Informatisches Kolloquium Wintersemester 2007/2008

Montag, 30. Januar 2009

Prof. Dr. Lutz Prechelt
Freie Universität Berlin

Methodology and Study Design in Empirical Software Engineering: Two Case Studies

Physics is usually considered the role model for empirical science. This model suggests controlled experimentation as the one methodology of choice. However in software engineering, controlled experiments are frequently NOT the best tool for producing insights.

This talk presents two examples where this is the case:

• Studying pair programming, where controlled experiments have brought forth some results but remarkably little insight and
• Comparing web development platform technologies where full control is not just infeasible but even nonsensical.

I will present the considerations that lead to the choice of alternative methodology, the resulting study designs, and a glimpse of the results obtained.

Kontakt

Prof. Dr. Christiane Floyd
Telefon 2415