Do you really want Linux phones
blog.brixit.nl
blog.brixit.nl
I'm maybe not the guy to ask, I have curiousity and professional interest in open source phone stuff but I've had an iPhone since the 4, and plan to spend a total of less than 5 hours a year debugging my phone for the foreseeable future.
I'm not sure that pruning apps off my homescreen or rebooting a few times a year to solve some heisenbug really counts as debugging, but it's already closer than I care to be.
Android also comes with a bunch of anti features like an api to block the user from taking screenshots. Its better to completely replace the OS with something Google doesn't control.
When it came out there's a malware campaign that tries to extract MFA codes from Google Authenticator by taking screenshots, there was a lot of criticism towards Google for not setting the flag that activated that API and made screenshots impossible, and it's a good solution, there's no viable scenario in which you'd want to screenshot your MFA TOTP codes. I admit that i'm sometimes annoyed with it's use in banking apps, but IMHO it's a needed feature for security.
Lineage support will hopefully get better with the new Android split between system with drivers and userspace.
This is the opposite. The device is stopping the owner from doing something that someone else doesn't want.
You can say "so what, it still works!" but I think there are weird psychological effects at play when you can make a direct comparison like "Android 37 was released 8 weeks ago but LineageOS is still on 36!" It makes Lineage feel like a second class citizen.
In contrast, if you have a phone that is designed specifically to run a custom non-Android-based OS, there's nothing to directly compare it to. Sure, you can say things like "I wish my FoobarOS phone had Google Pay like Android does", but you still inherently get the fact that they are completely different platforms and won't have app/feature parity.
I used to run CM years ago, but quit even before the LineageOS fork/transition, because I was always looking at the latest-and-greatest Android releases (even for my particular phone hardware) and feeling left behind. And it was even worse when things should have worked but didn't because of some peculiarity of CM, or things like apps refusing to run on rooted phones.
I'm really considering getting a PinePhone just to try it out. I know that I will miss Google Pay and some other things, but I'd be going into it not expecting those to be there, and expecting it to be a different platform with a different experience and different features.
>You can say "so what, it still works!" but I think there are weird psychological effects at play when you can make a direct comparison like "Android 37 was released 8 weeks ago but LineageOS is still on 36!" It makes Lineage feel like a second class citizen.
Eight weeks, really? That's a blink of an eye. I didn't even know we're on Android 37, thought it was 9 or something.
If you want commercial support have a look at https://esolutions.shop/ (/e/ is based on LineageOS).
Do they really? iOS isn’t just not open source because we don’t have the source.
It’s also not open source because it isn’t developed in the open or intended for a community to work with.
It seems like the thing you are saying people want just isn’t real.
This isn’t to say you are wrong - but if you are right, it just means the people who are buying these phones don’t just don’t understand open source.
I couldn't give two flips if it is the Linux kernel that accomplishes that, but I think it's pretty obvious that Linux is our only reasonable hope.
I had imagined that because KDE was so solid on my desktop that it would be really solid on the phone, but I think the phone is not just fast enough.
I was really happy with my feature phone, but had to give it up because I kept pocket dialing the emergency services. I wish I could find a shell that was as simple.
I would love a shell that was as fast and responsive and had the features of my Nokia 3310. I love these phones https://www.nokia.com/phones/en_in/feature-phones
The only thing they are missing in my opinion is a browser.
I just wanted a phone that works as a phone and then I can plugin to a dock and I get a full linux desktop experience.
It's still a thing, but nowhere near what it was when they were trying to launch an actual phone.
I have even ran a full Plasma Desktop on an external screen this way - and it worked surprisingly well even despite of GPU acceleration not being available there because of a bug.
I like open source too, but I'm not serious about it, I want to be able to fix things or file an issue on GitHub or whatever if necessary, but I don't have the energy to worry about proprietary blobs.
Having control of my phone and its configuration like I do my Linux desktop is what I want, but I'm slightly resigned to perhaps not getting it because of app support (I tried Anbox on my desktop, it was buggy and slow, with a fraction of the RAM I dread to imagine) - so I've been thinking about and occasionally working on a Terraform provider for Android instead. So far I'm using it to install all apps, no Play Store at all; I'd like to have it handle settings too.
So yes, it is Linux more than OSS that appeals to me here, because it's control more than openness that interests me in it. (I suppose some wouldn't see a distinction, and that's partly why it is more important to others.)
I didn't understand why it's so important for it to be Linux, as opposed to any other operating system that's as hackable/customizable. So I wanted to know if I had missed something.
I should apologize for my response then. I'd thought you were alluding to iOS, and thought that it was strange to bring that up when discussing configurable OS's.
As it is now, the PinePhone is officially supported by Manjaro and that makes it the most obvious option. Also, all attempts thus far to bring open source to mobile has been some form of Linux, so it's natural that when people think "open source phone", it's probably going to be using Linux. PinePhone itself already has 18-ish different distros, including ports of Debian and Ubuntu, so it's certainly well on its way.
I wish something like CloudAPI/Capsicum would take off to be more secure, while also being an open platform, instead of Linux ubiquity rendering standards irrelevant.
Android is a Linux phone but it’s decidedly not open.
Another Unix would be ok as long as it's not hidden away under a forced UI, with some things unconfigurable, required, etc.
But ultimately I want to share dotfiles between my desktop and phone (sure I might have to accept some amount of conditional guards..) - and that pretty much means Linux and ideally Arch. But only because that's what I happen to use on non-phones.
I like in principle that Android is open source (minus all the proprietary junk getting jammed into Play Services). I or someone could verify that it's not sending my data to a sketchy third party, but only to a point. Unless the OS and all apps are completely open source, any closed source component could be secretly betray me. Ironically, though, I trust Apple a lot more to write privacy-respecting code even though it's all proprietary. But on Android I do what I can to mitigate these issues, by running as much tracking- and ad-blocking software as I can (a thing that isn't really possible on iOS, at least not to the same extent). And I do have some apps sideloaded that I'd miss if I had an iPhone. But I still assume that Google is not being a good steward of any data it gleans from my Android usage, and that sucks.
I don't care too much about the ability to make modifications. The hurdle to jump to go from a stock to custom ROM is pretty high nowadays, as I expect most financial apps (and probably some others) to aggressively detect enabled root access and refuse to run. And the process of building your own OS images to make tweaks is not particularly fun, and can be a mess to clean up if you make a mistake. It's critical to me that my phone doesn't have downtime, so I'm less likely to mess around with it.
The problem with the current crop of "Linux phones" is that (while they do respect privacy, don't try to monetize every interaction, and are open enough to run what I want on them) they don't have anywhere near the polish of iOS or Android, and are (understandably) missing key applications that I use daily. So anything I use will have trade offs. For better or worse (probably worse) I've chosen the easier path of Android, at least for now.
Having said that, I do think I want to get a PinePhone, not to use as my primary mobile device, but as something to tinker around with. Maybe it's something that eventually could be a primary device, at least for some situations, but I don't see that as being the case without a lot of work, and a lot of customization that I have to do myself, which I don't really care to do all that much.
...back when I felt the need to jailbreak at least - now with altstore, ios improvements, or my own dev-ish - I am good.
So I would say that is the long term goal: convergence. Linux has a major advantage over other operating systems in this regard.
edit: Hangover lets you run x86 programs via Wine on ARM platforms: https://github.com/AndreRH/hangover
iSH almost looks usable and then you try compiling something or cloning a git repo and your iphone will get hotter than a pinephone watching youtube (and whatever you do can be fairly precisely measured in ~15 minute increments because it takes ages.)
I want access to the same applications; I want to be able to run strace on random processes to understand what they're doing; I want to be able to organize my files under $HOME the same way I do in my other machines.
I want a Linux phone (ideally of a similar distro). It being open source is just part of that.
This is also the reason why I'm not really attracted to PineTime, despite it being open source.
If by polish you refer to aesthetic stuff like animations and whatnot, I don't really care about it.
My Android phone doesn't work that way. The company behind it expects me to do certain tasks with the phone. Some tasks are open for gradual adoption, other aren't, meaning they are mandated. After all those years, I'm beginning to dislike that.
Pretty sure AOSP does not do this.
Similarly, the sooner we get over the trope that any handset maker must be forced into selling a tiny general purpose computing device instead of an iPod PDA communicator, the better. It is what it is, let makers just make it the best, without unnecessary and less reliable modularity. Yes people want that, it’s a market segment, and so other makers can make things for that category.
The reality of this is much different. To the average person buying a Linux phone, their new sofa has 3 legs on Tuesday, they can only invite 2 friends to sit on it if it’s an odd number day, and sometimes the fabric glows in the dark. 1% of the time it works exactly how they dreamed and 99% of the time it’s frustrating.
But they own it and can fix all those problems if they’re so inclined! Or they can be at the mercy of people who can fix those problems and hope they don’t see the random glow in the dark fabric as a feature and that they don’t prioritize the problems they’re having over the problems they’re having.
Or they can buy an iPhone or flagship Android phone and have it work exactly like they want it to 99% of the time and be frustrating 1% of the time.
The average person’s iPhone/Android is the sofa in their living room just waiting to be used.
Except, it isn't.
Because at random times the sofa company comes in and reorganises the way the cushions are positioned. And then they take away the ability to use the tilt function.
Then, because they really need to know how they use the couch, they'll start tracking when you sit and lie on the couch. So when you have your day off and its outside the scheduled expectations, when you sit on the couch they suddenly come over and put locks and chains on the couch because it clearly wasn't you who was using the couch.
Most people just accept this, because what are you going to do? Buy a couch that doesn't have tilt or cushions? You want those! You don't want a wicker chair, you want a comfy couch, so you have to deal with the couch company making life hell every so often. Because there isn't an alternative for the average person.
I field calls from my parents at least once a week about something that has "broken". The tech support is like pulling teeth, because everything is hidden from the user. They don't know what they did wrong and all that they want is for it to "work". And half the time I can't help them - because the phone doesn't belong to them and the OS or apps has decided that X feature is no longer worth supporting, and force-updated them to the latest version. Rolling back to the version that did work is difficult or impossible.
I want a Linux kernel, because even though today I'd list the modern niceties like namespaces, containers, WireGuard, NFS 4.2+, back in the Maemo days, having a Linux kernel still paid dividends. All of the standard networking and WLAN debugging tools you had on desktop Linux were available in your pocket, FUSE meant you could use sshfs, etc. The userland was the same GNU userland you'd find on most Linux distros, you could use X11 forwarding via ssh, GTK apps worked, and you had apt as the system package manager.
I want that again. Android doesn't come close, and neither does iOS.
NFS on a phone? What's the use case for that? Isn't it easier at that point to use small linux laptop?
Also, if I'm at a friend's place who has a Roku, Chromecast etc, I can already cast media from my NFS collection easily from my Linux laptop via VPN. I want to be able to do the same thing with my phone.
You can recreate the "ability to make modifications" (although you'll struggle to come to anywhere near the seriously integrated layers of technology Linux environments so robustly build). You can offer "Freedom and transparency" to be able to tinker. But it's the co-participation with other tinkerers, with many venerable but separate & adjustable/hackable layers, that makes Linux so rifely diverse & compelling.
In short, my strong belief is: people want Linux phones. The technical ecosystem is more interesting as a whole than any array of boxes/capabilities one might tick.
I'd love to live in a world where BSD or Minix became The Thing™, but since it didn't I basically live and breath Linux for work and that spills over into my personal usage too. Interoperability between my phone and my laptop is huge. I recently started using KDE Connect (and okay - that doesn't have to be Linux-specific) and being able to seamlessly copy / paste between devices, etc. is a game changer for me.
And sadly, pine is still not my daily driver. At this time, it is a toy.
I want control over my devices, so linux phone is not a bad idea. We keep missing on execution though.
It was never sold as a daily driver and this was always made abundantly clear in every Pine store page since the beginning of time. Even now they only call their latest version "beta" for a good reason.
