Validity Conditions Impact of a Software Measure

Despite real progress in software engineering, the implementation of a successful measurement program in a software organization is still a serious challenge. Most of the problems are mainly due to the maturity level of the organization and to the credit given by the software engineering community for the software measure. Software measurement, like other disciplines, must be based on the measurement theory; as such a scientific basis allows gaining a widespread acceptance among the community. Validation of a software measure is an important task, but not an easy one. This task aims to prove that measures are actually measuring what they claim to do. One of the main problems is the relative validity of a software measure. It is often observed that a software measure is strongly dependent on environmental factors. The goal of this paper is to highlight experimentally the impact of these factors, called validity conditions, on the validity measure.