While visualization is a very important feature of information retrieval (IR) systems, existing visualizing schemes have not been found to be effective, especially for domain novices. The authors propose visualization solutions for three types of classification schemes: collocation classification schemes, selection classification schemes, and navigation classification schemes. These schemes are created for the three ways in which users interact with an IR system. In this paper, the authors describe the navigation schemes.

Based on the theoretical belief that a users interaction with an IR system consists of a series of discrete moments, stimulated by shifting frames of knowledge, the authors navigation classification schemes use the metaphor of sailing an inland sea, to help novice users interact with the IR system, and to identify their evolving information needs. Using these schemes, users cognitive knowledge is translated by the navigator into conceptual terms, and mapped onto the systems classification schemes by representing such terms with lighthouses. The lighthouses also bear broader, related, and narrower terms in the domain. By allowing users to navigate via the lighthouses, and by managing the series of discrete interactions between the user and the system, the navigation classification schemes are able to help users identify their information needs in a broader domain environment.

The papers primary contribution is the proposal that classification schemes consider a users information search process as being ecologically situated. During the search process, the users frame of reference shifts, and the users needs change. This paper also provides an approach for visualizing such classification schemes, using the navigation metaphor, which is already familiar to many novice users, and is relatively easy to manipulate.