通过检查性能度量评估软件团队实施有效检查过程的特征
Estimation of Characteristics of a Software Team for Implementing Effective Inspection Process through Inspection Performance Metric
摘要：
任何软件行业的持续存在都取决于其开发几乎零缺陷产品的能力，而这一点可以通过有效的缺陷管理来实现。检测已被证明是一种有前途的缺陷管理技术。引入诸如检查深度（DI，过程度量）和检查性能度量（IPM，人员度量）之类的度量使人们能够对检查技术进行适当的测量。本文阐述了一种数学方法来估计IPM值，而不依赖于车间缺陷计数在每一个时间。通过应用多元线性回归模型，评估了一组团队的特征系数。这些系数是从基于产品和基于服务的IT行业的团队中抽样的经验项目中计算出来的。三个验证项目的样本表明，从缺陷计数（IPMdc）获得的IPM值与使用数学模型（IPMtc）使用团队系数获得的IPM值之间非常匹配。现场观测的IPM值和我们的模型产生的IPM值具有很强的匹配性，通过团队系数支持IPM的预测能力。在确定了企业在项目中应达到的IPM值后，可以调整检验影响参数，以实现IPM的预期质量水平。团队系数的评估解决了几个与缺陷相关的问题，这些问题与管理层、利益相关者、外包代理商和客户有关。此外，系数向量还可以进一步辅助PSP和TSP的策略
英文摘要：
The continued existence of any software industry depends on its capability to develop nearly zero-defect product, which is achievable through effective defect management. Inspection has proven to be one of the promising techniques of defect management. Introductions of metrics like, Depth of Inspection (DI, a process metric) and Inspection Performance Metric (IPM, a people metric) enable one to have an appropriate measurement of inspection technique. This article elucidates a mathematical approach to estimate the IPM value without depending on shop floor defect count at every time. By applying multiple linear regression models, a set of characteristic coefficients of the team is evaluated. These coefficients are calculated from the empirical projects that are sampled from the teams of product-based and service-based IT industries. A sample of three verification projects indicates a close match between the IPM values obtained from the defect count (IPMdc) and IPM values obtained using the team coefficients using the mathematical model (IPMtc). The IPM values observed onsite and IPM values produced by our model which are strongly matching, support the predictive capability of IPM through team coefficients. Having finalized the value of IPM that a company should achieve for a project, it can tune the inspection influencing parameters to realize the desired quality level of IPM. Evaluation of team coefficients resolves several defect-associated issues, which are related to the management, stakeholders, outsourcing agents and customers. In addition, the coefficient vector will further aid the strategy of PSP and TSP
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
