A Replication and Reproduction of Code Clone Detection Studies

Chen, X., Wang, A. Y. and Tempero, E.

Code clones, fragments of code that are similar in some way, are regarded as costly. In order to understand the level of threat and opportunity of clones, we need to be able to efficiently detect clones in existing code. Recently, a new clone detection technique, CMCD, has been proposed. Our goal is to evaluate it and, if possible, improve on the original. We replicated the original study to evaluate the effectiveness of basic CMCD technique, improved it based on our experience with the replication, and applied it to a 43 open-source Java code from the Qualitas Corpus. We confirmed the effectiveness of the original technique but found some weaknesses. We improved the technique, and applied our improved technique. We found that that 1 in 2 systems had at least 10% cloned code, not counting the original, indicating that cloned code is quite common.

Cite as: Chen, X., Wang, A. Y. and Tempero, E. (2014). A Replication and Reproduction of Code Clone Detection Studies. In Proc. Thirty-Seventh Australasian Computer Science Conference (ACSC 2014) Auckland, New Zealand. CRPIT, 147. Thomas, B. and Parry, D. Eds., ACS. 105-114

(from crpit.com) (local if available)