A formal method for assessing architecture model and design maturity using domain-independent patterns

Abstract

Design patterns have been used as a formal or systematic means for extracting and patterning knowledge about good design choices, as well as capturing lessons learned associated with poor design choices (or so-called anti-patterns). Yet little attention is devoted specifically to pattern languages that are based on the fabric of architecture models – the conceptual data model – to capture reusable design knowledge and architecting best practices that can be applied in more than one domain at a high level of abstraction. This paper demonstrates a simple model-based method for identifying and patterning architecture design aspects that are domain-independent, and thus transferable and reusable in any system design with a comparable data model. The use of this method in formally documenting good and poor patterns in an abstract way is demonstrated by example. Discovered patterns such as those presented herein can be distributed, codified in a tool of choice, and sought out in actual architecture models of systems using automation. Since there may not be universal agreement on a common set of “good” and “poor” patterns, individual architects or organizations can use this method to state their particular practices as formal axioms, and structure them to assist in the assessment of model and design maturity against their own specific standards.