Two colleagues in the DataPLANT consortium participated in the workshop "Forschungssoftware managen" held at Stuttgart University mid September. The workshop was initiated by the DINI/nestor e.V. and brought together software developers and research data management specialists from various institutes, experts from service providers and university libraries.
Software plays a central role in everyday research in order to generate, manage and use research data. Often, however, the software used is not simple standard software, but has complex dependencies, is itself a product of research, and is indispensable for its reproduction and understanding. For example, in the DataPLANT consortium, in addition to self-developed tools and customized components, many software packages are used in different constellations. Examples include Swate/Swobup, the ARC Commander, the DataHUB, code in plant science tool stacks, or analysis systems for bioinformatics data. Questions arise as to how such research software can be developed and maintained in such a way that it will be re-used over the long term. This is where, for example, functional long-term access comes in, as shown by Rafael Gieschke as a tool for reproducing workflows "Long-term use of research software".
Another important issue is software licenses, which on the one hand ensure a reasonable reuse of self-written software and on the other handplay a role for reproducibility and reusability, especially in the case of proprietary software ("Ensuring reproducibility of research results by preserving access to research software" a talk stemming from experiences in the BioDATEN project). In many further presentations, version control (mostly via Git) played a central role in managing software and workflows. As in DataPLANT, this way of version control is used for deploying and sharing code and data in a variety of ways. Another larger issue was ensuring the sustainability of research software and associated funding models. Software publications are an important building block for reusability and reproducibility and should increasingly become part of the reputation system in science. This has just be emphasized by the DFG.This is part of the culture change that DataPLANT is helping to drive. For this purpose, DataPLANT provides a DataHUB from GitLab and soon InvenioRDM to be able to publish software DOI referenceable.