Enhancing architectural mismatch detection with assumptions

Detecting software architecture inconsistencies is a critical issue in software design. Software systems are described in terms of components, component behavior and interaction and mismatch detection is explored through techniques based on behavior analysis. Integration problems, however, are not only caused by behavioral mismatch: components make assumptions about their environment to guarantee functional and non-functional properties. If the actual deployment environment of each component does not satisfy its assumptions, component and system properties may not hold. In this work we propose to extend the idea of architectural mismatch to include the notion of assumption. We concentrate on a subset of possible assumptions and show how software architects can benefit from using them. We also present a discussion on how architecture description languages (ADLs) can be extended to include assumptions.