Empirical assessment and improvement of fault detection techniques for conurrent software

保证并发程序的质量是一个非常困难的问题，由于硬件技术的进步，这个问题变得越来越重要。像多核处理器这样的进步导致了对并发软件开发的需求的增加。不幸的是，开发正确的并发代码比开发正确的顺序代码要困难得多。建议的研究计划的目的是提高我们的理解不同的故障检测技术的并发软件，并利用我们的理解，以改善和定制这些技术的使用。该研究分为三个主题：1）经验评估和比较不同的故障检测技术的并发系统; 2）特定领域的故障检测技术，如模型检查和测试的定制; 3）提高不同的故障检测工具在大型软件项目的组合使用。在拟议的研究计划中的所有主题的基本主题是使用实证研究方法，以促进更好地使用工具，以及更有效和更高效的故障检测工具的开发。这项工作的最大意义是软件工程领域，特别是软件质量保证。在短期内，我们的研究计划将提供更好地了解今天的故障检测工具。这项研究的长期利益是开发综合质量保证方法，使用自动优化到特定领域或特定类别的故障的工具。

Ensuring the quality of concurrent programs is a very difficult problem that is becoming increasingly pertinent due to advances in hardware technology. Advances like multi-core processors have resulted in an increase in the need for concurrent software development. Unfortunately, developing correct concurrent code is much more difficult than developing correct sequential code. The objective of the proposed research program is to enhance our understanding of different fault detection techniques for concurrent software and to leverage our understanding to improve and customize the use of these techniques. The research is divided into three topics: 1) empirical assessment and comparison of different fault detection techniques for concurrent systems; 2) domain-specific customization of fault detection techniques like model checking and testing; 3) improving the combined use of different fault detection tools in large software projects. The underlying theme of all topics within the proposed research program is the use of empirical research methods to facilitate better tool usage as well as the development of more effective and efficient fault detection tools. The greatest significance of this work is to the field of software engineering and specifically software quality assurance. In the short term our research program will provide better understanding of today's fault detection tools. The long-term benefit of this research is the development of combined quality assurance approaches that use tools automatically optimized to a particular domain or a particular class of faults.