NSF Workshop: Towards an Open Source Model for Data and Metadata Standards

机器学习和人工智能的最新进展有望在广泛的领域和活动中推进研究和理解。与此同时，人们对开放数据在可重复性和科学透明度方面的重要性的认识不断提高，这也在传统上没有产生大型公开数据集的领域取得了进展。来自出版商和资助者以及其他利益相关者的数据共享要求也产生了压力，要求通过数字存储库提供具有研究和/或公共利益价值的数据集。然而，为了充分利用现有数据，并促进创建有用的未来数据集，稳健、可互操作和可用的标准需要随着时间的推移而发展和适应。开源开发模型为标准的创建和调整过程提供了重要的潜在好处。特别是，标准的开发和调整可以利用长期存在的社会技术过程，这些过程一直是管理开源软件开发的关键，并允许将广泛的社区投入到这些标准的制定中。本次研讨会旨在在广泛的研究领域之间建立跨学科的联系，从而为知识交流提供肥沃的土壤，并创造有关将开源模型应用于数据和元数据标准的新的和广泛有用的知识。此外，将产生的综合将有助于决策者和资助者确定有价值的政策和资金投资途径，以最好地利用开源生产和治理原则来支持广泛的社会目标。通过坚持正式描述的开源标准（例如，通过实现标准规范的模式，和/或通过实现自动化的标准验证），在开源软件开发中很重要的过程，如自动化测试和持续集成，也可以在定义数据和元数据标准时采用。类似地，开源治理在标准开发中为一系列涉众提供了发言权，潜在地启用了在标准开发的自顶向下模型中不会考虑到的用例和关注点。另一方面，开源模型也有需要考虑的独特风险。本次研讨会的目标是讨论一些例子，在这些例子中，标准开发的开源模型对某个领域内的实践产生了重大影响。重要的是，研讨会还将讨论该模型在过去不起作用的情况，以及该模型不适合的情况。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Recent progress in machine learning and artificial intelligence promises to advance research and understanding across a wide range of fields and activities. In tandem, an increased awareness of the importance of open data for reproducibility and scientific transparency is making inroads in fields that have not traditionally produced large publicly available datasets. Data sharing requirements from publishers and funders, as well as from other stakeholders, have also created pressure to make datasets with research and/or public interest value available through digital repositories. However, to make the best use of existing data, and facilitate the creation of useful future datasets, robust, interoperable and usable standards need to evolve and adapt over time. The open-source development model offers significant potential benefits to the process of standard creation and adaptation. In particular, development and adaptation of standards can take advantage of long-standing socio-technical processes that have been key to managing the development of open-source software, and allow incorporating broad community input into the formulation of these standards. This workshop aims to create interdisciplinary connections across a wide range of research fields, thereby providing fertile ground for exchange of knowledge and the creation of new and broadly useful knowledge about the application of the open-source model to data and metadata standards. Furthermore, the synthesis that will be generated will be useful for policy makers and funders in determining worthwhile avenues for policy and funding investment to best make use of the open-source production and governance principles in support of broad societal goals.By adhering to open-source standards for formal descriptions (e.g., by implementing schemata for standard specification, and/or by implementing automated standard validation), processes such as automated testing and continuous integration, which have been important in the development of open-source software, can be adopted in defining data and metadata standards as well. Similarly, open-source governance provides a range of stakeholders a voice in the development of standards, potentially enabling use-cases and concerns that would not be taken into account in a top-down model of standards development. On the other hand, open-source models also carry unique risks that need to be taken into account. The goal of this workshop is to discuss examples where an open-source model for standards development has had significant impact on the practice within a field. Importantly, the workshop will also discuss cases where this model has not worked in the past, and cases where this model is not a good fit.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.