Mining patterns to support software architecture evaluation

In this paper, we present an approach to improve the software architecture evaluation process by systematically extracting and appropriately documenting architecturally significant information from software architecture and design patterns; we are interested in only two pieces of information found in software patterns: general scenarios and architectural tactics. General scenarios distilled from patterns not only assist stakeholders in developing concrete scenarios during a scenario-based architecture evaluation, but can also help an architect select and calibrate a quality attribute reasoning framework. Architectural tactics in patterns are used as a means of manipulating independent parameters in the reasoning framework to achieve the desired quality. Moreover, we believe if we use general scenarios and tactics extracted from patterns in an architectural evaluation, the results of that evaluation can be used as an evidence to validate the pattern's claim with respect to the quality attributes. We demonstrate our approach by using EJB architecture usage patterns. We contend that this approach can be used to analyze and validate any architecture pattern.