Workflow Analysis of the Marcive Process

Interim Report 2/4/04
prepared by the MARCIVE Load Work Group, a subgroup of the Government Publications Committee

Background

In January of 2002 the Work Group for Adding Catalog Records for the Federal Depository Collection recommended that RUL subscribe to the MARCIVE enhanced GPO (Government Printing Office) database service and the MARCIVE shipping list service. The shipping list service (SLS) delivers brief, place-holder MARC records that match our depository shipments. The enhanced GPO database service delivers full bibliographic records cataloged by catalogers at GPO, that match the RUL depository item selection profile. These subscriptions together with GPO authority files were instituted in August 2002. The Work Group also recommended purchasing a customized retrospective extraction with matching authority records for the cataloging backfile, following a pilot project to determine best treatment practices. This has not been implemented.

In the months preceding the start of the subscriptions, the Work Group reviewed several test files from MARCIVE. As a result of this extensive review, the files load as expected and without incident. The profile of items selected by the New Brunswick libraries (6292 item selection numbers selected from 6705 available) was modified for the MARCIVE shipping list service to exclude records for serials and topographic maps. Topographic maps are also excluded in the MARCIVE enhanced database service; however serials are included in separate files.

The full bibliographic records in the enhanced database service have been grouped in the following files:

These 8 files, plus one file for authority control records, are received monthly. Two report files, for new and changed authority records, are received along with the bibliographic record files. The brief temporary records from SLS are provided weekly.

A chart of the numbers of records loaded to date, provided by Bob Warwick, is appended to this report. The over 8700 brief records loaded in the SLS subscription to date represent print, microform, or CD-ROM items immediately accessible to RUL patrons. Over 3400 of these brief records have been overlaid by full bibliographic records; many include online access. The 681 records in the MON-NEW-OL total represent titles that GPO did not distribute in print. We rely on the PURL included in the catalog record provided by our MARCIVE subscription for online access to these titles. To date, we have added 8854 records with PURLS to IRIS.

In the past few months, the MARCIVE Load Work Group has charted the workflow processes for cataloging the federal document collections. It is important to note that analysis of the process has been a constant factor in the implementation of the MARCIVE service. We have made modifications, as more efficient procedures became apparent, and have been diligent in including documents in systemwide standards, as for example, those developed for MARC holding statements. It is important to keep in mind the labor-intensive nature of this undertaking, given that in the past, cataloging efforts for the federal document collection were minimal at best. The collaborative effort evident in this endeavor assures that all three depository libraries at RUL will be in compliance with GPO's mandate for piece level holdings as we transition to a primarily electronic documents environment.

Accomplishments

People in all three areas of the Rutgers University Libraries - systems, cataloging, and the units, are responsible for the successful transition from an uncataloged, underutilized documents collection to the highly visible and accessible collection made possible by the MARCIVE subscriptions. A very significant benefit of the MARCIVE project was the cultivation of a seamless perspective for all those working on the implementation. Each representative gained from the revelation and analysis of all workflows. The following outline highlights some of these accomplishments:

Systems:

Staff retrieve files from MARCIVE on a regular schedule.

For the SLS, Marcive announces the file availability on Thursday afternoons, and the files are loaded by the following Monday at the latest, and usually within 24 hours. The monthly authority files are loaded on a regular basis as announced. The full bibliographic records for monographs and serials, referred to by MARCIVE as the GPO database, are retrieved each month as announced, and loaded as requested by the Work Group.

Loader scripts written for all 10 files, and reloads for some.

Reports for unit process and the Cataloging Department are created following each load.

Cataloging:

Authorities processing normalized.

Multiple formats "root record" established.

Resolved the occurrence of multiple OCLC 035 fields:

In order to avoid unintended overlays and subsequent recataloging, procedures were established for choosing among multiple 035 OCLC record numbers (the match point for change records) and recording this information appropriately in bibliographic records. Such procedures are essential if we wish to maintain the integrity of our root records.

Procedures established for Workflows review files updates:

Special reload scripts were written and methods of handling records explored.

Establish new sublocations: DOCUS-MAP for SuDoc classed maps:

Sublocations for documents were normalized for all units in RUL.

