FortranFuture Network+
FortranFuture is a 3-year, £2M project funded by the UK’s Engineering and Physical Sciences Research Council (EPSRC) to investigate the question,
How do we sustain sciences currently dependent on Fortran so that they continue to advance knowledge – and contribute to society, the economy and the environment – for the next 70 years?
FortranFuture is a collaboration between Queen’s University Belfast (QUB, Northern Ireland) and University of Southampton (UoS, England). The project will formally start on 1 September 2026 with an official launch at a satellite event to the UK’s Research Software Engineering Conference (RSECon26) The co-investigators are: Prof. Austen Rainer (Lead, QUB), Dr Andrew Brown (co-lead, QUB), Dr Rebecca Taylor (UoS) and Prof. Simon Hettrick (UoS). The co-investigators are supported by an international Accountability and Advisory Board (AAB). The project also has over 25 partner institutions.
If we don’t sustain the Fortran codebases that encode the science, we can’t sustain or further advance that science. The consequence is that these sciences will stall or decline, both in terms of advancing knowledge and in terms of their ability to continue to contribute to society, the economy and the environment.
We chose a timeframe of 70 years because the Fortran language is, by some measures, 70 years old this year (2026).
Given the motivating question, FortranFuture has the aim of ensuring that the existing science and engineering disciplines currently dependent on Fortran can continue to be pursued and applied for decades to come, for the benefit of those disciplines and, beyond that, for their positive impact on society, the environment and the economy. FortranFuture is committed to a strategic, evidence-based, actionable and impactful programme of work to sustain the many disciplines currently powered by Fortran codebases.
FortranFuture’s objectives are:
- To raise awareness and appreciation, and correct misunderstandings, amongst those outside the national and international Fortran community, of the continuing significance of Fortran codebases. There are widely held, but misleading and even factually incorrect, assumptions and preceptions that the Fortran ecosystem – the language, the technical ecosystem, the codebases, and the community – are outdated.
- To bring together and strengthen the national and international Fortran community, which comprises a wide range of stakeholders, including research software engineers (RSE) and computational scientists. One insight emerging from a recent workshop (Back to the Fortran Future I), was that the Fortran community is invisible, even to itself. But surveys show that over 40 countries continue to have active contributors to Fortran codebases.
- To strengthen connections between the Fortran community and other communities and stakeholders, e.g., with research technical professionals (RTPs) and with the HPC community.
- To develop a national and international interdisciplinary research network that will co-enquire, in collaboration with the Fortran community and other stakeholders, on the challenges confronting the Fortran ecosystem, and propose ways to address those challenges. Whether one decides to invest in, or divest from, Fortran there are many challenges common to investment and divestment, e.g., the comprehension of very large codebases, the development of automated test suites, refactoring complex code, the need for domain expertise.
- To develop a roadmap for the future of Fortran-based science and engineering disciplines, and to (begin to) secure resources to support the achievement of that roadmap.
- To advocate for the Fortran community, e.g., with policy makers and funders.
FortranFuture builds on several already-successful activities. In September 2024, Austen Rainer and Andrew Brown organised a one-day workshop, Back to the Fortran Future, as a satellite event to the UK’s Research Software Engineering conference, RSECon24. 50 participants attended from the UK, Europe and the USA, spanning academia, government, hardware and software vendors, and other sectors. Subsequently, Back to the Fortran Future 2 was organised as a one-day satellite event for RSECon25 and Back to the Fortran Future 3 was organised as a one-day satellite event for the Software Sustainability Institute’s Collaborations Week 2026. Back to the Fortran Future 4 is planned for RSECon26. In addition, Austen Rainer joined the FortranCon 2025 Organising Committee.
Several independent but aligned initiatives have emerged from these workshops. Participants at the workshops have subsequently developed Fortitude, a linter essential for best practices in modern Fortran SE, Fortran Index, a curation of resources and guidance for Fortran SE, and Fortran Carpentries, a training resource within the carpentries framework. Fortran Index now works closely with the maintainers of fortran-lang, an online space to support the international community for the Fortran language. We are also working with the Society of Research Software Engineers to reinvigorate the UK’s Fortran Special Interest Group.
Further information
Fortran codebases have a long and distinguished history of contributing to world-class science and engineering. It is both the oldest third-generation programming language in active use and development today, and the only standardised language with a native parallel programming model (Curcic, 2020). Fortran codebases remain crucial for many current scientific disciplines (see Table 1 for examples of research areas that have benefitted from Fortran codebases), and are mission-critical for some of the most significant national and international institutions and initiatives. Fortran has, therefore, over the last 70 years, made an immense and sustained contribution to science. And, in turn, that science has made, and continues to make, an enormous contribution to society, the economy and the environment. As one illustrative example: when the UK’s Met Office warns of severe weather disruptions to travel and logistics it does so on the basis of 2M lines of Fortran code. Overall, since its inception in the mid-1950s, Fortran has had an remarkable (albeit indirect and hidden) impact on society, the economy and the environment.
