Enhancing Requirements Engineering Through Transparent Collection and Monitoring - A Metadata Schema and GitHub Repository Blueprint for a Requirement Collection Process

For the creation of appropriate Research Data Management (RDM) services, it is important to collect requirements from as many relevant stakeholders as possible. Common strategies to involve the community include mailing lists, workshops, face-to-face interviews, and surveys. We found them to be rather cumbersome to use because they make it hard to track the changes and the respective discussions in the long term. Instead, we propose using GitHub repositories to collect and visualize requirements, while the discussions can be tracked inside GitHub issues that are referenced when making changes. Using GitHub also enables easy input of requirements from external stakeholders. As requirements need to be tracked for different projects, we use multiple repositories that all follow the same repository blueprint on GitHub. This ensures a unified experience of requirement tracking, while also compartmentalizing the collection effort. The blueprint is intended to be reused not only by NFDI4Energy, but also by other NFDI consortia and beyond. We have defined a minimal metadata schema for requirements that can be adapted according to the specific needs of the target project or topic. The metadata schema consists of 14 fields, and where applicable, the DC Terms equivalent is listed. Any further development of the blueprint can be integrated into a repository that was created from the presented blueprint, using a git merge across both repositories. This provides an easy way to add new features, so all users of the blueprint can benefit. The blueprint brings several features that allow for the quick setup of a new requirements collection repository. An MkDocs configuration automatically provides a GitHub Pages website that can be used to easily browse the requirements and their metadata schema. Furthermore, the requirements can be visualized in an interactive table. This table is based on the DataTables.js library to bring features such as search, filtering, and downloading. The requirements themselves are managed in a CSV document within the repository, so they can be properly version-controlled. Any change requests shall be tracked with GitHub issues, so the repository blueprint provides the respective issue templates. It also comes with a proper first draft of the README file, as well as a quick guide to modifying the repository to get started. The blueprint is currently used for two requirement collections within NFDI4Energy: first, the requirements that were already collected and will be collected in the future for all the services that are or will be provided by NFDI4Energy, and second, a slightlycustomized collection that focuses on technical metadata requirements that shall form the foundation of NFDI4Energy metadata recommendations. For the latter collection, the requirements schema has been extended by a classification on the life cycle of metadata and to what part of metadata (schema, record, or tool) the requirement belongs to. This extension is visible in both the table and the metadata schema on the GitHub Pages website. A third repository that collects requirements exists within NFDI4Energy. It focuses on the requirements for a simulation-as-a-service hub. However, it has not used the blueprint yet, but will do so eventually.