Poor Performing Patterns of Code: Analysis and Detection

Abstract

Software maintenance is a way to optimize software performances, but for huge code source's softwares, this maintenance may be very complex. The aim of our work is to detect suspected poor performing patterns in source code and to measure their performances, in order to state in which contexts they need to be refactored. We present the context of our work and our pattern analysis approach. We explain the reason why we have to establish a link between static analysis for pattern detection in source code and dynamic analysis for performance measurement in the corresponding executed instructions. This link will allow us to identify which set pattern instances really need to be refactored, expected to be restricted with respect to the large set and are instances encountered in the source code.