The Homologous Software Detection Technology Study Of Structural Information Based And Token-Based Comparison

With the rapid development of computer software technology, there are more and more excellent software, and the computer software market has changed dramatically. We pay more attention to the value of computer software because of the huge economic and social benefits brought by computer software. Therefore, plagiarism becomes more serious. Especially, many open source codes are commonly copied by other software. So the detection of plagiarisms for the software becomes more important.How to maintain the rights of computer software developer, which is to protect software intellectual property rights, has become a worthy of study more and more. Homologous software detection technology can provide a vital support for intellectual property protection. The homologous software detection will find out whether the two software are identical or similar by comparing both of them from source code to the functionality layer, and then taking the result as certificate of software intellectual property. The basic source code is the most important evidence for the comparison and identification.The homologous software detection technology plays a very important role in the work of intellectual property protection by identifying code plagiarism. Plagiarism mainly happens as copy-and-paste of the code, replacing the name of functions or variables, reordering the sequence of the statement, type redefinition, and so on. At present, there are two homologous software detection technology methods on the market:binary executable -based similarity detection and source code -based similarity detection. And there are mainly three homologous software detection technology methods based on source code: text-based similarity detection, token-based similarity detection and syntax structure-based similarity detection.At first, this paper describes the theories of software testing technology homology, introduces the relevant technology and application tools in the market. This article mainly researches homologous software detection technology from two points:binary executable -based similarity detection, source code -based similarity detection.Then, this paper describes the homologous software detection technology based on binary executable, including the steps of the technology, key technology, test results and related technical modules. The paper designed and implemented a binary executable program homology detection system, and researched the similarity measure of the binary executable program in-depth.Token-based similarity detection technology can find the plagiarism of copy-and-paste of the code, replacing the name of functions or variables, reordering the sequence of the statement but type redefinition. In order to detect code plagiarism more effectively, we present a detecting algorithm based on type redefinition plagiarism in this paper. It could detect any level of simple type redefinition plagiarism, repeated type redefinition plagiarism and type redefinition with pointer plagiarism.Finally, the paper researched token-based homologous software detection by experiment. Experiments show, the algorithm can detect type redefinition code plagiarism effectively, increasing accuracy of detection, performing well in the code comparison field.