CRII: SHF: Analysis and Automation of Global Systematic Changes

CRII：SHF：全球系统性变化的分析和自动化

• 批准号: 1565827
• 负责人: Na Meng
• 金额: $ 17.5万
• 依托单位: Virginia Polytechnic Institute and State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-03-15 至 2018-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1565827&HistoricalAwards=false
• 关键词: CRII; SHF; Analysis; Automation; Global

Software version history captures how software evolves, records application-specific domain knowledge, and tracks success as well as failure in the software maintenance. By learning from software evolution history, researchers can leverage the knowledge to help improve programmer productivity and software quality. However, the ?big data? aspect of software version history presents challenges. It requires sophisticated program analysis and automated tools to identify project-specific knowledge and to suggest coding accordingly to developers. Existing solutions are only capable of identifying and automating simple code changes. The objective of this research is to advance the state-of-the-art programming knowledge discovery and suggestion through innovations in automatic inference and application of program transformations.This project will characterize recurring global edits in software, and investigate tools to fully automate the crucial process to generalize, locate, and apply complicated program transformations based on code change examples. The unifying theme of this research is the judicious application of inter-procedural program analysis and constraint solving techniques. Inter-procedural analysis will identify relationship between applied changes by developers to cluster relevant changes. Constraint solving techniques will reason about the rationale behind the applied changes to infer the manifested general program transformation patterns. The PI will incorporate the results from this research into both undergraduate and graduate software engineering courses with the aim of maintaining the competitive vitality of the national technology workforce.

软件版本历史记录捕获软件如何演变，记录特定于应用程序的领域知识，并跟踪软件维护的成功和失败。通过学习软件演化历史，研究人员可以利用这些知识来帮助提高程序员的生产力和软件质量。然而，“大数据”？软件版本历史方面提出了挑战。它需要复杂的程序分析和自动化工具来识别特定于项目的知识并向开发人员建议相应的编码。现有的解决方案只能识别和自动化简单的代码更改。这项研究的目的是通过自动推理和程序转换应用方面的创新来推进最先进的编程知识发现和建议。该项目将描述软件中反复出现的全局编辑的特征，并研究使关键的工具完全自动化的工具。基于代码更改示例来概括、定位和应用复杂程序转换的过程。这项研究的统一主题是明智地应用过程间程序分析和约束求解技术。过程间分析将识别开发人员应用的更改之间的关系，以对相关更改进行聚类。约束求解技术将推理所应用的更改背后的基本原理，以推断所表现的一般程序转换模式。 PI将把这项研究成果纳入本科生和研究生软件工程课程中，以保持国家技术劳动力的竞争活力。