Both of those are missing the point IMO. What people really want, is a phone that is on their side. iOS doesn't support open standards like Vulkan mainly because it's not in Apple's interest to support them. Not because it's in the user's interest, necessarily.
What people really want is a phone that they can trust to be on their side implicitly. The nugget of brilliance in Free Software is saying "if you do all this stuff, then the power will be so tilted towards you that you will be able to trust your software to be on your side".
People want something that's both polished, and won't [forego polish in one particular area for the sake of the company's interests/$$$].
Think of it like this: if I'm actively using an app that's giving me turn-by-turn directions, then of course I'm not going to want to spoof GPS, since it would make the app useless to me. But if your app demands GPS data just for ad analytics, then yes, I would absolutely and without any remorse spoof GPS data for that app. In that latter case, I don't care that a developer might be unhappy with it, because as a user I'm unhappy that the developer is slurping up data they don't need!
That's the issue with augmented reality games like Pokemon Go.
If the app wants to force a user to not modify their phone just so they have a secure enclave for which they can implement DRM, then i can't agree to it. This includes things like spoofing, but also things like memory edits or hacking save files, or anything else client side. If the app needs secure storage that's unchangeable, they can save it on their servers.
https://developer.android.com/guide/topics/sensors/gnss on android now apps can access the raw gps data, which is much harder to fake
Considering that most of the world has some form of travel restrictions over the past year, and I believe that being able to interact with friends in the game requires you be be in very close proximity. Can you explain how spoofing GPS data so that people are allowed to play together even though they are not close is bad?
If you can just spoof, the whole nature of augmented reality falls apart and someone could simply spoof themselves anywhere and take those positions. It was a big problem before the android os made it harder to spoof.
There's no point/ fun in playing augmented reality games if enough people simply cheat and the real world limitations aren't a factor.
not to say how it you are anywhere not in a medium size city you need either spoofers or endless grind to unlock significant parts of the game
[1] to be fair I left before the pandemic, so I cannot comment on how they handled that
It really stopped being a large scale issue after Android started making it harder to spoof.
You could attack a pokemon gym at your convenience (from your couch, whatever time of the day). This negatively impacts other players because a legit, non-cheating spoofer will have to expend far more effort to do the same thing.
>Can you explain how spoofing GPS data so that people are allowed to play together even though they are not close is bad?
Friend interaction or just interaction in the game has to do with raiding and trading. You can already league battle (pvp) over any distance.
Niantic has made some changes along these lines - main one is the remote raid pass. One person close enough to a gym can invite friends to raid, no spoofing needed. They also upped the trade distance, how close you need to be to interact with a gym/pokestop, etc.
So the answer to your question is the game itself has made changes to address actual in-game scenarios. That leaves spoofing GPS as a way to asymmetrically cheat against gyms, or perhaps trivialize collecting pokemon (some are region locked and so on).
If your app needs to get info that benefits the user (as the mapping app GPS example), nobody will deny it because it makes sense.
If your app wants to get info that benefits you but not the user, it only makes sense for that data to be denied (preferably) or faked (as a backup plan). You have no right to it.
That's the whole issue - it's not the case anymorez now it's the ads that pay
One of the frustrations people have with Google is that they're making it increasingly closed and problematic to make modifications. Given that nowadays some banks for example are mobile only and everyone is waiting for Google to completely shut down modified boot systems(probably in response to Qualcomm TrustZone being untrustworthy), people are re evaluating whether it makes sense to invest in the Android ecosystem.
EDIT: google could have capitalized on this and made it a part of their offering.
On iOS the app installs and works and THEN it might ask you for GPS permission and you can deny that and the app keeps working, just without the feature that needs GPS.
I just installed Mi Home, a chinese app which requested every permission there is. I just denied all of them, could still use the app and am 0 afraid that the app somehow accessed my data.
But it is true that Android won't reject apps that refuse to work without a permission while iOS mandates that basic functionalities of an app work even when permission is denied.
A phone that would truly value the user would allow the user to provide a fake GPS location. Then you could use features that need GPS and not give up your privacy. But Apple doesn't allow that since it would make the phone less useful for companies who want to use phones for tracking.
A phone that truly valued the user would allow users to install multiple copies of an app, and allow the user to log in with different accounts. But Apple doesn't allow that, because they don't want people to share devices.
So you, I suspect, don’t want to “avoid tracking” so much as you want to avoid geo-gating of apps and content.
Content distributors rely on such features to satisfy legal obligations about what content they can show where. If the OS started adding features like “pretend I’m in location X” or “download movie regardless of app restrictions” (as an example of another feature that could be argued is pro-user) then at some point you’ll have a platform that isn’t commercially supported or viable.
Honestly, Apple is by far the most pro-user company in this field and there are countless examples of this.
But a very small minority chafe against the restrictions that are actually in most users interests (eg not allowing side loading of apps, which for most people would simply be an attack vector).
Many apps require location data for access to features. If you can't provide a fake location, your only choice is to either accept tracking, or not use the service.
Lots of companies use phone apps for tracking (eg. for tracking location of agents). The company will immediately realise when you turn off GPS access. If you had a way to spoof a location (eg. tell the phone to pretend you stayed in the current location), you could go to the toilet without your employers knowledge and without leaving your phone behind.
Abusers often use phone apps to track their victims. If you are in an abusive relationship, you may be too scared to turn off Find My Friends. If the phone was actually on your side, you would be able to enter a fake location, and your abuser wouldn't have to know that you visited your parents on your way home from work.
Apple wants location data from iOS users to be hard to fake, because it is more valuable this way. Apple's most important asset is the value of their ecosystem. It's why they can charge companies 30% to access it. The user's needs are secondary.
That's not to say I wouldn't want it either but I'm pretty privacy conscious, and a tech enthusiast. I just don't think you can say Apple is putting user's needs second, I don't imagine most users need this.
Oh and as for tracking, at least for the company they can already track you by what towers you’re connected to. So there’s no technical limitation on this being provided to apps regardless of what you might spoof.
Strava refuses to let you track your ride if you don't allow always-on background tracking. (They changed this about a year ago. Previously "while using the app" was sufficient)
A lunch menu app that I used before Covid allowed you to show nearby restaurants based on location. But there was no way to manually enter a location (without GPS it was just a huge unsorted list).
These are two examples that come to my mind (I'm not a person who uses a lot of different apps, I'm sure there's more out there. Try rejecting location access for a few apps and see what happens)
Edit: I'm also not sure why you are so hung up on geo restrictions. I don't really care about geo restrictions, and even if, you could avoid them trivially with VPNs that are widely available. That just has nothing to do with keeping my GPS coordinates private.
This exact same feature you are asking for would be useful for people abducting children, who could make it look like the child was still at school while they took them away.
This is far more likely to be why Apple hasn’t implemented it.
Are you an Apple employer with access to confidential memos or meetings?
I can’t see how you would know this otherwise.
https://www.androidpolice.com/2021/03/22/android-12-isnt-sim...
Caveat: It's been almost 4 or 5 years since I used it. I used to prank colleagues on Facebook by posting from a local prison or other weird places. Have had no use for it since I deleted my fb account. So I don't know if it can still do what you want.
> What people really want is a phone that they can trust to be on their side implicitly.
Absolutely! These days our computers only do things that align with big company interests. I want a phone that will obey me even if it means ruin for its maker.
(Am I misinformed?)
There's a weird situation with mobile phones where for most people who aren't going to go out of their way to make their devices private, Apple's ecosystem is the right choice; it has better defaults and better sandboxing. But once you do decide to go out of your way to make your device private, Apple's ecosystem stops being the right choice.
On a rooted Android phone I have network-level adblocking on my device that directly works with the firewall. I also run Ublock Origin in my phone browser, which blows every single Safari adblocker out of the water. I can also cut off Internet access to specific apps and games, not just to domains. I also have access to a litany of Open Source apps that wouldn't be allowed on app stores like NewPipe that improve my privacy by replacing services. And LineageOS imposes additional sandboxing on top of Android's built-in features.
But if a friend buys a phone, they're not going to do any of that. They can't get NewPipe because they're not using F-Droid. They're probably going to stick with Chrome, which has next to no adblocking or extra privacy controls. An iPhone will be more private and more secure for them in almost every way.
So it's a weird situation where which device is more private depends a lot on who you are. For most average consumers, it's likely iOS. But it's not quite so simple to say that iOS has the best privacy overall.
The shortcoming of using Blockada is that I can't use another VPN. I don't really need it yet, but if I would that could be a big problem.
I don't know why Google doesn't allow chaining VPNs. Maybe the UX would be too difficult on a phone, maybe it's a real edge case not worth spending money on, maybe it would be used only to block ads.
https://developer.apple.com/documentation/safariservices/cre...
Quick appreciation regarding the intellectual humility required to add this to the end of your comment. In my humble opinion, this kind of attitude and approach represents some of the best of the HN community.
Plus it applies to Safari View Controller for in-app browsing.
https://www.imore.com/ios-9-content-blocking-extensions-expl...
Not only does Safari allow for 3rd-party adblock extensions that are pretty good, there are iOS browsers available, like iCab Mobile[0] for instance, that allow for extremely fine-grained, granular and tweakable ad-blocking, filtering, domain-blocking, and CSS and javascript rules - all of which can be user-edited and refined to fit exactly the user's own requirements.
There are always so many comments like the above in these threads: often characterized by people banging on iOS because it "can't be customized the way I want it after I paid so much for my iPhone". But these seem to be just standard consumer-mindset complaining from users that don't even have the intellectual curiosity to investigate whether there's actually anything already available that solves their imagined problem, or even try to fix it themselves - just as the article says.
[0] https://apps.apple.com/us/app/icab-mobile-web-browser/id3081...
Did you read my entire comment?
Can you see that there are entire browser apps that implement really granular and user-controllable ad-blocking.
In other words exactly the thing the poster I was responding to is complaining about.
Android also randomly breaks things now and then, but I can at least direct users to Firefox, which never seems to have any problems.
Please understand redirecting users to another browser is only ever a last resort. If I can fix the code I will, but sometimes something just stops working and I can't even reproduce it.
...and I have a teacher who is relying on my app for their classes this week and "try firefox" could save them...
This. No matter how convenient and useful current smartphones are in daily life, there's always the nagging discomfort by knowing you are carrying a half-nefarious spying device. One where even the so-called reputable vendors can't be trusted to not put in malicious adware, planned obsolecense, and do everything in their power to ensure you don't really own the thing.
About a year ago I switched from Windows to Linux for my home media PC, and it's such a relaxing experience. My computer just does what I tell it to do, and nothing changes unless I want it to.
By contrast, whenever I do have to boot into Windows to run some specific application, it's like I've invited a corporate representative from Microsoft into my home. Things are changed for me on updates. The OS suggests things to me I have no interest in. When I want to use Microsoft software, I have to log into accounts for some reason, which often means a round trip through email to reset the password. Once Windows even changed firmware settings without asking me which broke things in Linux.
The best things about Linux are not what it does, but what it doesn't do. I want this for my phone as well.
Do you? I don't. Microsoft is exclusively profit-oriented, they'll sell all your information in a heartbeat if they believe to have a legal way to do it and that it won't kill them in the next 5 years.
A group of open source maintainers (the more radical the better) will not. They will (usually) not polish the product as well, and they might be difficult to deal with if you found an issue (but you _can_ talk to them, unlike Microsoft).
I definitely trust open source people more, because they do their thing for ideological reasons and based on principles, not to maximize engagement and ad-spend and their yearly bonus.
This is why it's way more dangerous to give data to even open source apps that are encrypted and secure because if they ever do change your privacy policy and you don't pay attention to it, they could say they start selling your data and you continue to agree to it.
Have you really thought through that all the way?
Google has actually went even further in making it transparent where they show every bit of data they have on you. You can delete it all review it all have it auto delete, they don't need any of it
I'm running about 10 windows 10 machines, various builds, auto updates enabled. Some in a domain, some in a workgroup. The last time an update bricked a system, or even broke a feature, must be over a year ago.
