Final talk of the first day of #scam2012 by Saman Bazrafshan, who will talk about near-miss clones. He studied the evolution of near-miss clones to investigate how they change. This is useful, since irregularities in clone evolution can help to distinguish between potentially harmful clones and the less interesting ones.
Saman builds upon the work of Nils Gode, who did a similar study on exact clones and presented that on SCAM 2009. His analysis shows that near-miss clones are more common and more persistent than identical clones.
Even more interesting is the study of the relation between the two types of clones. They detected several cases of near-miss clones to identical ones and they other way around.
In many cases unfortunately, it wasn’t that interesting, since there were many style fixes (such as adding curlies around a statement) They located a large number of inconsistent changes, but both to identical and near-miss clones, there was nothing to conclude about that.
The presentation spurred many questions, for instance: how many near-miss to near-miss or identical to identical changes were made, did you see identical clones being removed and could you monitor near-miss clones from their incarnation (when they are most similar) to track their behavior. There seems to be a lot of research to be done here.
The paper is not available online yet.