基于软件网络错误传播分析的软件质量量度

潘伟丰¹，李兵^{2, 3}

(1. 浙江工商大学 计算机与信息工程学院，浙江 杭州，310018；
2. 武汉大学 软件工程国家重点实验室，湖北 武汉，430072；
3. 武汉大学 计算机学院，湖北 武汉，430072)

摘 要：基于软件网络错误传播的分析，提出一种量度软件质量的新方法。首先，将软件系统抽象成软件网络(方法调用网络)，节点代表方法，有向边代表方法间的调用关系；然后，通过错误植入的方法(随机植入和受控植入)，研究错误在方法调用网络中的传播过程，最终构建新度量SQC量度软件质量。真实软件系统的实例研究结果表明：SQC可以有效地量度软件的质量，并且与软件网络边数、结构熵等参数存在明显的负相关性，这对于指导高质量软件开发具有一定意义。

关键词：软件网络；复杂网络；软件度量；错误传播；软件质量

Software quality measurement based on error propagation analysis in software networks

PAN Wei-feng¹, LI Bing^{2, 3}

(1. School of Computer Science and Information Engineering, Zhejiang Gongshang University,
Hangzhou 310018, China;
2. State Key Laboratory of Software Engineering, Wuhan University, Wuhan 430072, China;
3. School of Computer, Wuhan University, Wuhan 430072, China)

Abstract:Based on the analysis of error propagation in software networks, a novel method was presented for quantifying the quality of software. Firstly, software systems were modeled as software networks (i.e., method calling network) with nodes representing the methods and directed edges representing the calling relationships between methods. And then errors were injected into software networks randomly or under control to study propagation dynamics of errors and proposed a new metric (SQC) to measure the quality of software. Empirical results on several real software systems show that SQC can be effectively used to measure software quality, and it has a negative correlation with other software networks metrics such as number of edges and structural entropy. All these results have many implications on the high quality software development.

Key words:software networks; complex networks; software metrics; error propagation; software quality