A farewell to RSEs: reflecting on our progress and remaining barriers
By SSI Deputy Director Simon Hettrick.
Last September, I stood down as a trustee of the Society of Research Software Engineering. It was a huge decision for me. I was the first Chair of the UK RSE Association in 2013 and I had been a committee member of the Association and the Society ever since. From the very beginning, I knew that success for me would be marked by the day that I could step down. That this has taken seven years says something about the difficulty of setting up a sustainable organisation and, without a doubt, how difficult it is to leave an organisation that I feel so passionate about. Nostalgia aside, this moment seems like a good time to reflect on the challenges we still face.
The history of the RSE campaign is detailed in other posts and in innumerable talks, but a quick summary is necessary. We’re winning the war for recognition. Far more people are aware of RSEs and understand why they are vital to modern research. The growth of RSE Groups, which now exist at 29 universities and research organisations across the UK, demonstrates the rapidly growing awareness of research software engineering. I have found researchers to be wholly supportive of our community: they’re not just willing but eager to hand over the development of software to professionals who understand both software engineering and research. As groups of RSEs have grown in size they have, almost accidentally, created the hierarchy of positions needed to produce an RSE career path. Financial support for RSEs has grown from something experimental that only the EPSRC would dare provide, to something approaching the norm across a range of funders. The RSE concept has travelled incredibly well, from Australia to the US, across Europe, and now into South America. The RSE Community is a truly global phenomenon.
In summary, the prospects for an RSE in 2021 are far brighter than we could have dreamed of when we started in 2013, and this success is entirely attributable to the hard work and tenacity of our community and the people who run our events and our organisations. Yet our position as first-class research citizens is still not universally accepted. In my opinion, most stakeholders have accepted our arguments, and we are now held back only by the glacial pace at which academic policy changes. We must keep up the pressure of our campaign if we are to keep the turgid juggernaut of academic bureaucracy moving in the right direction. While we wait for the universities to catch up, there are issues we need to resolve, sometimes on our own, and sometimes in partnership with the wider research community.
We have too great a reliance on volunteers. Think of an RSE activity and you will find its existence is reliant on volunteers. The people who run the RSE conference, a significant amount of the training we conduct, the trustees of the Society, and almost everything else we do. It is amazing and heartening that people are willing to donate their time for the benefit of the community, but it’s a sticking plaster on a problem that needs a more fundamental solution.
The volunteer effort of hundreds of RSEs has hugely improved the research environment in this country, which means the research community has an obligation to pay for it. We cannot expect an end to volunteering, but we can expect institutions and funders to pay for the time that RSEs invest into running training, creating communities and, ultimately, solving problems for research. RSE activities that help train new RSEs must also be supported because, without them, the colossal demand for research software engineering will soon deplete our relatively small community.
Currently, RSEs who work in RSE Groups are more likely to have their role recognised and have a greater chance of continued employment than RSEs who are embedded in traditional research groups. Numbers are partially to blame: it’s harder to ignore a large group of RSEs who work as an independent unit than it is to overlook a lone RSE working in a sea of traditional researchers. I have seen this problem used to criticise the RSE campaign, but I do not agree with this logic. We may have had more success in gaining recognition for one subset of RSEs than others, but this is not a failure. It is a step towards the greater success of gaining recognition for all RSEs.
With RSE Groups flourishing, now is the time to focus more effort on fighting for consistent rights for RSEs wherever they are based. The first port of call is job descriptions. There are far too many people conducting research software engineering under the wrong job title who are then doomed to be judged by irrelevant success criteria. We need to encourage more people to ask for their job descriptions to be changed to reflect their work and support them in doing so. The RSE career path must be accepted by all UK universities and research organisations, and university policies must change to ensure that RSEs receive the same employment benefits and have the same expectations of career progress as researchers. This change entails fighting for the recognition of software as a first-class research output, and ensuring that it is this metric that is used to judge RSE career progression. This has been a campaign issue for the Software Sustainability Institute since its foundation, and is also a focus of my latest campaign, the Hidden REF.
The research world has woken up to the importance of research software engineering, but its appetite is quickly draining the pool of experienced RSEs. What’s more, RSE Groups are finding it difficult to recruit people to lead them, because RSE Group leaders are even rarer than RSEs. To date, RSEs tended to learn their profession through experience of writing software as part of a research career. Although this provides great experience, it’s not an efficient way to train the thousands of RSEs the research community will soon demand. It is time to invest into the training of RSEs, definitely at Masters and PhD level, and possibly at undergraduate too. There have been some excellent examples of the type of courses that are needed, such as those taught to UCL Masters students and at Oxford’s SABS R3 CDT. We will not see significant expansion of these courses until funding is made available to support RSEs in creating and teaching them.
There is an abundance of demand, and hence funding, for research software engineering, but it is almost exclusively focussed on RSEs who spend their time developing software. The same cannot be said for the other activities that are fundamental to the creation of an environment in which RSEs can work. The result is that many RSE Group Leaders engage in a form of quasi-volunteering: whilst they are paid a salary, it does not support the time they must invest into running their group. It’s a painfully obvious thing to write, but people and project management are not optional extras. The groups that have best addressed this issue are those based in professional services where it is understood that managers must manage. This indicates the solution: like any other senior manager, the management of RSE Groups should be paid for by central university funds.
The RSE campaign had one aim: to ensure that RSEs were treated like any researcher. We have come far in achieving that goal. Indeed, recently I’ve found myself in arguments where I have been told that our demands would lead to RSEs being treated better than researchers. I find this race-to-the-bottom argument entirely fallacious. RSEs should fight for the career they deserve without reference to anyone. We should not apply the brakes if our goals turn out to be greater than what researchers expect. Instead, researchers should join with us and campaign to improve the situation for everyone employed in academia.
I’ve come to the end of my tenure and to the end of this post so I shall allow myself a little bit of pride. When I was first putting together ideas for the RSE campaign, I was told by some senior academics that, although they understood the inherent unfairness of the RSE situation, academia would not or could not change to improve it. I persisted because I knew they were wrong. They didn’t understand the power of a community working together to achieve a shared goal. This power has only grown with our numbers, which leads me to believe that our community is now unstoppable. You may think that the main beneficiary of this progress would be the RSEs themselves, but it is not. The main benefit of a recognised and supported community of RSEs will be felt by the entire research community and their much advanced ability to conduct valuable research.
I have enjoyed every second of my time working with the RSE community, and I will continue to campaign for RSEs in the future, just not as a trustee of the Society. Good luck to the new committee and to the many generations of new committees that are to come.