The Software Sustainability Institute is ten years old this month!
Director Neil Chue Hong and Deputy Director Simon Hettrick took some time out from worrying about Covid-19 to log onto Zoom from home, and looked back ten years to the start of the Institute.
Congratulations! Ten years is impressive. Can you cast your mind back a decade and tell me how it all began?
Neil: The impetus came out of the e-Science programme. In the 2000s we had a lot of people from academia and industry working on middleware projects for science. They were large scale projects, with a focus on things like industry best practice and robustness. And I think out of that there came a realisation that you can’t take industry practices wholesale and apply them to researchers who don’t have the same background.
People were aware that better software was a good thing to do – but we weren’t sure how to do it. Then EPSRC ran a survey asking people what their experience had been of things like the e-Science programme, and it turned out a key concern was software quality. So that led to a call out to the community to propose what they felt was needed to support software development.
And that was the call that we bid into – ‘we’ being the existing partnership of Edinburgh, Southampton and Manchester Universities. We’d come together to run the Open Middleware Infrastructure Institute, OMII-UK, but we pitched this differently: if OMII-UK was supporting the specific sector of middleware developed by large teams, what we were doing was flipping it round and focusing on the researcher level, at end user applications across a very large set of domains.
Simon: The e-Science programme was an eye opener. Coming in as a domain scientist, the thing that instantly hit you was that it was completely interdisciplinary, people came from lots of different backgrounds and domains, and the thing joining them together was software. And we began to see that it was software in general that was vital – not software for a particular discipline, just software itself.
Neil: Yeah, I think that’s key - it was clear that software was the glue that held together research, but it was also becoming increasingly clear that the way it was being supported was wrong. There was a trend for creating generic tools, which was looking at things the wrong way round. Working with different researchers and different domains, we realised that the culture of the way people do research is very important, and you have to understand their requirements. There are very few tools that are truly generic. But there are many ways of doing software development so that you can reuse a lot of the components. So big projects aiming to create software for everyone got switched for a cooperative, codesign approach.
That had been tried in a smaller way, but no one had tried to do it on a large scale until this call came out for a national institute.
Where were you at this time, Simon?
Simon: I was there from the start, though I wasn’t part of the bid writing team – I wasn’t senior enough at the time. I was communications lead for the first three years, until I became the deputy director in 2013. I was always there, though, being a colossal pain for Neil, asking questions and making suggestions and stuff! So when things went wrong I would say ‘we should do this…’ and I guess Neil agreed with me often enough that he decided to give me the promotion.
Neil: Well, also, Simon’s a better ‘enforcer’!
Simon: He actually used to call me that: “The Enforcer”! I quite liked it.
How did those first few years compare to the Institute today?
Neil: There was a different feel to it, and we were still finding our feet. There were two notable things about the focus in the first year and a half. The first was that it was very much structured around doing projects with people, working in collaboration with research groups and doing direct consultancy. The second thing was that we didn’t do any training at that point. We’d deliberately left training out of the proposal, making it clear that it wasn’t because training wasn’t important, but because we felt there wasn’t enough funding and effort available to do it properly. We were very much focused on identifying key projects to work with, and using the experience of working on those as the basis to create guides that we could then publicise.
Simon: There was a change of philosophy with the Institute. Back in the OMII-UK days it was all about ‘we’ll take the software off the researcher who doesn’t know how to deal with it, and we’ll do the work because we are the software engineers’. Neil and the team knew that wasn’t working, and they wanted to try a different approach, by keeping the software with the researchers and providing them with advice so that they could engineer better software themselves. But it’s not easy to make that philosophical leap in one project! So the idea of the software engineering projects stayed with the Institute for a while.
In the end, it’s not scalable. You can only work with a handful of projects and there are, what, 210,000 researchers out there? But those early projects gave us credibility and we evolved them as the Institute developed.
Neil: That evolution worked because we learned that a lot of research software is driven by one person who has the idea and who wants to keep control over the direction it goes in. They’re happy to accept external help, recognising that they aren’t expert in parallelisation, say, or setting up a test suite, but after that they want to continue developing the software themselves, or in their group. Recognising that need to keep control was a key difference between OMII-UK and the Institute: it became about instilling better practices in the people themselves, and recognising that they were the only ones who could change how they were doing things.
How did you start to broaden what you could offer?