Now on Ubuntu, I have the same experience as some other posters. I'm reasonably confident that I can update if I do it semi regularly. But the number of times I've experienced breakage that lead to hours of bugfixing during $DAYJOB, is way too high. Now everything is in its own tightly controller docker, no more random breakage (unless docker breaks, or a key expires...).
A debian 9 apt upgrade actually managed to render my install inoperable once. But that might be semi-unfair as it had to do with graphics drivers.
Are you running windows systems on semi-recent hardware? Or is it all decades old thinkpad laptops with shitty thirdparty drivers?
The same paranoia applies to my MacBook Pro as well that was provided by work. All user generated content gets backed up regularly to aid recovery if things go north.
If anything Google is so far ahead they are advocating more privacy and less data being sent if you choose, because they don't sell it directly to people like a local grocery store with its own app or Amazon does.
Federated learning AI is basically making it where Google can continue to have all the benefits of massive data mining without any personal information ever having to leak out
The parts involved with spying on you will update regardless of what you want or say. Unless of course you're not using the play store (or any other google services).
It's a huge advocate of keeping all my data private but now knowing how useful it is to anonymize, encrypt and use it all on my device with things like Google Pay, I strongly advocate if you're using Google or Samsung services to update.
Plus these updates usually are Android level and not related to Samsung or Google
My experience with Linux on the Desktop is that it wants to install updates. It doesn't force me to install updates, which I appreciate. However, eventually I also want to install updates. Unfortunately, on numerous occasions these updates have rendered the system unbootable. This is not a worthwhile tradeoff.
This has never happened to me with Windows. I'm aware that it does happen to some users, but I have trouble believing that I'm just really unlucky with Ubuntu/Debian/PopOS and Fedora.
If you want modern-looking LTS, try KDE Neon.
Across well over a decade of using Apt on Debian, Ubuntu & Mint, the amount of times I have seen Apt crash, is never. By contrast I have seen Apt tasks fail leaving packages partially installed. If relevant, it may seem a trivial, semantic difference, but I think it is a little more important than that.
Following the typical Microsoft paradigm of 'just reboot and hope for the best' will rarely result in the desired outcome, and in the case of something like grub could well end up with a system that will not boot.
For me a key difference in this case, is I have never encountered an unbootable Linux system (outside of hardware issues), that I could not fix with some basic tools. I can't say the same of Microsoft products (Personally, I don't consider a fresh install a fix :) ).
Are these "updates" or "upgrades"? If the problems are with updates, I would say that your experience is not typical of most Linux users.
Its a shame that you are having problems because the modern Linux experience is pretty slick.
I don't think you can blame linux, it's just a specific distro trying to do a task which is too hard and fail.
Not sure if it's still the case with ubuntu, I just run Arch which is a rolling release distro; I decide when to upgrade and what it implies + their info on breaking updates is always punctual.
No, of course you can't but that's also the problem. It's always someone else's fault.
If you try to upgrade Ubuntu you're told to RTFM. If you try to upgrade Raspbian, you're told you should have just installed a new image.
Linux distros routinely offer up options and features that do not work and there's no responsibility to be had.
With Windows, it's Microsoft's fault and they fix it (eventually).
I've never once ran into an issue with Fedora.
Fedora (at around version 20 iirc) was the absolute worst in my experience. I haven't used it since.
Saying "linux" has a poor update process or stability issues is meaningless.
> If, for example, you are running Arch Nightly and constantly pulling in the latest and greatest features, you're likely to see some stability issues. If you're using something like Ubuntu LTS, you'll miss out on a few new bells and whistles, but you'll get a very stable OS.
"You're doing it wrong, if you did it right it would work"
That's true of software in general. You can use (often older) more stable tech and lose access to some newer features, or you can used the latest and greatest and get a few more features and accept a reduction in stability in exchange.
IME, there's a lot of talking past each other when it comes to OSes. The vastly superior experience that Linux fans describe requires a couple of one-time boxes to be ticked, like constraining yourself to known compatible hardware platforms and stable software versions. There are plenty of people who don't realize this dynamic exists and trip over it, and it's not dodging the question to make them aware of these limitations and let them decide whether the cost/benefit fits their situation. Eg, in my case, I like Thinkpads, don't game, and need a stable, reliable, and performant system that I can be productive on. The pitfalls of Linux don't affect me personally, and I obviously can't meet the needs I describe with a Windows system (or to a lesser extent, OS X).
[1] I hear this is getting much better with things like Valve's Proton, but I assume that Linux is still far worse than Windows for a dedicated PC gamer that wants to play new games.
I'm not trying to bash on Linux, merely point out that for everything it does right there's just as much it does wrong. The same is true of Windows and MacOS.
And to be fair, updates broke my Windows systems multiple times in the past as well. This isn't really a problem by itself since errors and mistakes happen, the problem is that the state of recovery tools in both systems is somewhat lacking. Linux can be repaired from Live USB and has Timeshift, but it doesn't have an easily accessible built-in rescue environment that's actually useful. Windows has DaRT, but all decent Live USB's for it are third-party and it's very hard to repair its internals if the existing tools don't work. FreeBSD is probably the closest to being bulletproof, but it has its own problems.
"Unlucky" is probably the wrong word, since I doubt all those failures were independent, but you're presumably in some niche with a high probability of update failure (just as the users failing their Windows updates are). For a counter anecdote, I've used Linux laptops/workstations (primarily Debian/Ubuntu, but dabbled in others) exclusively for 15 years and have never had an update break _anything_.
If I had to guess, I'd say that it's a hardware compatibility issue? Linux can still be a _terrible_ experience relative to Windows if you have strong opinions about the hardware you want to use. IME, all the people who talk about Linux being a vastly superior user experience to OS X/Windows are those whose hardware preferences line up with good Linux support.
Lucky you! When this happened to me with Windows last time, I had no idea what to do and how to diagnose it even though I spent a lot of time on it and had the knowledge of all of the Internet available. I had to reinstall and painstakingly redo my whole setup.
When it happened to me with GNU/Linux, it gave me enough rope to actually debug what happened, so I could fix it and get back to where I was before. And I went through several Ubuntu and Debian dist-upgrades, and now I'm even using Arch which is a rolling distro that's somewhat expected to break from time to time.
So sure, it does happen more often than it did on Windows - but I don't think I actually waste more time with that than I did when using Windows.
Its time to upgrade your PC. Now. We don't care you're running a presentation. yeah We'll let you delay it... for now.
Its time to dim your screen. We don't care if you're reading a recipe while you cook. No we won't put a widget to allow this temporarily. We think its better buried in the settings. And you'll have to set it back later.
Its time for feature X to not work or not exist anywhere. We don't care that it worked in the old version. We don't like it.
> We think its better buried in the settings. And you'll have to set it back later.
A criticism of the Settings App as it stands today based on your usage of it a year ago is very unfair as it's continuously improving and evolving.
How do you know his/her opinion is based on year-old usage?
As MS has slowly whittled down and removed features from the Control Panel, they have simultaneously added analogs where appropriate in the Settings App. The Settings App in 20H2 is very different and far from useless when compared to what shipped a year ago, or when Windows 10 launched.
You shouldn't need a third-party utility for re-disabling settings after updates, and it's pretty clear at this point that Microsoft just doesn't give a damn about respecting the user's choices.
There are settings that only apply to some "apps", there are "app wide settings", there are important settings hidden 4 layers deep under some category where one would never suspect them to be.
Then there's the fact that MS has the tendency to just silently revert settings, often privacy and telemetry related, after updates. On that end, Windows 10 very much feels like it's actively working against the user, it feels like you are only a tolerated guest on your own system.
This is so user-hostile it's not even funny. Imagine if you had someone in your life who silently went back on agreements when you weren't looking. You would work to remove them from your life as soon as possible.
My roommate in college refused the Windows update request enough times that it just took the decision out of his hands and _told_ him it was rebooting...in the middle of a LAN Starcraft game we were having. (Hilariously, Starcraft also ran perfectly on my Wine/Ubuntu and had all sort of weird bugs and visual artifacts on his Vista/7)
I truly can't imagine how anyone can consider a system that reboots without your consent, no matter whether you're playing Starcraft or giving a presentation or launching nuclear warheads, to be appropriate for serious work. It seems like one of the very basic jobs of an OS to _not turn off while you might be doing something important_. Perhaps the answer is as simple as "there's a setting buried somewhere that all technically-literate Windows users change", but the default behavior is beyond insane.
With all the editions and versions of Windows 10, there doesn't appear to be a single definitive answer to that.
I for one settled on https://github.com/marcosbozzani/Win10ActiveHours , which prevents reboots by scheduling a task to update Windows' own "Active Hours" twice a day.
But wow, phones and computers nowadays are like a giant steamroller, squashing users common sense, interests and privacy.
Maybe, even probably, no one have any nefarious motives behind the latest changes, but someone decided that if I want to upgrade my system now I have to start using systemd-resolve, and while I can and probably will alter it so it won't bother me, I wouldn't be installing it on my own.
Also, Firefox is the king of unwanted changes with little to no warnings, and while technically it isn't Linux, one day an innocent yum or dnf update install just killed my customized setup, flagged all my beloved addons as unsupported, and left them in disabled state.
>whenever I do have to boot into Windows to run some specific application, it's like I've invited a corporate representative from Microsoft into my home. Things are changed for me on updates
Yeah, it is really nice to have a consistent interface. But there is a flip side. Sometimes, people have to change in light of new security problems, and these sometimes need to be done quickly. The most obnoxious of these updates are interface changes (both UI and API). I think Windows does this too much and maybe for non-security related issues. The only real example I can think of is how OpenSSL uses envelope functions because ppl would abuse the primitives [Source: Alexandra Boldyreva, Christopher Patton, Thomas Shrimpton: Hedging Public-Key Encryption in the Real World].
May I inquire about the name of this app? Is it desktop-environment specific?
Definitely a refreshing feeling compared to the use of a million cloud services.
This will never happen. Personal radio communication is innocuous compared to what a modern cell phone is, yet personal radio communication is heavily restricted since decades.
Maybe we will eventually see an open source radio stack and truly open source apps. This will only happen once surveillance and kill switches have been moved to another layer.
A cell phone will never be on our side, it is far to dangerous.
If you’re contrasting it with the HAM/Shortwave community, there’s an obvious difference in transmission power. The Electromagnetic Spectrum is a public resource we have to share intelligently. All the core tech in a cell phone is about how to use spectrum efficiently, delivering high speeds with minimal interference. Shortwave is much less sophisticated and operates at much higher power. There are real costs to amateur radio enthusiasts blasting transmissions with incorrectly configured equipment.
You could use proof-carrying code or reproducible artifacts to keep the radio stack out of the users' control while still making it obvious what it is that radio stack does.
The BIOS analog would then only accept radio code that behaves in a particular way, but it would be clear from the (open) source that the radio code is not spying on you.
Radio firmwares can be open source. It's a red herring.
You nerd-sniped me there. I'm now trying to come up with some consistent metric of the nefariousness fraction of a product/service...
What I feel when using Android.
It's all about for who's benefit the product is built. I like the phrasing of having a phone "on my side".
In a recent thread here on obsolete hardware I was commenting how up to the fairly recent past products were built with the idea of competing on features the user wanted (or at least presumably wanted).
But no longer - in the last ten years that has shifted in too many industries to be all about features that the company wants (spyware, tracking, DRM, cloud lock-in, and on and on). The phone duopoly is a poster child of this evil trend. Neither platform puts the user first and into the drivers seat.
So no, I don't want an open source phone nor a Linux phone, specifically. (Not saying I'd mind if one or both are true, that'd be cool, but not a deciding factor.) I want a phone where I the customer are in full control of what it does. I must be able to override anything the OS wants to do and to override anything any app wants to do.
Or maybe because Vulkan was a standard so long and late in the making that Apple released their own graphical stack a year before the Vulkan was even announced?
100% this.
Why is it so difficult to copy files to/from my phone?
Why is it so difficult to list all the files in a folder?
Why is it so difficult to get a list of what's running?
_Why is it so difficult to know that my phone is working for me_?
