| |
|
Current Research Goal (subject to randomized variations dictated by a mind's search probability distribution): Why aren't Analysis, Design and Modeling Tools more used? How can Modeling be more effective and how can modeling tools achieve a higher market acceptance? We are interested in studying the usability problems faced by interface development tools and by software analysis, design and modeling tools. We argue that in order to achieve a stronger market acceptance of modeling tools, a new generation of user-centric tools will have to emerge, tools capable of clearly supporting user interface specific activities. TaskSketch TaskSketch is a UML-Based tool that focuses on linking and tracing use cases to the conceptual architecture of a system [Nunes, 2000]. It uses a shared display, a speech recognition system and a thesaurus to help leverage the requirements elicitation process as well as better clustering of tasks. CanonSketch CanonSketch is a tool for the creation, design and editing of Canonical Abstract Prototypes [Constantine, 2003]. It combines the modeling capabilities of modern UML editors with an abstract prototype view and a concrete prototype view (in HTML). A collaborative version allows team development and discussion. Past Projects: Abalearn Abalearn is a program that learns to play Abalone using a new risk-sensitive version of the Temporal Difference Learning algorithm. The main contribution stems from the observation that risk-sensitivity may help speed up learning. |