Neil: Around 2012, we were looking at how we could get a better grasp of what was happening in different communities. We couldn’t go and talk to everyone. We came up with an idea that was modelled on Microsoft’s valued professional programme, where you look for people in each community who are passionate about what they do and willing to help people, and then you recognise them for that work. We looked for early career people, mostly, who were interested in the things we were trying to achieve, and would be willing to act as our eyes and ears in their community; originally they were called our “Agents”. That’s what turned into the Fellowship programme, which is one of the most successful things we’ve done. It started as a way for us to get more information on the changes in different communities, and it’s turned into a network of people actively supporting each other. They do things on their own, independently of us, but still furthering our aims. And increasingly, are ending up in positions of power and influence!
Simon: There’s another aspect to the Fellows programme too. When we first went out to events, talking about the importance of software, there was a lot of pushback. “How can one institute possibly deal with all software? My software is different from their software!” – we were fighting against a hell of a lot of scepticism. People thought we were trying to come up with a set of guidelines that must be adhered to without deviation. What we were actually saying was the opposite: we wanted to work with communities and then adapt the guidance based on their requirements.
It turned out one of the best ways to spread this idea was to have domain ambassadors: people within the domains who would also say that software was important. And that’s where the Fellows really came in, because people’s scepticism was immediately overcome when they heard the message from ‘one of them’.
The same period must have seen the start of the RSE campaign?
Simon: Yes, that started in 2013, and I’m incredibly proud of the Institute’s role in starting it. Its success has been down to the efforts of a vast number of people in the community, and it’s been amazing to see that community develop from a small grassroots movement to the tens of thousands of people who now identify as Research Software Engineers. It’s got a very active life of its own nowadays, with new things happening in countries across the world, and I think that’s the real mark of success.
Neil: The Carpentries also started around that time. If Greg Wilson hadn’t emailed me asking for funding I probably wouldn’t have heard of Software Carpentry, we wouldn’t have given our support, and I think that helped grow The Carpentries a lot at that time. We took the decision to support the running of lots of workshops in the UK because it was the right thing to do. It enabled us to finally address the training needs we’d cut out the proposal, by building on someone else’s work.
Simon: It would have been so easy for us to just have developed our own materials and pushed them instead, but instead we took the risk of pushing someone else’s material and brand, and it worked really well.
It sounds like a lot of decisions were being made on the run, as you went along?
Simon: The communications team’s unofficial slogan was ‘beg, borrow and steal’, because a lot of the ideas we came up with, things that worked out pretty well, were either free or very cheap! The Better Software Better Research slogan, for example, was something Mike Jackson came up with and I designed the t-shirts myself.
Neil: And then there’s the stickers! Having something people can rally around is really important. I got accused by the head of computing at a large research organisation that all we did was give out bloody stickers! And I said, ‘I’m happy with that! If people know us, even for just having given out lots of stickers, then it shows we’re doing a good job of getting our message across.’
It’s that weird thing where it becomes a sort of clan badge. We were lucky to come up with something short and snappy, and that people agreed with. They don’t necessarily have to agree with us to agree with the slogan.
The community is often seen as one of the Institute’s greatest strengths. Did you set out with an open, inclusive community as a goal?
Neil: Yes and no. Yes, in that personally I’ve had bad experiences in the past with research being very cliquey and ego-driven and I was keen not to have that. No, because I think a lot of it has been sheer luck in getting the right people into the community and getting them to stay engaged. And not fall out! I’m surprised we’ve not had more fallings out. I mean, there are some – people argue over what editor to use! There’s a whole spectrum of people. But we’ve managed to nurture a friendly, supportive community.
Simon: Up to the point where we started, the whole philosophy of research seemed to be that you carve out a territory and then you fix gun emplacements and prevent anyone else from moving onto it – because it’s yours! And if you allow anyone to know your trade secrets or your operating methods then all you’re doing is reducing your own potential for success.
From the get-go, the Institute was all about collaboration over competition. We were determined to be seen as honest, as doing the best job we could to advance research without hiding anything. Everything was open from very early on; meeting minutes, all the processes. We’d blog about the things we were thinking about, get feedback – and then blog about whether they worked or not. That was a real sea change to the environment that we came from.
There were dissenting voices, saying that if you’re open about your processes then people will come in and do it better. But if they do, then the fault’s on us for not doing it well enough in the first place! If we’re doing everything openly and collaboratively, and we’re good at it, all we’re going to do is amplify our success.
