Machine assisted browsing for the naive user

The purpose of this paper is to demonstrate how a radically different approach to the storage and retrieval of information can result in: (1) a reduction in the need for user sophistication in the use of information systems, and (2) the support of a browsing approach to information system searching. Our approach promotes the view that information system databases should be structured for people, not machines. Many of the problems associated with information systems occur precisely at the interface between the user and information storage. The purpose of the interface is to map user requests onto the database structure. The more "machine-like" the interface, the faster the mapping. As interfaces become more sophisticated (i.e., allow the user to express requests in a more natural form such as natural language), more processing is required to carry out the mapping. And as processing time increases, certain types of search processes (e.g., browsing) become increasingly difficult to provide at a reasonable response rate. Until the time when processing power can meet the real-time needs of system users, information system design and construction should conform to user needs, meaning that the physical structure should be tailored to the user's view (logical structure) of the database, reducing mapping and search time. In the following discussion, we will describe the BROWSE system, a database browsing system for computer- naive users. The primary application of the BROWSE system is to allow browsing access to the Carnegie- Mellon University Computer Science Department library collection, but it can be used for other applications, such as the automated dictionary.This research was sponsored by the Office of Naval Research under contract No. N0001 4-76-0874, and, in part, by the Defense Advanced Research Projects Agency (DOD), Arpa Order No. 3597, monitored by the Air Force Avionics Laboratory Contract 533615-78-C-1551.published or submitted for publicatio