Advancing research software through global collaboration: Highlights from IRSC26

Oct 13, 2026

This post is being cross-posted by Better Scientific Software (BSSw) and the Netherlands eScience Center.

Introduction

Research software sits behind an increasing share of modern research, but the systems needed to fund, maintain, recognise, govern, and develop it have not kept pace with its importance. Generative AI is adding new urgency: as producing code and other research outputs becomes easier, questions of trust, verification, provenance, maintenance, and long-term stewardship become even more important. Addressing these challenges requires coordination not only among research software engineers, but across the funders, publishers, policymakers, infrastructure providers, trainers, and research organisations that shape the wider research ecosystem.

These interconnected challenges provided the backdrop to the inaugural International Research Software Conference (IRSC). Across two days, discussions ranged from AI challenges, sustainable funding, and software policy to publishing, metascience, skills and training, and international community building. By creating a dedicated space for these communities to connect, exchange practices, and address shared challenges, IRSC advanced dialogue on research software and its intersections with funding, publishing, policy, metascience, skills and training, and generative AI, while building on work underway through ReSA’s forums, task forces, and other international collaborations. A recurring message was that many of the challenges facing research software cannot be solved by individual projects or organisations: they require stronger coordination across the international research software community.

Co-located with RSECon26 in September 2026 in the UK, IRSC brought together more than 110 in-person and remote participants from across the international research software ecosystem. The program featured 20 talks, 10 lightning talks, 10 Birds of a Feather (BoF) sessions, and two plenary discussions, along with keynotes from Riva Quiroga (RSE Chile), What We Talk About When We Talk About RSE, and Stefano Zacchiroli (Software Heritage), What Public Code Can Teach Us: Research Software meets Software Research.

IRSC highlights

Funding

A BoF on Evaluating research software funding, led by Neil Chue Hong (Institute for Research Software and University of Edinburgh) and Dario Taraborelli (Renaissance Philanthropy), provided an opportunity to explore approaches to understanding the effectiveness and impact of research software funding.

The conversation continued on the second day, with talks exploring lessons from a new multi-donor fund, Open Source for Science Fund (OS4Science), and approaches to improving how research software funding is allocated.

These discussions connect closely with the Research Software Funders Forum, which brings together funding organisations to share practices, identify common challenges, and explore opportunities for greater coordination around research software funding.

Publishing

Lauren Cadwallader (PLOS) presented Beyond Computational Biology: Making Code Sharing Policies Work Across Research Fields, exploring how approaches to code-sharing policies can be applied across research disciplines.

This was followed by a BoF on Separating out data and code in open science policies, which introduced a new Working Group (WG) within the Publishing Research Software (PubSoft) Forum focused on improving publisher policies for sharing research software and code. The discussion also connected this work with related initiatives across the research software ecosystem. Jeremy Cohen (Imperial College London) shared insights from the joint ReSA and Research Data Alliance (RDA) WG on Policies in Research Organisations for Research Software (PRO4RS), while Neil Chue Hong brought perspectives from discussions with research funders on code policy.

The new PubSoft WG complements other work underway through the forum, including a WG focused on peer review of code.

Policy

The BoF Committing to Sustainable Research Software with ADORE, led by Lieke de Boer (Netherlands Organisation for Health Research and Development - ZonMw), Marta Teperek (Dutch Research Council - NWO), and Carlos Martinez-Ortiz (Netherlands eScience Center), explored how organisations are putting the recommendations of the Amsterdam Declaration on Funding Research Software Sustainability (ADORE) into practice, the practical and strategic benefits of signing the declaration, and opportunities to support wider adoption.

Examples from signatories, including the Digital Research Alliance of Canada, ELIXIR, and the German Research Foundation (DFG), demonstrated the different ways ADORE can support institutional change.

Other sessions highlighted complementary approaches to supporting sustainable research software. In a lightning talk, Ishwar Chandramouliswaran (U.S. National Institutes of Health - NIH) highlighted how the NIH Office of Data Science Strategy is supporting sustainable, open, and interoperable research software ecosystems through investments in best practices, workforce development, infrastructure, standards, and partnerships. While NIH does not currently have a single standalone research software policy, its programs and initiatives collectively promote FAIR software practices, open-source dissemination, sustainable maintenance, and collaborative development.

IRSC also highlighted efforts to translate shared principles into practical action across the research ecosystem. The BoF From Principles to Practice: Driving Adoption of FAIR Research Software Across the Ecosystem introduced practical, step-by-step guidelines developed by ReSA’s Actionable FAIR4RS task force to help translate the FAIR Principles for Research Software into practice.

These discussions reflect ReSA’s broader work to strengthen research software policy internationally. The Research Software Policy (RSP) Forum provides an ongoing space for this work, bringing together policymakers to share practices, address common challenges, and support greater coordination in policy development, implementation, and alignment.

Metascience

The Research Software Metascience BoF, led by Wilhelm Hasselbring (Kiel University), Daniel S. Katz (University of Illinois Urbana-Champaign), Rena Bakhshi (Netherlands Organisation for Applied Scientific Research - TNO), and Simon Hettrick (University of Southampton), brought participants together to consider research software through a metascience lens.

The session built on conversations taking place through the Metascience and Research Software (MaRS) Forum, which have resulted in a paper on Metascience, Research Software, and the Case for Research Software Metascience. The forum is a collaboration of metascience groups, communities, and organisations working to better understand the implications of research software for the research ecosystem and research process. The MaRS Forum provides a space to identify gaps and opportunities in research software metascience, strengthen research software as an area of research, and increase awareness of research examining the role and impact of software across the research ecosystem.

