Software Process Analysis: Integrating Models and Data

9804067管理开发和维护软件的过程已成为需要以具有成本效益的方式交付可靠软件的组织的重点。为此，已经寻求了两个相互垂直的方向：行业制定了标准，而学术研究努力集中在正式模型上。本研究的目标是通过数据分析和测量将这两种方法结合起来。在这个项目中，设计了将过程数据分析与过程模型相结合的新技术，包括对过程的并发行为、属性关系和实时性能的分析。此外，对现有的过程数据分析框架进行了扩展，以支持这些新方法，并构建了体现这些新技术的工具。对新方法的评估是通过将它们应用于工业、真实世界的数据，并与工业合作一起进行的。提供这些技术将使软件开发人员能够保持真正的优化流程-正式模型的严格框架允许数据分析以明确的方式衡量流程，从而向流程反馈清晰和有洞察力的信息。除了这一实际影响之外，该项目还将增强和扩展开发可靠软件系统过程的科学知识。*

9804067 Managing the process of developing and maintaining software has become the focus of organizations that need to deliver reliable software in a cost-effective manner. To this end, two orthogonal directions have been pursued: industry has developed standards, while academic research efforts have focussed on formal models. The goal of this research is to integrate these two approaches through data analysis and measurement. In this project, new techniques are devised for combining process data analysis with process models, including analysis of concurrent behavior, attribute relations, and real-time performance of processes. In addition, an existing framework for process data analysis is extended to provide support for these new methods, and tools are constructed embodying these new techniques. Evaluation of the new methods is undertaken by applying them to industrial, real-world data, in conjunction with industrial collaboration. Providing these techniques will empower those who develop software to maintain truly optimizing processes---the rigorous framework of a formal model allows the data analysis to measure the process in an unambiguous manner, thus feeding back clear and insightful information to the process. In addition to this practical impact, this project will enhance and extend the scientific knowledge of the process of developing reliable software systems.***