Getting back to basics: Promoting the use of a traceability information model in practice

It is widely assumed that following a process is a good thing if you want to achieve and exploit the benefits of traceability on a software development project. A core component of any such process is the definition and use of a traceability information model. Such models provide guidance as to those software development artifacts to collect and those relations to establish, and are designed to ultimately support required project analyses. However, traceability still tends to be undertaken in rather ad hoc ways in industry, with unpredictable results. We contend that one reason for this situation is that current software development tools provide little support to practitioners for building and using customized project-specific traceability information models, without which even the simplest of processes are problematic to implement and gain the anticipated benefits from. In this paper, we highlight the typical decisions involved in creating a basic traceability information model, suggest a simple UML-based representation for its definition, and illustrate its central role in the context of a modeling tool. The intent of this paper is to re-focus attention on very practical ways to apply traceability information models in practice so as to encourage wider adoption.