Searching a database based web site

Abstract

Currently, information systems are usually supported by databases (DB) and accessed through a Web interface. Pages in such Web sites are not drawn from HTML files but are generated on the fly upon request. Indexing and searching such dynamic pages raises several extra difficulties not solved by most search engines, which were designed for static contents. In this paper we describe the development of a search engine that overcomes most of the problems for a specific Web site, how the limitations put to indexing dynamic Web pages were circumvented, and an evaluation of the results obtained. The solution involves using a locally developed crawler, the Oracle Text full text indexer, and meta-information automatically drawn from the DB or manually added to improve the relevance factor calculation. It has the advantage of uniformly covering the dynamic pages and the static Web pages of the site.