Homology Detection Based on Abstract Syntax Tree Combined Simple Semantics Analysis

Abstract

Nowadays, the research of software homology detection is more and more important in the flourishing software market. Most of the existing detection tools are based on text structure but ignore the syntax framework structure of program codes. A few tools based on syntax structure do not consider detection on semantics level, which can be avoided by some plagiarism. The comparison system put forward in this paper combines comparison methods based on abstract syntax tree and simple semantics, such that being able to detect the homology software more accurately and comprehensively. It not only analyses the syntax structure of program codes, but also researches some simple semantics changes to achieve detection on both syntax and semantics level. Based on these, a homology detection system Code Compare is developed. Compared with other detection tools, Code Compare is capable of recognizing more code plagiarism types, thus can detect homology software more effectively.