摘要
一个项目的开发活动是很多软件过程的集合,不同软件过程之间关联性很强,成功地分析特定软件过程质量的关键是确保软件过程分析的独立性,剔除来自于其他过程的影响。传统Shewhart控制图基于统计假设检验理论,能够区分软件过程中的偶然因素和系统因素,但Shewhart控制图是全控图,无法区分过程之间的影响。为解决这种问题,定义软件过程的总质量和分质量,把系统因素细分为外部系统因素和内部系统因素,并总结软件过程质量诊断表,以使用控制图和选控图来帮助诊断导致软件过程质量异常的偏差源。
Software development consists of many related software processes; the key of successfully analyzing the quality factors of a specific software process is to eliminate the effects from other processes. Traditional shewhart control charts can only distinguish common causes from assignable causes, but it can＇t be used to identify assignable causes from other processes. Total quality and partial quality of software development processes were defined. The assignable causes were divided into process-inner assignable causes and processouter assignable causes, and shewhart control charts and select cause control charts were used to identify variation sources that result in software process variation. A diagnosis table was summarized that helps to resolve every variation scenario of software process quality.
出处
《计算机工程与设计》
CSCD
北大核心
2005年第7期1793-1796,共4页
Computer Engineering and Design
关键词
统计过程控制
控制图
软件过程
statistical process control
control charts
software processes
