How significant is the effect of fault interactions on coverage-based fault localizations?

The effectiveness of coverage-based fault localizations in the presence of multiple faults has been a major concern for the software testing research community. A commonly held belief is that the fault localization techniques based on coverage statistics are less effective in the presence of multiple faults and their performance deteriorates. The fault interference phenomenon refers to cases where the software under test contains multiple faults whose interactions hinder effective debugging. The immediate research question that arises is to what extent fault interactions are influential. This paper focuses on verifying the existence of fault interference phenomenon in programs developed in programming languages with object-oriented features. The paper then statistically measures the influence and significance of fault interactions on the performance of debugging based on coverage-based fault localizations. The result verifies that the fault interleaving phenomenon occurs. However, its impact on the performance of fault localizations is negligible.