Experimenting with Category Partition´s 1-Way and 2-Way Test Selection Criteria

The Category Partition (CP) black-box testing method has shown to be effective in a number of situations. There is however little support for automating its use and little is known about the cost effectiveness of its associated selection criteria. In this paper, we report on a tool to automatically create test frames, i.e., test case specifications, for three wellknown criteria associated with the CP method, including the application of 1-way and 2-way interactions. We then report on the cost, in terms of number of test cases, and the effectiveness, in terms of mutation score, of adequate test suites for these criteria. The main lesson learnt is that, in addition to the intuition that the CP specification does impact effectiveness, a significant part of the effectiveness is also due to the test input selection procedure for test frames.