Modeling Clone Evolution

Abstract

During the maintenance of a program, not only the source code but also the code clones contained in it evolve. Some recent studies focused on detecting patterns in the history of code clones to evaluate the harmfulness of clones. Since clone evolution is a rather novel field of study, there is still need for more comprehensive models and improved methodologies. Current approaches are limited to detecting only a small specialized set of patterns in a clone’s evolution, they generally lack scalability and the way they collect data has to be discussed. We provide an overview of existing methods to model clone evolution by describing the patterns that can be observed and the respective detection procedure. We discuss their shortcomings and point out open questions towards more detailed models of clone evolution. 1