A Study on PDG-based Source Code Similarity Checking Tools

Abstract

Modern software development uses OSS to reduce development time and costs. However, non-compliance with OSS licenses can lead to various problems, including lawsuits. These issues can hinder the growth of the software industry, so it's important to check OSS licenses. This can be done through source code similarity checking to detect the OSS used. However, simple source code modifications can circumvent text-based source code similarity checks. To solve this problem, this paper proposes a PDG-based source code similarity-checking tool that compares similarities in the structural form of programs. Text-based source code similarity-checking tools are measured to have an average similarity of 96.5% when checking the same source code. In comparison, the proposed method can achieve a more accurate similarity check with an average of 100%. In addition, it is more accurate than text-based source code similarity checks for modified source code. This allows you to identify OSS in your source code accurately.