Ceased initial review of MON-NEW file in Cataloging:

Duplicates or problems are reported from the unit when discovered.

Developed process for Robeson Library use of LC classification for documents.

Campus Libraries

All three federal depository libraries use the MARCIVE records to record holdings:

Library staff from ALEX, LSM, DANA, and ROBESON collaborate to insure the uniform standards of the integrated library system are applicable to the document collection.

Modified the SLS to cancel temporary records for congressional hearings:

The full bibliographic record for print hearings is accepted for microfiche copies of hearings in the collection, conforming to the RUL merged format "root record" standard.

Use the reports generated by the Systems Dept. to produce files for library staff to work with:

We have standardized management of text report files for monographs and serials, and produce a table for DANA library profiled for item selection for MON-NEW files.

Everyone who formerly spent time updating card catalogs and performing other manual functions, are now working to understand the impact of an integrated library system and learning how the melding of software produces more efficient workflows. Their technical and communication skills have benefited enormously from the project.

As much as possible holdings for Depository material are evaluated with a mainstream perspective and mainstream standards are applied resulting in the following:

MARCIVE records are incorporated in standard unit procedures, as for an example,

MARC Holdings statements now include elements needed for document serials.

Non-depository federal titles are included in procedures.

A process is in place for retrieving documents with a status of "pending" for Rutgers Request Service.

Decisions

Many decisions have been made since the inception of the process, to facilitate streamlining and automating procedure wherever possible. A few of the more notable include:

• Share a common listserv feddocscat_rul@email for communications to all groups.

• Person discovering a problem notifies MARCIVE directly, and shares solution with the group.

• Accept NASAT (NASA Thesaurus) as "secondary subjects" for NASA documents after consultation with IPAC.

• Depository videos located at MEDIA Library (ALEX/LSM items) or DANA MEDIA:
  • The physical item does not need to be sent to cataloging, rather, a request for an accession number is e-mailed to Cataloging/Special Formats and the material is processed in the units before it is sent to the MEDIA Library.

• Identified key elements for manual review of change records and merged formats, both monographs and serials.

• Accept PURLS from GPO, preferred over agency URL, and rely on the GPO check system. 856 update reports for all files loaded disclose duplicate entries for deletion.

• Accept the MARCIVE record as the bibliographic record for monographs. If a duplicate IRIS record already exists for the title, we will merge holdings and delete the existing IRIS record. If there is only a shipping list record from MARCIVE, and a duplicate full IRIS record, the same procedure will apply; only the shipping list record will be retained, with any existing holdings, in expectation of receipt of a full record from MARCIVE.

• Accept the MARCIVE record as the bibliographic record for serials unless an IRIS record already exists for that title. The existing IRIS record is maintained and updated if needed with elements from the MARCIVE record.

• If a serial title is "born digital", the unit does not check for holdings until or unless a tangible item is received.

• Serial histories are not pursued beyond the immediately preceding or subsequent titles. Holding statements are closed when a MARCIVE record indicates that a title has ceased or changed.

• Use MARC Holdings statement for monographs with unusual housing, e.g. slip laws, maps.

• Remove "tmp" numbers (temporary identifying numbers) from all full records, both newly created and overlays.

• Use SERMARPEN (SERial MARcive PENding) note in 500 field for new serial titles with a brief record in IRIS. When the brief record is overlaid with a full bibliographic record, this note is deleted.

Pending

There are a few things left on our list of things to do before we reach our goal of normalizing the entire MARCIVE process into the technical service workload. The list to date includes:

• Profiling serial lists for DANA by item selection.

• Standardization of item holdings (consistency in bar-coding pieces):
  • This has been referred to the Holdings Work Group for input.

• Normalization of all load schedules.

• Review of "hanging" records:
• For example, SLS tmp records that were not overlaid, records that never got a date cataloged, records with a status of pending, etc.

• Review of brief records with SERMARPEN notes.

• Review of item profile with MARCIVE (re: GPO item update each October).

• Automation of MON_NEW_OL "date cataloged" with load.

• Development of Campus library process for reloads of "new" titles.

• Tracking GPO's use of separate bib records for indexes and supplements to incorporate into RUL best practice.