Utilizing Z39.50 to Obtain Bibliographic Copy: A Cost-Containment Study

Article excerpt

This paper looks at one approach to controlling costs when seeking cataloging copy. A small task group in the University of Minnesota Libraries Technical Services Department conducted a study to devise the most cost-effective strategy for searching for and importing bibliographic copy, by compiling costs and benefits of importing records from the OCLC Online Computer Library Center database, the Research Libraries Group Union Catalog (RLIN), and the Library of Congress (LC) catalog. Results of the study indicated that, although the LC database is smaller than the other two utilities, a sufficient portion of needed records were available from LC to more than offset the cost of re-searching in the other databases for records not found. In addition, due to differences in pricing structure, searching RLIN second was found to be more cost effective than going next to OCLC, even though a slightly larger proportion of items were found in OCLC than RLIN. This study may prove useful either as a research method or in terms of its findings for other libraries wishing to compare sources of cataloging copy.

**********

Many libraries and their technical services units face similar challenges, including steady state or declining operating budgets, implementation of new integrated library systems (or version upgrades), and continuing interest in improving processes to gain efficiencies. One goal in cataloging units often is to locate and use as many copy cataloging records as possible in order to speed movement of items to the collection and reduce the number of items that require original cataloging. This paper reports the results of a research project at the University of Minnesota Libraries (UML) Technical Services Department that explored costs and workflow issues related to obtaining bibliographic copy from the OCLC Online Computer Library Center (OCLC) database, the Research Libraries Group (RLG) Union Catalog (RLIN), and the Library of Congress (LC) catalog.

The authors describe the record import process and workflow prior to the project, present the questions that the project sought to answer, and report the research method and findings. A literature review places this project within the context of other cost studies. Implemented changes and the resulting financial impact are reported. Finally, the authors draw conclusions for other libraries considering a similar cost-benefit analysis.

Background

In spring 2003, a small group, the Source of Records Task Group (SRTG), was formed at UML to explore issues concerning cost and workflow for obtaining bibliographic copy from various sources. UML's technical services department was in the midst of revisiting and redefining workflow as a result of a July 2002 conversion from the NOTIS system to the ExLibris Aleph500 integrated library system, and the department was also seeking to reduce expenditures as part of a libraries-wide budget retrenchment.

UML has a longstanding policy of cataloging in the local system rather than directly in the OCLC database or RLIN and of using batch loading to contribute to the utilities. RLIN (now RLIN21) properly refers to the set of database services provided by RLG. In this paper, following common usage in the profession, RLIN refers more narrowly to the RLG Union Catalog. Because of the variety and breadth of UML's collections, the libraries have found using and contributing to both services beneficial. At the time of the system conversion in 2002, system implementers decided to discontinue using OCLC Passport and RLIN for Windows client interfaces (hereafter referred to as native clients) for the purpose of searching and downloading records into the local system. Instead, implementers explored the option of importing records through Z39.50, a protocol that allows a client to search multiple remote servers using a single search interface. One reason for this was a desire to provide staff members with a uniform interface for bibliographic searching so that they need only be trained in a single set of commands and search string formulation criteria. …