Analyzing Clone Evolution for Identifying the Important Clones for Management

Abstract

Code clones (identical or similar code fragments in a code-base) have dual but contradictory impacts (i.e., both positive and negative impacts) on the evolution and maintenance of a software system. Because of the negative impacts (such as high change-proneness, bug-proneness, and unintentional inconsistencies), software researchers consider code clones to be the number one bad-smell in a code-base. Existing studies on clone management suggest managing code clones through refactoring and tracking. However, a software system's code-base may contain a huge number of code clones, and it is impractical to consider all these clones for refactoring or tracking. In these circumstances, it is essential to identify code clones that can be considered particularly important for refactoring and tracking. However, no existing study has investigated this matter. We conduct our research emphasizing this matter, and perform five studies on identifying important clones by analyzing clone evolution history.

In our first study we detect evolutionary coupling of code clones by automatically investigating clone evolution history from thousands of commits of software systems downloaded from on-line SVN repositories. By analyzing evolutionary coupling of code clones we identify a particular clone change pattern, Similarity Preserving Change Pattern (SPCP), such that code clones that evolve following this pattern should be considered important for refactoring. We call these important clones the SPCP clones. We rank SPCP clones considering their strength of evolutionary coupling. In our second study we further analyze evolutionary coupling of code clones with an aim to assist clone tracking. The purpose of clone tracking is to identify the co-change (i.e. changing together) candidates of code clones to ensure consistency of changes in the code-base. Our research in the second study identifies and ranks the important co-change candidates by analyzing their evolutionary coupling. In our third study we perform a deeper analysis on the SPCP clones and identify their cross-boundary evolutionary couplings. On the basis of such couplings we separate the SPCP clones into two disjoint subsets. While one subset contains the non-cross-boundary SPCP clones which can be considered important for refactoring, the other subset contains the cross-boundary SPCP clones which should be considered important for tracking. In our fourth study we analyze the bug-proneness of different types of SPCP clones in order to identify which type(s) of code clones have high tendencies of experiencing bug-fixes. Such clone-types can be given high priorities for management (refactoring or tracking). In our last study we analyze and compare the late propagation tendencies of different types of code clones. Late propagation is commonly regarded as a harmful clone evolution pattern. Findings from our last study can help us prioritize clone-types for management on the basis of their tendencies of experiencing late propagations. We also find that late propagation can be considerably minimized by managing the SPCP clones. On the basis of our studies we develop an automatic system called AMIC (Automatic Mining of Important Clones) that identifies the important clones for management (refactoring and tracking) and ranks these clones considering their evolutionary coupling, bug-proneness, and late propagation tendencies. We believe that our research findings have the potential to assist clone management by pin-pointing the important clones to be managed, and thus, considerably minimizing clone management effort.