This project is focused on highly concurrent source code static analysis based on the Map/Reduce programming model. Though the use of Hadoop, the metric extraction from very large code bases can be transparently parallelized to achieve a highly scalable and performant implementation.

 

The implementation is being conducted by two first year M.Sc. students taking the course on Software Engineering and Formal Methods at Minho University.

  

The Software Improvement Group (SIG) has software tools capable of performing (language independent) analysis of source code. As the systems that the SIG has to analyze tend to get larger and computers to have several cores, it becomes evermore relevant to investigate to which extent static analysis of source code can be parallelized to improve scalability and performance. For the students, it is an opportunity to apply the knowledge in Formal Methods in Software Engineering they acquire during their studies to a real life problem that they would not otherwise encounter in their academic path.