A while back I visited the 2016 International Conference on Testing Software and Systems. The conference was held at the Graz University of Technology in Austria. The university is just over 200 years old and right at the city center, very different from the TUT campus.
As is often the case, the keynote presentations had some of the most interesting ideas. Jeff Offutt proposed that test cases should not be inert clumps of code and data as they usually are, but smart actors who know why they exist, when they should be executed and so on. Nowadays this kind of information is in separate test management software, or only in testers’ heads. Gordon Fraser talked about gamification of testing and its teaching. In his Code Defenders game, some players add bugs into code and others try to write tests that detect them. At least early experiences were good: students enjoyed the game and managed to create high-quality test cases. Perhaps something to try out in the next year’s testing course?
Surprisingly many of the papers were related to mutation testing, which I had previously considered a somewhat marginal approach. Probably something I should take more seriously in the future.
My own paper outlined a method for producing test models automatically based on the GUI of the system under test and some pre-existing test cases, a hybrid of two earlier approaches. The outline was all I had when I submitted the paper, but by the time the conference was held, I had a prototype tool ready and was able to present some early results. Unfortunately, my presentation was at the very final session of the conference, so there wasn’t much chance to discuss the topic with other people afterward. The tool is available at http://www.cs.tut.fi/~jaaskel9/murphy.html if anyone’s interested.
Text by Antti Jääskeläinen