Measuring the impact of different categories of software evolution

Abstract. Software evolution involves different categories of interven-tions, having variable impact on the code. Knowledge about the expected impact of an intervention is fundamental for project planning and re-source allocation. Moreover, deviations from the expected impact may hint for areas of the system having a poor design. In this paper, we inves-tigate the relationship between evolution categories and impacted code by means of a set of metrics computed over time for a subject system