Bridging the Gap Between Slicing and Model-based Diagnosis.
Franz Wotawa · Software Engineering and Knowledge Engineering · 2008
Fault localization is considered an important and difficult task in the software engineering process. In the last decades several approaches to fault localization have been published. Some of them are based on either static or dynamic program slicing. In this paper, we present an approach that combines program slicing with the computation of hitting sets. Hitting sets are used in model-based diagnosis to compute diagnoses from conflicting assumptions. We introduce the underlying definitions and algorithms of the approach, and show that the combination of slicing and hitting set computation reduces the number of statement to be considered. The presented approach does not rely on a specific slicing methodology and can be used in combination with static or dynamic slicing.