Against the blasphemous habit of reinventing the software wheel time and time again (Poster)

Research software is essential for the generation and analysis of research data and should be findable, accessible, interoperable and reusable (FAIR), however, the challenges in findability and reusability lead to redundant development. There is a growing need across scientific disciplines of the cultural sciences, humanities and social sciences, engineering and the natural and life sciences to ensure the sustainable use and further development of research software, which is enabled by its description with metadata, including its intended use, license, and requirements. However, this research software metadata is hard to discover, as it is scattered across various repositories, related publications, and code repositories. Our aim within the nfdi.software project is to improve access to research data by supporting specific software packages and enabling better (re)use and harmonization of metadata and software. By establishing a central access portal, the project aims to link and coordinate independent developments within a federated data infrastructure. The networking and contextualisation of research software unlocks the potential for the sustainable use of research data and expands the scope for its analysis and processing. Project development started in November 2024 with a clear requirements analysis and design for a Research Software Engineering (RSE) software marketplace. This formed the basis for the product vision, which is continuously evaluated and improved. By combining several approaches, a solid software architecture was established at the beginning of the project, which now provides the foundation for prototyping. Due to the various technical and content-related requirements, the team has formed working groups and is establishing cooperation with other NFDI consortia on specific topics. Joint workshops are being held on areas such as deep dependency, dependency graphs, reproducibility, monitoring, software KPIs, and sustainability. From the life-sciences domain, we have a comprehensive set of well annotated software, including corresponding training materials, benchmarks, packages, containers, and execution environments. The nfdi.software proof-of-concept service includes a website and an API, providing an interface to search and discover research software via the NFDI and beyond it, employing its integrated meta search engine. Identification of needs and community specific requirements will be collected in collaboration with consortia and basic services through workshops, discussions, and prototyping. The 'Prototype for Integration' will feature functionalities for harvesting and retrieving information, offering detailed software data along with related publications, people, and projects. Existing tools such as the Research Software Directory, bio.tools, the Research Software Ecosystem, Betty´s Research Engine, the Galaxy CoDex, and physics.tools will be integrated. An advanced search and recommender system will also be added. Various metadata ontologies and vocabularies will be explored and discussed to develop a common cross-sector vocabulary. Metadata-related efforts like knowledge graphs and Jupyter4NFDI will be evaluated, with results documented for feedback and improvement. Integration with IAM4NFDI is also planned to foster synergies. nfdi.software works closely with the task forces RSE of the Common Infrastructures section and Research Software Metadata of the (Meta)data, Terminology, Provenance section. The establishment of a German node of the EOSC under the responsibility of the NFDI and the onboarding of production-ready services will result in new requirements for an EOSC software marketplace.