Effect of slice scope on data slice-based class cohesion metrics  

切片作用域对基于数据切片类内聚性度量的影响(英文)

作　　者：周毓明[1] 徐宝文[1] 

机构地区：[1]东南大学计算机科学与工程学院,南京210096

出　　处：《Journal of Southeast University(English Edition)》2008年第2期174-177,共4页东南大学学报（英文版）

基　　金：The National Natural Science Foundation of China(No.60425206,60633010);the High Technology Research and Development Program of Jiangsu Province(No.BG2005032)

摘　　要：This paper suggests that a single class rather than methods should be used as the slice scope to compute class cohesion. First, for a given attribute, the statements in all methods that last define the attribute are computed. Then, the forward and backward data slices for this attribute are generated by using the class as the slice scope and are combined to compute the corresponding class data slice. Finally, the class cohesion is computed based on all class data slices for the attributes. Compared to traditional cohesion metrics that use methods as the slice scope, the proposed metrics that use a single class as slice scope take into account the possible interactions between the methods. The experimental results show that class cohesion can be more accurately measured when using the class as the slice scope.提出应该使用类而不是方法作为切片作用域来计算类的内聚性.首先,对一个给定的属性,找出每个方法中对该属性最后一次定义的语句.然后,以类为切片作用域计算这些语句相对于该属性的前向和后向数据切片,通过合并得到该属性的类数据切片.最后,在所有属性的类数据切片的基础上计算类的内聚性.与传统的以方法为切片作用域的内聚性度量相比,以类为切片作用域的内聚性度量考虑了类中方法之间可能存在的交互.实验结果表明,当使用类为切片作用域时,所得的度量值能更准确地量化类的内聚性.

关 键 词：COHESION class data slice metrics OBJECT-ORIENTED 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]