|
|
Subscribe / Log in / New account

Linux-for-Rust or Rust-for-Linux

Linux-for-Rust or Rust-for-Linux

Posted Sep 2, 2024 23:38 UTC (Mon) by anthm (guest, #173221)
In reply to: Linux-for-Rust or Rust-for-Linux by corbet
Parent article: Rust-for-Linux developer Wedson Almeida Filho drops out

> There is not much opposition to Rust in the kernel community — far less than I had expected, honestly

But there is opposition, and it's loud and aggressive as evidenced by the video.

Loud aggressive and seemingly not particularly caring about the technical merit of their arguments.


to post comments

Linux-for-Rust or Rust-for-Linux

Posted Sep 3, 2024 0:32 UTC (Tue) by pizza (subscriber, #46) [Link] (3 responses)

> But there is opposition, and it's loud and aggressive as evidenced by the video.

There is "opposition" to nearly everything that is submitted for inclusion in Linux, for any number of reasons, technical or otherwise. Historically this opinionated maintainer gatekeeping activity has been lauded as a GoodThing(tm), and has been one of the primary contributors to steadily raising the overall quality bar for the past 30-odd years.

Meanwhile, on what basis do you assume that video is representative? Or is it just an example of confirmation bias?

Linux-for-Rust or Rust-for-Linux

Posted Sep 3, 2024 7:24 UTC (Tue) by Wol (subscriber, #4433) [Link] (2 responses)

> Historically this opinionated maintainer gatekeeping activity has been lauded as a GoodThing(tm), and has been one of the primary contributors to steadily raising the overall quality bar for the past 30-odd years.

The problem, and history will also tell you this, is that "opinionated gatekeeping" from the "young turks" can be a damn good thing. The problem is when they become "old turks", they are a damn liability!!!

Robert Hooke (or was it Boyle?) is, iirc, a particularly good example, but it seems to be the norm. "Science advances one funeral at a time" or whatever the quote is. Why should Computer Science (and linux) be any exception?

Unfortunately, all the serious pushback from opinionated maintainers seems to be of the form "I don't like it so it's not going in". That's a VERY strong indicator that the maintainer in question should step down and retire! If the reaction was "I don't like this - I don't understand it but it's ringing alarm bells" then fair enough. That at least opens the door for dialog.

I'm afraid (hoping?) things could easily start going the way someone else suggested. All it will take is for Asahi to do what she's threatened - rewrite drmbuf or whatever it was. Linus then says "okay, it looks good, it's going in" - as he should if it lives up to its promises, and next minute all new graphic drivers will be in Rust.

At which point, all those opinionated gatekeeper maintainers will find their fiefdoms deprecated to "legacy" as new people come in with Rust versions and shove them aside. That, imho, would be an extremely successful manoeuvre for linux, even if tragic for the likes of Ted. (And there's plenty of alternative disk formats. No reason why it should always be ext2/3/4. The momentum for change already seems to have been building for a while ...)

Cheers,
Wol

Linux-for-Rust or Rust-for-Linux

Posted Sep 3, 2024 13:17 UTC (Tue) by pizza (subscriber, #46) [Link] (1 responses)

> Linus then says "okay, it looks good, it's going in" - as he should if it lives up to its promises, and next minute all new graphic drivers will be in Rust.

Linus has a lot of "soft power" but at the end of the day, he can't actually _force_ anyone to do anything. Other than enforcing the "Linux" trademark, anyway.

> Unfortunately, all the serious pushback from opinionated maintainers seems to be of the form "I don't like it so it's not going in".

Their objections have considerably more meat than that. Even Ted Tso's.

(And I recall numerous Rust folks having strong opinions about what should or shouldn't go in to stuff they maintain. So let's not pretend that opinionated maintainers are okay for me, but not for thee!)

> as new people come in with Rust versions and shove them aside.

(1) there isn't a legion of "Rust developers" just waiting at the gates, and (2) nearly everyone seems to vastly, vastly underestimate just how much _work_ these core maintainers actually do, day in and day out. Rust doesn't magically make most (if any) of that ongoing work become unnecessary.

Linux-for-Rust or Rust-for-Linux

Posted Sep 5, 2024 1:16 UTC (Thu) by raven667 (subscriber, #5198) [Link]

> nearly everyone seems to vastly, vastly underestimate just how much _work_ these core maintainers actually do, day in and day out

this seems to be a major sticking point, for someone who learned a Rust, understands the code and is past that experience it can be hard to understand the fear and trepidation from a maintainer who may feel like they are burnt out and just hanging on at the edge of their competence and time, who are told that they need to stretch and grow and spend even more mental effort learning when they don't feel like their current workload will take a holiday to let their brain relax and absorb new thoughts. It just looks like pain. Like being at mile 20 in a marathon and having someone start jogging alongside saying "what's wrong with you,why are you breathing so hard, can't you go faster".

The Rust design seems to do what it was designed to do well, which can clear up whole classes of hard to find bugs, but being right isn't the same as winning hearts and minds and its supremely frustrating to be right and still lose. It's one thing to be confident you are technically correct, it's another to make changes which affect millions of people when you are nagged by doubt that the house of cards will topple of you touch it wrong, from somewhere outside your awareness


Copyright © 2025, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds