CAREER: Dataflow Analysis for Modern Software Systems

提案编号：0546040提案标题：职业生涯：现代软件系统的数据流分析PI：Atanas Rountev，俄亥俄州州立大学跨过程数据流分析在软件维护、测试、验证和优化工具中发挥着核心作用。 现代软件具有传统方法无法处理的特征：它通常使用多个分布式组件，并且经常采用难以静态分析的动态机制。现有的分析在存在这些特征时会失败，从而难以为现实世界的软件系统提供复杂的工具支持。反过来，这降低了程序员的生产力，并导致软件质量下降。这个项目的重点是现代软件所带来的三个挑战：可重用组件，如标准库;分布式软件;通过动态类加载和反射的运行时适应。这一努力是朝着构建在软件行业中真正可用和有用的功能强大的软件工具迈出的重要一步。广泛使用的分析（例如，指向分析、MOD/REF分析、常数传播和对象命名）适用于分布式基于组件的系统。系统地概括了这些分析，以处理动态语言特征。传播是通过开源分析实现和两个程序理解工具。该项目的更广泛的影响包括：（1）研究基础设施，它提供了几个基本静态分析的可扩展的现成实现，（2）工具，它为程序理解提供了高质量的支持，这将提高生产力和软件质量，（3）研究与教育的结合，这将提高学生对当前软件开发方法和工具的熟练程度。并将加速他们的职业发展。

Proposal Number: 0546040Proposal Title: CAREER: Dataflow Analysis for Modern Software SystemsPI: Atanas Rountev, Ohio State UniversityInterprocedural dataflow analysis plays a central role in tools forsoftware maintenance, testing, verification, and optimization. Modernsoftware has characteristics that cannot be handled by traditionalapproaches: it typically uses multiple distributed components, and itoften employs dynamic mechanisms that are hard to analyze statically.Existing analyses fail in the presence of such features, making ithard to provide sophisticated tool support for real-world softwaresystems. In turn, this reduces programmer productivity and leads tolower software quality. This project focuses on three challenges posedby modern software: reusable components, such as standard libraries;distributed software; and run-time adaptation through dynamic classloading and reflection. This effort is a significant step towardsbuilding powerful software tools that are truly usable and useful inthe software industry.The theoretical foundations of dataflow analysis are generalized toachieve precision and scalability in the presence of reusablecomponents. Widely used analyses (e.g., points-to analysis, MOD/REFanalysis, constant propagation, and object naming) are adapted todistributed component-based systems. The analyses are systematicallygeneralized to handle dynamic language features. Dissemination isachieved through open-source analysis implementations and two programunderstanding tools. The broader impacts of the project include (1)research infrastructure which provides scalable off-the-shelfimplementations of several fundamental static analyses, (2) tools thatsupply high-quality support for program understanding, which willimprove productivity and software quality, and (3) integration of theresearch with education, which will increase the students' proficiencyin current methods and tools for software development, and willaccelerate their career progress.