The Ohio State University

The goal of OCLC's Cooperative Online Resource Catalog (CORC) is to provide access to intellectual content available on the World Wide Web. CORC is a database of Web resource descriptions, contributed by participating libraries. OCLC provides the software for the description, and maintenance of the records. The CORC software is designed to accept input using either MARC or Dublin Core metadata.

The Ohio State University Libraries (OSUL) experimentation with the CORC project began September 1, 1999 and will end by June 30, 2000. The closing date corresponds with OCLC's move from offering CORC as an experimental prototype to offering CORC as a paid subscriber service. This report outlines the OSUL's involvement with the experimental project, and presents issues that arose during the experimentation that affect OSUL's future involvement with CORC .

Overview

OSUL's participation in the experiment can be divided into two phases. The goals of the first phase (Sept-Dec 1999):

• to have OSUL personnel gain experience working with Dublin Core (DC)
• to learn the CORC software in order to evaluate its strengths and weakness
• to have Collection Managers gain a comfort level in creating preliminary records for resources
• to try out different work flows for identifying and creating records and pathfinders
• to evaluate the process of exporting CORC records to OSCAR

The second phase (Jan-June 2000) was brought about by concern for the substantial workload of creating CORC records. The goals of the second phase:

• to experiment with minimally revised short records
• to concentrate on pathfinder creation
• to evaluate CORC as a public service

Initially, four working groups consisting of collection managers and technical service personnel were created. Participants in the project are from Agriculture (Connie Britton, OARDC; Ella Copeland, ATI; Susan Logan, Agriculture Library); Health Sciences (Lynda Hartel); Library Science (Carol Diedrichs); Reference/Information Resources Division (Nancy Courtney); and, Technical Services (Tschera Connell, Trisha Davis, Jose Diaz, Magda El-Sherbini, Jim Murphy, Cheryl Obong, and Noelle Van Pulis). Mary Rider from the OSU Law Library joined the project later in the fall.

Decisions regarding process throughout the project were made on two levels, project level decisions and individual group level decisions. Project level decisions were relatively few and included a project-wide status code to indicate records ready to be exported into OSCAR, and the establishment of four DC fields that were defined as mandatory for all records created for the project (Title, DescNote or DescSummary, ID- URL, and one occurrence of SubjLCSH). MARC fields necessary for the maintenance of records throughout the project were also prescribed.

Individual group decisions included which sites to enter into CORC, whether to enter pathfinders, and group workflow. In all cases, Collection Managers chose the resources, but the workflows varied greatly among the groups in terms of the point at which technical services personnel examined the records, and the degree to which the records were edited. In order to gain expertise with Dublin Core all participants were encouraged to enter and edit data using Dublin Core.

During phase 2, two groups experimented with Collection Managers entering records into CORC (and OSCAR) without cataloger review. For these groups the automatic CORC-software conversions of DC data into MARC fields was not evaluated. One group experimented with pathfinders. Additionally, conversations were held with Collections Advisory Council (CAC) to help evaluate CORC from a public services point of view.

An analysis of records shows that Law produced over 200 records during the CORC project. Other groups produced approximately 90; however, due to system problems, only 50 records made it into OSCAR. Of those, 38 were created by Collection Managers. Further analysis shows that there were no differences in the numbers of access points assigned. Both groups (Collection Managers and Technical Services personnel) assigned 4.2 access points per record, 2.5 of which were subject access points (see appendix). However, although the numbers of access points are the same, Collection Managers were more likely to create a field for keyword subjects, and catalogers were more likely to assign Library of Congress subject headings. It is encouraging that almost cases of access errors on records were due to improper tagging of the MARC title (245) field. If it is decided that Collection Managers will create brief preliminary records in CORC, the tagging of this field is not difficult to learn, nor difficult to correct. The experiment with Collection Managers entering brief preliminary records is a success.

Observations, Questions, etc.

The product, CORC:

• CORC is improving, but is still not as efficient as Prism. OCLC, in preparation for the upgrade of CORC to product status, is putting a great deal of effort into improvement of the software. During the time of OSUL's experimental project, however, CORC has been often down, and when operational, slow. As well, the harvester software (including the harvester for automatically created Dublin Core and MARC fields, and the harvester of potential sites for pathfinders) needs refinement.
• The pathfinder creation software for CORC is not the best mechanism for creating web pages. There is much easier and more flexible web page development software available. On the other hand, if there were more involvement of libraries using CORC for pathfinders, then CORC has great potential for enabling libraries to incorporate enhanced records in their pathfinders and for sharing resources.
• OCLC has already made the decision to combine the CORC database with the WorldCat database, so there is no need to decide upon CORC as a separate product. OCLC has also stated that the CORC software will become the cataloging interface of OCLC products in the future. Therefore, it is important for OSUL to maintain knowledge of the software

The process, CORC:

• The project participants like the idea of having Collection Managers create brief records in CORC and DC and then send them on to Technical Services for completion. For Technical Services, it is a great advantage to have Collection Managers identify subject headings that they would like to have included on the record.
• However, in terms of time, the question has been asked if it is it realistic to expect Collection Managers to enter records into CORC.
• What degree of training would be needed to assist Collection Managers in creating brief records? The experience of the experimental project indicates that minimal training is needed to create the level of record accepted for the project.
• Additional resources would be needed to add this workload to Technical Services. As the database grows more records for resources will be found and the cataloging process may become more similar to copy cataloging than it is now. As the software is improved, the difficulty of creating records should decrease. However, adding web resources on a large scale is not a workflow which could be absorbed by existing staff.

The context, CORC:

• The major issues of CORC for OSU involve how OSUL will present web resources to the public. Will intellectual access to web resources be provided through a digital registry of web sites, through the catalog, or through some combination of both?
• If the catalog is used to present web resources, what resources should be in the catalog? How should web resources be prioritized in terms of processing?
• What level of completeness should be used for records describing web resources? Should the OSUL standard be brief records, full records, or PCC Core?
• What is the role of Pathfinders? Should pathfinders be included in OSCAR? Should OSUL pages and/or pathfinders be cataloged as individual resources?

Recommendation

With OCLC's decision to fold CORC into the WorldCat environment, CORC has become primarily a web-based cataloging tool for web resources. From the perspective of process, continuing involvement with CORC becomes part of the OCLC package of services. Within that framework some level of continuation with CORC will continue in order to monitor the development of the product. The degree of involvement must be decided within the larger framework of public service decisions regarding how web-based resources should be made available to the public. It is recommended that OSUL's experimental project with CORC be brought to a close, and that future involvement be guided by decisions about the presentation of web resources to the public as well as workload issues.

Subject:
• Catalogers:   30/12 = 2.5
• CMs:        91/38 = 2.5

Overall:
• Catalogers:   50/12 = 4.2
• Cms:        153/38 = 4.2