Requirement boilerplates: Transition from manually-enforced to automatically-verifiable natural language patterns

Abstract

By enforcing predefined linguistic patterns on requirements statements, boilerplates serve as an effective tool for mitigating ambiguities and making Natural Language requirements more amenable to automation. For a boilerplate to be effective, one needs to check whether the boilerplate has been properly applied. This should preferably be done automatically, as manual checking of conformance to a boilerplate can be laborious and error prone. In this paper, we present insights into building an automatic solution for checking conformance to requirement boilerplates using Natural Language Processing (NLP). We present a generalizable method for casting requirement boilerplates into automated NLP pattern matchers and reflect on our practical experience implementing automated checkers for two well-known boilerplates in the RE community. We further highlight the use of NLP for identification of several problematic syntactic constructs in requirements which can lead to ambiguities.