Plagiarism is defined as an activity to use someone’s work without the author’s agreement or without a proper citation about the reference[1]. To detect plagiarized programming source codes should be performed by instructors to promote the quality of school education. However, it is virtually impossible to detect those source codes completely within the limited time by only using human’s ability. Therefore, it is natural to try to adopt computing power to this operation.

Most of the existing plagiarism detection systems use an algorithm to find Longest Common Subsection to measure the similarity between two program sources, but does not provide a way to compare algorithms used for those program source files[2-6].

In this paper, we discuss how to build a new detection system named IRIS that uses Strict Binary Tree structure that was introduced in JK system[7] and execution function call sequences in order to determine the algorithms used in program source files, which will be the major factor to measure the similarity of two compared files by applying Software Metrics additionally.