An Automated Transformation Approach for Requirement Specification

Abstract Use cases are often useful in capturing requirements by defining goal-oriented set of interactions between the system and its environment. Formalization of precise requirement is then important for context-aware verification based on use cases scenarios in the form of contexts. In this paper, we propose a high-level formalism for expressing requirements based on interaction overview diagrams that orchestrate activity diagrams automatically transformed from textual use cases. Our approach is qualified as context-aware model-checking; it supposes the availability of the system model as concurrent communicating automata and a specification language for describing requirements. Specification of requirements is performed through transformation phases to generate intermediate artefacts able to reduce the semantic gap between informal and informal requirement. The transformation is based on meta-models implemented in Ecore environment, algorithm and rules are defined using QVT Relational language, and primarily illustrated on an example.