Integration of Scripts in Reliable Programs

JAVASCRIPT编程语言无处不在：超过94%的网站使用它，它是GitHub和Stackoverflow等社区网站上使用最多的语言。随着它在服务器端应用程序和物联网应用程序中的使用越来越多，它不再仅仅是Web语言。然而，开发人员在用JavaScript维护相当大的软件系统时遇到了困难，因为该语言没有配备用于“大型编程”。TypeScript是微软开发的Java脚本的扩展，它提供类型和类型化的组件接口，以更好地支持大型项目。研究表明，类型带来的额外文档和工具机会将开发人员的生产率提高了15%。然而，由于现有的Java脚本程序、库和框架的类型化接口必须手动构建和审查，这阻碍了从Java代码到类型脚本的迁移。该项目将通过探索、开发和实现用于从现有的Java脚本代码库自动推断精确的类型脚本接口的新技术来改善这种情况。与以前在这一领域的工作相比，我们将使用一种简单的静态和动态分析的新组合，结合符号执行和测试生成的技术。这种组合的目标是直接从程序执行中获得真实和准确的类型信息，而不是从单独实现的构件(如精细的和易出错的分析系统)中获得。

The JavaScript programming language is ubiquitous: it is used inmore than 94 percent of all web sites and it is the most used language on community sites like GitHub and Stackoverflow. JavaScript is no longer just the language of the Web, as it sees increasing use in server-sideapplications and in IoT applications. However, developers struggle when maintaining sizable software systems in JavaScript because the language is not equipped for ``Programming in the Large''.TypeScript is an extension of JavaScript developed by Microsoftwhich provides types and typed component interfaces to better support largescale projects. Studies show that the additional documentation andtooling opportunies due to types boost developer productivity by upto 15 percent. However, the migration of JavaScript code to TypeScript ishampered by the fact that typed interfaces for existing JavaScriptprograms, libraries, and frameworks must be constructed and vetted manually.This project will improve on this situation by exploring,developing, and implementing new techniques for automaticallyinferring precise TypeScript interfaces from existing JavaScriptcode bases. Compared to previous work in this area, we will makeuse of a novel combination of simple static and dynamic analysis,combined with techniques from symbolic execution and testgeneration. The goal of this combination is to obtain authentic andprecise type information directly from program execution rather thanfrom separately implemented artifacts like elaborate and error-pronestatic analysis systems.