Beyond Clone Detection

Code duplication, or cloning, is a common programming practice that can lead to serious maintenance issues if not properly managed. The solution to managing clones has traditionally focused on clone detection and subsequent removal. Unfortunately, detecting clones is a difficult problem where even the state of the art (e.g., CCFinder [16]) achieves precision and recall of only 72% [6]. In practice, this means that roughly a quarter of all clones are never detected at all, and of those returned, a quarter are not actual clones. Consequently, other tools are often needed to filter through the results (e.g., CLICS [19]). Moreover, the notion of clone detection implies that clones are implicit, hidden artifacts waiting to be uncovered. However, in reality almost all clones are deliberately and explicitly created by developers through some form of copy-and-pasting.