SHF: Medium: Combining Speculation with Continuous Validation for Software Developers

SHF：媒介：将推测与软件开发人员的持续验证相结合

• 批准号: 0963757
• 负责人: Michael Ernst
• 金额: $ 75万
• 依托单位: University of Washington
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-09-01 至 2015-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0963757&HistoricalAwards=false
• 关键词: SHF; Medium; Combining; Speculation; Continuous

Unprecedented computational power is available from multi-core processors and cloud computing. To date, this power has been used primarily to make programs run faster. However, in many cases the bottleneck to solving users' problems is in the challenge of creating the software, not in the time to run it. This project will apply computational power to the real bottleneck, providing developers with new types of feedback. As a key broader impact, the research will enable developers to create software more quickly, more cheaply, and with higher quality.The key technical idea is to inform developers, in advance, of the consequences of their likely actions. The development environment speculates about developer actions, evaluates the effect of each action (on compilation, tests, version control conflicts, etc.), and unobtrusively makes this information available to the developer. By knowing which choices are good and which are bad, developers can avoid bad choices that cost time or reduce quality. The project's intellectual merits include algorithms to quickly create and evaluate many possible developer actions, UI design for developer awareness, and evaluation of how increased awareness of contingent information, about possible actions, affects developers. This also leads toward an answer to the question: If developers had infinite processing power, what fundamental software engineering research problems would remain?

多核处理器和云计算提供了前所未有的计算能力。 到目前为止，这种能力主要用于使程序运行得更快。 然而，在许多情况下，解决用户问题的瓶颈在于创建软件的挑战，而不是运行软件的时间。该项目将把计算能力应用于真实的瓶颈，为开发人员提供新类型的反馈。 作为一个关键的更广泛的影响，该研究将使开发人员能够更快，更便宜，更高质量地创建软件。关键的技术思想是提前告知开发人员他们可能采取的行动的后果。 开发环境推测开发人员的动作，评估每个动作的效果（对编译、测试、版本控制冲突等），并且不引人注意地使该信息对开发者可用。 通过了解哪些选择是好的，哪些是坏的，开发人员可以避免浪费时间或降低质量的错误选择。 该项目的智力优势包括快速创建和评估许多可能的开发人员操作的算法，开发人员意识的UI设计，以及对可能的操作的偶然信息的意识如何影响开发人员的评估。 这也导致了一个问题的答案：如果开发人员有无限的处理能力，什么基本的软件工程研究问题将仍然存在？