Metadata-driven infrastructure

Abstract

A Metadata-driven Infrastructure Supporting FAIR and Open-Source Practices Throughout the Research Software Lifecycle Research software is an essential part of research. It includes standalone tools, scripts, notebooks, and workflows for tasks such as data preparation, analysis, and visualization. When research software is not developed according to FAIR and Open Science principles, it can be difficult to maintain, discover, and reuse. We developed research software management recommendations and a modular tool suite that translates recommendations into practical outcomes. It includes a flexible Software Management Plan (SMP) template compatible with the Data Stewardship Wizard. The machine-readable information captured in the SMP can be exported as structured metadata, edited, and reused to generate documentation, citations, registry metadata, community files, and configurable repository scaffolds. A continuous integration tool validates repository metadata against a defined profile and identifies inconsistencies across metadata files and package manifests. The tool suite is designed for broad applicability, including those with limited programming experience. We piloted the SMP tool at Leiden University Medical Center. Users reported increased awareness of relevant practices and a willingness to implement them. Their feedback informed the tool's development. Our approach connects practical guidance with machine-readable metadata and repository tooling. It helps researchers adopt FAIR and open-source practices and supports research software that is sustainable, understandable, maintainable, discoverable, and reusable.