A Multilanguage Source Code Retrieval System Using Structural-Semantic Fingerprints

Source code retrieval is of immense importance in the software engineering field. The complex tasks of retrieving and extracting information from source code documents is vital in the development cycle of the large software systems. The two main subtasks which result from these activities are code duplication prevention and plagiarism detection. In this paper, the authors propose a multilanguage source code retrieval system based on two-level fingerprint representation, respectively the structural and the semantic information within a source code. A sequence alignment technique is applied on these fingerprints in order to quantify the similarity between source code portions.