| Even though software requirements and software architecture tremendously impact the success and quality of a software project, few tools exist to support the transition. The desired qualities in a software project that are expressed in the software requirements can be realized with software architectural styles. This thesis proposes an approach towards an automated tool that recommends software architectural styles given natural language software requirements. The approach, Automated Architecture Scoring Method (AASM), transforms software requirements into a model which is analyzed for possible software architectural components, and possible properties of those components. AASM then analyzes those properties and recommends an architectural style. This thesis focuses on the architectural style selection, with a further emphasis on software architectural connector properties. |
