alt Hacker News> ability to run other operating systems on phones
> building those alternatives is basically impossible
For smart people it is not impossible. Just few years ago, few folks wrote complicated drivers for completely closed hardware, and I'm talking about M1 Macbook.
Google Pixel, on the other hand, was pretty open until very recently. I might be wrong about specifics, but I'm pretty sure that most of software was open, so you could just look at the kernel sources in the readable C to look for anything. You can literally build this kernel and run linux userspace and go from there to any lengths of development. Or you can build alternative systems, looking at driver sources.
I don't understand why mobile systems do not attract OS builders.
Replies
> I don't understand why mobile systems do not attract OS builders.
They're graphical consumer devices, the quality bar is so high nobody can reach it except huge well funded teams. It's like asking why desktop Linux doesn't still attract OS builders, or for that matter, why the PC platform doesn't attract OS builders. Occasionally someone makes an OS that boots to a simple windowed GUI as a hobby, that's as far as it gets now.
A lot of these HN discussions dance around or ignore this point. When people demand the freedom to run whatever they want, they never give use cases that motivate this. Which OS do they want to dual boot? Some minor respin of Android with a few tweaks that doesn't disagree with Google on anything substantial (Google accepted a lot of PRs from GrapheneOS people).
Nobody is building a compelling new OS even on platforms that have fully documented drivers. There's no point. There are no new ideas, operating systems are mature, it's done, there's nothing to do there. Even Meta gave up on their XROS and that was at least for a new hardware profile. Google did bend over backwards to let people treat phones like they were PCs but it seems regular Android is in practice open enough for what people want to do.
> I don't understand why mobile systems do not attract OS builders.
Cellphones are not very useful as programming tools (too small), which is what Open Source excels at.
Also, cellphones need to handle some annoying things, like it should always be possible and easy to call emergency services. Which is to say, the UI work seems stressful.
I’m fairly sure the modem firmware on the Pixels was never open. There’s some hardware that will never have open firmware to it. Especially when that firmware deals with regulated airwaves like cell signals.
With the right trusted computing modules, it will be impossible. As far as I am concerned, the asahi developers are building on a foundation of sand because Apple could just lock down the bootloader for the iMac laptops or whatever next generation
>I don't understand why mobile systems do not attract OS builders.
My guess would be that it's a continuously moving target. There's no point in spending years working to support some weird integrated wifi adapter+battery controller when by the time you're done the hardware is already obsolete and no longer being manufactured. Repeat that for every device on the phone. The only ones who can keep up with that pace are the manufacturers themselves. It'd be different if there was some kind of standardization that would make the effort worthwhile, though.