I've had a bit more time to play with @postmarketOS.
It is a joy.
I must look at how I can contribute, as this is a massively useful project, which I can see liberating numerous devices from outdated OSs in the future.
@neil Not just outdated devices. It's the key to liberating us all from Apple and Google. Nobody else is doing a real viable alternative like @postmarketOS is.
@neil @postmarketOS And I don't know if this is intentional or not, but I *love* the double entendre where "postmarket" can be interpreted either as "you install it after you buy the device" (as *all* software/OS's should be, by law, for consumer safety) or as "post-capitalism".
@dalias @neil @postmarketOS it was unintentional originally but it is very fun :D
unfortunately it will be going away with the new name, but it will be worth it to have branding with a broader appeal
@neil if you figure out how to get camera and Wi-Fi of any phone to work, please tell me
(such feature must be previously broken)
@neil Thanks to a kind loaner from a reader, I have a testbed phone, and flashed it with pmOS Edge + Plasma Mobile on Thursday.
TBH my impression was that the UI is shockingly flakey. I can't even open the browser. Great deal of flickering & redraws.
I wanted to try something non-Gnomic as I already have a GnomePhone: https://www.theregister.com/2025/02/03/furiphone_flx1/
But this ain't gonna be it. Wondering if I should try Lomiri instead, or Gnome Mobile in place of Phosh (which is what Furilabs uses)...
We need lighter, faster FOSS phone OSes. GNOME & KDE are behemoths. 😞 In market share, in resource demands, in RAM & CPU consumption.
We had attractive, responsive GUIs in 0.5 MB of RAM and single-digit MHz of CPU 30 years ago. If it takes a gig of RAM, it's broken, simple as.
@lproven I am using gnome-mobile, if that helps.
I can't say that experience is as slick / fast as GrapheneOS on my Pixel 6, but nor was I expecting it to be.
It is (for me) most definitely usable though, even if that is just "firefox is working fine for me"?
@cas @neil @postmarketOS Awww that's disappointing. You should totally keep it.
@cas @dalias @postmarketOS @neil YESSS I love the “post market” as in “post all the bigtech bullshit” name too!!
I really hope the new name reflects a similar concept as well
And honestly y’all really should keep the postmarket name for something, like a Postmarket Foundation that takes care of [new name for postmarketOS here] or something
@iacore Wi-Fi works fine on mine (OnePlus 6), without any tinkering.
Camera works but not well!
@neil do you know why does so many models' camera and wifi broken?
@postmarketOS @dalias @neil @cas Like, really, PLEASE don’t trow the awesome name away please use it for something good
@lproven @neil I too had trouble with the UI. Even when it's not flaky, it's hard to use. The media player, in particular, is a complete mess of non-obvious (well, to me) user flow. And there's not enough integration between the phone/text apps and the contacts list -- you can't add the details of a new incoming text to the contacts lists (or at least, I hadn't worked out how to do it).
(To anyone reading this: don't bother telling me how -- I stopped using pmOS a couple of months ago).
@iacore Lack of specifications for the chipsets, I'd have thought, or lack of people with those devices and the right skills, time, and incentive to dig into it.
@neil @postmarketOS I'm using postmarketos on a pinephone to host my gitea system, works really well and otherwise would just be a wasted device
@smallsolar @neil @postmarketOS I had major problems with pmOS on the PinePhone -- it kept forgetting how to use the modem (usually less than an hour after booting), so no calls, no texts. The only solution was a reboot. Kind of made the whole thing useless.
@darkling @neil @postmarketOS oh I sadly gave up using it as a phone, I've connected a usb-c ethernet dongle and have it permanently powered and the installed pmOS and gitea. Only challenge was getting the nftables sorted but once that was done it's worked well
@neil @postmarketOS @n_dimension I'm looking forward to trying this on a couple of android tablets on my workbench
@lproven @neil @darkling Not quite true. #SailfishOS with official support is only on the Sony Xperia 10 range (but there's more coming).
