Requirements and Their Impact Downstream: Improving Causal Analysis Processes Through Measurement and Analysis of Textual Information

Abstract : Requirements documents, test procedures, and problem and change reports from a U. S. Army Software Engineering Center (SEC) were analyzed to identify, clarify, and begin categorizing recurring patterns of issues raised throughout the product life cycle. Semiautomated content analysis was used to identify underlying patterns in the SEC documents. Automated tools and techniques were used to support efficient search and related semantic analysis that would not be possible manually. Discussions with Army personnel were used to confirm and elaborate initial findings and interpretations. The same analytic methods can be used as a basis for novel, proactive causal analysis processes. One of the patterns identified here suggests that usability is not sufficiently articulated and quantified early in the product life cycle. While the SEC has established exemplary processes to handle usability-related issues when they arise, some of them might be mitigated or prevented by documented consideration upstream.