Comment by neilv
4 months ago
No matter how this turns out, I'm sure GrapheneOS will make a smart effort. https://grapheneos.org/
But long-term, Android is such a massive code base, and was designed more for surveillance and consumption, than for privacy&security and the user's interests.
I think getting mainline Linux on viable and sustainable on multiple hardware devices is warmer, fuzzier foundation. (Sort of a cross between Purism's work on the Librem 5, and PostmarketOS's work on trying to get mainline Linux viable on something else.)
> think getting mainline Linux on viable and sustainable on multiple hardware devices is warmer, fuzzier foundation.
You just have to somehow speedrun the decades of development that went into Android to make it decently run on mobile hardware.. never really understood this "throwing out the baby" direction - the UNIX userspace model simply doesn't work on mobile (I would wager it also doesn't work on desktop anymore), has no security (everything runs as your user which made sense when you ran some batch job on a terminal with multiple other users, but nowadays when a single user has as many processes as all the user had back then it effectively means no security between any of those programs), there is no real resource control, no lifecycles, so the device will burn scorching hot and have terrible battery life.
On Android (and iOS) apps were always living in a world with lifecycles so if they wanted to operate correctly, they had to become decent citizens (save state when asked, so they can be stopped and resumed at any moment). This also fits nicely with sandboxes and user permissions, etc.
So without developing an alternative user-space for "GNU-Linux", it's simply not competing with android in any form or shape.
And even if you do, now every GNU app has to somehow be ported to that userspace API (you can't just kill GIMP or whatever Linux process)
The closest I got to Linux mobile is GPD Pocket 4 with LTE and regular apps. Since I can get it to cap at 5 watts, it can give 9 hours of battery life. It does most things I care about, but it is just a mini laptop (which is good enough for me).
> You just have to somehow speedrun the decades of development that went into Android to make it decently run on mobile hardware
Isn't this mainly due to proprietary drivers and firmware?
No, just take a look at how long and smooth does a pinephone run with "GNU Linux" vs stock android.
Android devs actually backported a bunch of work to the mainline kernel with regards to low-level energy management, but that's only one half of the story. The other is your phone stopping unused apps gracefully, and being able to go back to sleep regularly.
1 reply →
flatpak has existed for a decade, and namespaces even longer. They could stand some improvement, but work well enough.
The problem is for developers. Abandoning Android for Linux is not viable for software developers who need to eat. Sure, we can use Linux smartphones ourselves, but if the software we make has a grand total of three people who ever lay eyes on it, that's less than ideal. And given how The Year of the Linux Desktop has gone, I think it'd be strongly preferable if we managed to stave off the tightening of control over Android rather than placing bets on the future Year of the Linux Smartphone.
The Year of the Linux Desktop is kind of happening. Not at the scale that the meme implies, but I've never seen anywhere near as much adoption of the Linux desktop as this year. The combination of Valve's efforts, more usage of Linux gaming handhelds, distributions like Bazzite that have strong selling points for Windows gamers, and Microsoft pissing everyone off with everything that is Windows 11, the Linux desktop has some legitimate momentum for once
Especially considering how much software these days on Windows are all Electron/Web. So is not a hard switch as it once was.
I switched from Windows to Linux it's been 2 years. One of the few things I missed on Windows, was the native WhatsApp app, as the Web WhatsApp it's horrible. Then a few months Meta killed the native app and made into a webview-app :)
1 reply →
And I can just take about any Linux distro, install it to about any computer and have an extremely nice device to work, play games, and handle almost any daily task with. I call that a huge success.
13 replies →
Not really, because Proton is Win32, kind of.
1 reply →
It really isn't. This is a temporary sugar rush that comes after pretty much every time Microsoft does something awful. After a while the buzz will fizz out and the majority of those PC gamers that looked to switching go back to Windows.
IME a lot developers don't even use Linux on their desktop machine. I've met three developers that use Linux professional IRL. A lot of devs have a hard time even using git bash on Windows.
I am always called up by people at work because I am "the Linux guy" when they have a problem with Linux or Bash.
Sure, there are a lot of people that use Linux indirectly e.g. deploy to a Linux box, use Docker or a VM. But if someone isn't running Windows, 9 times out of 10 they are running a Mac.
More generally the thing that has paid the bills for me is always these huge proprietary tech stacks I've had to deal with. Whether it be Microsoft's old ASP.NET tech stack with SQL Server, AWS, Azure, GCP, what pays the bills is proprietary shite. I hate working with this stuff, but that what you gotta to pay the bills.
12 replies →
I know it's been tried before (eg by Mozilla), but perhaps now the time is right for a web apps-only OS.
Many developers would need some help to get offline functionality and updates right though.. And it would be really nice if these apps didn't require parsing megabytes of JavaScript libraries on startup.