There's a lot of ports though, some of which run really well: https://forum.sailfishos.org/t/community-hardware-adaptations/14081
@darkling @lproven @neil I recently started to play with postmarketOS on one of the better supported phones (OnePlus 6) and it's just brilliant. Literally all I ever wanted [*] from a phone was a decent terminal emulator and a full strength web browser, and I got all this - plus I can make duck facts appear on the home screen with just a simple shell script! https://martinh.net/@m/114920751495302303
[*] Apparently some people also want to do something called a "phone call". Weirdos! It'll never catch on...
@neil I might try that, then. Phosh looks slick and it's quick on the Furi phone, but the launcher is tiny and there's no drawer. So I put all my apps in groups... Only to discover you can't pin groups to the favourites bar. The phrase "WTAF?" sprang to mind.
It has way too few gestures (basically 2, swipe in from left or up from the bottom) and it doesn't support a button bar, so it basically reduces a capacitative touchscreen to a single button mouse. That was OK on a monochrome Mac in System 7 in the '80s but it's not really acceptable 40 years later.
It makes more sense on a phone than on a desktop, but it still feels very limiting to me. Like GNOME everywhere, it's claustrophobic: it makes me feel trapped and most of my familiar controls don't work any more.
@lproven FWIW, I use Gnome as my desktop environment on my laptop, so there is a sense of familiarity for me, which probably helps.
@darkling @neil The same chap lent me 2 Sailfish devices. One works. 😉 I really like it but I do feel a bit lost even after several months of using it. I'm starting to get used to it. It has almost too many gestures and it's very unlike iOS or Android -- because it predates them both, of course.
GNOME feels like a straitjacket, as it always does. Sailfish feels like being lost in a foreign country where I don't speak the language and barely read the alphabet. (Which I've experienced.) Disorienting but not frightening, just... Strange and alien and everything is unfamiliar.
@m @darkling @neil I'm talking about a OnePlus 6, as it happens.
KDE mobile is a disaster area and unusable. I didn't try GNOME yet because of the Furi phone.
But it sounds like you want something very different from me.
I am a UNIX professional of nearly 40 years standing, but I'm a very unusual, atypical one. For instance I unequivocally detest both Vi and Emacs and I don't like any UNIX shell. I preferred DEC VMS and even DOS to any UNIX shell.
I find this confuses a lot of Linux people. 😁
@lproven I don't mind XFCE at all, but I haven't used it for a while.
@lproven @darkling @neil I'm kind of joking-not-joking - on the one hand I really do love being able to run what amounts to "my whole computing environment" on a phone natively without any nonsense. On the other hand, I am a weird nerd. But for regular folk I think there is something important here - phones don't have to be a black box, there is room for other kinds of UI and control/transparency over who/what your phone is talking to. And duck facts, don't forget the duck facts
@cthululemon @neil @lproven TBH, I'm pretty disillusioned with the whole small-device thing these days. So many exciting-looking OBCs and phones and the like, and pretty much all of them have an initial partly-working port that rarely, if ever, gets updated. Then you get a bunch of slightly-better-working-but-not-by-much community ports done as a one-off by a single developer who drops development shortly afterwards. So you're left with a half-working device that's never really usable.
@darkling @cthululemon @lproven
Yep!
GrapheneOS works well for me, although I have some concerns with which I'll just have to live.
But for a non-Android Linux FOSS phone, postmarketOS seems like the best bet for me at the moment.
@neil @dalias @postmarketOS in the works for a while:
https://gitlab.postmarketos.org/postmarketOS/pmcr/-/blob/main/0003-name-change-plan.md
we have picked one and are currently uhh figuring out how certain we are with it i guess
@cthululemon Yep! I wouldn't buy another of their devices / a future project from them.
@darkling @smallsolar @neil @postmarketOS I had the same trouble with the modem, but with multiple other OS's. I sold the pinephone... not fit for purpose.
