Mastersthesis,
Empirical Evaluation of a Visualization Tool for Knowledge Engineering
(2003)
Abstract
The lack of methods and models for evaluation has long been a challenge in software development. Without empirical investigations into real practice, the potential of many new technologies may remain unexplored. This thesis demonstrates the feasibility of using requirement elicitation techniques, including contextual observation and introspective case studies, to guide the exploratory evaluation of an information visualization tool to support knowledge engineering, in the absence of established evaluation theory in this domain. The contributions of this thesis are: " the creation of the Shrimpbib tool, built during an introspective case study; " a descriptive taxonomy of tasks performed by users of frame-based knowledge representation systems, including knowledge modeling tasks, model verification tasks, model usage and understanding tasks, and co-ordination and collaboration tasks; " suggestions for cognitive support for some of these tasks, including visualization support; and " suggestions for future evaluations of information visualization tools for knowledge engineering.
