Digital RI in materials science should include several interconnected components.
The first component is computational workflows. These are prepared environments for typical materials science tasks: DFT, molecular dynamics, MLIP, structural modeling, calculation of material properties, analysis of results. At the first stage, it is advisable to focus on the tools Quantum ESPRESSO, AiiDA, ASE, LAMMPS, Python, PyTorch, NequIP and Jupyter.
The second component is the user's digital working environment. This can be JupyterHub, containerized environments, access to prepared notebooks, examples of running calculations, workflow templates and basic instructions. It is important for the user not only to have a server, but also to have a clear route: what to open, which example to take, how to run, where to save the result, how to describe it.
The third component is the FAIR data layer. This is the layer responsible for the data: DataverseUA or another compatible repository, DOI, DataCite metadata, README, licenses, minimum metadata profiles, preservation and citation rules. It is this component that transforms the results of calculations into data suitable for reuse.
The fourth component is the reproducibility service. This is critical for materials science. It is necessary to store not only the final result, but also input files, parameters, program versions, execution environment, scripts/notebooks, workflow description and the relationship between the calculation, dataset and publication.
The fifth component is user training and support. Many researchers need not only technical infrastructure, but also practical help: how to design a DMP, how to prepare a README, how to describe a DFT calculation, how to publish data, how to work with Jupyter, how to prepare a dataset for an article or grant application.
The sixth component is a public catalog of services and competencies. It should show what services are available, who supports them, who the demonstrators are, which groups are participating, which datasets have already been published, and what training materials are available.
