Phosh 0.56.0
phosh.mobi
phosh.mobi
> About Phosh
> The Phosh project aims to provide a daily-usable, robust and easy to use graphical user environment for mobile devices running mainline Linux. The name is a portmanteau of phone and shell as phosh was one of the first components developed by the project. It hence coined the whole project’s name and is still one of its core components. All of Phosh is entirely Free Software.
As it is I didn’t find it at all difficult to find the answers to your questions by going to the “about page”.
Like the person you replied to (I assume), I clicked on the home page, pondered why one would not describe the app's purpose on the home page, failed to scroll horizontally with a mouse and closed the page before noticing the about section.
They didn't even lose me with the original link I couldn't make sense of.
App support and camera quality I can understand more. I'm on a Linux phone using Phosh (FLX1s), and there's Android app compatibility, but it is a little rough (and of course things that rely on Play Integrity won't work). I've managed to avoid tying myself to anything that requires Google for now, but I acknowledge that I'm lucky there.
It can be a lifestyle difference.
I personally don't being my wallet in most daily trips, have no use of it. I used to stick a credit card in my phone case but also got rid of it as more stores reliably offered wireless and QR code payments.
Mind you this comes with a specific environment I don't expect everyone to live in or long for, I'm just explaining.
IMO it's not a matter of tap to pay and camera quality, rather a matter of whole system paradigm. Having millions of disconnected services in the "do one thing and do it right" spirit and using text based communication and hundreds of python and shell scripts is relatively maintainable and relatively easy to use, but very inefficient when it comes to CPU cycles - and on a handheld every cycle counts.
And of course every app is optimized for desktops/laptops... but I guess that's a chicken-or-egg problem: once there is a working distro, there will be apps too. And once there will be apps to use, there will be a working distro. Maybe.
If I read correctly, you identified the lack of smarthpone-level apps and distro as the limiting factor. In my opinion, the lack of sufficiently powerful but still open hardware is what we miss. Mainline linux with proper hardware support is pretty good but not complete on the Snapdragon 845, a 9 y/o platform. Anything newer? Nearly impossible without some android-specific layers (such as libhybris in the case of Droidian). Currently mid-high level hardware with PinePhone-like openness would probably let Phosh (and Plasma-Mobile, SXMO etc.) distros strive. The smart features, as in the apps, are mostly there.
My experience today is that I have all the applications I need daily, and the odd ones I don't have natively I can run in Waydroid. There are currently some limitations when it comes to access to hardware from within the container, but nothing that couldn't be overcome with some moderate effort. The only real blocker is remote attestation, but I see it as a threat to the whole fabric of society, not just Linux phones, and it should be opposed regardless of platform one uses.
I don't really get tap-to-pay. I never thought it's something I want or need to have on my phone (which doesn't even have NFC, though it could be added as an extension). I just use a card. I get it that it may be slightly more convenient, but definitely not worth changing the OS over, let alone giving others control over my phone. That said, I have full access to my bank account from my phone anyway. Practically everything except TOTP payments (Blik) can be done from the Web (and with Epiphany's webapps it's just there as an app to launch), and Blik can be used from the Android app in Waydroid. Frankly, if I couldn't access the bank from my phone I would rather change the bank than the phone, and would make sure to let the customer service know very clearly why I'm doing that.
And when it comes to camera quality, matching the state-of-the-art overprocessed mess would be hard, but having played with camera processing a lot recently I can say that perfectly adequate quality is absolutely well within the reach (if not there already for daylight photos). I only researched absolute basics of photo processing and implemented some essential stuff while completely ignoring others and I'm already quite happy with the photos I get (https://social.librem.one/@dos/tagged/shotonlibrem5) and there's enormous room for improvement still that only needs someone who knows what they're doing to sit down and implement things that are missing. Some of that stuff is already happening around libcamera as more people get involved there thanks to laptops with webcams that also need similar software handling. The particular phone I'm using doesn't have hardware video encoder (only a decoder) so video recording will stay low res there forever unless limited to short clips, but it's not a limitation inherent to the platform, just to this specific device.
