Towards a software diagnosis method based on rough set reasoning

Software diagnosis for finding faults based on the test results is one of the most time-consuming and labor-intensive activities in large scale software development. Revealing the potential knowledge hidden in the test results or program constructs to assist this activity is a rational solution. In this paper, we propose two kinds of debugging applications based on rough set reasoning. One is to select key input parameters which will affect programpsilas behaviors to facilitate diagnosis. The other is to extract association rules between program input and its behaviors. The inputs of the above two rough reasoning applications are all the test results of functional testing. Our work is the first attempt to utilize functional testing information to help software debugging. The feasibility and effectiveness of our approach is validated by some examples and experiments. In addition, some on-going research issues are also addressed.