Process Improvement Experiments in Perspective


In the context of European Communities funded programs, several software companies and research institutions have recently started collaborations aimed at conducting empirical studies. The industrial objective is ususally the evaluation of a process improvement in a ``scientific'' way. On the other side, academy is interested in building knowledge about software engineering practice from such studies. In this paper empirical studies are considered in perspective, starting from the experience made during the process improvement experiment ITALO.

The objective of ITALO was evaluating the benefits coming from test management automation. In a first phase commercial tools were selected by following a methodology based on {\em Feature Analysis}. Then, a controlled experiment was designed to measure the effects of tool adoption in terms of changed quality and effort. Results confirmed the experimental hypothesis that an increased defect removal can be achieved by tool adoption. The associated increased effort is expected to disappear in the next iterations of tool usage.

Postscript version of the paper.