NFDI4Ing

Engineering sciences play a key role in developing solutions for the technical, environmental, and economic challenges imposed by the demands of our modern society. The associated research processes as well as the solutions themselves will only be sustainable if being accompanied by a proper research data management (RDM) that implements the FAIR data principles: data has to be findable, accessible, interoperable, and re-usable. NFDI4Ing brings together the engineering communities to work towards that goal. As part of the German National Research Data Infrastructure (NFDI), the consortium aims to develop, disseminate, standardise and provide methods and services to make engineering research data FAIR.

An overarching goal of the NFDI4Ing consortium is that scientists of all disciplines are able to retrace and reproduce all steps of engineering research processes. This requires careful research data management, research software development, and documentation.

Within the task area Betty in NFDI4Ing, our main focus is on research software development, with the vision that researchers are equipped with the tools and the knowledge required to produce validated research software of high quality, which is able to reproduce published research results.

To this end, we have compiled teaching material around sustainable software development, whose scope and depth have been adjusted to scholars with a background in engineering. It comprises basic software design principles, design patterns and an introduction into software tools that facilitate the software development process, for instance, for version control, automated testing or continuous integration.

In order to facilitate quality-assurance, we have developed a tool that helps researchers with implementing regression tests for research software that performs numerical simulations. In addition, we are currently developing a C++ library that enables research software developers to easily write their data into standard file formats for numerical results.

A key component of sustainable research software is that it is able to reproduce published research results, and that it can be reused and extended. Oftentimes, research results are the product of an entire toolchain of computations, possibly done with different software, arranged in a particular order. In a recent project, we have investigated how workflow management systems, which allow for the composition of such software toolchains, can help to increase the transparency, reproducibility and reusability of research workflows. The result of this investigation  documented in a public repository  and a journal submission can help researchers to decide which workflow management system is most suitable for their particular needs.

A complementary ingredient for improving the interoperability and reusability of research data and the reproducibility of computational results is the provision of Jupyter notebooks. As part of the Base Services Measure S-2 of NFDI4Ing, the TIK offers a JupyterHub instance which allows researchers to create, adapt and execute Jupyter notebooks inside the browser without requiring any local installation. A particular strength of our instance is the provision of a Matlab kernel in addition to the standard Python kernel. 

Within the framework of the Base Services Measure S-3, we are also significantly involved in the development of the ontology m4i, which enables machine-readable and semantified documentation of research processes and their results.