Nuanced term-matching to assist in compositional safety assurance

Increased complexity in the design, technology and supply chains for software-intensive safety-critical systems has resulted in a growing demand for a compositional approach to safety assurance. Assurance data relating to independently-derived components must be melded together into a compelling case for overall system safety. One of the barriers to composition is the lack of consistency in the terminology used to describe and share assurance data. Linguistic mismatches highlight various problems for the composition of peer modules and their integration into an overall case. In this paper, we propose the application of a linguistic model of understanding to identify mismatches and to provide guidance on composition and integration. The approach is illustrated using a simple example.