SemaLink: an approach for semantic browsing through large distributed document spaces

Hypermedia retrieval in combination with querying is a very powerful document access metaphor for digital library systems. Conventional distributed hypermedia systems, however, suffer from some well-known problems that make effective, goal-directed document retrieval and maintenance impossible. On the one hand, relationships between documents are modeled on a very low level of abstraction, preventing knowledge re-use and user-adapted navigation through the document space. On the other hand, huge static web structures are prone to redundancies, inconsistencies and costly to maintain. We present an approach of a scalable distributed resource discovery and delivery system that offers access to information through querying and knowledge-based hypermedia browsing. Relationship knowledge is partitioned into handy, independent structures, modeled on a high level of abstraction apart from documents. Navigation paths are computed by the system, combining semantic networks with information retrieval