基于UML的软件构件内聚耦合性度量工具设计与实现  

Design and Implementation of Software Component Cohesion and Coupling Metrics ToolBasedon UML

作　　者：王桐[1] 

机构地区：[1]天水市天水师范学院,甘肃天水741000

出　　处：《电脑与信息技术》2013年第3期34-38,共5页Computer and Information Technology

摘　　要：基于构件的软件复用被看成是提高软件生产率和软件质量的有效途径,被称为是有效解决软件危机的方法之一。近几年来,随着基于构件的软件工程的发展,构件的度量方法有了很大的发展。但多是研究可复用性,对构件的内聚耦合研究较少。而有效的系统分解是构件获取的主要手段,它为构件的复用提供了强有力的支持。文章在对系统分解后,以构件的内聚耦合度量为研究重心,利用其度量结果,再对构件可复用性进行研究。文章以有向带权依赖图为基础,运用图的谱分割技术,及结合信息熵技术的构件内聚耦合度量方法,实现对构件的自动聚集和自动度量。实验结果表明,图的谱分割技术能够较合理、有效的分解系统,文章中的度量方法是一种较准确的软件构件度量方法,为权衡构件的设计质量提供了可靠的数据。Component-based software reuse is regarded as an effective way to improve software productivity and software quality, and is also known as an effective way to solve the software crisis. In recent years, with the development of the component-based software engineering, the measurement of component has a great development. But more is to study the reusability of components, the study for component cohesion and coupling is few. And effective method of system decomposition is a primary means of getting components, it provides a strong support for the reuse of components. In this paper, after decomposing the system, we focus on the study of component cohesion and coupling, then we can study the reusability of components by the measurement results. In this paper, based on directed weighted class dependence graph, we apply spectral graph segmentation, and combine the measurement of the cohesion and coupling of component based on technology of information entropy, to realize the automatic aggregation and automatic measurement of component. Finally, through the evaluation of the reusability of the components, we can classify components by their quality. The results show that spectral graph segmentation can reasonable and effectively divide a system into components. And the metrics in this paper, is an accurate way to measure the software component and it is also a good way to offer the reliable data to design quality of software component.

关 键 词：图的谱分割技术 构件度量 内聚耦合 独立性 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]