Test suite reduction in good order: comparing heuristics from a new viewpoint

A new perspective in assessing test suite reduction techniques based on their rate of fault detection is introduced in this paper. This criterion, which is standard in assessing test-suite prioritization, has never been used for reduction. Our proposal stems from the consideration that under pressure testing could be stopped before all tests in the reduced test-suite are run, and in such cases the ordering in the reduced test-suite is also important. We compare four well-known reduction heuristics showing that by considering the rate of fault detection, the reduction technique to be chosen when time is an issue might be different from the one performing the best when testing can be completed