Prof. Dr. Andreas Zeller

"The Impact of Synthetic Software Changes"

(Institutkolloguium am Max Planck Institut für Software-Systeme)

When we talk about software evolution, we usually think of humans constantly changing and adapting the software towards a greater goal. In this talk, I present recent work in which we _automatically_ change software: * We seed _artificial defects_ into software to assess the quality of test suites -- if the test suite does not find the defect, it needs to be improved.
* We mine specification differences to synthesize _artificial fixes_ for common programming errors, such that the program effectively fixes itself. In both these applications, it is important to assess the _impact_ of these synthetic changes on the execution: For defects, one wants to maximize impact; for fixes, one wants to minimize impact. Experiences with impact measures on coverage and dynamic invariants on large-scale programs like AspectJ demonstrate the feasibility of our approaches.

Bio: Andreas Zeller is professor for software engineering at Saarland University, Saarbrücken, Germany. His research is concerned with the analysis of large software systems, in particular their execution ("Why does my program fail?") and their development history ("Where do most bugs occur?"). In 2006, his book "Why Programs Fail" received the Software Development Magazine productivity award. In 2009, his work on delta debugging got the ACM SIGSOFT Impact Paper Award as the most influential software engineering paper of 1999.