[Scaevolus]

Marcan's career as a developer includes lots of development on hostile systems where he's jailbreaking various consoles to allow homebrew.

Asahi Linux is similar, given how hostile and undocumented Apple Silicon is, but it has a great amount of expectations of feature completeness and additional bureaucracy for code changes that really destroys the free-wheeling hacker spirit.

[bayindirh]

I understand. While I'm not as prolific as him, I've grown with systems which retrocomputing fans meticulously restore and use, so I had to do tons of free-wheeling peeking and poking.

What I found is being able have this "afterburner mode" alongside "advanced communications" capabilities gives the real edge in real life. So, this is why I wish he can build his soft skills.

These skills occupy different slots. You don't have to sacrifice one for the other.

[caycep]

Why not develop a distro based on BSD/darwin kernel then?

[tcmart14]

Probably a few reasons. For Darwin, there are a few small projects but I think they are all functionally dead. The benefit with Linux, or even the BSDs here is, sure you gotta port to the hardware, but you should get a good set of user land stuff running for 'free' after that. Lots of programs just need to be compiled to target arm64 and they will at the very minimum function a little bit. Then you can have package maintainers help improve that. I don't think any of the open source Darwin based projects got far enough to build anything in user land. So you'd probably just get the Darwin code from Apple, figure out how to build it and then build everything else on top of it.

The BSDs. You can fork a BSD. Maybe he could try to mainline into the BSD, but would probably face a similar battle with the BSDs. Right, one again, the benefit mainlining into linux, and there is some (maybe limited) support to include Rust, is you can narrow your scope. You don't need to worry as much about some things because they will just sorta work, I am thinking like upper layers of the kernel. You have a CPU scheduler and some subsystems that, may not be as optimized for the hardware, but at least it is something and you can focus on other things before coming around to the CPU scheduler. You can fork a BSD, but most would probably consider it a hard fork. I also don't think any of the BSDs have developers who are that interested in brining in Rust. Some people have mentioned it, but as far as I know, nothing is in the works to mainline any kind of Rust support in the BSD kernels. So he would probably meet similar resistance if he tried to work with FreeBSD. OpenBSD isn't really open to Rust at all.

[darthrupert]

Why insist on developing in Rust? I mean, I see how it's much cooler and actually better than something like C, but people are hugely underestimating how difficult it is to change the established language of a 3 decade old project.

If Rust is the point you get up from the bed in the morning, why don't you focus on Redox and make it the new Linux? Redox today is much more than Linux was in 1991 so it's not like you would be starting from scratch.

You're probably not as good as Linus in, well, anything related to this field really. The only way to find out whether you actually are is to do the work. Note that also he spent a lot of time whining to people who were perceived as the powerful in the field. But in addition to whining he went and did the work and proved those people wrong.

[maples37]

I remember reading one of the earlier Asahi Linux blog posts on HN a while ago: https://asahilinux.org/2022/11/tales-of-the-m1-gpu/

Mind you, I'm a PHP developer by day, so this Rust-vs-C debate and memory management stuff is not something I've had experience with personally, but the "Rust is magical" section towards the bottom seems like a good summary of why the developer chose to use Rust.

Discussion at the time: https://news.ycombinator.com/item?id=33789940

[tcmart14]

Oh no, I totally agree. I am just saying from the perspective of the Asahi Linux project and wanting to use as much Rust as they can, that is what they are facing and the associated trade offs.

I personally fall a little more on the side of the Linux kernel C devs. Inter-oping languages and such does bring in a lot of complications. And the burden is on the Rust devs to prove it out over the long haul. And yes, that is an uphill battle, and it isn't the first time. Tons of organizations go through these pains. Being someone who works in a .NET shop, transitioning from .NET Framework to .NET core slowly is an uphill battle. And that's technically not even a language change!

But I do agree, Redox would probably less friction and a better route if you want to get into OS dev on an already existing project and be able to go "balls to the walls" with Rust. But you also run into, Redox just has a lot less of everything. That is just because it's a small project.

[caycep]

would be kinda neat if they made redox a drop in kernel replacement, kind of like MkLinux back in the old Apple days

[inkyoto]

> Asahi Linux is similar, given how hostile and undocumented Apple Silicon is, […]

«Undocumented» – yes, but «hostile» is an emotionally charged term that elicits a strong negative reaction; more significantly, though, it constitutes a flagrant misrepresentation of the veritable truth as stipulated within the resignation letter itself:

  When Apple released the M1, I realized that making it run Linux was my dream project. The technical challenges were the same as my console homebrew projects of the past (in fact, much bigger), but this time, the platform was already open - there was no need for a jailbreak, and no drama and entitled users who want to pirate software to worry about.

Which is consistent with marcan's multiple previous blog posts and comments on here. Porting Linux (as well as NetBSD, OpenBSD) onto Apple Silicon has been no different from porting Linux/*BSD onto SPARC, MIPS, HP-PA and other platforms.

Also, if you had a chance to reverse-engineer a closed source system, you would have known that «hostile» has a very specific meaning in such a context as it refers to a system that has been designed to resist the reverse-engineering attempts. No such resistance has been observed on the Apple Silion computing contraptions.

[bayindirh]

> No such resistance has been observed on the Apple Silion computing contraptions.

I think they even left a "direct boot from image" (or something similar) mode as a small door to allow Asahi Linux development, if not to accelerate a little bit without affecting their own roadmap. Even Hector tweeted about it himself!

[tcmart14]

I also think calling it hostile is a little far. I recall Hector making comments of, "yea, even though is not greatly documented, it does things quiet a few things the way I would expect" and I believe even applauded Apple on a few thing. I wanna recall it was specifically around the booting.