The swMATH Project

Finding the right research software is a common hurdle for mathematicians. While databases excellently catalog academic papers, they often fall short in helping researchers discover and compare the specialized computational tools they need. This project aims to fill that gap by significantly upgrading swMATH, a key service that already catalogues thousands of mathematical software packages.

Our approach is to move swMATH from a static list to an intelligent discovery tool. The core of this upgrade is a deep integration with Software Heritage, the global archive for source code. By linking swMATH's detailed records directly to the preserved source code, we can build a much richer understanding of each software tool. With this foundation, the project's central goal becomes possible: "we envision equipping swMATH with a source code recommender system (RS) tailored for research software, helping users discover the most suitable software."

To build this system, the work will progress in connected phases. First, we need to understand exactly what makes a software recommendation useful by defining clear, practical success metrics. Next, we'll merge swMATH's data with other software registries to create a more complete picture. Finally, we'll develop and test a prototype recommender that analyzes both a tool's metadata and its actual code to provide personalized suggestions. The result will be a service that actively guides mathematicians to the best software for their specific problem, saving time and improving research outcomes.

• Duration: 01. Juni 2025 -31. Mai 2028
• Project sponsor: DFG
• Responsible for the project: Dr. Moritz Schubotz