Analyzing Change Impact in Object-Oriented Systems

The development of software products consumes a lot of time and resources. On the other hand, these development costs are lower than maintenance costs, which represent a major concern, specially, for systems designed with recent technologies. Systems modification should be taken rigorously, and change effects must be considered. In this paper, we propose an approach, both analytical and experimental; its objective is to analyze and predict changes impacts in object-oriented (OO) systems. The method we follow consists first, to choose an existing impact model, and adapt it afterward. An impact calculation technique based on a meta-model is developed. To evaluate our approach, an empirical study was led on a real system in which a correlation hypothesis between coupling and change impact was advanced. A concrete change was done in the target system and coupling metrics were extracted from it. The hypothesis was verified with machine-learning (ML) techniques. Obtained results are interesting; they are presented and commented