Inadequate communication of domain knowledge in natural language (such as English textual descriptions) is a major source of requirements defects in high-confidence software. Such defects can threaten lives, property, and the dependability of critical infrastructures. This research develops innovative, multi-disciplinary techniques designed expressly to identify and cope with the properties of natural language that lead to these problems. It analyses the domain-knowledge communication problem from the perspective
of current linguistic theory in order to generate models and techniques that reduce communication breakdowns between domain experts and software developers. On the basis of this analysis, techniques for the elicitation, recording, and propagation of domain knowledge in requirements activities will be designed. A set of tools will also be developed to support the practical use of these engineering techniques. Several case studies will be used to assess and refine the techniques and tools on actual, high-confidence software systems in the aviation and spacecraft application domains. The research effort is directed at the reduction of requirements defects through the application of techniques that take into account the linguistic difficulties particular to domain knowledge communication.