一个设计模式自动识别技术研究框架  

New Research Framework for Automatic Recovery of Design Pattern

作　　者：周晓宇[1,2] 徐宝文[1,2,3] 

机构地区：[1]东南大学计算机科学与工程学院,南京211189 [2]江苏省软件质量研究所,南京211189 [3]武汉大学软件工程国家重点实验室,武汉430072

出　　处：《计算机科学》2009年第5期124-128,共5页Computer Science

基　　金：国家杰出青年科学基金项目(60425206);国家自然科学基金资助项目(60503020);江苏省自然科学基金(BK2006094)资助

摘　　要：在程序代码中自动识别设计模式可以为面向设计模式的软件理解、维护和重构等活动提供自动化支持。比较系统地分析了该领域十余年来的研究状况和存在的问题,提出了一套以准确和通用的设计模式识别技术为目标,以层次化识别特征的确定、基于静态分析技术的语义特征发掘与识别技术、基于语义特征和排除性约束的软件框架描述技术、基于描述逻辑的设计模式形式化技术等为内容的研究框架。介绍了各项技术的基本思想和主要内容,论述了其可行性。Automatic recovery of design patterns in codes can promote automation of program comprehension, maintenance and refactoring. Researches on design pattern recovery during the past decade and more were summarized, and shortcomings of existing techniques were pointed out. To solve the problems and achieve more accurate and general recovery technique,a research framework was introduced that is made up of researches on multilevel recovery target, precise definition and identification of semantic features based on static program analysis, software framework description based on semantic features and exclusive rules, formal design pattern description technology based on description logic. The main idea was presented and feasibility of the framework was demonstrated.

关 键 词：设计模式 软件分析 逆向工程 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]