Because they want you to have no choice but use their cloud storage
> Why is it so difficult to list all the files in a folder?
Because they don't want you to know what is a folder, just to type everything in their cloud storage search engine so you are even more locked in
I couldn't agree with "people want a phone that is on their side" more.
I also agree that Linux matters as much as iOS's BSD heart—that is, it doesn't. But this is also why the Vulkan example doesn't serve your point. As with Linux, technical implementation details aren't why people are using iPhones.
Most people use iPhones because of the complete experience. For the few who care whether the iPhone is "on their side", it is clear that Apple's business model is generally aligned with their interests. This is either unknown or untrue for alternatives.
This is where open source phone vendors will fail. Trust is not simply a technology problem.
A better example might be places where it's hard to switch away from Apple's systems, or perhaps how they only support hardware that doesn't last very long and are notoriously hostile to attempts to repair your device.
But it's also not going to, for example, let me install a web browser which is anything other than a Safari reskin.
There are other factors to 'being on the users side' than just privacy.
I have a phone, I don't use it for too many purposes, whatever software that is installed on it doesn't bother me much because there are larger pie to fry.
Now desktop OS and laptop choices/ cloud native stuff is a hill I'm willing to die on, because I'm a lot more invested.
Same reason I don't have an issue using iPad as development machine with termux.
Because your argument makes no sense at all as to why I might want a Linux phone. I want a Linux phone for the same reasons why I use a Linux laptop and PC. Freedom, control over my computing environment and my data, that sort of thing.
Taking a step back, there is no mobile OS out there that provides the things that I want. The closest that I've seen is Linux for the PinePhone. So, yeah, I want a Linux phone. It's not that my initial premise was "I want a Linux phone". It's "I want control and I want to have the freedom to do what I want with my device, so what system currently offers that kind of freedom while being feasible on technical and usability levels?"
Maybe people also do just want general open-source computing, and Linux just happens to be the one that's furthest along, its philosophy and ideology be damned.
Personally, I'm hoping Haiku eventually gets further along to be usable as a Linux alternative, and it's fun to wonder what it would be like as a mobile OS.
Yes, KDE has copied Windows features.
But: they made it better (for people like me). I didn't get Windows, but something that happened to look similar but work better.
And the latest big Windows update copied a good deal of styling from KDE Plasma so they are clearly doing something interesting :-)
I imagine it's because, when the year of the Linux desktop didn't arrive, developers everywhere decided they could make it happen if only they tried hard enough to put together the perfect UI, and so the kitchen-sink approach proliferated.
> Or back in the '90s/'00s, attempts towards recreating the Windows UI/UX?
I take exception to this one. In fact it was Microsoft with Windows 7 (released in 2007) which copied much of the UI of older Linux desktops like KDE and GNOME versions 1.0 (circa 1997). See:
https://linuxgazette.net/165/misc/laycock/gnome.png https://images.channelpartner.de/bdb/490847/1066x600.jpg
Linux desktop environments were better back then. XFce is the only tolerable one I can think of, today. Even there, I hate the defaults and preferred it in the XFce3 days when it was just an (improved) CDE clone. Fluxbox or bust!
There's also ReactOS, though admittedly it's a non-Linux open source OS.
I personally use Elementary daily and I really don't like OSX, they only superficially look similar but it's a very different mindset and UX.
Why the hell do you think those things are related? I can easily make my OS look macOS, windows, something weird inbetween or some pink vapourwave hell from the 90's or some minimalist tiling WM and that's the beauty of it. That's exactly one of the things I want.
I don't use Linux because of a particular DE with presets that let's me mimic windows. If that was all why wouldn't I just use windows.
> I don't think most people who buy a Linux phone actually want a Linux phone. They want an open-source phone. The Linux philosophy isn't important to them, they essentially want open-source iOS
Seems like the phenomenon of people choosing Linux, only to choose a windows manager that makes the entire thing look like a non-open OS, fits this same line of thinking. At the end of the day, most people don't care about the particulars of Linux, or its ideologies. They just want an open and customizable OS, and if there's a variant that looks like a more widely-used OS, so much as the better.
If there were more open OS's that were as far along as Linux, it would likely be less popular. The important thing is open-source phones or computers, not Linux ones.
Which, sure. I could (and do) use LineageOS right now. It's not as good. I still need to go through the Android app development process, I'm still fighting the system every step of the way. It's still a pain to de-Google things, I still have to deal with an architecture that if fundamentally designed to work best with Google services. I still have to deal with what I see as design flaws in Android itself.
So I don't technically need this to be Linux. I recently bought a smartwatch that runs on an ESP32 chip and is programmed using Arduino code/C++. It's not running Linux, but it fits a lot of my needs and it will communicate well with my Linux computer. And you're right, I didn't have serious qualms about that because of the kernel. In fact, the simpler dev process was what made it attractive.
However, there isn't that kind of thing for phones. If you want an Open Source phone that is easy to modify and develop on, Linux is the option that's pushing in that direction.
> but they don't necessarily want to open a PR against the camera app if they want it to more closely match the polish of iOS.
It's deeper than that: I want the ability to fix my own problems. I don't want to open a pull request, I want software that is understandable enough and an ecosystem that is open enough and broad enough that it is reasonable for me to fix my own problems. I (personally) can do that on Linux, but I can't (currently) do it on Android devices, they're a pain to work with and I hate them.
I am used to Linux on my computers, including the fragmentation of software. Sometimes the fragmentation is an advantage because it means there's a diversity of software that is hyper-specialized rather than one or two solutions that are built to kind-of satisfy "most" people's needs. More than Linux itself, what I am used to is the idea that if something goes wrong or if I want my computer to do something, I can make it happen without asking anybody else's permission, without building a giant project that I need to invest serious time into, that I can pull things apart and build pipelines with Unixy tools.
I want that on my phone. It doesn't have to be Linux, but nothing else is providing it, so it does have to be Linux unless someone else is going to build a better OS. I'm not holding my breath for that.
Some of these devices already have 12 GB RAM and octacore processors, that's more than a Raspberry Pi.
When these devices reach their EOL as a phone/tablet, they would make wonderful SBCs with integrated touchscreen, camera and microphone, and in the future also good AI co-processors for good on-device speech or face recognition. All this with an up-to-date OS, and not an Android version which is over 4 years old.
Whatever I try with Android, sooner or later the processes get killed.
What's worst is that when Chrome gets updated, any app using a WebView gets killed. Why not just let it continue running and make it use the new WebView after it gets started again manually?
That whole process lifecycle thing that you have to jump through with activities and bundles and what have you is so that your app disappears not just out of RAM, but out of the scheduler too from a piece of code that can make such decisions at a complete system level. Yes, it kills processes essentially arbitrarily, but the processes should have all of the hooks and places to squirrel away state to come back from the aether as if nothing happened. Programming Android code should embrace that rather than fight it.
As for webview, I think they're generally pushing a security fix and think that your app should be able to pull itself back together, the same as if a browser was refreshed.
I can't think of work that needs to be performed that requires a GUI to constantly be loaded, but doesn't require that the GUI be in the foreground.
It shows stuff like weather, clock, calendar, land-line phone calls and has buttons for light scenes.
I used to use Chrome for this, but I always had to manually maximize the page, and hosting it in a custom app has additional benefits, like loading and maximizing the page as soon as the device boots, and offering a JavaScript interface so that the page can interact with Java functionality. It's much better that just hosting the page in Chrome.
Each time Chrome gets updated, the app gets killed. And even though the app has set up an alarm which makes Android call into the app every minute to ensure it is running, for some reason this doesn't bring up the UI (probably something which I need to debug).
But I know that if this app were running on Linux directly, in some Python Qt app with a Qt WebView, it would never get killed, as it has plenty of resources available.
I've also considered integrating Mozilla's rendering engine into the app, but decided not to, I can't remember why, but there were some drawbacks in doing so.
It would be hard to imagine using such an app in an industrial setting, where the app would need to be visible all the time, displaying gauges and stuff like that. The hardware could do that, with a proper OS.
Try postmarketos, it's designed to run vanilla Linux on old devices such as phones.
CPU and USB OTG generally works, which already make for a pretty general-purpose machine. The touchscreen also works most of the time. The rest is generally hit-or-miss, depending on how much time you or the maintainer spends on making it work, or if there has been a mainlining effort.
Would be very strange that in the vast group of Linux users you can make a claim that "most of them want X". Seeing the amount of fragmentation happening out there, a lot of them have different priorities.
> They want the ability to make modifications, and they want the freedom and transparency that open-source...
Look at all the answers about caring about the user's freedom, privacy and security.
This is exactly what Free Software is about.
That's true, but to some degree. I'm OK with LineageOS, wouldn't care if that has Linux or BSD. As long as I can switch distro, reinstall and do not lose warranty or permanently damage bootloader.
Linux gives that option. None of the other phones do.
Linux's rough patches on the desktop/laptop is one thing, and for me - easy to deal with.
on a mobile device, the polish matters in usability in big way. Running into a problem on a phone is infinitely more of a PITA than a machine with a sizeable screen and a full keyboard attached, where remote access is easy if necessary.
I did all of this back from December to February.
I still have no idea when Debian will bother to look at the mmsd packaging to include it into their repository, nor when my changes for chatty will be integrated. The ofono mailing list and IRC channel (i.e. the original authors of mmsd, but i think they abandoned it as they have said it hasn't been maintained in 8 years) have ignored me, so I doubt my changes will ever go back into upstream mmsd (I just consider what I have a fork now). I'm waiting on the chatty dev to finish off a current project to then get to integrating my MMS patches into chatty.
I still really want to help, but there's not much for me to do except wait for upstream devs to get around to integrating my work.
Being frank, I'm frustrated that I spent all of that time to implement a feature I want and I know a lot of other folks want, but now I'm just in a limbo hoping that others will get around to it.
So I'll ask, if you want folks to help on the boring things or missing functionality, how do you prevent that from happening?
The mms stuff in postmarketOS is also still in progress as far as I know, I'm un europe myself, like a lot of the pinephone developers and mms just isn't really a thing which makes integrating these thing quite hard.
Do you have a link so I can look it up?
I get it, and I get your frustrations too, that's why I just picked it up myself. I really do want to help.
https://source.puri.sm/kop316/mmsd/
There is my mmsd fork.
https://source.puri.sm/kop316/chatty/-/tree/wip/sadiq/mm-acc...
This is where I have my reference MMS implementation for chatty.
You are also welcome to get a hold of me on Matrix, I am on the pmOS matrix channel!
The issue right now is that Debian maintainers are all pretty busy trying to nail down Bullseye as they're currently in their hard freeze[1] period... lots of bugs and dependency problems to deal with to get to release.
My suggestion would be to wait until Bullseye is released and then contact the maintainer and see if you need to be dealing with them or upstream to get your changes in. If you're impatient, you might have more luck contacting upstream directly right now. Either way, the odds of getting new features in for Bullseye are likely slim to none right now.
So I may be in a weird spot then. I am trying to get mmsd (not modem manager) into Debian proper. Mmsd itself is not in Debian at all.
Per my original comment, I have tried to contact the ofono (mmsd) maintainers via IRC and their mailing list no less than four times to get my patches upstreamed (every time I have been ignored), and at this point I have given up trying to contact them. One of the ofono maintainers has mentioned that mmsd hasn't been worked on in over 8 years, so I think its abandoned.
The Debian on Mobile (i.e. the Mobian and Purism devs) maintainers think the best course of action is to package upstream mmsd, then update with my changes.
