The university is home to a centralized group of Research Software Engineers, or RSEs, who lend their expertise to a variety of research projects across campus.
Since its formation in 2018, the US Research Software Engineer Association has grown to nearly 1,000 members and is rapidly transforming scholarly research in the United States. But it has yet to host an in-person event.
That's about to change, as the US Research Software Engineer Association (US-RSE) is set to hold its first face-to-face workshop on April 26th and 27th.
Held at Princeton University and sponsored by the Alfred P. Sloan Foundation, the US-RSE Community Building Workshop brought together research software engineers from universities, industry, and research laboratories across the country, plus a few more from abroad, to chart a path forward for the swiftly expanding group.
"I'm just so excited to actually have this happen," says Ian Cosden, the director of Princeton’s Research Software Engineering group and the chair of US-RSE's steering committee. Cosden says that the gathering was originally scheduled for April 2020, but was canceled because of the COVID-19 pandemic. Instead of turning the workshop into a virtual meeting, the workshop organizing committee postponed it three more times, holding out for a moment when they could harness, in Cosden's words, "the synergistic, whole-is-greater-than-the-sum-of-its-parts effect that happens when people come together in person."
"When we have virtual meetings, it's hard to do anything that's concentrated," says Daniel S. Katz, Chief Scientist at the National Center for Supercomputing Applications and one of US-RSE's founding members. "You need the social interaction, eating together, going for walks together. I think that's a necessary part of having this kind of meeting work. We need that level of comfort."
From Slack channel to professional association
US-RSE is one of a number of professional organizations around the world for research software engineers, or RSEs. The first group, now called the Society of Research Software Engineering, began in 2013 in the United Kingdom with the aim of creating a defined career path for software developers working in academia.
Many research projects take an ad-hoc approach to software, relying on graduate students with little or no formal training in software development. These roles tend to be transient, and much of the code they write isn't reusable once the research project ends.
The RSE movement hopes to change that, by encouraging researchers to involve professionals with up-to-date knowledge of the software industry's best practices, fluency with the research topic at hand, comfort in a research environment, and the ability to build reliable software tools that can be repurposed and maintained beyond the scope of any particular project.
Katz emphasizes the value that an RSE's input can add when it is included at a project's very beginning.
"When you ask people how to do something, they limit their answers to what they think is possible," he says. "The RSE brings in is a different set of assumptions that builds a larger potential solution space."
In 2016, the UK group held its first general conference and opened it to research software specialists from overseas.
In the three years that followed, similar organizations sprung up in Germany, the Netherlands, the Nordic countries, Australia and New Zealand, Belgium, and the United States.
Cosden points to two events that led to the creation of the US group. The first was an international survey created in 2017 by the UK group that sought information on the state of research software engineering in other countries. Katz volunteered to customize the US version of the survey, and added a question asking if respondents would be interested in organizing an RSE community in the United States. Ten people volunteered.
The other event was a January 2018 international gathering in London hosted by the UK RSE Association. Cosden and four other research software engineering leaders attended, and they returned from the meeting energized.
"We saw the clear benefits of this level of commitment in the UK, and we thought we had to have this in the US," Cosden says.
Within days of returning to the United States, the volunteers had created a Slack channel and a bare-bones website. But the nascent group's growth remained minimal.
"We had the idea of a community," says Katz, "but we didn't have the tools for a community."
That began to change in 2019 following the creation of a mailing list and a more complete website. By that summer, the group had about 100 members. They approached the Sloan Foundation to discuss funding an in-person meeting for the following April.
"We found that there was some interest, so we decided to throw gasoline on the fire," Cosden says.
The Sloan Foundation, a philanthropic nonprofit whose stated mission is "to make the world a better place through the advancement of scientific knowledge," agreed to support the workshop.
“Research software engineers are essential for creating useful and durable software in science, which is why we’re delighted to support US-RSE,” says Joshua Greenberg, director of the Sloan Foundation's technology program. “The upcoming workshop marks a fantastic opportunity to engage and strengthen the professional network of engineers—and we’re excited to see what emerges from these thoughtful and fruitful discussions.”
Like most gatherings scheduled for April 2020, that workshop didn't happen. But the pandemic didn't stop the group from growing nearly tenfold over the next three years.
Getting name recognition
As US-RSE's membership steadily increased, the term "research software engineer" has begun to gain currency among members of the scientific community, particularly in universities. Cosden and other members of the RSE movement see this development as critical for forging a group identity.
When Cosden was establishing his team at Princeton in 2016, nobody knew what to call its members. "We had a posting for 'computational research application analysts,'" he says. "Over the course of the next two years it went from 'the programmers,' or 'the ninjas,' or 'Ian's group,' to finally, ‘the RSE group’" he says.
Even without a name for its members, Princeton's centralized team offered a new paradigm for collaboration between researchers and software engineers.
"There aren't a lot of other people whose job it is to lead a central group of RSEs, though it is growing," Cosden says. A handful of other universities, including Notre Dame, Harvard, MIT, and the University of Washington, have also launched centralized RSE groups in recent years. The National Center for Supercomputing Applications is home to about 40 RSEs, but are supported by externally funded projects.
"Princeton is very good at listening to faculty requests," says Cosden. "And the overwhelmingly positive feedback about the RSE program to the central administration has been game changing."