Portal:persona lucas story
Lucas Schmidt is a PhD student in algorithm design and computational mathematics. He developed a novel algorithm for his thesis and then faced a familiar problem: he needed to find comparable approaches, reproduce their results, and eventually make his own work open and verifiable. Here is how he did it with MaRDI.
Finding comparable approaches and source code
Lucas needed open-access implementations of similar algorithms along with the datasets used to benchmark them. The MaRDI Knowledge Graph gave him a structured starting point — linking algorithms, software packages, and associated publications in one place. Through its connection to swMATH and OpenML, he found existing implementations and benchmark datasets he could actually run.
For algorithm-specific discovery, MathAlgoDB was particularly useful. It connects mathematical algorithms, problems, and software in a structured knowledge graph, making the landscape of scientific computing searchable and comparable — exactly what Lucas needed to position his own contribution.
Packaging his software for reproducibility
Once Lucas was ready to share his own algorithm, he used the MaRDI Packaging System (MaPS) to package his implementation with all its dependencies into a self-contained runtime. This meant anyone could install and run his code with a single command, without worrying about version conflicts or missing libraries.
Making his work citable
Lucas documented his algorithm and workflow using MaRDMO, which guided him through a structured interview to capture the relevant metadata and publish it to the MaRDI Knowledge Graph. His work now has a persistent identifier and is linked to the datasets and publications it builds on — making it both citable and discoverable by others.
Storing his data and notebooks
To ensure his datasets and Jupyter notebooks remain available long-term, Lucas deposited them via MaRDI Portal Storage, which provides content-addressable, IPFS-based storage with permanent, verifiable links.
Not sure where to start?
The MaRDI Help Desk is happy to help you get started and find the right approach for your use case.
