Object-oriented cataloging

Abstract

Catalogs have developed from lists of physical items present in particular libraries into computerized access and retrieval tools for works dispersed across local and national boundaries. Works themselves are no longer necessarily constrained in one physical form. Cataloging rules, however, have not evolved in parallel with these developments. This paper reanalyzes the nature of works and their publication in an approach based on object-oriented modeling, and demonstrates the advantages to be gained thereby. It suggests a strategic plan to enable an organic transformation to be made from current MARC-based cataloging to object-oriented cataloging. Finding Books in Libraries Catalogs began as listings of particular books on specific shelves. My own institution, the Bodleian Library, compiled the first such published catalog in the English-speaking world nearly four centuries ago (Bodleian Library 1986). Over the centuries since then--but slowly--ideas such as relative location, consistent description, and catalog headings (i.e., access points) have been developed. This process came to full flower with the publication of the first edition of the Anglo-American Cataloguing Rules in 1967. In the following year, MARC was initiated. Here were a complete set of cataloging rules and a comprehensive way of coding catalog entries for computer use, arriving more or less one right after the other. The revision of the rules for the second edition in 1978 and its official adoption in 1980 by the two major sources of bibliographic records in the United Kingdom and the United States spread that florescence yet further. As a result, AACR2 has taken firm root all over the Anglophone world and beyond. The revision of the second edition of AACR in 1988 was not felt to be radical enough to warrant its identification as a full new edition. It really does seem that with the 1988 revision of AACR2 we have reached a plateau of maturity, with comprehensive rules based on firm theoretical foundations. In reality, of course, we have reached no such plateau. The problems arise from the fact that library catalogs do not function solely as descriptive lists of books: they function also as elements in library management systems and, in the electronic age, as sophisticated information retrieval tools. The catalog as a list of books is centered on the physical objects contained in a library. The structure of AACR2 makes it very clear that description is the first priority and that, once the description of an object is complete, one's attention may be turned to attaching "handles" to it so that readers may gain access to the material. The handles--the access points---arise from the description. The description itself is a stage on the road leading to the full text. In some environments in which AACR2 is used, different aspects of it may be considered important. In most booksellers, catalogs, there is only one access point for each work, and the bibliographic description is the main element; after all, attention is being focused on the physical item that is being offered for sale. In a company research library, classified access may be more important than access by personal author. How valid is AACR2 for all of these situations? Is the code really equally suitable for all environments? Access based on bibliographic description is historically grounded in the development of library catalogs. Until this century each library was self-contained and each catalog truly reflected the library's contents. If a requirement could not be met within the library, that was the end of the matter. The user had to be satisfied with the library's contents. But how many users enter a library looking for particular physical items. In the case of subject searches, the answer is obviously "None." Even in the case of what are generally called "known item" searches, most users are looking for any item which satisfies the request "an edition of Hamlet," rather than "the Macmillan edition of Hamlet. …