Could A New Linux Base For Tablets/Smartphones Succeed In 2017?

[jukk]

Time to time chinese vendors do release source code for the phones.
https://forum.xda-developers.com/and...10000-t3341486

Source code links do not work. And leads only to mega.nz. Doesn't appear very serious.

What is most interesting, do they upstream their kernel drivers?

[Min1123]

What about the intel chip in Asus Zenfone 2? Wasn't that supposed to be an x86 processor? Is this supposed to work out of the box then?

Unfortunately no. x86 as an architecture is one thing, an IBM PC-compatible is quite something else. The Zenfone 2 is quite an Android device, it has an Android kernel, partitioning, fastboot, and drivers. The Atom chips used have PowerVR GPUs on them, and the Android is actually running a 64-bit kernel (to support the amount of RAM), but a 32-bit userspace on top of it, presumably because they could only get 32-bit drivers for the hardware within userspace.

There is no BIOS/UEFI to these because they are handled by the fastboot bootloader and the changes baked into the Linux kernel for this particular device (in Android the manufacturer takes Google's base for that version of Android's kernel, then modifies it with their drivers and if they're proper rereleases the source for each device, then they never update it again).

So, like all other Android devices, the device is locked to its device-specific. android-version-specific, kernel. A vanilla kernel just doesn't know where the devices are or how to access them to even attempt to use drivers on them, if it can boot. And ARM chips are in the same mess.

[GreatEmerald]

Well, you need a bootloader unlocked device, patched kernel sources for that device, fingers crossed. Usually the kernel sources provided by the manufacturer are gonna have compile errors and/or missing hardware support. So the trick really is buying a phone that other people already have working kernels for.

Uh, no, not in my case. It's an amd64 Cherry Trail device that works with vanilla kernels and has UEFI. It also reportedly runs at least Ubuntu with only minor issues (audio and perhaps suspend). I do my research before I buy.

Hence why my question is indeed about what DE would be most suitable to use. All this talk about architectures and kernels do not apply.

[duby229]

Uh, no, not in my case. It's an amd64 Cherry Trail device that works with vanilla kernels and has UEFI. It also reportedly runs at least Ubuntu with only minor issues (audio and perhaps suspend). I do my research before I buy.

Hence why my question is indeed about what DE would be most suitable to use. All this talk about architectures and kernels do not apply.

Yeah, no doubt that's the sticking point, but truthfully, I boot Android to use it as a phone. I did get gpm working on an older phone, but it just isn't that useful. So no DE really. But it's useful for having full gentoo os that fits in my pocket.

[Maxjen]

It would be really great to have a successful cross-platform development framework. An analogue to what Unity3D is for games. And I think the closest thing that we have are frameworks based on web technologies. The current state of the art example would probably be Electron, but there is still lots of room for improvements. It needs to work on mobile and have comparable performance to native applications. You probably also shouldn't be required to use JavaScript. WebAssembly, WebGL-Next and Servo are pieces that could be used to enable all of that. Then with a healthy ecosystem of truly universal applications it would be much easier to move to a new OS.

[bregma]

The thing that's going to render this dream stillborn is the margins.

See, the margins on personal computing devices is so low that nobody makes them for profit any more. The hardware is essentially free, but you can't buy a phone or a tablet or even a PC really, you can only buy a personal consumer tracking device. The whole point of the hardware is to enable the financial backers to track your personal habits and meta-information on everything you do so they can analyze it, crunch the numbers, and sell it. That's where the money is, and they're going to insist the devices be locked up tight and use a fully-integrated, controlled stack so they can leverage every last drop of revenue.

The real market is Asia, where device penetration is higher, data is cheaper, the consumer market is growing incredible rapidly, regulatory restrictions are non-existent, and a quarter of the world's population lives, losing the US market in the noise.

Trainspotting hobbyists might be able to hack their favourite OS on to a handful of devices for lulz, but the personal device market is lost already. The next battlefield is IoT, and the same market forces (commodity hardware, zero margins, levarging your personal information for revenue) are at play in that arena.

[GI_Jack]

I think what you meant is "Can a GNU base for tablets/smartphones Succeed in 2017".

There already is a Linux based cellphone OS, and that is android.

The answer is "No" by the way. Tablets, especially hybrids mabey. Laptops that can also make cellphone calls/handle SIM cards, mabey. Phones no.

The more I use android, the more I see android open source project as the base for future mobile OSes. There is really no need for GNU in the phone space. You can, by the way, already run a lot of GNU on top of android with termux, and the myriad of debian installers that install debian in a chroot. I am really not sure why you'd really want a full distro based on GNU for your phone. Android makes more sense for mobile, especially its security model takes into account how phones are actually used. The need for me running desktop and server apps on a phone is at best a novelty 99% of the time. The rest of the time, termux suffices. if anything, we should help make a community OS revolving around ASOP. Perhaps putting some work into the Lineage OS, the successor to Cyanogenmod

[GI_Jack]

Ahhh....spoken like a true TALIBAN religious Linux Phreak and Geek. Look....if marginalization is so attractive to you go work on HaikuOS. You'll be "really cool" then.

Linux needs to be adopted by the wider computing audience. The moms and dads and uncles and aunts and grandmas and grandpas and basic folk who don't give a shit about how cool and "unmainstream" they are simply because they use Linux and know the "command line". Linux folk like you are actually a hinderance.

If you don't like the community, then leave. I do believe there is a train with your name on it waiting for you to jump in front of.

[ldo17]

That's why the world can only accommodate no more than two of anything ...

Except for the Nokia 3310, currently enjoying a comeback, I believe.

Except for the Nokia Asha, which was actually selling well enough to make a profit, until Microsoft took it over and killed it.

Except for the Hiptop from Danger, until Microsoft took it over and killed it.

[ldo17]

It's all about the apps.

It was never about the apps. Android came out from behind IOS, without the benefit of the apps. IOS long had the overwhelming advantage in apps and in developer support--some might claim it still does. Yet that didn’t stop Android from taking over.