Enhancing the discoverability of mathematical source code

• Applicant: Dr. Moritz Schubotz
• Subject Area: Mathematics
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 561181416

swMATH, launched in 2013, is an open-access platform for comprehensive metadata on mathematical software. It was created by FIZ Karlsruhe - Leibniz Institute for Information Infrastructure in collaboration with the Mathematical Research Institute Oberwolfach. It systematically collects references to relevant mathematical publications, enhancing the visibility and impact of research software. Since 2022, swMATH has been integrated as a core long-term service to zbMATH Open. As of November 2024, swMATH contains 47 057 software packages referenced by 296 479 documents, making it the world's largest collection of mathematical research software. As a result of the FairCore4Eosc project, the source code referenced in swMATH is announced to the software heritage. Software heritage ensures the long-term preservation of the sources and creates an ecosystem around software source code. This project seeks to leverage the added value generated by the software heritage source code ecosystem to benefit the users of swMATH. For instance, we envision equipping swMATH with a source code recommender system (RS) tailored for research software, helping users discover the most suitable software. In addition, the swMATH editors can be supported in curating the metadata by AI-generated suggestions extracted from the software source code. The project involves three components: Based on our preliminary landscape analysis, we elaborate specific user-centric metrics to measure the performance of our service, define success criteria for the envisioned recommender system, and thus foster the discoverability of mathematical source code. Second, we enrich swMATH data with data from software heritage and other software registries. Moreover, we align general purpose (AI) services for software source code to the swMATH catalog so that those services can be more seamlessly used by mathematicians. Third, we develop an initial version of a recommender system and evaluate its effectiveness. The proposed RS will enhance swMATH by providing personalized recommendations based on metadata and source code analysis, including options for open-source alternatives and different algorithm implementations. The ultimate goal is to combine forces in the research software ecosystem and combine subject-specific services like swMATH and cross-disciplinary offers like software heritage and the associated ecosystem. We expect the mathematics community to benefit from this new service.

• DFG Programme: Research data and software (Scientific Library Services and Information Systems)