Received wisdom says that Fortran is an old and therefore obsolete language [1]. But there continues to be a very wide range of sciences that actively depend on Fortran codebases. This includes 6 of the top 10 codebases running on the UK’s Archer2 HPC service: VASP, CASTEP, SENGA+, CP2K Open Source Molecular Dynamics, NEMO Community Ocean Model, and the Met Office Unified Model. Beliavsky actively maintains a list of several hundred Fortran codebases, one on Fortran (scientific) codes and one on Fortran tools.
New technologies are being developed in Fortran. For example, Rouson and his CLaSS group at Lawrence Berkley National Lab (LBNL) are developing a deep learning library (Fiats) written entirely in Fortran, as well as a parallel runtime library (Caffeine, part Fortran).
Promising developments
There are already promising developments and opportunities around the periphery of Fortran-based science. There is increasing national and international recognition of the importance of professional research software engineers, e.g., with the UK’s Society of Research Software Engineers and the United States’ equivalent. Thus there is a growing appreciation of the importance of software engineering for science, and investment in the requisite skills base. But these initiatives do not have a specific focus on Fortran or Fortran-based science. By contrast, the focus tends to be toward ‘modern’ languages that are increasingly adopted in science, e.g., Python.
Furthermore, as of April 2024 (see also January 2025), Fortran has returned to the top 10 most popular programming languages, according to the respected TIOBE Index. This return is attributed to the growing importance of numerical and mathematical computing, e.g., required for machine learning applications. This is encouraging for future recruitment of RSEs, however these indicators are not focused on science or on Fortran-based science, and may more accurately reflect the state of the computing industry, as well as industrial research. Furthermore, whilst the increase in popularity of Fortran is encouraging, there is also the need for application domain knowledge in the respective scientific discipline.
Fortran as a language is in a stronger position today than at any time previously in its 70yr history. But again, these developments do not related specifically to Fortran-based science. The Fortran community has established Fortran-lang as the ‘home’ of the international community, and has made substantial progress on the Fortran language’s technical ecosystem, with the implementation of a standard library, a package manager (with an fpm registry under development), and an open source compiler, LFortran (also under development), that can be used interactively with Jupyter. The Fortran standards are now scheduled for regular five-year updates, first in 2018, then 2023, with the next update scheduled for 2028. As of April 2024, Fortran-lang is now an officially fiscally-sponsored project of NumFOCUS, whose mission is to promote open practices in research, data, and scientific computing. The membership of “J3”, the USA’s national body for the Fortran International Standards Committee, includes wide representation from many corporate hardware vendors (e.g., NVidia, Arm, Intel, HP, IBM, Fujistu) indicating the importance of Fortran to these vendors, in part because of the importance of Fortran to national science and national security. In the UK, the BCS (British Computer Society) Fortran Specialist Group has approx. 200 companies and institutions listed as members.
Given these independent and complementary initatives - relating to the Fortran language, its uptake in industry, and the increasing appreciation of the importance of research software engineering in general - now is the time to exploit these initiatives and undertake a programme of research dedicated to sustaining the future of Fortran-based science.
Indicative list of research areas in the UK using Fortran codebases
Indicative lists of research areas are summarised in Table 1 and Table 2 (there is a degree of repetition and overlap).
Table 1. Indicative research areas aligned with this programme.
| Topic | Areas | UK groups (e.g.) | International Groups (e.g.) |
|---|---|---|---|
| Computational Fluid Dynamics | Coastal management (Wave structure interaction) | CCP-WSI | |
| Nuclear Thermal Hydraulics | CCP-NTH | ||
| Turbulence modelling | UKTC | ||
| Astrophysics | NASA | ||
| Computational atomic, molecular, and optical physics | Laser matter interactions | UK-AMOR | AMOSgateway |
| Plasma diagnostics | York Plasma Institute | Exomol | |
| Astrophysics | CCPQ | HEAVYMETAL | |
| Computational Chemistry | Materials modelling | CCP9 | |
| Quantum Chemistry | CCP5 | ||
| Petrochemicals | Computational Chemistry Consortium | ||
| Climate | Climate modelling | Institute of Computing for climate science | |
| Clean energy | Culham Centre for Fusion Energy | ITER | |
| Weather forecasting | MET-office | ECMWF |
References
[1] Matsuoka, S., et al., ‘Myths and legends in high-performance computing’, Int. J. High Perform. Comput. Appl., vol. 37, no. 3-4, pp. 245-259, Jul. 2023, doi: 10.1177/10943420231166608.