Classification and Hypermedia

Douglas Tudhope, Carl Taylor, Paul Beynon-Davies


This paper discusses a research prototype demonstrating an architecture for a hypermedia system, in which the index space has semantic relationships between terms. Information items in a social history museum application are indexed according to spatial, temporal, and subject-based classifications. The spatial and temporal indexes are inter-linked. Measures of semantic closeness over the different classification dimensions allow access to information, which might be missed by retrieval tools that require exact matching of terms used in a request. Illustrative examples of hypermedia navigation tools that make use of semantic closeness are described. A method of integrating different index dimensions drawing on research in numerical taxonomy 1s discussed. The paper goes on to discuss possibilities for a richer set of semantic primitives, that could profitably draw on concepts in the classification research literature.

Full Text: