Context: The Mutant Analysis test selection criterion, also known as mutation testing, is an important option for the evaluation of both software systems and test sets that are used to validate them. However, mutation testing imposes high application cost. Due to this, cost reduction techniques are on focus, aiming at reducing significantly the cost of this criterion, without compromising the quality of derived test sets.Problem Definition: Results yielded by such cost reduction techniques are valid only for restricted groups of programs, as highlighted in related research. The question that persists regards the usefulness of such results to be applied to untested programs.Goal: The general aim of this proposal is the definition of a strategy for the application of cost reduction techniques for mutation testing that reuses empirical knowledge gathered from previous studies.Method: Initially, we will characterise the state-of-the-art with respect to cost reduction techniques for mutation testing. We will rank the best techniques in order to define a strategy for their application. The intended strategy will tank into account the similarity between tested programs and programs to which cost reduction techniques will be applied. Results will be assessed with respect to the quality of the newly created tests.
News published in Agência FAPESP Newsletter about the scholarship: