Enhancing program dependency graph based clone detection using approximate subgraph matching

Software code clone detection techniques and tools play a major role in improving the software quality as well as saving maintenance cost and effort. Program Dependency Graph (PDG) based clone detection techniques have a key advantage over other techniques as they are capable of detecting non-contiguous code clones in addition to contiguous clones. We propose further enhancement to current state of the art PDG-based detection to identify all possible (exact and approximate) clone relations from the obtained clone pair (PDG-based) results using Approximate Subgraph Matching (ASM). We obtain clone results of our proposed technique on three subject software systems, and validate the results on eclipse-ant from Bellon's benchmark. We also present a new ASM-based distance measure to represent the similarity between software code clones.