What is it about?

Fortran has been continuously updated for over 60 years and remains the dominant language for high-performance numerical computing. The language itself is modern (Fortran 2018 has coarrays, submodules, and improved interoperability with C), but its ecosystem -- package management, build tools, documentation, community infrastructure -- lagged behind languages like Python and Rust. This paper surveys the state of the Fortran ecosystem as of 2022. We document the Fortran-lang community's efforts to build a package manager (fpm), a standard library (stdlib), a package registry, and modern documentation tools. We also assess compiler support for recent language standards across the major implementations (gfortran, ifort, ifx, flang, NAG). The paper reflects the perspective of community members actively building these tools, rather than an outside assessment.

Featured Image

Why is it important?

Fortran's share of new scientific code has been declining not because the language is slow or outdated, but because the surrounding ecosystem made it harder to use than alternatives. A researcher starting a new project had to write their own build system, manage dependencies manually, and navigate fragmented documentation. The tools described in this paper (fpm, stdlib, the Fortran Discourse forum) directly address these frictions. fpm gives Fortran a cargo-like build experience. stdlib provides tested implementations of common operations. The Discourse forum replaced scattered mailing lists with a searchable, moderated community. For the HPC community, Fortran's performance advantages only matter if people continue writing Fortran. This paper documents the infrastructure that makes that more likely.

Perspectives

My involvement with Fortran started as a PhD student needing raw performance for saddle point searches. I saw the computational power but also the tooling gap. That gap is what led me to become the maintainer of f2py (NumPy's Fortran-Python bridge) and to contribute to fpm and stdlib. Writing this survey with the Fortran-lang community was an exercise in honest assessment. We documented both the progress and the remaining gaps. Compiler support for Fortran 2018 features remains uneven, and some tools are still early-stage. The paper also served as a call to action: Fortran will keep losing ground unless the ecosystem improvements continue.

Rohit Goswami
University of Iceland

Read the Original

This page is a summary of: The State of Fortran, Computing in Science & Engineering, March 2022, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/mcse.2022.3159862.
You can read the full text:

Read

Contributors

The following have contributed to this page