T.J. Ostrand and E.J. Weyuker (USA)
: Software Faults, Fault Prediction, Fault-prone, Empirical Study, Software Testing.
This case study investigates characteristics of files that are among the most fault-prone in twelve releases of a large industrial inventory tracking system. The study examines the relationship between fault-proneness and the age of a file, the relationship between a file's change status and its fault-proneness, and to what extent the entire system's age affects its fault density. It is a follow-up to a study that con sidered the relationship between fault-proneness and other properties of files including size, whether files were new or old, as well as their behavior during earlier stages of de velopment and earlier releases. The ultimate goal of this research is to allow developers and testers to identify par ticular files that are likely to be fault-prone. This will allow testers to prioritize their testing effort appropriately and tells developers which characteristics to avoid in their code. It will also alert developers about which files might benefit most by being replaced before release, or require extra care when modifying.
Important Links:
Go Back