@lproven @neil you were messing with a OnePlus 6 right? what UI were you using?
pmOS is definitely missing a lot polish, plenty of folks have produced content about it where they just look at the surface level, as if it were a product
What we need is content that actually dives in to the project, our ethos, how development actually works (decentralised!), the permacomputing aspect, kernel development, etc. We are creating an entirely new way to approach embedded development on a broad scale that has simply never been done before, and none of this will be apparent if you only mess with a few apps
@cas @dalias @postmarketOS Cool! I look forward to finding out what it will be, one day :)
@cas > pmOS is definitely missing a lot polish
So you’re saying I should nerdsnipe my friends into helping with the project? I believe I only know 3 polish people (myself included) involved
Sorry for the terrible joke, I’ll see myself out
@neil @postmarketOS been daily driving for three or four months. Love it. My wife not so much.
Very recently upgraded from OP6 to FF5, big improvement but the OP6 is viable
@neil @cas @lproven Just joining this thread to point out that, in my experience, the main performance issue with postmarketOS is down to a lack of hardware graphics acceleration support in key software components. For most non-graphical software I've tried, the IO becomes a bottleneck long before the processor is saturated.
This is not me saying that all the good points about sustainability that @cas made aren't also valid, but rather that I strongly believe the performance issues aren't because the software isn't "lightweight". It's lightweight alright, it's just that it's not optimized yet for the type of hardware it's running on.
@cthululemon @neil @darkling I have Gemini, and I found it very usable indeed. Until $JOB provided me with a MacBook Air, the Gemini was my go-to device for taking notes at conferences -- an important task for me. I wrote entire articles on it on buses and planes, and filed them from the device itself. Nothing else could do that and go in a jacket pocket.
I believe Planet still has some inventory. I have tried to tell them: donate a few to the pmOS project and get a port. They could win some friends back if there was an upgrade path from elderly Android, I think.
@neil It's the only Linux desktop than can do a vertical taskbar *well* and with the preponderance of wide screen displays these days, for me, that's an absolutely essential feature.
I don't give a stuff about Wayland. X.org works very well for me. It did all I need, does it well, and all the current Wayland environments are not slightly but markedly inferior to the best X11 ones, such as Unity and Xfce. It is very clear that neither the GNOME nor KDE teams know how to drive a standard desktop entirely through the keyboard using the standard Windows keystrokes. But the Xfce team does and does it well.
This is an essential feature for most blind users, as well.
@northernscrub @neil I think it does, but I'd have to do a custom build and I'm very new to this area. Much as I love Xfce I don't think I want it on my phone.
I think the default options are Phosh, GNOME Mobile, Plasma Mobile, and SXMO. Maybe Lomiri (fka Unity 8.)
@lproven @neil maybe not. Postmarket claims the use of tinydm, but the compatibility of lightdm. Ostensibly this implies the ability to install any conceivable desktop env, provided aarch64 packages exist, without the requirement of compilation. In fact, I'm going to test this today in a vm I think.
@neil @postmarketOS I’m still waiting for someone to liberate the pile of outdated iPads in my drawer. No idea why that is such an issue but I’m an admin not dev. Phones I get to dispose usually seen some hard time long before getting obsolete. People tend to care more about tablets. AFAIK it’s still just a branched iOS? Maybe @cas knows?
The locked bootloader, so the need for an exploit to get code onto the device, probably doesn't help - plus someone needs to be motivated to spend the time doing it!
@neil @Unreed @postmarketOS @cas Well PostmarketOS is supported on iPX and older so I assume it’s using the same checkm8 hack to bypass bootloader lock that should be applicable to iPadOS too…
And yes. There should be some easier way to support open source project funding. Something like buying a share on sprint with defined task. Buy me a cofee is somehow too abstract for me.
@Unreeden @neil @Unreed @postmarketOS
basically a case of who wants to actually do the work. As with everything in postmarketOS
With funding we are able to motivate folks to work on tasks that are generally less glamorous but important (like the recent wireplumber work for call audio switching), and this is the approach we take with allocating funds
We have discussed the idea of bounties and letting people allocate money to a specific task, the main concern with this is actually finding appropriate tasks. It's easy to put a bounty on "make usb type-c work on X device", but if the end result is a set of potentially dubious patches that we have to carry in a kernel fork then it doesn't really align with our goals
in less words: collaborating with upstream is super important to us, and it is much harder to find folks who are willing to do that additional work, and for them money is typically not much of a motivator (although we are increasingly trying to award folks who do this work)
@lproven Just out of curiosity, which device do you have?
