Search results

Transaction logging has been used extensively in Okapi‐related projects to allow search algorithms and user interfaces to be investigated, evaluated and compared. Logging software may be independent of the system being examined, or integrated with it. Logging is undertaken for various purposes: recovery, playback or analysis, and the methods and formats used will vary accordingly. Logging techniques have developed over time to take into account the increasing sophistication of user interfaces and the requirement for more detailed analyses of the recorded data. A series of examples is presented, illustrating logging for both character‐based and GUI systems, and demonstrating the usefulness of a relational DBMS for storing and analysing log data. Future work will involve developing logging techniques for distributed versions of Okapi.

The need for thesauri to help users in their search for information in online information systems has been discussed for several decades. Many wide‐ranging contributions have been made to solve this problem. Nevertheless, investigation is needed to design a thesaurus structure based on what is relevant for users and generators of information within a specific subject domain. This paper explores the possibility of creating a thesaurus from the cognitive viewpoint. This approach is based on a system (in this case represented by a thesaurus) that organises its representation of knowledge or its classification as closely as possible to the authors‘ and users’ images of the subject domain with the objective of increasing the interaction between users and texts, and thus the communication in a given information retrieval system. From this point of view, the thesaurus structure is considered as the essential foundation on which to base such an interactive thesaurus. Furthermore, this structure is conceived as representing the merging point for both the generators‘ and the users’ models of the subject domain and for their information needs. This paper is dedicated mainly to the generators‘ side involved in this process. It demonstrates how an author’s writings can be used to identify the generators‘ model and perception of the subject domain, and how these can later be inserted in the thesaurus structure. Discourse analysis is used as a main method to identify the categories and its relevance for building such a structure is discussed. It also outlines a general approach for the user side to set up different methods of getting the users’ information needs into the thesaurus structure.