面向代码设计质量监控的软件度量指标集研究

随着软件的演化,软件规模和复杂性的上升往往会造成代码设计质量的退化,进而导致代码可维护性下降。已有大量软件度量指标用于量化代码设计质量,但是由于数量众多,不同指标体系度量结果可比性较差,使得开发人员难以找到存在设计质量问题的模块。系统调研并整理归类了现有的规模、耦合、内聚、封装、继承和多态等6方面软件度量指标。结合度量指标关系以及实验分析,从每个方面的指标集中发现了与代码设计质量关联度最高的6个指标,从而提出一种面向代码设计质量监控的软件度量指标集。实验表明,该指标集可以有效挑选出存在代码设计质量问题的类,可作为开发人员监控和定位代码设计问题的重点关注指标。

SOFTWARE METRICS SET FOR CODE DESIGN QUALITY MONITORING

Along with the evolution of software,the increase of software scale and complexity often leads to the degradation of code design quality,which leads to the decline of code maintainability.A large number of software metrics have been used to quantify the quality of code design.However,due to the large number and poor comparability of metrics results of different metrics systems,it is difficult for developers to find modules with design quality problems.This paper sorted out the existing software metrics and divided them into 6 types,including scale,coupling,cohesion,encapsulation,inheritance and polymorphism.Combining the relationship between the metrics and the experimental analysis,6 metrics with the highest degree of correlation with the quality of code design were found from each metrics set.Thus,this paper proposed a software metrics set for code design quality monitoring.The experiment result shows that the metrics set can select classes with code design quality problem effectively,and can be used as the key metrics for developers to monitor and locate code design problems.