(翻译)组织和融合仿真数据的元数据目录A metadata catalog for organization and systemization).docVIP

A metadata catalog for organization and systemization of fusion simulation data 组织和融合仿真数据的元数据目录 Abstract 摘要 Careful management of data and associated metadata is a critical part of any scientific enterprise. Unfortunately, most current fusion simulation efforts lack systematic, project-wide organization of their data. This paper describes an approach to managing simulation data through creation of a comprehensive metadata catalog, currently under development. The catalog is intended to document all past and current simulation activities (including data provenance); to enable global data location and to facilitate data access, analysis and visualization through uniform provision of metadata. The catalog will capture workflow, holding entries for each simulation activity including, at least, data importing and staging, data pre-processing and input preparation, code execution, data storage, post-processing and exporting. The overall aim is that between the catalog and the main data archive, the system would hold a complete and accessible description of the data, all of its attributes and the processes used to generate the data. The catalog will describe data collections, including those representing simulation workflows as well as any other useful groupings. Finally it would be populated with user supplied comments to explain the motivation and results of any activity documented by the catalog. 对任何科技企业来说，精心管理数据和相关联元数据都是十分关键的部分。不幸的是，当前大多数针对数据的融合仿真工作都缺乏系统性，以及项目的组织性。本文介绍了一种通过全面的元数据目录来管理仿真数据的方法，目前这种方法还在开发中。这种目录原打算记录所有过去和现在的仿真行为（包括数据源）；通过元数据的统一规定使全球数据定位成为可能，并提供便利的数据访问、数据分析和可视化数据。元数据目录将捕获工作流程，掌握至少包括数据导入和分级、数据预处理和输入准备、执行代码、数据存储、数据后期处理和输出在内的每个仿真模拟活动记录。该系统的总体目标是，在目录和主要数据档案之间，应持有对数据完整易懂的描述，它的所有属性和方法都用于生成数据。该目录将描述那些代表仿真工作流程的数据集，就像描述其他有用的组合数据集一样。最后，它收集用户提供的评论，解释目录记载的任何活动动机和结果。 Highlights 强调 ? We find that modeling and simulation data need better systemization. ? Workflow, data provenance and relations among data items need to be captured. ? We have begun a design for a simulat