USI - Email
Automating Non-Formalized Testing Techniques via Quantification and Optimisation: The Case of Boundary Value Testing
Host: Prof. Carlo A. Furia
USI East Campus, Room D1.15
Chalmers University of Technology / Blekinge Institute of Technology
Many testing techniques, even when required or recommended by standards, have been only informally defined and thus are typically still done manually. A notable example is boundary value analysis/testing which is simple to understand conceptually and important in industry but have little to no support for automation. In this talk we summarise some of our work on quantifying diversity in software testing, how the distance functions involved can be used to define program derivatives, and such derivatives can then help automate boundary value testing. We present also empirical evaluation of the approach and how the automated boundary detection it enables can find problems either in the specification, implemented software, or help extend the test suite. Finally, we outline future work both to extend and generalise the approach.
Robert Feldt is professor of Software Engineering at Chalmers University of Technology and Blekinge Institute of Technology, both in Sweden. He is passionate about a range of topics from human factors and automation to statistics, causal analysis, and applied machine learning (ML). His primary focus is on software testing and quality, along with human-centered software engineering. Lately, he's also been working on using ML in medicine/healthcare and materials science.
He often works alongside companies in Sweden, Europe, and Asia, while also leading more basic research. In 2002, he earned a PhD in Computer Engineering from Chalmers University of Technology. He also studied Psychology at Gothenburg University in the 1990s. Beyond academia, he brings over 30 years of experience as an IT, software, and ML consultant. Dr. Feldt contributes as the co-Editor in Chief of the Empirical Software Engineering journal and serves on the editorial boards of two other journals, STVR and SQJ.
INF Newsletter Archive