Hardware. Most Linux phones don’t use an SoC intended for a phone, or use outdated/slow hardware. This is a big reason why devices like Librem 5 and PinePhone aren’t viable in the long run.
Software. Currently the software side of things still needs a ton of work.
From what I can see, there seems to be a bit of a bit of a focus (obsession?) on Linux phones only running mainline Linux kernel and as FOSS as possible from the very beginning, rather than building something that works and then improving things over time.
This is why I like the approach that FuriLabs is taking. Yes, they don’t use a mainline kernel (they currently use the MediaTek kernel). And they use Halium and (probably) some proprietary drivers. But they are building something that (mostly) works like an actual phone, which from what I have seen/heard is a lot more than what can be said of other phones. And they have been putting in quite a bit of work improving the UX, which makes up a lot of the software side usability problems. They can try to mainline later.
I should also probably mention I own an FLX1, which may make me slightly biased when discussing this topic. There are some things I don’t like about it though, such as the size and weight. LCD screen instead of OLED is also a bit of a bummer.
Have you tried Phosh? I haven’t, but apparently it runs on a variety of devices. I guess the devil will always be, as you say, in hardware support. It’s like decades ago when you had to be really careful about your PC components and whether or not they’d work under Linux. Nowadays, I’m more surprised when I come across something that doesn’t work OOTB, but these compatability tables for phones really takes me back to the early 00’s and similar tables for laptops.
Well, not really. The FLX1 has a reasonably good phone SOC (Mediatek Dimensity 900). The big problem is that the phone SOCs typically use more proprietary software and use kernels that aren’t mainline. The massive obsession of using mainline Linux kernel only leads most Linux phones to use SOCs that aren’t suited to phones, basically trying to fit a square peg into a round hole. Furilabs has proven with the FLX1 that you can get competent Linux phone hardware, if you concede using mainline Linux kernel.
Although the FLX1 has some hardware downsides (size, weight and screen technology), these are more down to the implementation and goals of this specific device. The ODM of the FLX1 intended the hardware as a rugged phone, hence the size and weight. Price was likely a factor in the choice not to use OLED.
With the hardware of the FLX1 being competent, it highlights the shortcomings of the software. Phosh still needs a ton of work to be great, most Linux software isn’t designed to be used on a phone, etc. I especially hate the phosh keyboard and it needs a ton of work to be even semi-decent. Phosh should honestly just copy the design of the Android keyboard.
I’m all pro FSF and OSS, but not dogmatically so. I really don’t much care if the kernel is mainline or contains proprietary modules; ideally, it wouldn’t, but not at the cost of not working properly.
The rugged thing, though; that’s what’ll keep me from getting one. I’m already bitter enough about phone size creep. The screen tech is not a deciding factor, for me; in fact, if I got better battery life, I’d take an e-ink screen.
So, maybe not the FLX1 for me, but hopefully they’ll do well enough to build something more slim.
Thank you for the info, though. Honest opinions are so hard to come by.
Damn you, Voice of Truth! Curse you for ruining my dreams and crushing my hope.
When will I be able to get a daily driver Linux phone that doesn’t suck?
There are 2 main problems with Linux phones.
Hardware. Most Linux phones don’t use an SoC intended for a phone, or use outdated/slow hardware. This is a big reason why devices like Librem 5 and PinePhone aren’t viable in the long run.
Software. Currently the software side of things still needs a ton of work.
From what I can see, there seems to be a bit of a bit of a focus (obsession?) on Linux phones only running mainline Linux kernel and as FOSS as possible from the very beginning, rather than building something that works and then improving things over time.
This is why I like the approach that FuriLabs is taking. Yes, they don’t use a mainline kernel (they currently use the MediaTek kernel). And they use Halium and (probably) some proprietary drivers. But they are building something that (mostly) works like an actual phone, which from what I have seen/heard is a lot more than what can be said of other phones. And they have been putting in quite a bit of work improving the UX, which makes up a lot of the software side usability problems. They can try to mainline later.
I should also probably mention I own an FLX1, which may make me slightly biased when discussing this topic. There are some things I don’t like about it though, such as the size and weight. LCD screen instead of OLED is also a bit of a bummer.
Have you tried Phosh? I haven’t, but apparently it runs on a variety of devices. I guess the devil will always be, as you say, in hardware support. It’s like decades ago when you had to be really careful about your PC components and whether or not they’d work under Linux. Nowadays, I’m more surprised when I come across something that doesn’t work OOTB, but these compatability tables for phones really takes me back to the early 00’s and similar tables for laptops.
Yeah FLX1 uses phosh
So, like you originally said: it’s the hardware that has to catch up.
Well, not really. The FLX1 has a reasonably good phone SOC (Mediatek Dimensity 900). The big problem is that the phone SOCs typically use more proprietary software and use kernels that aren’t mainline. The massive obsession of using mainline Linux kernel only leads most Linux phones to use SOCs that aren’t suited to phones, basically trying to fit a square peg into a round hole. Furilabs has proven with the FLX1 that you can get competent Linux phone hardware, if you concede using mainline Linux kernel.
Although the FLX1 has some hardware downsides (size, weight and screen technology), these are more down to the implementation and goals of this specific device. The ODM of the FLX1 intended the hardware as a rugged phone, hence the size and weight. Price was likely a factor in the choice not to use OLED.
With the hardware of the FLX1 being competent, it highlights the shortcomings of the software. Phosh still needs a ton of work to be great, most Linux software isn’t designed to be used on a phone, etc. I especially hate the phosh keyboard and it needs a ton of work to be even semi-decent. Phosh should honestly just copy the design of the Android keyboard.
Very interesting, thanks.
I’m all pro FSF and OSS, but not dogmatically so. I really don’t much care if the kernel is mainline or contains proprietary modules; ideally, it wouldn’t, but not at the cost of not working properly.
The rugged thing, though; that’s what’ll keep me from getting one. I’m already bitter enough about phone size creep. The screen tech is not a deciding factor, for me; in fact, if I got better battery life, I’d take an e-ink screen.
So, maybe not the FLX1 for me, but hopefully they’ll do well enough to build something more slim.
Thank you for the info, though. Honest opinions are so hard to come by.
I’ve been looking at the FLX1; it’s disappointing to hear about the size. I still prefer smaller phones.