Every time I bring up Fortran in a discussion with engineers who don’t specialize in it, someone always wants to talk about COBOL.
Thing is, they don’t really belong in the same bucket of elder languages.
COBOL is mostly a legacy business language. It still exists in production, sure, but it survives as infrastructure and inertia. In 2026, the only surviving COBOL apps are the ones that are too important to be replaced.
But that doesn’t mean COBOL doesn’t have value. It does, but that value feels more instructional than practical. For one, I like how Englishy it is. It’s great for teaching engineers, especially young engineers, why systems work the way they do and why business software evolved the way it did.
A modern machine room
Fortran is a very different animal.
For one, it’s actively modern, and has been under continuous development for 70 years.
Today, it’s a high-performance language with a clear niche, in supercomputing, numerical methods, scientific computing, simulation, HPC, the guts of a lot of systems people casually lean on every day.
It’s everywhere. You’ve probably used apps that are at least partially written in Fortran without realizing it, especially if you spend time in ML.
Like COBOL, it’s useful for historical understanding of older systems, but it’s useful because it’s still very much part of the present.
There are mathematical performance benchmarks you can hit in Fortran that you can’t hit in Rust.
That’s the difference.
COBOL is a museum piece.
Fortran is a machine room.
Part of why I like it is that it resists your vibes. You can’t charm it with abstraction and hope for the best.
You have to know what you’re building. You have to mean it. Otherwise the whole thing collapses under the weight of its brutalist idealism.
I was a BASIC kid. The grammar of it feels like home.
Nice and comfy.
I think data scientists and HPC people get it.
Other engineers?
Not so much.