Historically, the library catalog has described a library collection. All those catalog cards in all those cabinets collectively described and provided intellectual access to the physical objects held in the library. With the advent of online catalogs, the format and structure of the data changed, and the catalog interface changed, but the scope of the library catalog (“my” collection) was the same. The introduction of online resources (specifically nonlicensed or open-access resources) into our service environment pushed many librarians to question the context for the “local” catalog. Some thought online resources were not appropriate in the local catalog, while others envisioned the evolution of the library catalog into an information portal that described a larger set of information objects of interest and use to the populations we serve.

But no matter whether you believe the library catalog is merely an inventory tool or an information portal, the individual library catalog is a local database that is in local control. The records added to the database (through download or direct input) are “my” records. The library does postcataloging authority control, record cleanup, maintenance, and enhancement in “my” library system on “my” records, and typically these changes are not propagated to the WorldCat record. Libraries around the country are maintaining the same bibliographic records in all our local systems, and the result of this individual silo maintenance is a huge, unnecessary duplication of effort with very little of this effort reflected in the WorldCat record. In the integrated library system, all data is local.

Authority processing is one example of silo cleaning typically performed by academic libraries. The University of Washington is not unusual in that we have two FTEs that spend their time primarily on authority processing and database cleanup. In doing the math, it becomes obvious that ARL libraries are collectively spending millions of dollars on authority processing alone. That's only 123 ARL libraries, and that's just one database function. How many more millions are spent by North American libraries in creating and maintaining local catalog data?

Changes and enhancements made to the WorldCat record are not propagated to the local record (unless a library extensively uses OCLC's bibliographic notification service to identify updated records that can be brought into the local system). And as mentioned earlier in this article, local enhancements are not reflected in WorldCat and thus are not part of the user's discovery experience. To most effectively support discovery, these enhancements should be brought into the WorldCat record.

As previously described, WorldCat Local is a discovery/delivery interface that provides appropriate services based on information both in WorldCat and in the local ILS. So what data should be held/maintained at the “network” (WorldCat) level, and what data should be held/maintained locally? The assumption of those involved in WorldCat Local development is that discovery data (primarily bibliographic data, but also other data that is used in relevancy, limits, and scopes) should be held at the network level and that “inventory” data about resources (e.g., item availability, circulation status, holdings) and customer data (e.g., authentication/authorization) should be held and maintained at the local level. There is no consensus in the library community about what data is best maintained at the network level versus the local level, but it is obvious that while having all discovery data at the local level is severely redundant and inefficient, having all service data at the network level is not practical or workable.

Because libraries have been working with local data for many years, a number of efficiencies have been developed to support local maintenance of bibliographic data. Services such as regular authority processing and non-OCLC record sources (e.g., vendor record sets) are examples. Tools have also been developed to assist in our silo “cleaning.” These include the tools within the integrated library system (e.g., global updating, list creation, headings processing) and data editing tools such as MarcEdit. One of the primary barriers in maintaining discovery data in WorldCat is the lack of these features within the OCLC cataloging environment. Related barriers include the centralization of record enhancement to a relatively small number of participants and the limited ability to perform batch processing in WorldCat. We are heartened to see that OCLC is now beginning to perform batch, postcataloging authority processing (at the time of writing, approximately 400,000 headings are being linked nightly in OCLC records). But more work will be needed in several areas (including the addition of network-level maintenance tools to support batch operations and more granular and automatic bibliographic notification services) for OCLC to provide catalogers with the same types of efficiencies they currently have in the ILS environment.

And what of the cataloging community itself? If the community reaches consensus that WorldCat (rather than the local ILS) can be the repository of metadata used to support the users' information discovery, then what is required of the community to support that commitment? Will record maintenance be a community responsibility, and what role do the Program for Cooperative Cataloging and RLG Programs play in this community? Can end users be considered part of the community, and what role can they play in enhancing records? How can the OCLC master record be developed as a shared network resource? All these questions need to be addressed before WorldCat Local can realize its full potential as a discovery and fulfillment service.