Analysis on Identification Technologies of Program Code Similarity

摘要

The technologies for identifying program code similarity are significant in college practice courses and software design. This paper introduces the definition of code similarity, summarizes eight means of plagiarism of program code, and presents the Pyramid of Code Plagiarism. It also compares some identification technologies and systems suggested by previous scholars, and analyses identification technologies in the past five years which are based on Abstract Syntax Tree, Program Dependence Graph, BP Neural Network and XML. The paper further points out the lacks of identification accuracy, test sets and an official definition of similarity in this field. It also looks into the trend of this technology, such as a combination of different techniques and application of new algorithms.