ECAI-2000 Logo

ECAI-2000 Conference Paper

[PDF] [full paper] [prev] [tofc] [next]

Modeling Java Programs for Diagnosis

Cristinel Mateis, Markus Stumptner, Franz Wotawa

A key advantage of model-based diagnosis is the ability to use a generic model for the production of system descriptions that can be used to derive diagnoses for differently structured individual systems from a domain. This advantage is nowhere more apparent than in the software error diagnosis (or debugging) area, where given a model, system descriptions can be automatically derived from source code. However, effective models for diagnosing programs have so far been limited to special-purpose languages. We describe a value-based model for Java programs that enables us to explicitly deal with imperative program execution (including loop execution), and compare the outcome of our approach to the results obtained by using program slicing, a traditional technique from the software debugging community, and a simple dependency-based model for Java.

Keywords: Diagnosis, Model-based Reasoning

Citation: Cristinel Mateis, Markus Stumptner, Franz Wotawa: Modeling Java Programs for Diagnosis. In W.Horn (ed.): ECAI2000, Proceedings of the 14th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2000, pp.171-175.

[prev] [tofc] [next]

ECAI-2000 is organised by the European Coordinating Committee for Artificial Intelligence (ECCAI) and hosted by the Humboldt University on behalf of Gesellschaft für Informatik.