I have briefly carried an Android device as a secondary phone some 10 years ago out of necessity (N900 was starting to get too old to handle the Web and Librem 5 did not exist yet) and it felt quite miserable, it seemed like an appliance rather than a personal computing device despite of equipping it with microG, F-Droid, rooting it etc. When you're used to being able to script a simple thing right on your phone using whatever technology you already feel comfortable with, it's hard to give it up. Just try to patch a simple thing from the system's core on an Android phone - yes, it's possible if you're really determined, spend a lot of time on it and don't mind attestation issues, but I can do that with dpkg-buildpackage on my phone's screen within minutes, send a patch upstream and have it actually merged in time it would take my laptop to produce an Android image. I may not exercise this ability a lot in my daily use, but having such possibility is incredibly empowering. I can't see Android becoming competitive there, quite the contrary ;) The only actual obstacle I see on the path to growth is the threat of remote attestation gaining widespread use and this affects any existing or future platform from outside the duopoly, not just GNU/Linux in particular, and I believe this still isn't a lost game.
When it comes to Waydroid, I boot it up very rarely and usually just to do something specific and stop the container afterwards, so it's not a concern. But if you wanted to keep it loaded for fast access (booting it up takes about half a minute) you can either freeze the container or just suspend the entire phone and it shouldn't influence anything other than keeping RAM occupied.
Like the particular programs are no issue, but the whole UNIX-userspace as done in the mainframe era and still is. Like you definitely need cooperative program suspend/resume like on Android for any kind of sane battery life, but that's unfortunately completely missing in case of GNU/Linux.
I was looking at this and thinking maybe it would improve a cheap android phone. But now I know it's running gnome I won't even consider trying
Additionally, minimum requirements of Android 17 are way above what SXMO allows.
With Phosh you would have a point but SXMO is lighter than a modern Android.
Add a bunch of fat, semi-desktop binaries that actually provide some kind of functionality to make it remotely comparable, and then you just have a worse, fatter system that runs hot and wastes the battery.
Android is a unified system working together to make the device race to sleep as fast as possible, with as few wakeups as feasible (e.g. batch together events that would wake up the device).
In terms of resources I think it's roughly going like this, from lighter to heavier:
SXMO < Lomiri (Ubuntu Touch) < Plasma mobile < Android < Phosh < Gnome mobile
I could also add XFCE but they officially don't support mobile.
Android isn't the lightest mobile environment by any conceivable means, maybe it's the most featured but certainly not the most optimized for low-end devices, the days of Android 11 are over.
You need 4GB of ram, a good CPU and full hardware acceleration to run modern Android.
And if the next Android bumps the requirements again, I would put it at the end of the list.
Gnome is a bloated mess of a thing and I hate it. Why would anyone want their desktop to use over 1gb of ram. I have a 32gb laptop and I still loath the idea of throwing away memory on such a bloated awful thing.
Running gnome on a phone. Yeah... No
It uses some GNOME services, namely so it doesn't have to invent it's own. None of these services are memory heavy and all have a purpose (e.g. managing Bluetooth)
Without any different numbers, I think saying a massive memory hog is a little hyperbolic. Applications in use, especially browsers, are going to dwarf the desktop environment anyway. Having the polish is well worth it to many people, myself included.
I would definitely like to see less memory requirements for the various desktop environments, but at the end of the day I don't pay for any of this
Not in a million worlds. Android is by far the most optimized OS (as a whole, including user space, graphics stack everything) for mobile devices. It's almost like the most widely used mobile operating system has had quite a bit of dev hours spent on it.
edit: oops, the topic was RAM usage. in 2026. of course everyone, it's definitely the 200MB usage surplus that makes your phone feel heavy. let's add 1TB of RAM to a C2D and it will be ready to be paired with a 5090.
Why should software not adapt to reality?
As another data point, take a look at what "embedded" development means today. It is increasingly common to just simply have a device running full Linux, as it is getting cheap enough to do so.
It actually is though? My Pixel 9a has a perfectly serviceable CPU but I am often frustrated by its 8GB of RAM. Switching to the Revolut app to generate a disposable card number consistently evicts my browser tab from RAM. ~100% of the time this happens and I get extremely frustrated by it losing my state in the checkout flow.
Nothing. On my phone there's 770MB RAM consumed by the entire system with Phosh running right now. The rest is free to be eaten by the browser.
If anything, it's GTK that can be rather inefficient when it comes to rendering (both GTK3 and GTK4, but for different reasons). There are some tricks that phoc (Phosh's compositor) could still learn too, but it will get there.
> Seems like a very strange bundle to bring in for an extremely power constrained device, where every % of increased battery drain is noticed by the user
I'm sure you can make your Frankenstein version that would be 10% as usable and secure as phosh by removing everything but for most users, 100mb more ram and 1% more battery drain for an OS aiming to be a daily driver is something that's worth it.
Not that it would really succeed otherwise. You need Android app compatibility to stand a remote chance.
That's not phosh's problem; waydroid is a pretty much independent component.