Comment by uecker
10 months ago
Overall, this whole situation seems entirely weird to me. All this stuff such as Unix, Linux, and C ecosystem was build by C programmers and maintained for decades mostly voluntarily, while most of the industry pushed into other directions (with a gigantic influx of money). It is completely amazing that Linux become so successful against all the odds. Certainly it also then had a lot industry support, but I used it already before most of this and witnessed all the development. But somehow, C programmers are now suddenly portrayed as the evil gatekeepers, not stepping aside fast enough, because some want to see change. In the past, the people wanting to see something new in the open-source community would need to convince the community by building better things, not by pushing aggressively into existing projects.
I believe the Rust for Linux project was started by a Linux guy, rather than a Rust guy, and many of the Rust for Linux maintainers have come at this from a perspective of "we are Linux maintainers who want to use Rust" rather than "we are Rust users who want our code to be in Linux".
I think it's important to be wary of simplistic narratives (such as "C vs Rust"). Maintaining a complex piece of software comes with tradeoffs and compromises, and the fewer languages you have to worry about the better. On the other hand, the Asahi Linux team have been quite explicit that without Rust, they wouldn't have achieved a fraction of what they have. So clearly there is a lot of value in RfL for Linux as a whole, if implemented well. And that value is reflected in the decision from Linus that RfL should be supported, at least for now.
> many of the Rust for Linux maintainers have come at this from a perspective of "we are Linux maintainers who want to use Rust" rather than "we are Rust users who want our code to be in Linux".
This might be true, but do you have any actual quantifiable evidence for it? Because FWIW, from what I as an outsider see (mainly in threads like this), all the drama looks very much like "we are Rust users who want our code to be in Linux".
It is entirely unclear to me where the value actually is. It seems google is funding it for some reason. And some people clearly have a lot of opinions that this is "the future". People had similarly strong opinions about various other things in the past.
It's worth reading Asahi Lina's posts about writing a driver in a Linux - she is very explicit that what they've achieved would not have been possible without Rust.
See, for example:
https://xcancel.com/linaasahi/status/1577667445719912450?s=4... https://vt.social/@lina/113056457969145576 https://asahilinux.org/2022/11/tales-of-the-m1-gpu/
In fairness, this is one team working on one project, but if they're attributing much of their success to Rust, it's probably worth listening to and understanding why, particularly as I don't believe they were particularly evangelistic about Rust before this project.
I have no idea about the Google funding, but Marcan's blog post is very explicit they they do not have any corporate sponsorship. If you believe that to be untrue, please explain your reasoning rather than spreading unsubstantiated rumours.
1 reply →