Database querying on the World Wide Web: the design and implementation of domain-specific database search engines

The World Wide Web can be considered to be a very large semi-structured database that holds vast amounts of useful information. However, our ability to query, search, and reuse the information on the Web is limited at present. Existing search techniques suffer from critical deficiencies with respect to robustness, flexibility, and precision. This research attempts to develop domain-centred alternatives to keyword and subject directory search engines. By focusing on well-demarcated logical domains whose ontologies can be modelled using third-generation, object-relational data models, we demonstrate how relevant Web data can be restructured using object-relational databases, against which SQL-type queries can be issued. A prototype implementation for the 'universities' domain entitled 'UniGuide' is presented in this dissertation.