN=650 or 690; Orbis=690
Geographic Name: RLIN=651 or 691; Orbis=691

[Note: IS&P later indicated that our only option is to run reports (mid-spring 2006 at the earliest) and do the cleanup manually. We cannot complicate the conversion to UTF-8 with custom processing.]

4. MARS Processing of our Chinese, Japanese, and Korean (CJK) Records

When we implemented Voyager in 2002, we loaded a snapshot of all of our RLG CJK records into Orbis (see above). However, there was no time to send the records through MARS processing. Consequently, many of those records have incorrect forms of headings. This processing is still on our list of things to do for the future though it will have to wait until after the Unicode migration. The East Asia Library has been sending its daily cataloging to MARS since we moved to Voyager.

5. Authority Control Workflow for Yale Catalog Librarians

The "Authority Control Workflow for Yale Catalog Librarians" document is being revised, primarily to recognize that RLIN is not our only option for searching the national authority file (LC and OCLC are also possible). However, we must still must search the RLIN pending file before creating a NACO record. Committee members were asked to send comments on the draft to Manon.

New draft:
http://www.library.yale.edu/~mtheroux/acworkflow.htm
Existing document:
http://www.library.yale.edu/cataloging/authorities/acworkflow.htm


Site URL: http://www.library.yale.edu/cataloging/authorities/acac.htm
Comments to: Manon Théroux  <manon.theroux@yale.edu>

© 2005 Yale University Library

Top of Page