Generic modelling of code clones

Code clones, i.e. instances of duplicated code, can be found in many software systems. They adversely aect the software systems’ quality, in particular their maintainability and comprehensibility. Thus, this aspect is particularly important to consider in software maintenance and reengineering. Many dierent algorithms detecting code clones have been developed. For various reasons, it is dicult to compare the results of dierent algorithms. Most notable among these reasons is that there is no conceptual model allowing description of code clones determined by dierent algorithms. Much more, each algorithm uses its specific concept of code clones, which is rarely made explicit. To overcome these problems, we have developed a generic model for describing clones. The model is generic in that it is independent of the programming language examined and of the clone detection algorithm used. It is flexible enough to facilitate various granularities of artifacts employed for selection and comparison, including inexact clones. The model allows separation of concerns between clone detection, description and management, which reduces the eort for the implementation of tools supporting these activities. On the basis of the model, we have implemented a prototype tool supporting these activities, tightly integrated into the Eclipse environment.