The search for academic software publication policies
Posted by j.laird
on 7 September 2022 - 10:00am
By Dr. Domhnall Carlin, Queen’s University Belfast (QUB).
When developing a Python course for university PhDs and Postdocs a while back, I wrote some slides on getting a publication for your research code. This created a nice little workflow from some toy code, through to getting a digital object identifier (DOI) and linking it to our institutional research output repository. Citable, versioned and credited code, just what every RSE gets up in the morning for. It occurred to me that I should probably check our institutional guidelines for publishing software, just to make sure I wasn’t training a cohort of researchers to break some fundamental rules. This has since led to a new avenue of research within my EPSRC RSE Fellowship, aiming to gauge the existence of and to develop such policies.
To make sure I was guiding the attendees at the training course along a university-approved path, I checked around with very experienced colleagues for any existing policies that might help. But they didn’t really know of any, and suggested I try another department, who did the same. This continued through to our expert open research staff, but there don't appear to be any such guidelines within QUB. So I checked on the RSE Slack with other university-based RSEs, but this again didn’t reveal much other than licensing recommendations.
We know that the publication of research code is a good, and essential, thing for reproducible science. Indeed, Jisc1 make the following recommendations:
“Treat computer code like any other output of your research... Share your computer code like you would any other research output...Computer code should have a URL or a DOI (digital object identifier)... Always include these when citing the code, including information on the version you used.”
Yet, while this research is in a very early phase, there doesn’t appear to be much in the form of guidance - Academic Software Publication Policies (ASPP) - for researchers to publish their code.
So why is this the case? Traditionally, software has been seen as an artefact of a ‘proper’ academic output (i.e. a conference or journal paper), rather than a ‘first-order’ output in its own right. When software is released, it is very often tied up with the research paper it supports, or is written as a specialised journal paper in its own right. The output then becomes the paper again. This may be a legacy artefact of journal papers being the gold of academic currency. However, some institutions have moved to include software as a first-order output with their research repositories, recognising the contribution and potential benefits of publishing research code.
In order to assess these questions, we have launched a survey (only takes 3 minutes) for anyone who writes research code in a UK-based academic institution. The purpose of the study is to assess whether any UK academic institutions have defined policies for the publication of research software and what they recommend. It will also solicit the opinions of those who write research software regarding the potential benefits of such policies to their work.
The main objectives are:
To survey RSEs at UK Academic Institutions on the availability of institutional academic software publication policies.
To survey RSEs on their attitudes to the benefits of have clear policies and the impact on their willingness to publish software.
To investigate any differences in attitudes, opinions and policy-availability on information provided.
The hopeful outcomes of this will be the promotion of software publication policies within research institutions, leading to increased publication of research software and credit for those who have developed it. Other facets of the research include polling university research output repositories for the presence of software and examining whether the infrastructure is in place for this to be effective.