Automated Fine Grained Traceability Links Recovery between High Level Requirements and Source Code Implementations

Abstract

Software Traceability has been a matter of discussion in the Software Engineering community for a long time. The process of keeping and recover traces among software artifacts in any system represents a fundamental aspect to properly perform software maintenance tasks and requirements compliance verification. Furthermore, there exist application contexts where this becomes a mandatory process, for instance, banking and healthcare. Software traceability has dedicated efforts in proposing alternatives to recover lost traceability links in a coarse-grained and middle-grained detail by so far, however, proposed techniques are not enough to meet the desired levels of granularity in critical contexts. In this work we propose a fine-grained traceability algorithm designed to recover traces between high level requirements written in human natural language and source code statements where they are implemented. We tested our approach in four open-source healthcare systems to trace constraints requirements specified by the HIPAA law, and we evaluated the results as presented is this paper.