A World Wide Web Resource Discovery System

As the WWW grows at an increasing rate, eeorts to make the technology more manageable are highly in demand. Applying advanced information retrieval techniques is one approach to such eeorts. Despite the potential beneet of these techniques in reducing users information overload and improving the eeectiveness access to on-line information, little research has been done on applying them to WWW environment. In this paper we present our attempt to apply the vector space retrieval model, relevance feedback mechanisms and a hypertext mapping technique as an integrated resource discovery system to the WWW. This paper discusses some design issues involved, as well as practical issues such as retrieval eeectiveness, usability and scalability.