An Ontology-Based Infrastructure for Creating Software Pattern Languages

点击翻译按钮获取中文摘要

Abstract0639164Scott Henninger - Univ Nebraska - LincolnTitle: An Ontology-Based Infrastructure for Creating Software Pattern LanguagesSoftware development is a knowledge intensive activity involving diverse technical and application domains. Software patterns, descriptions of recurring design problems coupled with proven solutions, have emerged as one means to systematically collect and disseminate software design knowledge. But as the number of available patterns proliferates, the expertise needed to find and effectively use patterns increases. Current pattern collections also contain no cross-references, missing important opportunities to capture and analyze interrelationships between knowledge embedded in patterns. The research objective is to investigate the efficacy of pattern "languages", tightly coupled organizations of patterns supporting systematic development methodologies. Semantic Web technologies will be investigated to federate pattern collections into Web-based ontologies supporting semantic pattern relationships and computational reasoning. Research efforts will empirically investigate existing software pattern collections, develop proof-of-concept demonstrations of ontology-based pattern languages, and support collaborative creation and refinement of pattern languages. The significance of this research is to address essential and elusive difficulties of using patterns to achieve significant quality and productivity gains. In addition, semantic relationships between patterns will help organize and refine software development knowledge. Broader impacts include Web-based access to federated pattern languages, creating building blocks for future pattern language research, and using research results for pedagogical purposes.