Statistical debugging using compound boolean predicates
Authors:
Open Access
- 9 July 2007
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 5-15
Abstract
Statistical debugging uses dynamic instrumentation and machine learning to identify predicates on program state that are strongly predictive of program failure. Prior approaches have only considered simple, atomic predicates such as the directions of branches or the return values of function calls. We enrich the predicate vocabulary by adding complex Boolean formulae derived from these simple predicates. We draw upon three-valued logic, static program structure, and statistical estimation techniques to efficiently sift through large numbers of candidate Boolean predicate formulae. We present qualitative and quantitative evidence that complex predicates are practical, precise, and informative. Furthermore, we demonstrate that our approach is robust in the face of incomplete data provided by the sparse random sampling that typifies postdeployment statistical debugging.Keywords
This publication has 14 references indexed in Scilit:
- Statistical debugging2006
- Path Optimization in Programs and Its Application to Debugging2006
- Supporting Controlled Experimentation with Testing Techniques: An Infrastructure and its Potential ImpactEmpirical Software Engineering, 2005
- SOBER2005
- Scalable statistical bug isolation2005
- Locating causes of program failures2005
- Gammatella: Visualizing Program-Execution Data for Deployed SoftwareInformation Visualization, 2004
- Bug isolation via remote program sampling2003
- Tracking down software bugs using automatic anomaly detection2002
- Data clusteringACM Computing Surveys, 1999