|
Abstract:
|
This research addresses the problem of locating software items in extensive libraries. It aims to increase the speed and accuracy with which a user may browse software libraries for reusable code. The method proposed for this is called active browsing. The system monitors user actions, made within a normal browser, to infer an analogue representing the user's search goal. A relevancy measure is constructed from this analogue and used by the system to scan the library independently of the user and to evaluate potentially interesting components. The results affect the browser display to emphasize relevant components and thus aid search. Although the main interest is software reuse, the approach has a much wider applicability. This is inherited directly from the broad applicability of browsing. Browsing is an important methodology particularly in search tasks where the target is not well defined. This thesis discusses a model of active browsing applicable in general to browseable libraries. An implementation of this model, based around a browser used to explore libraries of object oriented code, is described. In this implementation, an inference engine forward chains on rules to generate the analogue from the user's actions. A relevancy measure, in the form of a template, is constructed, used to assign a score to library items and to produce a ranked list of classes of potential interest to the user. This implementation is used both to illustrate active browsing in a particular application and to experimentally validate the model. |