If available for the device, please try v25.06 first, as Edge can be a bit like Fedora Rawhide: Things are always rolling, and depending on when you jump in, they may just be horribly broken.
(E.g., that browser issue is just due to some recent /usr-merge change, I think.)
You could also try the actual GNOME Shell Mobile on postmarketOS or Sxmo if you want to try something different - Sxmo is definitely the most light-weight of the bunch (way less than a gig of RAM). Lomiri on postmarketOS was rougher than Plasma Mobile when I last tried it, so I would not necessarily recommend it.
@neil
@linmob @neil It's mentioned elsewhere in the thread. It's a OnePlus 6, and I also have an Xperia with Sailfish, and a Jolla C2.
One of the problems of phone OS reviews is that so few OSes support so little hardware. I also have a Furiphone FLX1 and a Punkt MC02 I could use, and a Chuwi Hi9 Air tablet, but nothing else supports them.
@lproven Thanks - not always seeing all posts is on of the downside of being on a tiny GoToSocial instance.
Regarding device and ports: Yeah, porting devices to Halium and even more so to a close-to-mainline kernel is harder than writing words on socials to shit on the hard software work others have done, and people usually port the hardware they have. That's the one strength of the original non-Pro PinePhone: Every fun project was ported to it in its heyday.
At least the OnePlus 6 is a good device to try postmarketOS, Mobian or Kupfer on (unfortunately the Halium-based Ubuntu Touch and Droidian ports have been abandoned).
Plasma Mobile may be buggy, but when you look at really the small headcount working on something so featureful happen (in their free time, largely unpaid), it's amazing that it works at all, IMHO.
@neil
@linmob @neil Thanks for that.
There would be real value in someone writing an executive summary of the different approaches (halium vs AOSP vs upstreaming & mainline kernel) & the main OS offerings. I hadn't even heard of Kupfer before this. It's remarkably hard to get a high-level overview -- and I speak as a professional industry observer for 30+ years who specializes in doing exactly this.
Consider this a Very Big Hint.
@lproven I try to somewhat maintain https://linmob.net/resources/ and it already feels painful like a dayjob ;-)
In theory explaining the differences is not that difficult, as always, be mindful that the devil is in the details.
So: You have a device, it's powered by some ARM SoC it comes with "manufacturers Android", if the bootloader can be unlocked/has an exploit AND manufacturer publishes sources (e.g., is not Boox), you can run other things. AOSP is just the open-source bits of Android, turned into usable gold by projects like LineageOS, you re-use the vendor kernel (maybe patch it, if you can).
https://halium.org/ allows you to reuse the same kernel, but enables you to run 'desktop linux-y' user land/distro (remember Android is different, e.g., it has its own C library (bionic)) on top of it.
With mainline you abandon the vendor kernel, that usually was forked off way back when, carries loads (hundreds of thousands of lines) patches of spaghetti code, that often also calls proprietary binary drivers in user space to make that damn hardware work with Android quickly (time to market is everything!), and carefully re-do all these things, component by component, patch by patch so and try to get your patches, patch by patch, accepted in the proper kernel.org Linux kernel, with the end goal of eventually, just being able to compile that release tarball and have it 'just' run on your device.
Now, of course, we have multiple distributions that build on these core concepts, and the lines may get blurred here or there (e.g., with the PinePhone in Ubuntu Touch or Sailfish OS being a mainline device, or with downstream ports in postmarketOS).
Also, add in the whole 'why is booting Linux on a device not as simple on ARM as on my Intel/AMD-PC' thing that always comes up when this is dicussed.
@neil
@lproven Feel free to ask for clarifications, will try to help if I can/ask people who actually put in the midnight oil to run anything non-vendor provided on devices to chime in.
(I don't like doing free consulting generally, especially not my holidays (this is just my hobby, I do have a dayjob and a life), and I don't like wordings like "Consider this a Very Big Hint", but I managed to swallow my anger here, so let's try and do the right thing.)
