The requirements traceability matrix has proven successful as a way to predict quality before code is written. However, its tedious development process, which requires that analysts manually discover and vet links between artifact levels has stymied its widespread adoption. Our tracing toolkit, called RETRO (Requirements Tracing on Target), automates the information retrieval process, making RTM's development easier. Backed by empirical results, we describe RETRO?s advantages over other information retrieval approaches.
This article is part of a special issue on predictor modeling.