Manual debugging is expensive. And the high cost has motivated extensive research on automated fault lo- calization in both software engineering and data mining communities. Fault localization aims at automatically locating likely fault locations, and hence assists manual debugging. A number of fault localization algorithms have been developed in recent years, which prove effec- tive when multiple failing and passing cases are avail- able. However, we notice what is more commonly en- countered in practice is the two-sample debugging prob- lem, where only one failing and one passing cases are available. This problem has been either overlooked or insufficiently tackled in previous studies.

In this paper, we develop a new fault localization al- gorithm, named BayesDebug, which simulates some manual debugging principles through a Bayesian ap- proach. Different from existing approaches that base fault analysis on multiple passing and failing cases, BayesDebug only requires one passing and one failing cases. We reason about why BayesDebug fits the two- sample debugging problem and why other approaches do not. Finally, an experiment with a real-world program grep-2.2 is conducted, which exemplifies the effective- ness of BayesDebug.