THE DESIGN OF A HYPERTEXT INTERFACE FOR INFORMATION-RETRIEVAL

Hypertext for multimedia and fulltext documents is being rapidly developed on all fronts. For bibliographic-information retrieval the applications are promising but pose several difficulties. The first is that most bibliographic databases consist of millions of distinct records as opposed to hundreds or thousands of documents in most hypertext systems. One solution used in some hypertext systems is to restrict the domain by immediately performing a Boolean search. While this is effective, it just takes us back to the standard retrieval system. The objective of the current work is to add the capability of a browsing search to an existing retrieval system. The WHIZRD (Windowed Hypertext Interface for Zippy Retrieval and Display) system currently being developed uses nodes (documents or sets of documents) and links to other nodes. The links include authors, subject classification, titles, journals, series, thesaurus, and abstracts. The user can then browse within each node or choose a link to follow to another node to gradually restrict or expand the search in many different directions.