Towards a knowledge-based approach for creating software architecture patterns ontology

Abstract

Software architecture patterns present solutions for software architecture problems and help to document architectural design decisions. Complexity and variability of patterns, and the required expertise for selecting an appropriate pattern, would cause some difficulties in utilizing architectural patterns. Using an ontology for registering architectural patterns is an efficient step in solving those problems. However, the mentioned difficulties make the process of constructing the architectural patterns ontology even more complicated. This paper proposes an approach that considers the construction of the architectural patterns ontology from four perspectives in order to overcome this complexity. Each of these perspectives has certain functionality. The ontology engineer deals with large volumes of patterns that would complicate registration of them in the ontology. Thus, the using of a relational database as an ontology learning resource is considered in this approach and the learning process is explained in knowledge extraction perspective. Sometimes, all knowledge of patterns (the possible interactions of patterns and the architect decisions) is not explicitly stored. Knowledge storage and representation perspective is placed in this approach in order to describe knowledge of patterns and to infer further knowledge. How retrieving an appropriate pattern is discussed in knowledge retrieval perspective. Knowledge sharing is another perspective that discusses how making the patterns ontology available for software architecture stakeholders.