Frequent feature relationship tree-based coarse-grained business component design

Coarse-grained components improve the development efficiency of enterprise software and applications. However, the tradeoff between reusability and granularity of components is ignored by many methods of identifying and designing coarse-grained components. A frequent feature relationship tree (FRT)-based business component identification method is presented here, with the goal of identifying large-grained components that have a high reusability. A business model is described from the perspective of their functions and their compositions. An FRT-oriented component model and characteristic metrics are proposed. Furthermore, the mapping between component reusability and FRT frequency is established. An FRT mining method to find the largest-grained FRT satisfying a particular frequency is presented. The important steps of identifying frequent-edges, frequent single-level FRTs and frequent multi-level FRTs are presented, and a case study is also used to validate the method. © 2011 ISSN 2185-2766.