SHF: Small: Analyzing and Modeling Natural Language Usage in Software to Improve Software Maintenance Tools

该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。大型软件很难维护、修改和更新。迫切需要对软件系统导航、搜索和理解活动的自动化支持，以帮助软件开发人员和维护人员。该项目通过解决一些基本问题并提供创新方法来解决这个问题，这些方法为软件系统导航，搜索和理解活动提供了有用的自动化支持。这项研究将导致自动分析程序员的话和他们的关系，通过他们在代码中的使用，这将阐明编码在程序中的概念和行动。通过使用上下文和程序结构从“声音比特”构建概念模型，可以恢复更完整的画面，从而实现对软件维护和程序理解的自动化支持的变革性改进。分析在很大程度上由自然语言处理，信息检索和机器学习驱动。本文的研究将推动软件维护工具的理论和发展，提高软件维护人员的工作效率，降低软件维护成本，提高软件质量。在开发一个新的必修课程软件工程在小，PI将纳入学习如何使用和评估软件维护工具，包括那些在这个项目中开发的。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). Large scale software is difficult to maintain, modify and keep up to date. There is a dire need for automated support for software system navigation, search, and comprehension activities to assist software developers and maintainers. This project addresses this issue by tackling some foundational issues and providing innovative methods for useful automated support for software system navigation, search, and comprehension activities. The research will lead to automatic analyses of the programmer's words and their relationships, through their usage in code, which will elucidate the concepts and actions encoded in the program. By building the conceptual models from the ``sound bites'' using context and program structure, a more complete picture is recovered enabling transformative improvements in automated support for software maintenance and program comprehension.The analyses are heavily driven by natural language processing, information retrieval, and machine learning. The research will advance the theory and development of software maintenance tools, which will improve the effectiveness of software maintainers, decreasing software maintenance costs and raising software quality as software evolves. In developing a new required course on software engineering-in-the-small, the PIs will incorporate learning how to use and evaluate software maintenance tools, including those developed in this project.