基于执行模式和扇入分析的横切关注点识别  

Crosscutting concerns identifying based on execution patterns and fan-in analysis

作　　者：俞东进[1] 闫大强[1] 

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《华中科技大学学报（自然科学版）》2012年第1期45-48,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61002018);浙江省重大科技专项基金资助项目(2008C11099-1)

摘　　要：为提高遗留系统的横切关注点识别效率和精确率,提出一种改进的候选横切关注点识别方法.该方法选取覆盖系统核心功能的挖掘用例,根据横切关注点具有的方法执行模式特征和方法扇入度特征对活动代码进行动态分析,以便发现可供候选的方面.在识别过程中,特别引入了扇入阈值和横切阈值,对由于低于阈值而不可能成为候选横切关注点的方法进行预过滤,同时通过分析方法外语句对方法执行的影响,排除部分不可织入的候选横切关注点.与传统方法相比,该方法在保证查全率不下降的基础上,可有效提高识别效率和精确率.In order to improve the efficiency and precision of finding crosscutting concerns in legacy systems, an improved approach to identify the candidates for crosscutting concerns was proposed. The cases were selected to cover main functions of system, corresponding runtime codes were checked, and the candidates were found out according to the recurring execution patterns and fan-in counts of methods. In particular, the threshold of fan-in and threshold of crosscutting was introduced to filter out the methods whose fan-in or crosscutting counts were below the threshold. Moreover, some candidates were excluded which could not be weaved by considering the impacts on methods from outside statements. The experiment shows that this approach achieves higher efficiency and precision without sacrificing its recalling rate, compared with the traditional approach.

关 键 词：软件开发 横切关注点 重复执行模式 扇入分析 识别效率 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]