How do you get your changes to mmsd into the upstream mmsd repository on kernel.org? There have been commits to mmsd upstream in 2019. My suggestion would be to CC more folks on your next patch submission, at minimum: the Linux netdev mailing list, the oFono list, everyone in the mmsd AUTHORS file and all the folks who are listed as a committer in any of the mmsd git commits.
https://git.kernel.org/pub/scm/network/ofono/mmsd.git
How do you get mmsd into Debian? The usual procedure is to upload the package to the mentors site and file a request for sponsor bug report, XCCed to relevant people/lists, which in your case is probably the debian-mobile list, Debian oFono maintainers and any other Debian mobile related addresses (I guess Purism PureOS, Mobian etc). Of course, that doesn't guarantee that a Debian uploader will sponsor the package, but it is a good start.
https://mentors.debian.net/intro-maintainers https://mentors.debian.net/sponsors/rfs-howto https://www.debian.org/Bugs/Reporting#xcc https://lists.debian.org/debian-mobile/
Do both at the same time, if upstreaming the patches doesn't work out, you can add them to Debian while still figuring out a way to push them upstream.
So you can see the ITP packaging request here:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=982250
The debian-on-mobile team is aware of it, but even they have been commenting that getting upstream mmsd will take time. One has suggested to change the bug for a request for sponsor. It sounds like I need to do that.
The ITP comes before you have started the packaging, it is a notice of intent rather than a request and functions as a way to prevent other folks from spending time on something someone else is spending time on, a place to discuss and document any issues that come up during packaging and a way to advertise the potential new package such that others can get involved and co-maintain the package if they want to. Since you didn't XCC the ITP to debian-devel as is customary (and what reportbug wnpp does), that part of the purpose of ITPs was reduced.
The RFS comes after you consider the packaging complete and ready to be in the Debian archive. It is only for people who aren't yet able to upload themselves. It should be sent once each time you want to upload. It should be directed at debian-mentors (done automatically when using reportbug) and XCCed to the relevant team mailing lists and any other potential sponsors.
I did make am RFS:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=985893
and I want to say I did the XCC to the relevant sponsors? I know I included th Debian on Mobile mailing list (they are the ones to suggest how to submit debian packages).
I also did the RFS email.
Thank you for your help. I really do hope my package and my patches get somewhere.
I'll take a look at what pabs3 said, it looks like I have a couple of things to do today for this.
I know a lot of people were looking for MMS support.
At their suggestion, I have been working to get my mmsd package into Debian proper.
The problem is not with the Linux phones that are out there, it is that no apps are and likely never will be compatible with (non-Android) Linux phones. Until that is solved, most people (like me) could not switch out their phone for a Linux phone as a daily driver, however much they would like.
Having Linux as an alternative for Windows is a lot easier than having Linux as an alternative for Android/iOS. A lot of the Windows applications we use can be replaced by opensource applications on Linux. And browser based applications already work on both.
However, apps on Android/iOS are mostly tied to cloud services. Unfortunately services that most people need in this age. Messaging systems like WhatsApp and Signal are usually tied to your phone, and going without them would result in social isolation for a lot of people. But also for banking and sometimes government services, you likely need to use their apps. And there will never be opensource alternatives for such apps.
It is very unfortunate that so much of the Android ecosystem relies on a very proprietary and cloud-connected piece of software. Without it, most apps simply refuse to work. MicroG has done amazing things here, but unfortunately a lot of things still break when using it, to no fault of the MicroG developers.
Modern bank apps require not just Google Play Services, they also require the phone to pass Safety Net. Play Services substitutes like MicroG aren’t capable of that.
The truth is that service and app developers only care about one thing: audience size, which translates to actual money. And it takes a long time to build that audience, even more so in the mobile world where all metrics are orders of magnitude bigger than on the desktop. Until Linux phones can show they have sizeable audiences, developers will ignore them. Tens of thousands is nothing, we’ll need millions. Realistically, this is not going to change in 12 months or even 2 years.
But, something can be done in the meantime: making all popular app-development frameworks run on mobile Linux, and then putting pressure on developers to produce app builds. When portability comes essentially for free, it becomes easy to support an additional platform just for the brownie points.
download the apk from https://www.whatsapp.com/android/, start anbox (this takes a while), install android-tools, adb install WhatsApp.apk, done. It also works with stuff like Duo Mobile, which is the 2FA my organization uses (which is interesting because it doesn't agree with my rooted lineageOS phone).
Anbox comes by default on current versions of Manjaro
Replicant? https://replicant.us/
LineageOS? https://lineageos.org/
LineageOS is what I'm currently using on my daily driver, as a compromise. I would very much prefer a Linux phone, but in the current state it can't even fulfill half my needs in a smartphone.
If that's a problem, are you aware of microG? https://microg.org/
The web is already almost as closed (for apps) as phones are - you can choose Chrome, Safari if you're on Mac, or Firefox. Everything else is just repackaged Chrome. Edit: forgot Opera Mini, for what it's worth.
Firefox is dying out, and Google is pushing for less and less user freedom in Chrome, while Apple is extremely reluctant to implement new features that make Safari a better app platform.
Web Bluetooth
Web MIDI API
Magnetometer API
Web NFC API
Device Memory API
Network Information API
Battery Status API
Ambient Light Sensor
HDCP Policy Check extension for EME
Proximity Sensor
WebHID
Serial API
Web USB
Geolocation Sensor (background geolocation)
User Idle Detection
What? Nobody needs permission to publish a webpage, while every week we hear about google or apple blocking or unpublishing apps with flimsy justifications.
(Likely wrong terminology but you hopefully get the idea)
In a perfect world I'd have a phone that is analogous. As iPhone is to Apple, and android is to windows, my phone would be to whatever linux I use. Well, that's a tall order.
And because it's a tall order, the most efficient way to go through features is to start off with easy highly requested features and very slowly move toward the less important features that are harder to implement.
Phones with linux hits both sides of this equation. On one hand it turns out EVEN THOSE things are hard. On the other hand, people don't WANT to work on those things as much because free developers rarely want to work on other people's problems. I mean, some will, and those people are simply better people IMO, but rarely do they WANT to.
Sounds like linux to me!
Linux works exceptionally well in the areas where large companies with many developers pay to make it work well.
Put in the elbow grease, sponsor someone else to put in the elbow grease, or settle for non-free software where your itches seem to come pre-scratched. Personally I encourage you to scratch your own itch and contribute back.
I think the other part is that people see it as an opportunity to “make it big” so to speak. If you develop some small part of the UI or functionality for someone else’s OS you get a pat on the back and a commit message with your name on it. If you develop the distribution that becomes popular, you get to name it. You get to go to conferences and tell everyone how you started it. You give keynote speeches. You can start a company to make money off the distro. It opens a lot of doors.
So with that in mind, with only 18 distros, why wouldn’t I throw a 19th into the ring vs toiling away at fixing just my own annoyances?
The elephant in the room is that we desperately need better funding mechanisms. For instance, most distro repos don't support selling Free Software - either the package is available gratis to everyone, or it's not available to anyone. There are hacks to get around it perhaps, but no support at the OOTB click-a-button level.
The only way the Year Of The Linux Desktop will ever happen is if we actually make it a priority as a community to figure out ethical funding mechanisms.
https://www.networkworld.com/article/2224460/linux-users-don...
It's almost certainly profit. They have a sizable user base but not a lot of good ways of getting money out of them. I would say it's a direct example of the major unsolved problem I was talking about. This is not to condone what they did, as I found the Ubuntu advertisements particularly egregious.
I was in fact hoping that Ubuntu would be that distro a number of years ago until they pivoted away from easy, nice, clean, just works towards halfway bad copy of OS X.
Now another frustrating thing about this is that once many open source projects get to taste the sweet juice of cashflow it seems many will immediately abandon all the ideas that got me to desperately want to suppprt them and become just as hard to work with as their commercial counterparts.
Random thought: I wonder if the problem is personal version of "selling support" is home PC repair companies, ubuntu is trying to be MS but they should be GeekSquad or geeks4u.
There's room for a phone linux, a user/corporate client linux, an enterprise linux or two. It's pretty tough for a company, even the well funded ones to do all three. MS missed phone, google missed enterprise, redhat doesn't really do phone or client (fedora I guess but it's not really a paid product). Then, you got scrappy ubuntu that's gonna reinvent phones and debian and do server support, and they are kinda just bad at all of it . . ..
If you want to set up a browsing/email machine for something it's a great choice.
This is actually where I think ubuntu should focus, the server offering is kinda meh. The upstart and unity shit was a waste of time. Instead, they should focus all that energy on being a client OS. One that can run on grandmas old machine and one that can run in the enterprise.
It's just a little half done for my taste
Wouldn't you say the Freemium model is a successful business model for many? Or there are free software but we can do extra for you like Automatic/Wordfence etc. Ad driven 'free' software is quite common place. Also 'goodwill' supported.
Not saying its not challenging for many free software services but there are business models around 'free'.
Not when it comes to desktop software, it seems.
This is SaaS though, and much of it proprietary as I understand.
>Firefox
Seems to be sustained in large part by the Google search subsidy, and implicitly Google's desire not to be a true monopoly with Chrome.
>Opera
Isn't Opera proprietary?
I didn't mention it because it's uhhhh, antithetical to Free Software. Let me explain that:
Free Software says "put ultimate power in the users' hands, so that they can call the shots" - this is why they commonly use the GPL (sorry if obvious). Ads/tracking puts the devs at the mercy of the third parties who are paying their bills, despite said third parties not giving a shit about the quality of the software. It's why ads are so often unoptimized.
A more minor reason is that it's situational. Anything offline basically can't do either. Anything like libpng just doesn't really make sense.
But fundamentally, the crux is that devs have to do what they're paid to do. There are exceptions, but those are only exceptions and not the rule.
I have been using Linux for 10 years or so and a lot of my friends/colleagues use Linux and we use Linux because software for which we are paid to work - works best on Linux. I don't think beyond college students, professional programmers using Linux mind paying for software (I pay for several!). Most of us know - there are warts and all but I have also used Mac for 5 years and my wife uses Mac for her work and it does not look like fun wresting with brew everyday.
The problem isn't that people aren't willing to pay for software - IMO the problem is, Linux is really difficult to support, because of Linux Desktop being a continuous moving target. So, application development for Linux is hard because the base is continuously changing but no company has managed to invest enough money/resources to make the base stable. And given small Linux marketshare, resources required to make a stable base is astronomical (Like google did with Android).
So, It is a chicken and egg problem. Linux can't attract large install base without a wide variety of application software, but companies who have tried supporting Linux application development have burnt themselves in past (and why bother given small Linux install base) and hence nobody has incentive to make that solid base.
Plenty of Linux users I know do.
>People for whom cost is the primary concern can pirate Windows, buy cheap versions from online stores or install a legitimate version of Windows and never activate it(Microsoft makes it really easy these days). Heck already probably millions of people are using one of these options rather than installing Linux.
Not if they're programmers and prefer Unix. My experience with Linux users is different than yours. Yes, many of them end up caring about free software. Quite a few also have the "why pay for anything" attitude and pirate all their media, even when they could afford it. Sometimes these groups overlap. I don't think there's one single reason people come to Linux.
If people are willing to pay money for Linux, how come companies trying to support it haven't been successful? Would you pay $150 for a copy of a Linux distro? $200? How many of your Linux-using friends would? Would you be willing to do that every three years, like Windows?
The attitude I encounter frequently among Linux users is not just that they wouldn't pay but that they don't think money is a solution to the problems Linux has. So far, the most critical Linux development has been subsidized by proprietary software companies who use it for their infrastructure. Nobody's interested in supporting it for money because nobody wants to pay for it. The whole reason it's popular is because you don't have to. Linux would be nothing today if Linus had charged for it. There's no chicken and egg problem here. The egg is a community which doesn't value software in material terms.
If every one of those programmers paid $150 for their copy of Linux Desktop - do you think it would enough to fund a Company that could build/maintain a rock-solid-stable-base Linux Desktop? The problem is that developing/supporting an OS requires enormous engineering resources and you magically don't require less resources just because your install base is small.
Currently libinput is maintained by one man mostly and I have a suspicion that Mirosoft has a small team for input devices. Take X11 vs Wayland for example. What would this new distro target? If it targets wayland - users will complain about nvidia GPU support, if it targets X11 - users will complain about fractional scaling/tearing. These are hard engineering problems which require working with partner drivers and long term commitment. I personally think that - many companies have realized there is no money to be made in supporting desktop Linux(i.e desktop itself, not a cross-platform application running on top of it), even if every one of those users paid the one time fee.
