engleski

Analysis of programming code similarity by using intermediate language

The paper shows the possibility of analyzing programming code similarity by using intermediate language. Intermediate language that is generated by compiler is not sensitive to different kind of common code transformations, like changing variable names, modifications of some language constructs and loops. Therefore, the changes in original source code result in lesser changes in intermediate language code, that has positive effect on the results of code similarity analysis. Paper proposes a method for determining the similarity of programming code, and relies on existing and efficient string similarity algorithms. Method’s performance is shown by analyzing source code of applications submitted by students, as their homework. The paper describes method’s advantages and disadvantages, and suggests future steps to improve its results and accuracy.

plagiarism detection; source code; analysis; similarity; plagiarism detection algorithm

nije evidentirano