Finding a subset of test cases that satisfies the same testing objective as the original test suite is referred to as the problem of test suite reduction. Such a subset of test cases is referred 10 as a representative set. Recent studies show that the representative set and the original test suite have similar fault detecting capabilities. This paper reports an empirical study on whether representative sets found by different heuristics will still have similar fault detecting capabilities. As a result, software testers may he able to choose an heuristic which delivers smaller sized representative sets without rendering the fault detecting capability.
Proceedings of the 1st ACIS International Conference on Software Engineering Research and Applications (SERA'03), San Francisco, California, United States, 25-27 June 2003 / Roger Y. Lee and Kyung-Whan Lee (eds.),