Research software recommendations for policy makers

1. R2: Encourage research software to be open source and require it to be available

1.1. Require that all research software be released under a license

1.2. Encourage the use of open source code licenses and require a short explanation on why a particular license was selected

1.3. Require that the code of all research software be made accessible

1.4. Require software to be citable via persistent identifiers (PIDs)

1.5. Why is this important?

1.5.1. Accessible software is essential for reproducibility and understanding

1.5.1.1. It allows experimentation, enabling a better understanding of the software's functionality and properties

1.5.1.2. It allows others to re-run the code and therefore reproduce what has been reported

1.5.2. Open source makes it easier build upon, iteratively improve and maintain software, and further improves reproducibility and understanding

1.5.2.1. It helps researchers understand better what the software does and the reasoning behind it

1.5.2.2. It allows the code to be modified and distributed by others than the original team, making it more maintainable and allowing improvements by the community

1.5.3. PIDs contribute to findability of software, and thus are the first step to enable reproducibility and reusability

2. R3: Encourage the research community’s ability to apply best practices for research software, including training in software development concepts

2.1. Create financial incentives for increasing expertise in and adherence to software development best practices

2.1.1. Create programs and funding opportunities encouraging researchers and research support professionals, e.g., research software engineers and data stewards, to adhere to software development best practices

2.1.2. Encourage the development of research assessment systems enabling recording and sharing rewards for or recognition of software outputs

2.2. Why is this important?

2.2.1. Better software, i.e., developed following best practices, conduces to better research

2.2.2. Software is an important part of data-driven research and therefore researchers need to acquire appropriate skills in this area in order to conduct high-quality research

3. Who are these recommendations for?

3.1. Policy makers and funders, so their decisions support and increase recognition of research software and its importance to advance science

3.2. Raising the Profile of Research Software

4. R1: Support the funding of development and maintenance of critical research software

4.1. Provide grants for all stages of the software development life cycle

4.1.1. Development of novel software

4.1.2. Extension of existing softwares' functionality

4.1.3. Maintenance of existing software

4.2. Why is this important?

4.2.1. Without this financial support new software cannot not be developed

4.2.2. Without financial support, neither extension nor maintenance of software can be performed

4.2.3. Research software that is not maintained will be lost forever, depriving research communities of valuable resources and reducing the ROI on the initial development effort

5. R4: Support recognition of the role of software in achieving research outcomes

5.1. Allocate funds specifically for research software development and improvement

5.1.1. Create specific allocation of funds for research software development as part of existing funding opportunities

5.1.2. Create funding opportunities for software supporting research software reproducibility, e.g., benchmarking and testing frameworks

5.2. Reward software outputs alongside publications, data and other research outputs

5.2.1. Encourage publications on software research and inclusion of developers as authors

5.2.2. Encourage the development of research assessment systems that specifically reward software outputs

5.3. Why is this important?

5.3.1. Many researchers are reluctant to spend time on producing good software because this is not incentivized