2.1.1 Automated Requirements Quality Analysis through Requirements Comparisons

Abstract

Given the importance of key requirements documents such as the function and performance specification to the outcome of a systems engineering project, it is important to ensure that a system's requirements are captured and managed well. While contemporary requirements engineering tools are effective at the latter, they offer little assistance in preparing and assessing a set of requirements. This issue has been the subject of an ongoing research project within the Defence and Systems Institute at the University of South Australia that has led to the definition of a context-free grammar that underpins both a requirements entry wizard and a requirements assessment engine. This engine performs what we call Level 1 assessment in that it ensures that all the necessary components of the imperative statement are present and that the requirement passes elementary tests for reasonableness. This paper describes how through automated comparison of requirements statements much more powerful requirement assessments can be performed. The paper opens by describing a methodology to assess sets of requirements that can be applied at what we call the second and third levels of our three-tiered requirement assessment approach: namely, consistency checking across the requirements set and contextbased reasoning that seeks to apply a degree of real-world reasoning based on a comparison of the requirements for the system of interest against industry norms. To facilitate the comparison processes, this paper introduces and describes a finite set of comparison outcomes that can be used to characterise each comparison and become the input for further reasoning. The paper proceeds by showing how the second-tier assessment can be achieved by comparing each requirement within a requirements set against all of the other requirements thus identifying inconsistent and contradictory requirements. Similarly, the paper shows how the third-tier assessment can be achieved by comparing the overall meaning of the requirements set against that of benchmark sets. While the paper describes requirements matching concepts that are useful, it also identifies that there are numerous issues that need to be addressed to allow the concept to reach its full potential.

1.1.0 Is MBSE a Tectonic Shift in the Way We Perform Systems Engineering?

“…Some years later I accepted a position in systems engineering research and again became active in research in requirements engineering and support tools. This started with a project that commenced as a tools interoperability research task and progressed to a model-based approach to requirements engineering (Scott and Cook, 2008). In the early stages of this project, it was determined that the AP-233 Application Protocol for systems engineering data that was part of ISO-10303 (Version 5.1) was suitable for transfer of textual requirements between tools, however, it was considered no more suitable to act as a knowledge representation for artificial reasoning than the existing representations utilised by contemporary requirement management tools.…”