One can dream! :-)
My TV runs one, it isn't taking the world by storm.
https://webostv.developer.lge.com/discover
2 replies →
Some people don't care and build on top of Linux anyway. This lockdown will accelerate this. At some point a critical mass will eventually be reached, perhaps with the assistance of some corporate entity or organization of some sort that pushes it over the edge. Then there will be a real open competitor. Will take some time though.
so the thing is, as an Android dev if I get embedded linux experience then I have lateral career movement to the peripherals that I'm usually writing apps for. While the intersection of app developers to embedded linux developers is probably very small, there is a smidge of incentive there, and that can be a powerful thing for the community: a lot of the pain points on linux phones feel hardware oriented (I complain loudly about the pinephone battery elsewhere in this thread).
another tailwind might be in the gaming scene. I have the general sense that SteamOS has been an interesting gateway for technically-minded folks to be impressed by this Linux thing. A similar model for mobile phones might be a tailwind (like a SteamOS for ARM?) The reason why that's perfect is because it undermines the Google monopoly and creates an app ecosystem that people will absolutely flock to, at least for games ($$).
> Abandoning Android for Linux is not viable for software developers who need to eat.
We'll finally get our ecosystem diversity back when the next geopolitical happening happens and Google bans Chinese android apps on bullshit pretexts.
Wait a few years more.
I'd rather like to see AOSP development spun off to a separate non-profit entity. Either by Google doing it or by a hard fork (which will need a lot of funding). Traditional Linux misses the polish and especially the security layering to be a good phone OS. Better to start from an already good base that works.
Why would that affect anything? The Chinese Android ecosystem is already split from the Google one.
1 reply →
Waydroid does surprisingly well at running Android apps on Linux.
Sure some apps won't work for whatever reason & HN commenters will have incredibly scathing things to say about that, but I bet there's a lot of folks who'd be cool with missing an app here or there.
It sucks to be losing Android, but IMO it's an ecosystem in free-fall. Bootloaders are locked more and more, there's literally zero AOSP hardware buyable now, and the roms scene has diminished not grown over time.
I totally think theres a Steam Deck moment waiting around a corner, where what seemed impossible a year ago shows up and is dead obvious & direct, and we all wonder why there were so many doubts before.
> Right, but that's a choice from manufacturers, not a requirement of building a mobile platform.
IMO, I think Microsoft gave up on running Android apps on Windows because they read the writing on the wall: Google will use Play Integrity/Protect to ensure Android apps only run on Google-approved devices/operating systems and nothing else.
I think this is the ultimate fate for Waydroid, as well.
> Android is such a massive code base, and was designed more for surveillance and consumption
I disagree. I have been using de-googled / de-spywared Android for a decade now and I really love it. Once you remove google mobile services and rely on open source applications Android feels really good.
Also its questionable if projects such as purism or even the pinephone will ever offer such good security and privacy as a de-googled Pixel with GrapheneOS will.
https://grapheneos.social/@GrapheneOS/112712864209034804
The hope is lost for Android, there is no moving forward with google antagonizing its foss roots. Libre phone it is. We have to forcibly remove the bandage.
I wish you were wrong, but I don't disagree with assessment. I am on grapheneos ( edit: on pixel ) now, but even that should only be a pitstop now since google has decided to show its hand in such a nasty ( if not that unexpected ) manner.
Everyone is quick to ascribe malice without understanding why changes are made. It's never done for the reasons you think. Without a formal relationship between Graphene and Pixel, things were operating out of luck. This is why the next target hardware is starting with a business relationship. Even desktop Linux is most successful when business relationship between a vendor and the distro maker. Everything else is ripe for random breakage in support.
1 reply →
AOSP is open source so it could be forked.
Except many key features are nowadays delivered via APEX modules, distributed via PlayStore.
https://source.android.com/docs/core/ota/apex
6 replies →
> Android was designed more for surveillance and consumption, than for privacy&security and the user's interests
I disagree. The Android security model is better than the Linux one. I am very happy with GrapheneOS, I don't have much to complain about.
The problem is that Google sucks and nobody enforces antitrust laws. But it's not just Google: how many Android manufacturers don't suck, really? Do they contribute to AOSP at all? Probably not. Do they build reasonable devices that could run something like GrapheneOS? Nope. Just relocking the bootloader is often a problem.
> I disagree. The Android security model is better than the Linux one.
In some ways it probably is, but it still isn't that good in my opinion (although some of the problems have to do with the way the settings and controls are working rather than the security model itself, there are also problems with the security model itself too). (I think there are other problems with Android (and other operating systems) too.)
> although some of the problems have to do with the way the settings and controls are working
I was talking about the security model.
buy a used OnePlus 6 and load Mobian on it. quite functional these days running a mainline kernel.
(2018) makes me more than a bit sad. I have a OnePlus 6, and it was ok with the software I tried out ~3 years ago, and basically fast enough. But it's soul crushing how running mainline Linux is just so impossible for consumer mobile chips.
It felt at the time like there was positive progress, more bits getting mainlined at a trickle but at least steady trickle rate. But it feels dark now. At least the GPU drivers everywhere have been getting much better, but I get the impression Qualcomm couldn't even ship a desktop/laptop after years of delay, is barely getting that in order now. It feels impossible to hope for the mobile chips anywhere to find religion & get even basic drivers mainlined.
>than for privacy&security and the user's interests.
Even if that was true, AOSP is better for privacy and security than any other Linux distro.
By which criterion? This sounds wrong.
https://madaidans-insecurities.github.io/linux.html
10 replies →