Can Text Mining Assistants Help to Improve Requirements Specifications?

Software requirements specifications are commonly written in natural language, making them prone to a number of defects, such as ambiguity, inconsistency, or lack of readability. Natural Language Processing (NLP) techniques have been proposed as a means to (semi-)automatically improve requirements specifications, but so far have not been widely adopted. We integrated a number of text mining assistants into a wiki-based requirements engineering platform to investigate two key questions: Can software engineers without prior training in NLP effectively leverage these techniques? And are text mining assistants actually helpful in improving the quality of a specification? Results obtained during two software engineering courses demonstrate that both are indeed the case.