Is it possible to do controlled process experiments in software?
That was the question asked in the Kanbandev group over at Yahoo groups. We need evidence to support the practices we use in software development. After all we are talking about a serious business with an annual value of about 500 TUSD.