Defining and applying metrics in the context of continuing software evolution

This paper addresses a set of relevant issues in the context of the definit ion and application of measurement to long-term software evolution processe s and their products. It presents a practical example using empirical data from one of the systems studied as part of the on-going FEAST (Feedback, Ev olution And Software Technology) investigation. Such example demonstrates t he use of a sequential statistical test (CUSUM) on a suite of eight evoluti on activity metrics. The metrics are based on module and sub-system counts. The test permitted examination of whether productivity had changed over an 11-year period. Results were consistent in six of the eight metrics studie d, suggesting that sequential tests have potential in, for example, measuri ng the impact of process improvement (intervention analysis), identificatio n of stages of software evolution and, in general, in supporting evolution management.