The Software Deliverable

The software completed by the conclusion of the Cilks project comprises: We thus have a completely automated system for logging and analysing user behaviour during thesaurus navigation and query enhancement, and for evaluating different ways of adding thesaurus terms to an original query. The principle of using a relational database to log and analyse all interactions during the retrieval process has now been adopted for all current and future Okapi projects.

At present the Oracle (thesaurus) database and the Okapi (document) database are situated on two different machines - the software written to communicate between them uses remote procedure calls. This development provided useful experience in distributed processing, pending the adoption of a more general client-server architecture in future versions of Okapi.