Your events have also been a success. How did you go about that – did you plan it from the start?
Neil: We did have a planned strategy, although it’s not quite turned out exactly as envisaged. I think the plan was quite structured and then what’s happened has been more organic. We’ve come to understand different communities better, and how their levels of maturity vary. And we also realised that we needed events that brought together people who were interested in a common challenge, but not in the same areas.
Time to tell all: What are the secrets of your success?
Neil: I think the thing that can’t be underestimated in terms of the success of what we’ve done, is we were actually quite good at giving away money. A lot of our success has been by getting funding and then directing it to the right people!
Simon: Often people would tell us ‘oh, this is what the community needs’, and you’d scratch the surface and find that it was just their opinion. One of the strengths of the Institute is its philosophy of evidence-led policy making. That this is a good idea shouldn’t really be a revelation to anybody! But... you could argue that too much research policy is led by just listening to the silverbacks who have strong opinions. We got a lot of important people telling us what we should do – but we also went out and talked to the community and saw whether they thought something was a good idea, or whether it would work in their area. We did surveys to collect evidence where it was lacking, so that we could direct our efforts better.
Neil: I scared a lot of members of staff at a meeting about five years ago, when I said ‘I think our goal should be to go into an area, making people aware of it’s importance so that funding for that area increases, and then move out and move onto something else.” And my colleagues asked why not just stay in that area and keep doing it! But I think we should be improving things for everyone, and sometimes that means you have to withdraw, so other people can get a chance. When there’s limited resources, if you don’t step back then no one else can advance to the level you want them to be at.
Simon: You need fresh blood coming in, or you get stale. We’re good at going in, sorting out a new community or a new practice, and then moving on to something else. We’re good at evolving, and we’re good at working out the most scalable, high impact actions we can do at any particular time.
Oh, and on that note, if you’re going to name an Institute, never put Sustainability in the name, unless you want endless jokes about how sustainable you’re going to be…
How have you changed, yourselves?
Simon: I didn’t know anything about software when I started work for OMII-UK. I remember desperately trying to read up on grid software before the interview! So I’ve had quite an arc in learning about stuff I never knew anything about – it opens up this whole realm of knowledge that was completely closed before. It’s been interesting. And I never thought I would end up an academic! The idea was to stay in University for a few years until I worked out what I wanted to do with my life. It looks like I’m here for good now!
Neil: I think what’s changed is probably everything around me, and that’s been really nice to see. It sort of feels like I fit into the rest of the world better than where I was 10 years ago.
Simon: Neil was incredibly young when he took on the directorship of OMII-UK. It was this behemoth of a project, spread across three universities, and there was a lot of talk when the previous director stepped down – who would take his place? And when Neil got it, everyone was genuinely very happy. He was a good leader, and had this ability to bring everyone with him. Now he’s on every single advisory board going and listened to by people all round the world.
Neil: Back then I found my age to be a good cloak of invisibility because I could go into meetings and people would think I was just a postdoc researcher and I would hear a lot about what was really happening behind the scenes. And now I’m going to meetings and giving sage advice, the experienced voice – and I don’t consider myself that old yet! But actually, when I look at the community now, I am one of the senior members. And the people who are doing the really exciting stuff are in their late 20s, early 30s, just as I was back then.
What do you expect to see over the next five, ten years?
Neil: One thing that I’m actually quite glad that has come from this Covid-19 pandemic is an understanding of the need to share. For a few years we’ve been pushing two things: that you can’t do research without sharing stuff, and that you don’t always know what’s going to be important until you hit a crisis. And that really emphasises what we’ve been saying, that the job of Institutes like ours should be to make sure that everyone is 5% better rather than a few people 50% better. I’m hoping that in five years time, we will have managed to make the community aware of why it’s important to share their software and they know what to do to make it 5% better.
Simon: What I’d like to see, and I think things are moving in this direction, is different metrics being used to measure success in research. Obviously, we’re arguing for contributions to software as our main effort, but we’re joining with others and pushing for change in lots of directions. Hopefully as a community we can shift to judging people on the impact they have achieved, rather than the impact of the paper that they happened to be named on. A world that does that, is a world that will produce far better research.
Well, happy anniversary, and here's to the next ten years!