摘要
构件粒度在很大程度上影响着构件的复用性能,但目前尚缺乏严格有效的方法以支持对构件粒度的优化设计.以特征空间作为表达业务模型与构件语义的统一方式,给出了构件粒度的定量度量方式,并探讨了业务模型空间与构件空间之间的四种映射机制.通过分析构件粒度与业务模型稳定性之间的关系,提出基于模型稳定性的动态粒度构件设计思想,从异构体数目、稳定熵、相似度三个方面的稳定性进行度量,进而给出了基于最大稳定集的构件识别算法STCI M.实例验证与理论对比分析证明了STCI M相对于其它构件识别方法在目标构件复用性能优化方面的优越性.
Coarse-grained reuse has been the primary tendency in software reuse field, and it has been proved that granularity has great effect on component reuse performance. However, previ ous studies have ignored rigorous and effective methods to support coarse-grained component identification and granularity optimization. In this paper a business model stability based compo nent granularity optimization design method is presented to solve this problem. Firstly a featureoriented component model and the corresponding component granularity metrics are briefly pro posed, based on which, four different mapping strategies between business model space and component space are discussed for obtaining dynamic and optimal component granularities. According to some theoretic analysis on the correlation between business model stability and component granularity, i. e. , the more stable a business model is, the larger the corresponding component＇s granularity could be, a Stability based Component Identification Method（STCIM）is put forward in detail with the dynamic granularity mapping strategy, based on the metrics for business model stability with three independent factors, i. e. , isomer number, stability entropy and isomer similarities. The significant advantages of STCIM on component reusability optimization are validated by a practical case and the qualitative comparisons between STCIM and other component identification methods in literatures.
出处
《计算机学报》
EI
CSCD
北大核心
2006年第2期239-248,共10页
Chinese Journal of Computers
基金
国家自然科学基金项目"基于业务模型稳定性度量的软件构件获取与优化设计"(60573086)
高等学校博士学科点专项科研基金项目"大型企业管理应用软件的可重构模式
过程与实现技术研究"(20030213027)
国家"八六三"自动化领域十五重点项目"新一代企业资源计划系统研究与开发"(2003AA4Z3210)资助~~
关键词
构件识别
动态粒度
业务模型
稳定性度量
component identification
dynamic granularity
business model
stability evaluation
