Abstract
We comment on the experimental design and the result of the paper mentioned in the title. Our purpose is to show interested readers examples of what can go wrong with experiments in software research and how to avoid the attending problems.

This publication has 4 references indexed in Scilit: