Finding similar defects using synonymous identifier retrieval

When we encounter a defect in one part of a program, it is very important to find other parts of the program that may contain similar defects. In this paper, we propose a novel system to find similar defects in the large collection of source code. This system takes a code fragment containing a defect as the query input, and returns code fragments containing the same or synonymous identifiers which appear in the input fragment. Case studies with two open source systems and their defect data show the advantages of the proposed retrieval system, compared to the code-clone based retrievals.