User Interfaces for Accessing Information in Digital Repositories

Abstract

Technological advances have made more information available to a larger part of the population than ever before. At the same time, the ability to locate and retrieve relevant information has become much more central. Combined, these trends form the motivation for the work presented in this thesis.

My research focus has been to study how user interfaces for accessing information in large repositories can be designed to provide assistance to users without impairing usability. To this end, existing solutions have been examined and key design challenges identified. Most important among these were supporting a wider variety of users and handling complex data models and large repositories. In order to study these challenges, a methodology of repeated design-implementation-evaluation was used. To focus the research, the implemented solutions have all been based on four fundamental design ideas as to how a simple, usable interface for large information repositories can be made.

The first presented system, SESAM, was designed for textual metadata databases. It was later extended to handle images. The second system, Savanta, targeted temporal video annotation databases. Both systems integrated various user interface techniques to create a rich and powerful environment for search, retrieval, browsing, exploration and analysis, while not sacrificing ease of use.

Several usability evaluations highlighted two main contributions. They demonstrated the power of an iterative interaction model which integrates several methods for accessing information. It was also showed that using dynamic analysis to derive high-level information about properties of a collection of information objects, is a powerful technique that makes it much easier to get an overview of and navigate in such collections.