Collaborative Research: Implementing the GOLD Community of Practice: Laying the Foundations for a Linguistics Cyberinfrastructure

语言科学的经验部分以数字形式提供的数据量迅速增加。尽管最近在标记语言、Web协议和数据管理技术方面取得了进展，但作为一个整体，语言学还没有能够充分利用它们。例如，单独的语言数据集通常被封装在与其他数据集不兼容的形式中：语言数据通常不能互操作。这在一定程度上是因为语言学才刚刚开始开发管理其数据的全领域最佳实践资源，包括通用软件工具、网络基础设施和知识组件，如本体论。事实上，这些资源将成为任何全外地网络基础设施努力的支柱。为了实现这一目标，这一合作项目将实施GOLD实践共同体，这是一个将在线语言数据与语言描述通用本体论(GOLD)获取的语言知识联系起来的网络架构。该项目的组成部分由夏飞飞和威廉·D·刘易斯领导，将通过从Web上收集大量行间注释文本来解决遗留数据的问题。结果将转换为最佳做法格式，并存储在行间文本在线数据库(ODIN)中。其次，Helen Aristar-Dry和Anthony Aristar将专注于通过进一步开发field来直接创建最佳实践数据，field是一种允许领域语言学家产生高质量词汇数据的工具。最后，Scott Farrar将在GOLD框架中实例化产生的最佳实践数据，从而集成来自前两个组件的数据。然后，研究小组将通过实施本体驱动的搜索设施来展示项目的效力，该设施将语言学的一般知识与数据实例捕获的具体知识结合起来。为了确保由此产生的体系结构得到广泛的曝光，我们将把这个项目的结果放在语言学家列表中，供整个语言学界查看和评估。这个项目将允许普通的工作语言学和任何对人类语言感兴趣的人搜索和查看大量语言数据的概括。它将直接解决比较和综合数据中涉及的关键问题，这些问题最初并不打算具有可比性。这些措施包括利用现有资源(即来自网络的遗留数据)、利用最佳做法数据标准和利用全外地知识。这些问题带来了重大的技术挑战，因为没有针对特定领域(如语言学)的通用现成解决方案。该项目的成功需要对语言数据对象和结构有深入的了解。事实上，该项目将展示如何在更广泛的框架中利用外地的基本数据结构。在世界即将失去大部分语言多样性的时候，这个项目将产生一个社区范围的资源，因为它作为探索各种人类语言结构的搜索工具的内在价值而可用。目前，还没有这样的语言数据搜索工具。当用户看到为此类工作做出贡献的价值时，他们将更有可能接受随附的数据标准和工具。因此，该项目将实现的是一个语言学家社区，致力于生产高质量的数据资源，以实现影响我们对语言结构理解的下一次重大进步的共同目标。

The empirical component of the linguistics sciences has seen a rapid increase in the amount of data available in digital form. Though there have been recent advances in markup languages, Web protocols, and techniques for data management, linguistics as a whole has not been able to take full advantage of them. For instance, individual sets of linguistic data are often encapsulated in forms that are not compatible with others: linguistic data are not generally interoperable. This is in part because linguistics has only begun to develop field-wide, best-practice resources for managing its data, including common software tools, Web infrastructures, and knowledge components such as ontologies. Such resources would, in fact, act as the backbone for any field-wide cyberinfrastructure effort. Towards such a goal, then, this collaborative project will implement the GOLD Community of Practice, a Web architecture for linking on-line linguistic data to linguistic knowledge captured by the General Ontology for Linguistic Description (GOLD). The component of the project, led by Fei Xia and William D. Lewis, will address the issue of legacy data by harvesting large amounts of interlinear glossed text from the Web. The results will be transformed into a best-practice format and stored in the Online Database of INterlinear text (ODIN). Second, Helen Aristar-Dry and Anthony Aristar will focus on the direct creation of best-practice data by further developing FIELD, a tool that allows field linguists to produce high quality lexical data. Finally, Scott Farrar will instantiate the resulting best-practice data in the GOLD framework, thus integrating data from the first two components. The research team will then demonstrate the efficacy of project by implementing an ontology-driven search facility that incorporates the general knowledge of linguistics with the specific knowledge captured by the data instances. To ensure that the resulting architecture gets wide exposure, we will house the results of this project at the LINGUIST List where it can be seen and evaluated by the linguistics community as a whole.This project will allow ordinary working linguistics and anyone with an interest in human language to search and see generalizations across large amounts of linguistic data. It will directly address the key issues involved in the comparison and integration data that were not originally intended to be comparable. These include the leveraging of existing resources (i.e., legacy data from the Web), taking advantage of best-practice data standards, and utilizing field-wide knowledge. These issues present significant technological challenges, as there are no general off-the-shelf solutions for given domains such as linguistics. The success of the project requires a deep understanding of linguistic data objects and structures. In fact, the project will demonstrate how the fundamental data structures of the field can be utilized in a broader framework. At a time when the world stands to lose much of its linguistic diversity, this project will result in a community-wide resource usable for its intrinsic value as a search tool to explore the structure of all kinds of human languages. At the present, there are no such search tools available for linguistic data. When users see the value in contributing to such an effort, they will be more likely to embrace the accompanying data standards and tools. Thus, what the project will achieve is a community of linguists dedicated to the production of quality data resources for the common goal of affecting the next great advance in our understanding of the structures of language.