Indexing in hypertext environments: the role of user models
A growing interest in the application of hypertext (hypermedia) to the task of indexing reflects the increasing availability of online electronic material (text and text/graphics/video; CD), and a recognition that considerable benefits might arise from the application of hypermedia techniques (both in academic and commercial contexts). This paper considers the relationship between indexing and hypermedia on two fronts: Conceptualizing adequate models of index structures and indexing procedures; and establishing what readers do when accessing indexes. There is a range of explicit and implicit models (or abstractions) underlying databases, documents and indexes. A fundamental starting point for the development of substructure indexing tools for use with hypermedia is the precise specification of such models. Analysis of substructure indexing models will allow the identification of largely technical constraints (fundamentally linked to the system-derived parameters of particular computing systems), and conceptual constraints (intrinsic to the ’semantics’ of indexes and indexing procedures). Related also are aspects of the problem deriving from how indexes (the electronic form) can, and should be, presented. The first part of this paper considers a framework for the identification of such models. The second aspect of the research reported here concerns index use. A field study was conducted looking at the use accountants (N = 10) make of indexing facilities in a large financial services document. Task analysis highlighted the range of problems encountered, strategies employed and the kinds of solutions arrived at. The results indicate that there are six key factors underlying the ‘rule of thumb’ procedure users employ when interrogating an index. The implication for this study for the design of indexing tools within hypermedia environment are discussed.