The application of established empirical methods (e.g., controlled experiments, replicated case studies, surveys) in software engineering is still limited. Because many new software engineering methods and tools are not formal enough for validation with sound mathematical proofs, systematic experimentation is necessary to validate their promised benefits. Although the use of empirical methods in software engineering has risen drastically during the last 10 years, many research proposals are still only advocated by their authors instead of validated systematically using experimentation. Existing experiments often do not exhibit sound research practises and lack in statistical analyses.
Introductory papers to empirical software engineering:
- W. Tichy: Should computer scientists experiment more? (1995)
- B. Kitchenham et al.: Preliminary guidelines for empirical research in software engineering (2002)
- D. Sjoberg et al.: A survey of controlled experiments in software engineering (2005)
- D. Sjoberg et al.: The future of empirical methods in software engineering research (2007)
My research applies empirical research methods on performance prediction and software architecture evaluation, rather than improving empirical methods themselves. For my diploma-thesis (MSc.), I conducted a small experiment with 30 students. For my PhD-thesis, I was involved in another student-experiment together with Anne and Steffen.
Selected contributions with my participation:
- The Role of Experimentation in Software Engineering
- From monolithic to component-based performance evaluation of software architectures: A series of experiments analysing accuracy and effort
(Journal on Empirical Software Engineering, 2011, Springer)
- Performance and Reliability Prediction for Evolving Service-Oriented Software Systems
(Journal on Empirical Software Engineering 2013, Springer)