Change, similarity and redundancy in software systems

点击翻译按钮获取中文摘要

Studies show that up to 80% of software development cost is spent on software maintenance. One aspect of software maintenance is dealing with duplicated code. Copy-and-paste of code sections with or without minor changes is a common software development practice, and over time large software systems develop many such "software clones". While such cloning is often intentional and can be useful in many ways, it can cause significant problems in the maintenance of large software systems because it increases the risk of update anomalies, increases program size and complexity, and contributes to high maintenance costs. Clone detection is thus an important problem in software maintenance research, and serves as the basis of a wide range of software engineering tasks. In the last several years, we have significantly advanced the state-of-the-art in the detection and analysis of near-miss software clones, those where minor additions, deletions and modifications have been made in the copied fragments.