Towards an Effective Analysis Pattern Retrieval

Since the introduction of patterns in Computer Science, a large number of libraries of patterns for different domains have been identified. In most cases these patterns are represented in a ‘flat’ fashion making their use difficult especially when there is a large number of patterns to consider in a particular application. In this paper we propose an analysis pattern retrieval (APR) approach under a classification schema in order to enable the automatic re-use of analysis patterns. The proposed classification scheme associates a problem to a set of pre-defined terms whereas the retrieval approach based on a Boolean model provides the necessary support to help a designer to locate and select analysis patterns that better solve the conceptual problem that he has in hands. An APR approach under a classification schema has the potential to make large analysis patterns collections easier to use and understand.