Detail research undertaken to determine the key differences in the performance of certain centralised (physical) and distributed (virtual) bibliographic catalogue services, and to suggest strategies for improving interoperability and performance in, and between, physical and virtual models.
Methodically defined searches of a centralised catalogue service and selected distributed catalogues were conducted using the Z39.50 information retrieval protocol, allowing search types to be semantically defined. The methodology also entailed the use of two workshops comprising systems librarians and cataloguers to inform suggested strategies for improving performance and interoperability within both environments.
Technical interoperability was permitted easily between centralised and distributed models, however, the various individual configurations permitted only limited semantic interoperability. Significant prescription in cataloguing and indexing guidelines, greater participation in the program for collaborative cataloguing, consideration of future functional requirements for bibliographic records migration, and greater disclosure to end users are some of the suggested strategies to improve performance and semantic interoperability.
This paper not only informs the library and information science research community and union catalogue administrators, but also has numerous practical implications for those establishing distributed systems based on Z39.50 and search/retrieve web services as well as those establishing centralised systems.
The paper moves the discussion of Z39.50‐based systems away from anecdotal evidence and provides recommendations based on testing, and is intimately informed by the UK cataloguing and systems librarian community.