Skills and training

International perspectives on sustainability challenges and opportunities in research software training networks, a BoF led by Eva Lekkerkerker and Fenne Riemslagh (Netherlands eScience Center), Jeremy Cohen (Imperial College London), and Danny Garside (Digital Research Academy) explored how research software training networks can become more sustainable and scalable, particularly as they remain heavily dependent on individual trainers whose contributions are often under-recognised and difficult to sustain.

Samantha Wittke (CSC - IT Center for Science) shared a practical example through CodeRefinery’s “bring your own classroom” model, which combines shared online teaching with local, community-led support. Extending the discussion to wider support for RSEs, Mike Simpson (Society of Research Software Engineering) highlighted the Society’s work to build community, strengthen recognition and career development for RSEs, and create opportunities for greater involvement in the RSE movement.

These themes are central to the Skills and Training for Research Software (STaRS) Forum, which connects organisations and initiatives working to strengthen research software skills and training globally through mapping of current offerings, sharing of best practices in curriculum design, approaches to demonstrating need and impact, and integration into broader research/open science/open source software training initiatives.

Generative AI and research software

Generative AI (GenAI) was a prominent theme throughout IRSC, with sessions exploring its implications for how research software is developed, maintained, evaluated, and sustained. In her welcome remarks, ReSA Director Michelle Barker highlighted how GenAI is amplifying existing challenges for the research software community, particularly around trust, verification, maintenance, and stewardship. As creating research outputs becomes easier and cheaper, greater investment is needed in the infrastructure and expertise required to verify and sustain them, as well as the shared knowledge, software, and communities on which AI depends. Rather than simply consuming this shared commons, the research community must also sustain and reproduce it, making the impact of GenAI not only a technological challenge, but a collective-action problem.

Several talks and sessions built on discussions from ReSA and Schmidt Sciences’ March 2026 Edinburgh workshop on Research Software Engineering in the Age of Generative AI: Building a Community Vision. The BoF Research Software Engineering: Evolving Competencies, Practices and Interfaces in the Age of Generative AI, led by Stephan Druskat et al., explored evolving RSE competencies, practices, training, and roles in the age of GenAI.

Other talks explored the practical implications of this shift. Joseph Shingleton (University of Glasgow) presented the Responsible AI Risk Register for RSE (RAIRR), a community resource for documenting AI risk. Arfon Smith (Schmidt Sciences) explored the evolving RSE role, including opportunities from AI-enabled workflows and interfaces, gaps in evaluation, provenance, and reproducibility, and new demands on software maintainers. This theme was continued in Arfon Smith’s keynote at RSECon, Where does the rigor go? Research software and the future of trustworthy science.

Carina Kemp (Amazon Web Services - AWS) also contributed to the discussion through AWS Software Development in the Age of AI, sharing practical lessons from AWS’s own AI-assisted development journey and exploring how the research software community can collectively establish shared standards, validation practices, and open tooling to ensure AI augments — rather than undermines — research integrity.

The talk Next Generation Software Ecosystems for Scientific Computing: Concerns and Calls for Action by Lois Curfmann McInnes (Argonne National Laboratory) and Anshu Dubey (RIKEN Center for Computational Science) explored what is needed to build AI-ready scientific software ecosystems, emphasising three priorities: making trust explicit, designing socio-technically, and investing in judgement. They posed a broader question to the community: What should our communities build together that no single project or institution can accomplish alone?

A plenary discussion on Sustaining research software under AI assisted development, led by Colette Bos and Faruk Diblen (Netherlands eScience Center), considered how GenAI is changing code generation and the roles and expectations of different stakeholders in research software engineering, highlighting the importance of machine-readable provenance and careful human verification.

Thank you to the IRSC community

The inaugural IRSC would not have been possible without the speakers, session chairs, remote facilitators, volunteers, Program Committee, RSECon26 organisers, and the ReSA and Institute for Research Software teams who contributed their time and expertise. We also gratefully acknowledge the Alfred P. Sloan Foundation for supporting the scoping report that helped shape IRSC, as well as the initial committee members and stakeholders who helped inform and guide this process.

IRSC26 was also made possible through the generous support of our sponsors and partners:

Gold Sponsor: Amazon Web Services (AWS)
Silver Sponsor: Schmidt Sciences
Bronze Sponsors: Netherlands eScience Center, NWO (Dutch Research Council), Society of Research Software Engineering, and U.S. NIH Office of Data Science Strategy
Partners: Institute for Research Software and National Center for Supercomputing Applications (NCSA) at the University of Illinois Urbana-Champaign

Looking ahead

IRSC highlighted the wide range of work taking place across the international research software community and the importance of greater coordination across these efforts.

By bringing these communities together, the conference created opportunities to strengthen connections, exchange approaches and experiences, and identify areas for future collaboration. We invite you to:

Watch the IRSC26 keynote recordings and explore the program.
Join a ReSA forum or task force to contribute to ongoing discussions and collaborations.
Subscribe to the ReSA newsletter for monthly updates and opportunities to get involved.

We look forward to building on the momentum of IRSC and continuing to strengthen collaboration across the global research software community.

Images: Dan Mogan Photography

This post is citable and FAIR thanks to Rogue Scholar. Browse ReSA posts on the Rogue Scholar.