CCRI: Collaborative Research: Planning for World Of Code (WoC): An Infrastructure for Open Source Software Census

虽然软件工程研究在过去的几十年里取得了很大的进步，但数据的不可用一直是一个主要的限制。开放源码软件(OSS)使大量数据可用，创造了在非常大的数据集上使用现代量化技术来了解如何创建更好的软件和提高生产率的前景。然而，这些数据以不同的格式存在，有许多错误和遗漏，位于世界各地的数百万个存储库中，需要进行广泛的处理才能以研究人员可以使用的形式呈现。计划拨款旨在开发创建代码世界(World of Code，WoC)的要求，这是一个加强软件工程研究的基础设施，并建立一个以从世界各地所有开放源码库收集的海量数据为中心的研究社区。这一规划过程将确保，如果实现了WOC，将使研究人员能够轻松访问所有开放源代码数据，解决复制问题，避免抽取不具代表性的样本，并降低基于错误数据得出不适当结论的风险。规划赠款将设立一个指导委员会和一个咨询委员会，以向研究人员社区提供最大价值的方式指导资源的开发和演变。黑客马拉松将为有经验和新手的研究人员提供一个机会，学习如何使用WoC的原型功能，添加工具，并具体表达WoC寻求服务的研究社区的愿望和优先事项。挖掘和维护整个OSS代码库和版本历史所需的资源远远超过传统假设驱动研究范围内单个研究小组所能完成的任务。调查整个OSS的迫切需求将通过为基础设施创建需求和构建社区来解决，该基础设施应包含：1)整个OSS的完整、快速更新的源代码数据；2)对此类数据必不可少的数据校正、增强和管理工具：例如，身份消除歧义和提取依赖项；3)支持依赖于整个集合的研究任务的服务；4)由研究人员、开发人员和维护、增强和运营该基础设施的公司组成的自治社区。为了得出对这一开放源码软件范围的基础设施的准确要求，将聘请广泛的软件研究人员。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

While software engineering research has made great progress over the decades, the unavailability of data has long been a major limitation. Open source software (OSS) makes large quantities of data available, creating the promise of using modern quantitative techniques on very large data sets to understand how to create better software and improve productivity. However, the data exists in different formats, has many errors and omissions, is located in millions of repositories around the world, and requires extensive processing to render it in a form researchers can use. The planning grant aims to develop the requirements for creating World of Code (WoC), an infrastructure for enhancing software engineering research, and to build a research community centered around massive data gathered from all open source repositories around the world. This planning process will ensure that WoC, if realized, will enable researchers to easily access all open source data, resolve issues of replication, avoid drawing samples that are not representative, and reduce the risk of inappropriate conclusions based on erroneous data. The planning grant will establish a Steering Committee and an Advisory Board to guide the development and evolution of the resource in a way that provides the greatest value to the community of researchers. Hackathons will give both experienced and novice researchers an opportunity to learn how to use WoC's prototype capabilities, add to the tooling, and concretely express the desires and priorities of the research community WoC seeks to serve.Resources needed to mine and maintain data on the entire OSS code base and version histories far exceed what individual research groups can accomplish in the scope of conventional hypothesis-driven research. The pressing need to investigate the entirety of OSS will be addressed by creating requirements and building a community for an infrastructure that should contain: 1) A complete, rapidly updated source code data for the entirety of OSS; 2) Tooling for data correction, augmentation, and curation essential for such data: e.g., identity disambiguation and extraction of dependencies; 3) Services supporting research tasks that rely on the entirety of the collection; 4) A self-governing community of researchers, developers, and companies who maintain, enhance, and operate this infrastructure. To elicit the precise requirements for this OSS-wide infrastructure a wide audience of software researchers will be engaged.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.