@neil
I am finding this thread difficult in multiple ways. I seem to have offended you already, and I did not intend that.
I was not asking you to summarise the state of the market or the main players here, now, in-thread. You said that you ran a mobile-Linux blog; in response to that, what I was suggesting is that a useful post for me, and I think for other people _on a mobile-Linux site_ would be an overview. An intro that laid out what the options were, what distinguishes them, and the pros and cons.
There are many such things for desktop Linux and I regularly write new ones myself. I also try to pierce through the layers of FUD by doing things like this:
https://www.theregister.com/2022/08/18/ubuntu_remixes/
People are forever saying "desktop N is heavier than desktop M, but desktop O is lighter." I measure it and publish the numbers.
You seem to be trying to answer in thread, which is interesting, and I am grateful, but it's not what I was attempting to suggest -- and I did not want to anger you.
There is value in frequent periodical updates on what's new, but if someone does not know _who the players are,_ that is not so helpful.
Does that make it clearer?
@lproven That helps a lot, thank you! I have been pondering to build something that fulfills that introductory use case, but I feel that - unless it's a blog post that's perpetually bumped up - it's not really a blog post, but also not good as a hidden page somewhere on a blog that maybe no one ever reads (I don't have analytics and thus no idea). Also, it feels a more difficult task than that Weekly Update, which is by now in habit territory. I'll find a way ;-) @neil
@lproven Ah, that's good! And, yes, this trying to keep track is something I've been trying to do, but I must admit that I do fail sometimes. I can't buy every device under the sun (e.g, I never got a FLX1), and attempting 'daily driving' all the distros or devices feels pretty much impossible.
@neil
@lproven I think that's very much like it is in 'desktop Linux'. You have more UIs than the world will ever need (and way more distributions than may feel strictly necessary), yet people will come up with additional ones to scratch their own itch - and these may be abandoned sooner or later or surprisingly get some traction. The additional complexity we add over here in Mobile Linux is device support - you'll also need $distro to support your device or be able to add support for your device to that distro.
@neil
@lproven I could try and list them all, but that does not usually reduce confusion.
So let me list some main projects, sorted by their rough starting age.
- Sailfish OS [libhybris (not using Halium, only parts of it), rpm-based, custom, proprietary UI (Qt 5-based), backed by Jolla (company)].
- Ubuntu Touch [Halium (these days), Unity 8/Lomiri as convergent Qt5, Ubuntu-based (20.04, soon 24.04), custom package format (clicks), read-only rootfs, desktop apps in Snaps or containers, graphical UBports installer],
- postmarketOS [mainline, various UIs https://wiki.postmarketos.org/wiki/Category:Interface, Alpine-based, adding systemd support, Alpine apk or flatpak for apps, difficulties in finding apps that work on phones (https://wiki.postmarketos.org/wiki/Applications_by_category, https://linuxphoneapps.org)
- Mobian [mainline, Debian-based, goal to 'disappear into Debian proper', Phosh and Plasma Mobile as main UIs (potentially Lomiri on Debian forky, and people have run Sxmo on it at some point at least, too)].
- Droidian [halium, Debian-based, Phosh as main UI, also experimental Plasma Mobile support (I think) + https://cutie-shell.org/)]
These are the most alive and impactful projects, I'd say. Before I continue: Is this what you had in mind?
@neil
@lproven From a users perspective, benefits of Halium are: Photo looking photos (likely a bit worse than they looked the original Android software), features like fingerprint unlock (never seen that on mainline, but it may also be a 'these people don't believe in using biometrics for security' thing, and not that it's technically impossible), Camera support in Waydroid (Android container).
The benefits of mainline are more intangible at first, it's proper long-term device support, less bugs, ... I really like how postmarketOS have described in their leaflet https://gitlab.postmarketos.org/postmarketOS/artwork/-/blob/master/leaflet/leaflet_en.pdf ("Why mainline matters", possibly that same copy is also somewhere on the website (cc @postmarketOS)
@neil
@cas @Unreeden @Unreed @postmarketOS
Absolutely!