It wouldn't be a bad place to start.
>The problem is that developing/supporting an OS requires enormous engineering resources and you magically don't require less resources just because your install base is small.
Well, you have to narrow the hardware you support. That's the main thing that makes Linux on the desktop suck. Only Microsoft has succeeded in writing a highly compatible operating system, as you mention with a lot of resources. But there isn't the will for this, because people want Linux to run on everything. In this area, System76 are probably doing the best. But their hardware is not very good.
>I personally think that - many companies have realized there is no money to be made in supporting desktop Linux(i.e desktop itself, not a cross-platform application running on top of it), even if every one of those users paid the one time fee.
The point is that the fee is off the table because nobody would pay it. That's all I'm saying. Your argument is a non-sequitur. You say these engineering problems require hard work, and that there's not enough money in it even if everyone paid the fee. In that case, one would need to raise the fee until it can pay for the development that needs to be done.
But the market would not bear it. It already will not bear a standard operating system price.
Hopefully they won’t repeatedly change their minds before then, like openmoko did.
I think my disappointment stemmed from this - I was expecting a (perhaps rudimentary, perhaps unpolished) platform that would make a base to start developing things on, but that wasn't there.
I'll keep an eye on pinephone, hopefully they've done their historical research in this area :)
It's not yet a mature ecosystem, but desktop linux is, and there's plenty to borrow from there.
And also: cryptocurrency. ~4,000 coins as of today.
Here we go again. App developers now have to 'define' support for a distribution target and test each of them to show it is actually 'supported' and not showing 'experimental support'.
Fantastic news for the technical consumer and the whole open-source crowd, Not so good for the app developers who need to test their apps on these numerous distros and bewildering for general users looking at this technical contraption that is likely not to have the same apps on their old phones.
If they want to solve this like it or not, supporting the Flutter ecosystem sounds like a good idea.
> Do you want Linux phones?
If it gets in the way of things and is harder to use than the alternatives, then no thanks and no deal. There is more things to consider than just 'open-source' or the 'free software' argument.
Probably close to that percentage are either better as a website or actually harmful garbage.
> Is the app ecosystem ready?
Here's the question you ask: Would you be happier if your thinkpad running Alpine Linux fit in your pocket and could do sms/calls? It has the same answer.
I was a Lumia Windows Phone user, and I loved the OS, and I'm not at all some avid app user on the phone (I don't like mobile games, I don't install random apps that are just a packaged web site). Still, every time I got out of a well-worn bubble, the phone felt second-class. I couldn't order a cab, couldn't scan that coupon, I couldn't install that dating app, etc etc. I used that phone for 3 or 4 years until the battery gave out or whatever, but moving to iOS and Android felt so empowering.
I can only imagine the experience on Linux, without any corporate pushing to get your app on their phone, would be much worse.
Back in the Nokia N900 days I did a great deal of hacking on that phone, but that device (even though its specs are much lower than a modern phone’s) had a much more responsive UI and it was actually enjoyable to test out what you had developed.
What's quickly become clear to me is that all of that stuff is futile until we have a better way to build and run applications. The existing shells and toolkits for building applications are simply too heavy for the hardware, and I don't see the situation changing much any time soon. The hardware ought to be powerful enough to do more than it can, but it's going to take a lot more effort to build more efficient underlying components.
Unfortunately, as an individual developer with several other large side projects going on at the moment I don't expect to be able to contribute much to this anytime soon, but I hope that the discussion will prompt some people with the spare cycles to pick this up and run with it.
The build time for megapixels for a clean build is maybe 5 seconds, way less on incremental builds. Editor runs fast, Test cycle is incredibly fast.
It's certainly a lot faster than I've ever been able to test android apps.
It's a bit odd considering phrases like:
> It's basically my first C project, first v4l2 project and first GTK3 project but I started it because nobody who actually knows any of these things want to, and so far nobody who knows these things has stepped in to do better (...)
(I probably do not know any of these things better - but if I did, and wanted to help, I'd probably click on the link to the github profile, not find megapixels source... And move on?)
sshfs gets executables automatically exported to the phone as they are linked.
Doing native builds on the phone seems like deliberately choosing to do it the hard way. [But somebody says GTK will not cross-build. Could that be possible?]
GTK 3 cannot currently be cross built due to some sort of dependency issue with Python, but random GTK apps can be cross-built just fine, and have been by Debian cross-build QA efforts:
http://crossqa.debian.net/src/gtk+3.0 http://crossqa.debian.net/src/xdemorse
Here are some links related to cross-building for Debian:
http://crossqa.debian.net/ https://wiki.debian.org/CrossCompiling https://wiki.debian.org/DebianBootstrap https://wiki.debian.org/HelmutGrohne/rebootstrap
None of them has been a great product. The software has limited features and there are few apps. The lack of manpower in development cannot be hidden. I cannot buy a subway ticket in several cities and I get difficulties with banks, because I have no Google playstore and we live in a world worse than Orwell's 1984 or Fahrenheit 451 could imagine.
When I want to tinker I have a shell, can become root and I know what I am doing.
Yes, I want a Linux phone.
>None of them has been a great product
The a1200 ming was pretty slick for its time. It had a 2MP camera when everything else shipped with 0.3MP at best. Touch screen, stylus, see through flip so you could see the screen when closed. That was a pretty cool phone.
You have read those books right?
For Fahrenheit 451 I have seen the movie twice and I remember it quite well. No idea how much it is different from the book, have not read it.
Of course not all aspects of those fiction works apply, especially outside of China, Russia and the like. But Orwell would be not have foreseen how Google/Apple/Facebook can use positioning info or face detection of their users or how Russian internet trolls can influence US presidential elections or German vaccination opponents.
It's such a shame we're locked into platform-specific apps. If only there was some way banks etc could make some kind off app that runs on any platform.
What if we had some sort of... app like thing - documents with interactive elements and media, even code that runs (safely sandboxed), but not tied to any one platform.
We could have one central app, a platform within a platform, that fetches and displays these documents, which are made according to a set of specifications and standards that can be implemented freely by anyone. It wouldn't be easy or cheap to do so, but the engines could be open source so that new OSes could still run them.
I don't know... it's probably a crazy, unrealistic idea.
Sadly, since no such thing exists, the banks have no choice but to lock us into the Android-iPhone duopoly.
> None of them has been a great product.
Likewise, and I disagree. The phones that support Sailfish X or later are all great. With Xperia XA2 or Xperia 10, you can install micro-g via f-droid, and then most android apps should work fine. Yes, it is not ideal that we need android/ios apps to function properly in a society. Blame it on SV. Oh, that's us?
Anyway, back to Jolla. A good example is that if the LCD screen breaks, I can just buy a replacement, ssh into the old phone, and scp from it to the new phone (this actually happened). That's a Linux phone.
>> None of them has been a great product.
> Likewise, and I disagree. The phones that support Sailfish X or later are all great. With Xperia XA2 or Xperia 10, you can install micro-g via f-droid, and then most android apps should work fine
I am still on Xperia X and since last fall I cannot even update Firefox any more. The few commercial apps I did try like my bank say they won't run on a rooted phone.
The models with the newer kernel/Android had their own share of issues what I read from the forums. So I did not feel any urge to upgrade to a hardware that does not fit into my pocket.
Of course the blame is on the kernel misery for phone SoCs, Google de-facto monopoly accepted by companies and governments publishing apps. So I could say I have a great phone and all the others are wrong...
> ssh into the old phone
Sure, you don't have to tell me, that's those reasons why I use it.
* everyone started to abandon the perfectly capable android 4.4 API, mostly due to the support of TLS 1.2
* payment / transport / banking / e-commerce / travel / food / coffee all started moving to mobile-first
Following that, there was a rather rough 2~3 months transition period when I upgraded to XA2 in early 2019. It has been smooth sailing since.
I once was a strong proponent of small form factor too, but I am afraid that boat has sailed, and well, XA2 Ultra actually still fits in mine :)
Exactly. When I read the headline, I immediately thought of my N900, and having pretty much a full computer in my pocket at all times. Yes, I really, really want a Linux phone.
It's hard to imagine what you're missing, given the present absolute dominance of user-friendly locked-down devices, but a proper Linux phone really is worth it.
Flagship phones are too expensive for me and they don't come with features like a good charger, headphone jack, big battery, etc. Most of them also suffer from Android ecosystem problems. Slow security updates and trash apps in the playstore.
iPhone is expensive too and iOS limits your freedom even more. You cannot customize the layout or app drawer too much. You cannot use non-web kit browser. Adblockers are crippled on iOS. Notch and suffers from the same problem as Android flagships. Lack of easy side loading.
All I want is a phone which respects me. I don't care if it's Linux based or not.
Xiaomi have an official bootloader unlock and are simple to put LineageOS on. Pixels are the only phones that are supported by GrapheneOS, which offers substantial security benefits over any other Android phone.
It's giving the consumer freedom and environmentally friendly for when the updates run out/planned obsolescence kicks in. I have a spare phone lay around from 2014 running up-to-date 2021 Android patches and it works perfectly fine, the battery went bad once but this was the days of replaceable batts and it cost $15 online for a new one. If you can accept the tradeoff with no further physical firmware updates it's a much better deal than the rampant consumerism and e-waste that plagues the mobile market.
Correct me if I'm wrong, but this isn't really a criticism of Android as a platform but that apps use DRM for tamper proofing. I suppose Android could try to make this harder but I'm sure hardware vendors would step in with their solutions. You're totally free to do as you please with android/rooting as long as you don't care about those apps -- which I have a strong feeling will never be available on more open platforms, unless somehow they gain market dominance.
Even games, social, and news apps use safety net now. Google should discourage blatant misuse of it which they don't.
If Google cared about security, they would push OEMs to provide proper security updates which none of them do.
> Almost all budget and mid range android phones come with invasive tracking, bloat, and ads now.
This is what tempts me to install a custom rom. Solving this could be great for security.
I'm sure they're not worried about lineageos and other aosp niche roms but they will be about Huawei and Amazon and their forks without play services.
This sounds somewhat pat and snarky, but I am trying to make a real point. Let's go to some of what the article says:
==================
Having an "one true way" to use the UI and tons of preinstalled apps is the reason I dislike the android ecosystem...
Megapixels has a fully user-configurable post-processing pipeline basically since the start. It allows you do anything you want after clicking the shutter button in the app since it's a shell script. Still people complain about how they don't like that there's an extra file they don't want (the dng) or photo upload is missing. ITS A SCRIPT, CHANGE IT. Do you really want a Linux phone?
==================
If you're defining "a Linux phone" as "a phone that uses bash scripts for configuration and where nothing works out of the box", then no. People don't want a Linux phone. But then, by that standard Ubuntu isn't catering to a "true" Linux audience. I don't think this is a useful definition of the word.
Another issue for apps might just be the lack of resources to get started. Kudos to you for all the work you're doing on Megapixel.
Let's keep iterating and true FOSS fashion share what we've learned so we all can benefit. And we can simply ignore all the "reviews" that complain about the lack of apps, compare to Android/iOS for polish.
Also for people complaining about performance, the code will be optimized, graphic acceleration will be used, and finally more powerful device will become available.
Of course a $200 Linux phone with immature software does not compare to $1000 subsidized spying devices
When pretending your app is running on a pinephone works properly-ish, I'd call it good. Any functional client or service is a boon to mobile Linux; and even if it's imperfect, so is everything else for now :)
Sure, I do not have the latest phone model, I have it on my Sony XA2 and Cosmo communicator. It has issues with some playstore things, but I also do not care. I can do 90% of what I want on them and the rest I save for my computer. I like the separation.
Something like Plasma Mobile should just use those ideas and make an open UI.
It is easier to use the programs which are already available on the desktop due to the fact that GNOME (and probably KDE) already support touch screens. The usage of a tablet only slightly shifts from using a desktop. So it is a more gradual change.
