Mining crosscutting concerns with ComSCId: A rule-based customizable mining tool

One of the first steps when reengineering legacy systems into aspect-oriented ones is to identify the crosscutting concerns (CCC) presented in the architecture of the former; a process known as aspect mining. However, this is a time- consuming and error-prone task when conducted manually. In this paper, we present a customizable mining tool, called ComSCId, which searches for the CCC in legacy Java systems in an automatic way. ComSCId owns a repository which stores all the rules used as base for the mining process. In this repository there are pre-defined rules for some ordinary CCC like persistence, buffering and logging. Moreover, the main characteristic of this repository is its flexibility, since it allows adding new rules or customizing the existing ones to specific contexts or domains. We conducted two studies to evaluate ComSCId and we have observed high percentages of identification coverage when using this tool in an incremental way.