[coliveira]

I also agree that Linux should close the door to Rust as a matter of principle, as it has done to any other language other than C. I don't believe in a mixed language kernel, it is just nonsense, specially with such a different language such as Rust which is closer to C++ in philosophy.

[preisschild]

I love Linux and have no preference to either C or Rust, but locking Linux in to a programming language that was designed in the relatively early stages of computing will lead to Linux dying in the long term.

There will be fewer and fewer new C programmers with people instead taking up newer systems programming languages like Rust or Zig.

[Keyframe]

There are zero indicators for that happening. In fact, if history is by any means an indicator for the future, we might see these dying out over the years in favor of new hyped ones and C standing against the time. Zig is nothing so far, personally I believe Rust has some staying power for now, but we'll see. As for C, kernel C is not all that C where C was led to in recent years.

[coliveira]

C and Unix have stood the test of time. It has been 50 years since they were designed, this is a long time in terms on computing. It is not unreasonable to think that they will stay for another 50 years.

[bbkane]

Yes, they SHOULD be adding Fortran, another language that has stood the test of time, to the Linux kernel.

Or... Maybe it's good that they add a new (still 10 years old) language with security and DX improvements occasionally. 50 years ago, that language was C...

[account42]

Adding languages is plainly a bad idea. Migrating to a different language might make sense at a time but IMO rust is still too young to consider that.

[coliveira]

How many operating systems are written in FORTRAN and still widely used?

[metaltyphoon]

If other, safer language would have existed the same time as C, I don’t think it would have stood the time as you say.

[chii]

that's just hindsight. You can also argue the same for the dvorak keyboard vs qwerty. And yet, qwerty stood the test of time.

Momentum and legacy is hard to displace, when the new fangled thing attemping to be an improvement is not proven, and has unknown unknowns.

[metaltyphoon]

> that's just hindsight.

Is it? Do you think it's hindsight to use the only "successful" story as "the only way forward?" Look at the gas vs electric vehicles. Yes, displacing legacy is hard but not insurmountable when there is a clear improvement. I don't think anyone can argue in good faith the Rust is NOT an improvement in C.

[account42]

Electric vehicles are also not a clear improvement over combustion engines. Which is why they need to be forced into use.

[_ph_]

They did. But they were disregarded.

[jorvi]

Interestingly enough the Windows kernel is slowly rusting. If they they see a significant drop in CVEs eventually (which is bound to happen, with how easy it is to footgun memory safety in C), it blows a huge hole in the stance of "no Rust in the Linux kernel". It'd be akin to saying "yes, more security issues please".

[account42]

Just like C++ has already led to there being fewer new C programmers and has caused C projects to die out? I don't think the programming language is a real problem for someone capable of working on the kernel. Some people might prefer to work in another language but with a large project you need to compromise in many ways and the language is just part of that.

[account42]

I agree. If it wasn't for the memory safety meme this would have never been considered.

[J_Shelby_J]

Then make it so, so that we can RWIR, and Linux can fade into the sunset with Linus.

[chikere232]

Nothing is stopping you

[mrits]

If the Kernel stopped all development it would still not fade into the sunset in our lifetime.