So by rolling out tablets first, development time could be spent on smoothing out hardware acceleration, supporting cameras, responsive UI, optimizing touch, etc. And then we could tackle phones.
For me as a PinePhone owner, the biggest problem is not apps but overall speed. The OS feels very sluggish. And although I know that it is really old hardware, doing basic things should be much faster. This holds me back from really using it. The other big problem is battery consumption. Any distro I tried drains the battery so fast that I always have to shut off the phone when I am not using it. Even if I don't do anything. On a fresh Manjaro+KDE, I can see the battery drop by 10% after picking it up and using the browser for a few minutes.
And wrt to linux tablets: The PineTab[0] is sadly having really bad resolution (one thing that is excellent on the PinePhone btw.) and it is too heavy. It is a far less attractive package compared to the PinePhone which has "pretty ok" specs througout. So I am really excited for things like the cutiepie[1] or the jingos tablet[2] (if it is not a scam). Although the latter will come with a custom linux kernel which kinda makes the advantage of a (mainline) linux tablet worthless since one still depends on some people maintaining some old linux fork.
[0]: https://www.pine64.org/pinetab/ [1]: https://cutiepi.io/ [2]: https://en.jingos.com/jingpad-a1/
I want a Visual Studio level IDE that I can develop Linux GUI application in. GTK is an amazing piece of software, and there are some beyond amazing GTK based applications that I found and started using.
As someone who has not done linux GUI development before, the biggest issue is getting started. Android has first class development support, and so does iOS. Linux has Glade for GTK and QtCreator for QT apps.
Vala is an amazing first step, and the code makes design & development super easy. But there is still a disconnect between design and development that hinders adoption. If there was a nice "all in one" application where development can be done like. You could see the hype, and then disappointment, when https://github.com/akiraux/Akira was announced and it looked like such an IDE. But alas.
The Rust GTK Bindings also look really good, but I think the biggest hurdle that needs to be overcome is making Glade either more user friendly and making it tie into an IDE like Builder, or have an IDE like IntelliJ support it.
There is now an active project Avalonia, that provides WPF-like GPU accelerated framework, that is cross-platform, and can potentially work well on PinePhone with .NET Core.
And although Flutter has historically been for ios and android, I actually prefer writing Flutter apps while running them on my Ubuntu box. It feels lighter and faster running them natively in Linux than running them in an emulator.
I'm a Linux desktop guy, but use an iPhone (I don't sign in with appleID). Privacy and libre arguments aside Linux on the phone sounds great at first until you need to make a call and it suddenly drops. Just my 2 cents
You must be holding it wrong.
I am with you about the Linux desktop, iPhone thing though.
Agreed, I had an N900 running maemo long ago. It was a novel idea to have a Linux device in my pocket, and being able to just drop in to a terminal or do some X11 forwarding over ssh was kinda neat.
But when you start missing phone calls because the screen becomes unresponsive, or the camera is failing to save a picture due to some driver bug, it's a little hard to convince yourself it's worth it.
A couple of years after the N900 launched and WhatsApp had taken over the (European) world, suddenly the lack of support elevated from minor annoyances to full blown disappointment and having awkward conversations about why I couldn't join the group chat. The much cheaper Nexus 4 would replace it with a much better experience all round.
Though I still have a soft spot for that slide out keyboard.
The gist is that most of us simply find unthinkable to use a piece of hardware that we don't 100% trust, for personal communications, writing or receiving mail, keeping personal data and documents, banking etc. therefore we need an open platform because allowing public scrutiny of the code base - device drivers and user apps included - is the only working method to prevent both software and firmware from doing nasty things behind the user. So far, a Linux phone is the closest we can get to that dream device.
In any case we all have more than one phone in our house. An extra Linux phone will give me and my family options. The junk stuff and games etc can go on my Android.
I think everything else is there. Also Linux games are generally better than Android ones IME. (although I haven't run any besides nethack on my pinephone.)
Spotify shouldn't be a problem [0][1] using librespot [2]. As for banking, I can't speak for all of Europe, but web apps are generally optimized for desktop use, apps have some convinient features, like scanning the QR code of a bill and pay it with one tap, instead of having to type in everything manually when it's on paper, or copy paste it from a PDF.
Additionally some banks require an app for 2FA for the web interface on PC too.
[0] https://github.com/kraxarn/spotify-qt [1] https://github.com/xou816/spot [2] https://github.com/librespot-org/librespot
I'm guessing it's the exact same problem that happened with microsoft: drivers are restricted and not open source, which prevents users to really have control over the mainstream devices they bought. It's the same loophole all over again: we don't own the thing we bought.
So in the end, the problem comes with hardware vendors, but most consumers really don't understand and don't care.
So why not just make an open phone, but with the same android OS everybody knows? Without the google play store? Why make a new OS, with worse hardware support?
The term isn't that accurate because the userspace might not be glibc and GNU coreutils etc. But I guess that's what people mean: a phone running X11 or Wayland, and the entire stack that comes with it, as opposed to the Android userspace (which does not use X11, and has a pretty different userspace structure).
What I want is a 100% iOS or Android phone, with a USB port that I can plug into a keyboard/mouse/monitor dongle on my desk, and get a desktop Linux environment that shares my same media.
No clever apps that work in both. No clever Linux UI nonsense. A Phone that's 100% android or iOS in my hand, and 100% Linux or MacOS when plugged in to my workstation.
I shouldn't have any clue that it's a Linux phone when I'm walking around with it.
I have no doubt that as ARM proliferates, the Linux phone experience will continue to improve, but it is going to take time (and as the article points out, it is going to take development effort from the community). Realistically, though, Linux phones will never reach Android/iOS levels of smoothness until you get a couple companies that want to disrupt the Android/iOS duopoly with a new Linux-based offering....
Until then, Linux phones are a no-go for mainstream users. That being said, mainstream users who just want a working phone that is not a massive data funnel back to FANG already have a really good option with https://e.foundation/
Basically tons of apps are crippled and non-functional unless you let Google's code run as a system app. Sure it can be tricked, spoofed, and mocked for the time being but doing so basically destroys the security model of Android in the process. There's also a lot of badwill between the 'secure' Android ROM groups which isn't great when the selling point is security and reliability.
A proper Linux phone with a Secure Boot/Verified Boot implementation would be great. Linux phone's ATM seem to be more concerned with userland stuff like camera apps.
In a way, a Linux computer in the form factor of a phone with a modem is probably best but it's the complicated nature of the ARM ecosystem getting in the way. e.g how many forms of a secure boot system for ARM exist between Android, Apple M1, Samsung Knox, et al.
Regarding the camera... I don't quite see why a camera system using MPV + standard webcam capture tool can't be used... I guess it's not user friendly but it would be simple. It's not like Linux+Camera is anything new.
Also, those pure Linux phones will take a decade to match the security of current Android and iOS phones. Have fun waiting for that verified boot and proper app sandboxing. insert the image of a skeleton in an armchair.
And I dont want apps or anything! Phone, SMS and Browser would be fine. (Maybe address book?)
Like, Xfce with those three apps on my com-set would be "done" and shippable to the eager market. Get the the core-com-set features working and let a marketplace/app-store mature around it. Don't try to 1=1 to iOS for mass market. It's not Lean.
No. This is also the reason why, at times, I loathe to use Linux. I would prefer to use something like FreeBSD, which has a different culture and attitude. It's a more stable and polished experience that doesn't aim to cater to every nerd's obsessions. This is what I want in a phone, which is why I use an iPhone, despite all the tradeoffs.
I don't know about others, but what I would want in alternative phone is something that is different in a different way. For example, a simple and open hackable phone that enables creativity. If Linux is part of that story, fine. But I'd be open to other options.
Basically it's the Raspberry Pi. And the arruino thing. Sure proving that I could do some sort of actual development on a pi might be fun for a while, but what makes the Pi fun is that it's simple and straightforward and approachable enough that it encourages creativity. People make cool and fun things with Pi's. I would want a phone platform that would enable that same type of exploratory creativity in the phone hardware space. Linux kind does that for the Pi.
If I get a Linux phone, I don't want to get a smartphone that tries and fails to be an Android phone.
I am perfectly fine if the phone only has a terminal interface, where I can type "call 212 555 5555" to make a phone call. Especially if I could then hack the audio input/output with some other terminal tool that would allow me to record, or change my voice.
You can already do that with PinePhone, right now. All that's needed for this use case is already there. (since 8 months ago, really)
The drivers are all there if you really want this, and it's a few weekends project to write a bunch of scripts that will make PinePhone do exactly what you want just by using common Linux tools.
You can use fbkeyboard and write a bunch of shell scripts. Calling is no more than a few AT commands that you can wrap in a shell functions or whatever. If you really want to be fancy, you can write something that will allow you to queue AT commands and call your scripts on unsolicited indications from the modem. You don't even need a modem manager or any other complicated SW like that, if you don't like complexity.
Mobile data can be entirely handled by the kernel too, with just some generic configuration commands to set it up. No complicated usersapce needed, no daemons needed, not even pppd, like with older GPRS modems.
Audio is solved at this point, and I documented all the controls and have a suggested setup for the very flexible calling setup where you can monitor, playback, record, modem audio, your mic/earphone, in whatever combination you want. There's even a tool that can do the setup based on some config flags, no need to mess with individual controls or ALSA tools. No audio daemon is necessary for basic use cases.
I even have this documented on my website: https://xnux.eu/devices/pine64-pinephone.html if you click through various links in the features table.
All you need is just start and implement your ideas. No kernel hacking needed, or any complicated stuff like that. It's just there for the taking.
People who actually did something like this can probably be counted on one hand. There's sxmo, and there are 1 or 2 people I remember writing to me personally, that seemed to have some similar low-tech but extremely flexible and hackable custom setup.
And yet, for me, you kind of answered another question with "a few weekends" and "a few AT commands". Because at first blush that isn't as approachable as I'd like. And what you described was doing things that I already do rudimentarily with phones for years. Does it have gpio that I can do weird things with? And what about graphics programming? Because handhelds have screens, and I'm 1000% sure I don't want to have to navigate the X stack with C as I have over the years, or GTK, and frankly, I don't really want to have to setup compiler tool chains or any of that. I realize in typing this up, that while the Linux on a pi is cool, it's actually that I can hack Python live in it right away that makes it so accessible. I think what I want in an open/creative phone is a live AND graphical programming environment like Squeak/Smalltalk or something of similar ilk.
Put phone calls in it if you want, but I'm probably never going to use it. The sim would be for mobile data access
Of course I can already SSH using my iPhone, the big thing I'm missing is a physical keyboard. Something like a Blackberry or that Nokia one with the slide out keyboard would be fantastic
My Samsung Galaxy Note 10+ kind of delivers this with Samsung Dex, but it feels sluggish and clunky and there are some minor but very annoying bugs and glitches.
I'm hoping that Linux phones will deliver on that.
I wouldn't mind dialing numbers from command line though.
I don't think we have the right definitions here, are we talking about GNU/Linux phones? Other distributions/Linux? Or about anything that is Open Source?
Labeling things is hard
i don't know that GNU is necessary but a open posix/unix complaint environment with root access is.
WhatsApp (not the web app, it's limited)
Skype and Meet (the Linux desktop versions would probably work but I wonder what their UX would be on a small screen)
Banking apps
Maps with satellite and street view
Location tracking / following tracks (in using OsmAnd for that)
Email (Thunderbird on a small screen?)
My country 2FA / digital signature app
A keyboard that does swype well
Then it should be small, let's say 14 cm tall (which used to be huge, oops) less than 1 cm thick, nearly no bezel, 130 g of weight.
There is no Linux phone with those specifications AFAIK. So yes, I want a Linux phone but I don't think I'll buy one in the near future.
Personally I don't really care if there's a browser, email, any messaging beyond SMS, maps are nice to have, but not required. I can even do without banking, but I do have a few special apps for work and interacting with the government.
Depending on where you live, we reached a point where you can choose to not have a SmartPhone, but good luck interacting with society, every new thing is "App first". I a few weeks we'll have Covid-19 passports, on our phones. Don't have a smartphone? I would count on the government having planned for that scenario.
