Research Computing in Pure
I often encounter researchers at the University who have not heard of Research Computing. Our team would like to improve our profile by ensuring our contributions to research are visible in Pure. Pure is the University’s research information system which brings together key information on all aspects of research at St Andrews, including research outputs, and links these to the people involved. These links are what will give our work prominence, connecting us through our work to the researchers and their projects.
I recently saw some output from an LLM that had been fed data from my Pure profile. It thought that I had a particular interest in researching commercial fishing. While LLMs (Chat-GPT, Microsoft Copilot etc.) can be prone to hallucinations where they invent and assert things that are not true, this was in fact a reasonable inference for it to make, albeit drawing on a very limited dataset. A human viewing my profile would reasonably come to the same conclusion.
I’ve been named as a co-author of several publications about fishing because I have written software, both mobile apps and web applications, that have been used by researchers to collect data from fishers. In some of the papers, I contributed a small section on the software I created, but for others I was not involved in the writing at all and my name was included because of my contribution to the project.
With the exception of a talk I once gave about Research Computing at a conference, everything else in my Pure profile is related to the Carpentries, mostly teaching workshops. I’ve worked with researchers from most of the University’s Schools, but my Pure profile (at the time of writing) makes me look like a biologist, or possibly a geographer (albeit one who is a member of the Institute of Behavioural and Neural Sciences).
My very partial Pure profile is a problem. A historian, for example, might get the impression that I specialise in biology projects and that I wouldn’t be able to work with them on a project they’re planning – which is not true.
The fundamental reason that our work isn’t properly represented in Pure is that we haven’t engaged with Pure as we might have. That we are seeking to change this reflects two developing ideas in research culture:
- Research outputs are diverse, extending beyond scholarly publications
- It is important to recognise the input of all who contribute to research
What is a Research Output?
When thinking about the outputs of academic research, the first things that come to mind will be things like journal articles, conference papers and books, i.e. scholarly publications. However, there is growing recognition that research outputs are more diverse than this.
The REF 2029 Guidance includes a section on “Diverse outputs” which starts:
5.6.1. Research involves a diverse set of research practices, which lead to a wide range of diverse outputs.
It goes on to note that such outputs may include a range of formats and modes of expression which “may depart from historical conventions” (5.6.3). Examples given include “non-textual (artefacts, audio, visual, multi-modal)”, “software and code”, “datasets” and “multi-component outputs” (5.6.4). Importantly, it goes on to state:
5.6.7 An underpinning principle of the REF is that all forms of research output and research practice will be assessed on a fair and equal basis.
The Hidden REF project has a mission to:
…achieve mainstream recognition for [non-traditional outputs] and hidden roles, because this will make the conduct of research more equitable and more effective.
The categories for its 2024 competition included “Research infrastructure”, “Design”, “Software”, “Website Content”, “Digital or Visual Media”, “Research Datasets and Databases” and even “Everything else!”.
Most of what Research Computing produces can be considered as research outputs. Pure has a particular record type for software, but we most often develop web applications to run on our own servers which aren’t intended to be considered as reusable, open-source software. However, the definitions above clearly cover the databases, interactive applications and websites which are the staples of our work.
Authorship and Recognising Contributions
This skew in my Pure profile partly reflects the nature of the work I’ve done – for most other projects, my work has enabled the public to access and interact with research data and other outputs rather than having a role in answering the research question. However, it also highlights the differing attitudes to authorship across disciplines.
There is a growing push to make sure that all who contribute to research get due recognition – not just research software engineers like us, but librarians, technicians and others. Co-authorship is only one way of recognising contributions, and it will not be appropriate in all cases. However, it does not currently seem possible to represent an “acknowledgement” or similar within Pure.
The cultural change required to ensure that all contributions are recognised and the modifications to Pure that may be necessary make this a much harder problem to tackle than highlighting non-traditional research outputs. It’s also something that can only be addressed for future outputs and not retrospectively.
The University has constituted a Short Life Working Group on Recognising Contributions to Research which will be looking at this.
Next Steps
While we wait on the SLWG to report, there are steps we can take to start to improve the visibility of our work in Pure
- We can go ahead and add records to Pure for the outputs – web applications, project websites, databases etc. – which we have previously created or substantially contributed to.
- We will seek to link those outputs to the research projects they’re associated with and the researchers with whom we worked on them. We will do this in consultation with the researchers so that they do not see unexpected changes to their Pure profile.
- Where there seems to be a clear case for co-authorship of future “traditional” research outputs, we can engage with the researchers on this issue.
When the SLWG does report, it should provide a firmer footing for these conversations.
This is going to take some time – we’re working on lots of other things – but we’re getting started.