:: Re: [DNG] An Anti-Initsystem
Top Page
Delete this message
Reply to this message
Author: Xenguy.fl9z
Date:  
To: Bruce Perens
CC: Rainer Weikusat, dng
Subject: Re: [DNG] An Anti-Initsystem
Rationale?

X.


----- Original Message -----
From: "Bruce Perens via Dng" <dng@???>
To: "Rainer Weikusat" <rweikusat@???>
Cc: "dng" <dng@???>
Sent: Tuesday, March 31, 2026 7:09:38 PM
Subject: Re: [DNG] An Anti-Initsystem

Please take this off the list.

On Tue, Mar 31, 2026 at 4:12 AM Rainer Weikusat via Dng
<dng@???> wrote:
>
> Steve Litt <slitt@???> writes:
> > Rainer Weikusat via Dng said on Mon, 30 Mar 2026 11:29:15 +0100
> >>Steve Litt <slitt@???> writes:
> >
> >>> Wait. Did I miss the notice of administrative mandation? Last I
> >>> heard, Linus said kernel code could be written in EITHER C or Rust.
> >>
> >>This is from a LWN article from 12/13/2025 which paraphrased the Linux
> >>DRM maintainer (David Arlie) as follows: [...] the subsystem is only
> >>""about a year away"" from disallowing new drivers written in C and
> >>requiring the use of Rust.
> >>
> >>https://lwn.net/Articles/1050174/
> >
> > Well THAT'S interesting. I'm still trying to figure out how I feel
> > about all that and the comments. The referenced article and comments
> > seem to imply that the long term goal is a pure-Rust kernel and
> > drivers.
>
> The long-term goal of the people pushing Rust is pushing Rust. One could
> also call it 1970sd, as it's still supposed to sort out turf wars of the
> 1970s. And to get rid of dynamic linking, obviously, because that's this
> awful new-fangled 1980s invention which lets user CHANGE (Gaaaaahhh!)
> running code and nobody hates change more than those who want to force
> it on others.
>
> >>No idea if the link can be accessed w/o a subscription. This has
> >>greatly reduced my enthusiasm for this language because it if was
> >>really good for something and not just some people's techno-partisan
> >>"true love", such mandates wouldn't be necessary.
> >
> > They'd be necessary if long term you want your kernel and drivers to be
> > written in only one language. Until 5 years ago, you could only do
> > Kernel/Driver work in C. Does that mean that C must be bad because if
> > it weren't, peoples' "true love" would make such a mandate
> > unnecessary?
>
> Some time in the 1990s, someone chose C as programming language for a
> hobby project. It wasn't mandated by a bunch of solutions pushers
> working for US megacorps. You have your history somewhat wrong.
>
> > Of course, more to your point, with an Ada-only kernel, the developers
> > of the Rust compiler would have an insane amount of power.
>
> ,,, and on to one of those tiresome political arguments. Not. I
> originally considered Rust and interesting language because they
> actually had a new idea for memory management, despite piggy-backing
> MUST GET RID OF DYNAMIC LINKING!!1 onto that dented my hopes somewhat.
>
> The abovementioned makes it pretty clear that the very people pushing
> Rust are convinced that it cannot successfully compete in a marketplace
> of idead on merit.
>
> Hence, end of the story, insofar I'm concerned. They must know this
> better than I do.
>
> NB: I will not participarte in any kind of "My programming language is
> longer than your programming language!" discussion.


--
Bruce Perens K6BP