This summer I wanted to hack with LineageOS or something like that. XDA developers said there was a good build for a new Samsung Android tablet so I bought it. I guess I should have read more as Samsung does not use the standard Android bootloader to flash but something called Odin, which is not available (although some people say they have a Windows tool that works). I could query some other non-Samsung Android devices I had with the Android flash tool, but there were no good builds for them or I was unable to build, or I didn't want to risk flashing that specific device.
I might try again this year. I'll look to see what non-Samsung tablets are selling near around here, then look online to see what might have a decent LineageOS or similar build, and then get it. If I flash and brick some cheap, new Android tablet it's not that big of a deal. It's one of the reasons I got into Android programming in the first place, that it is free/open and I can hack around on it like I can my Linux laptop. I thought by 2020 it would be easier to flash a tablet and install LineageOS but I guess it takes a little more effort. I did miss some things though like that Samsung does not use standard Android flash.
If I really needed it I would have done it already, but I was doing it on the weekend for fun, so when I ran into trouble and other things came up, I shelved it for a little while.
Browsing and editing files on PinePhone is substantially more cumbersome than doing so on a PC, unless you have SSH (takes time to setup, requires a separate computer, has issues finding the IP address, maybe laggy) or an external keyboard (tiny screen, USB dongle, can't easily charge at the same time).
Linux PCs are always be behind in supporting new hardware. Hardware components are always developed with the market in mind. If the market is commercial, Linux support is not a primary concern.
Linux PCs give you lots of options and customizability. Customization is nice but if you don't care this is a lost benefit. Customization also means the Linux community is fragmented based on preference. This means efforts are not unified.
OSS effort is mostly expended on improving data center computing. When it came to a desktop PC, running Linux only makes sense if you're a developer. Linux phones are the same. Improvements will double down on developer needs.
Upgrades are dangerous. Linux PC upgrades can have major issues and render your desktop unusable until you fix things. Linux phones will likely suffer from the same problem. Customization strikes again to make supporting upgrades a nightmare.
Finally, Linux's file orientation is the most granular level of control that is antithetic to an "app" model. The app model is convenient, fast and easy to manage. The file model results in more effort for the user to properly manage their device.
A Linux phone is something I don't want! Linux phones will take considerable time to get to present phone quality levels.
Smartphones are barely useful for anything to me these days. I thought about it and I basically want desktop experience but in a small form factor of a smartphone, because that's what I'm used to using most of the time.
Small portable computer that I can actually program and compile stuff on. Device where I am able to script and modify any aspect of it in case of need, without needing to install appstore apps for every little stupid thing I know how to shell script in 5 lines by combining existing tools. Something small enough that it would be more portable than a notebook, but just as generically programmable, and can host any of the usual scripting languages, databases, servers and tools I may want to run on the go, and already know all too well.
Something on the way between a dumb phone and a general computer. And the phone part would ever be interesting to me only if I could hook it, script it, and control it via some personal rules in a detailed manner. Otherwise just using a dumb phone has way better ergonomics than any smartphone ever had for me.
If people would document their stuff and would bother to give the slightest hint what should be done next for a newcommer to their project, people wouldn't have to restart everything from scratch.
"I want a truly libre, open phone! ... MY truly open libre phone that i own and brand and get rich on!"
With a linux phone, there is a solid chance that I won't be forced to replace my phone in order to continue security updates. I very much look forward to pulling my software from a DFSG-approved repository.
As long as there is enough interest to keep the manufacturing line open, the rest of the ecosystem will continue to grow.
I wish Microsoft hadn't bungled Windows Phone quite so thoroughly.
I don't care if Kernel reports version as Darwin 20.3.0 or Linux 2.6.26 or Linux 10.0.23456.0, I want some ... I don't know, something other than "it".
Current popular OSes are garbage - they lock the computer, turning it into an ad platform useful only for selfies and social media. You don't even own the device - you merely rent and are at the complete mercy of the vendor.
It's just about about linux. I'd take anything that gives me control over the device and turns into a real computer that i can plug to a monitor or two and use it as a desktop. Hardware is here, software isn't.
Yes, I want a Linux phone please! Especially one I can get parts and batteries for. Something with a removable battery even. Although I must admit, Motorola does a pretty good job with their unlocked Android phones.
Also as a side note, I bought an LG Stylo 6 from Boost and I wanted to smash the phone out of the box due to the bloatware. I know "install ADB bro, uninstall the bloat" but like the voice mail message literally forces you to choose an ad to use it like wtf is that.
Just today, as an example, I had to send someone a link from my desktop. I had to send it to their email. The next best option would have been to use Messenger on Facebook. Why can't I simply have the phone appear as a window on the my desktop (almost like a virtual machine) and have access to its full functionality, including, copy, paste, transfer, file system, etc.
It's 2021.
Microsoft might want to think about bringing back Windows Phone and doing it right this time. Call me. I'll design it for you.
Yet I can comfortably say that I don't care one bit about "Linux phones".
I want a handheld Linux computer, and it just so happens that these attempts to make "phones" are the closest approximation.
But if there was a mobile Linux distro that did not waste development time on features like calling, SMS or camera, and hardware that used the space wasted by camera, receiver etc for additional battery capacity, I would most certainly buy that.
I'm happy to use an Android or iOS device as a smartphone. I want a pocketable computer to do things those devices cannot (or rather: refuse to) let me do.
Even before doing that, I think the most important thing for Linux phones is getting HW-Accel to work. Any Android/iOS phone nowadays renders basically anything on the GPU and it makes the experience of using a computing device so much better.
Obviously this is very hard, as free hardware drivers (especially for high-end SOCs) are rarely available.
Does that mean that Linux in general is only for people who want to write their own code? I kind of thought some people thought Linux could be for anyone, not just for coders. Despite it not really taking off as a consumer desktop, I remember when people thought it could/should and it would be good if it did. Are those days gone? Now people think linux is just for coders, and if you don't want to write your own code you shouldn't use linux?
There are rough edges, there are dozens of distros, dynamic linking is popular even when it makes no sense which causes lots of programs to not work until you track down the right dependencies etc.
On the other hand Linux can be allright for people who don't do anything interesting, who just use firefox all day, that works fine in linux.
I expect that a free as in speech phone ecosystem would eventually overtake non-free ones like Wikipedia did Britannica.
The one killer app for me is phonecalls and text alerts, though. I can live without anything else, but those need to work. The main issue I have with the pinephone and the reason I still use my Android is because when the PinePhone turns its screen off, it also puts the 4G hardware to sleep.
I've tried it over and over, but the hardware just won't wake the phone up and make it ring. I'd love to ditch the Google shit, but unfortunately I really can't do that right now.
It should also have some low power built in core dedicated strictly to phone tasks that would have a priority over everything else.
Add to that decent hardware.
But it's still cool, and I'm happy I bought it, and I think that later in the future an OSS phone is going to be amazing.
However my unusual use case was that I want my phone to also be my 'computer', e.g. I arrive at work, plug my phone into the USB-C of the monitor that already has a keyboard, mouse and ethernet attached, and I get to work.
I already use linux day to day at work with an ASUS PN50 4500U and it's awesome.
Yes, there is a need for a 3rd reliable phone OS.
As a lay passenger, I wouldn't want to or have to know what planes are made of or how a flight management system works. That's not my business. Think of it from a user's perspective.
How? the existing handset manufacturers just need to make their phones functional enough for linux and/or provide the various blobs in a straightforward manner. And a suitable stretch goal? few or no blobs.
Alternative approach: release the blobs for old phones so we[2] can at least actually write our own updates.
There. No weird hardware needed. Just existing handsets. Just start there.
[1] next week is also good.
[2] or someone who is more clever, has more time available, or otherwise more motivated.
- I want it to be fast (as fast as last year's flagships), because
- I want to use it as my only computer. Mobile when it's in my hand, desktop when it's connected to my monitor or mobile lapdock
- I want the mobile experience to be good enough (phone, message, very good browser, and support for apps that run on Ubuntu) for daily use, but not on par with e.g. Android
That's already quite a planning. Thanks for your efforts on the camera app.
Why? I just want working hardware to be able to continue to work. It's still a great phone for me. If a "Linux phone" means the OS vendor doesn't isolate me from updates just because they have new hardware out, then yes, I want a Linux phone.
I'm not talking sandboxed inside a guest user within my iPad OS, I am talking full fat administrator access Ubuntu Gnome, MacOS or Windows.
If that happens to be a linux distro, open source, or otherwise, I'm not fussed.
Take my money Apple, ugh. Linux phones are cool - I had a Nokia N900 and N800, but polish in general is more important to me these days.
That's it. There's no another major player.
It’s for those who want extreme control more so than a polished ui. I use iOS by the way. Android is the worst of both worlds.
This is what I think people like about this.
> People even want to use Flutter to make UIs, was the phone lacking in google technologies for your liking?
If a company makes a tool that can repurposed for a system not by them, then why shame people for using that tool?
> But even weirder, there's tons of people asking for Android roms for the PinePhone. What's the point? Do you believe in the church of papa Alphabeticus, the pope of the internet?
People want to use apps that the environment current lacks. Again, there's no reason to shame them for it, FFS.
> It's the community that against "the other side". It's the developers that are actually working together to make something.
And yet, this is made after scorning users for not truly wanting to use a Linux phone on the basis of them possibly incorporating software that were from Google. How can the author hypocritically call for a Kumbaya within the Linux community, when they are already indulging in gatekeeping and confrontational behavior of their own?
And at what point is this arbitrary purism satiated? Is a completely ad-less Google-originated technology like Go to be rejected? Should Protobufs be considered harmful because of where they come from? When did Google become the '90s Microsoft of '20s Linux, anyway?
> People are still complaining to other people that they have chosen wrong. This is Linux, you can do what you want with your device.
And yet the author has the audacity to question people wanting to try different things with their mobile distros, even things they don't want, such as 'Having an "one true way" to use the UI and tons of preinstalled apps.' Just rampant naked hypocrisy. Do they really want a Linux phone?
I honestly don't mind when something is legitimately broken, especially when I have the tools to work around it. For the first few months the dialing app didn't work, fine, I just wrote some shell scripts to poke AT commands into the modem myself. The thing that makes Linux phones much nicer is the much weaker distinction between users and developers. If something is broken then you can just go fix it (or rip it out or hack around it.) I strongly dislike palm OS style UIs so I switched my phosh session file out for flux box. Many people won't like that and I don't have to care.
I used to dream of Linux phones and achieving the DIY freedom of desktop Linux. But just like desktop Linux stuck recreating XP era apps, what are we recreating, Ice Cream Sandwich era apps?
I'd love to see the benefits of shell modularity implemented so users can drag together scripts with their fingers.
Also, why is GTK and QT being used for these projects? It's already a drag that we are stuck using local GUI subsystems on desktops. They should be using layout engine platforms like Electron or React Native. That would increase portability, attract existing projects, and promote cross compatibility
GTK and Qt run slowly enough on the PinePhone's anemic processor and GPU, and Electron or React Native would perform even worse.
I don't care if it is linux, I don't care if it is open source!
And as such they don't have the economies of scale when paying for components, manufacturing, marketing etc.
Of things relevant, watches is the present and AR is the future. Both are a challenge from FLOSS point of view: watches and AR involve very latest hardware/tech which is very proprietary AND those devices are not very interesting from the "general-purpose computer" point of view. They have very narrow purpose/function, and they become obsolete very quickly (due to newer tech being objectively better and better). Very useful devices in everyday life, but not interesting to play with.
Unlike a laptop, or a desktop or a breadboard/rpi, which are general-purpose devices and are actually interesting to play with and learn from.
Would I like a truly open/libre Linux on my, say Apple Watch or Apple Glasses? Sure, but realistically it ain't gonna happen within the tiny time span of high-tech relevance. And it'd be a travesty if bright minds would spend their valuable time on that hopeless endeavor. There are so many more interesting and more valuable things to do: improve man pages, or cure cancer...