Research Software Engineers from the Geosciences assemble for the first time at EGU General Assembly 2018

Daniel Nüst, July 30, 2018


On April 12th 2018, the first Research Software Engineers (RSEs) for geosciences meeting was held at the European Geophysical Union (EGU) General Assembly (GA) in Vienna, Austria. The EGU GA is a huge event with over 15.000 people from more than 100 countries. It has a diverse programme with thousands of posters and hundreds of sessions, but what it lacked was an event to bring together scientists who contribute to research software. Daniel Nüst from the Institute for Geoinformatics, Germany, proposed the idea of such an event to a group of regular EGU GA attendees from the German RSE chapter. He was joined by Martin Hammitzsch (GFZ eScience Centre, Germany), Bernadette Fritzsch (AWI, Computing and Data Centre), and David Topping (University of Manchester) as co-conveners for a Townhall Meeting “Research Software Engineers in the Geosciences”. Townhall Meetings, or “townhalls”, are union-wide events. They allow participants to take part in open discussions covering a variety of topics. Townhalls take place in the evening after a full day of regular conference and poster sessions, so the motivation of people showing up is unquestionably high.

For the RSE townhall, it was difficult to stand out in the programme, it being a first time event at a large conference cutting across many domains and 24 divisions, from biogeosciences to seismology and paleontology, with over 600 sessions over the course of a whole week. But searching the submissions some weeks before the assembly for “software”, it seemed clear that software plays an important role for scientists attending EGU as it does of any researcher, so the conveners were hopeful to welcome a few people. And they did! About 40 people from 7 countries representing the wide range of EGU divisions joined the meeting. It was a good mix of seasoned RSE folks and early career scientists as well as senior researchers for whom the topic was relatively new.

The meeting kicked of with a short welcome by the initiator, Daniel, followed by Jens Klump (Science Leader Earth Science Informatics at CSIRO, Australia), Deputy President of EGU’s ESSI Division and advisory board member of the Australian and New Zealand chapter RSE-AUNZ. Both pointed out the relevance of contributions to science made by software and thereby the people contributing to that software in any way.

The motivations for national and international RSE activities were further detailed by three representatives from national chapters. David Topping from UK RSE, the oldest and largest RSE organisation, took a look at the definition of an RSE, at the community history, and its current state in the UK and beyond. Over 15 local groups already exist and more are forming at a high rate, sometimes even competing over members. Martin Hammitzsch presented the German chapter, de-RSE. Initiated only 1.5 years ago, he shared the group’s objectives, how they work to build a community, challenges they face and some lessons learned: a great resource for the attendees from countries without any organisational structure yet. Third up was Niels Drost (Netherlands eScience Center), who introduced the youngest European chapter NL-RSE and its core team, which already generated a considerable reach across the Netherlands.

After the short talks, we took advantage of the group size, had a complete round of short introductions, and enganged in a discussion. We found that one of the big challenges was to reach people engaged in Research Software Engineering (RSEng) activities at their work, especially outside of ESSI, and those for whom the “RSE” label does not ring a bell yet. The need for outreach also applies to holders of an office within the union, to reach better acknowledgements of the needs of RSEs in universities, scientific unions, and at scientific conferences.

Some good ideas came up and we plan to reach out to EGU and ESSI leadership and share them. For example, EGU could increase recognition of RSEs and their work by awarding a medal, by offering a special poster track for contributions to scientific software (allowing an additional “software submission” per author), or by tagging abstracts as “RSE” similarly to the “ECS” labels. These are ideas for “top-down” activites that we would like to advocate within the organisation. The usefulness and overall potential of such domain-specific (taking all EGU members as a group for a moment) actions, who are lateral to the national chapters, was commonly agreed on.

But there were also ideas for “bottom-up” activities which can support the RSE organisations’ causes, for example scientists organising sessions related to RSE roles and activities within their divisions, offering short courses specifically around RSEng capabilities (and also labeling the numerous existing courses as such), or organising software and data carpentry courses in the week before or after the conference. One participant suggested a session “Software development for professors - what do my students do with their computers?”, which seems funny at first, but at a second look goes to the heart of RSE outreach activities for raising awareness and teaching software-related skills. It was great to follow the lively discussions, which were enriched with a common understanding of the values and importance of diversity and openness. If you think about convening a session on scientific software yourself at an EGU GA, please get in touch!

An important aspect of structured RSE activities are surveys, because the people self-identifying as RSE are diverse and wide-spread and the roles that RSEs play in scientific research are manifold. We want to contribute to the process of understanding the needs and diversity of people involved in RSEng with the following survey on EGU attendees:

https://nuest.staff.ifgi.de/survey/index.php/598158?lang=en

The townhall meeting was a good start to spreading the word about the goals of RSE organisations and the activities to put Research Software Engineering on the map of all stakeholders in science, such as researchers, publishers, funding agencies, and scientific unions. What can we do better next year? Hopefully we can do what more “established” townhalls can offer: snacks and drinks! We should also announce and prominently place a sticker table. Apart from that, this first townhall did an excellent job, just like other long-running townhall meetings at EGU GA: It provided a place for like-minded people to connect and for newcomers to dip their toes into a new topic and be introduced to members of an international friendly community of dedicated people who “do science with code”.

The slides (download all slides here) include many links to further resources on the history and state of RSE-related activities. Please let us know what you think on Twitter: #egurse. See you at the RSE Townhall Meeting at EGU 2019!

RSE Townhall Conveners

Daniel Nüst
Martin Hammitzsch
David Topping
Bernadette Fritzsch