KDE Plasma Mobile On Track To End 2020 With Quite A Polished Linux Mobile Experience

[oiaohm]

ok, it may be a bit daft question- but why not write it as a some (deep) modification of open-source android, so it could run android app as well as those written especially with kplasma in mind? something like blackberry os did.

Really in some ways you come close to answering your own question. If you had started with why did blackberry OS fail doing this would have got you to answer.

One point to start of with most KDE applications are not built limited to kde plasma. Its a Linux Desktop class OS running gnome, kde.... and the applications work.

1) Google Play. There are a lot of applications that need google play for digital right management and other things that require approved device. Deep modifications of android you will never get approval.
2) Google controls the upstream. This is something blackberry run into google would add new features that new applications would need so it was basically back to the grind making the it work again.
3) Security model of Android and the security model Linux Desktop application and blackberry applications expect is totally different.
4) Linux Desktop will still want to keep on adding their own new features like blackberry did as well.

The security model of Android has changed version to version as well. Android is very much a moving target so is the Linux Desktop itself.

The only possible working solutions is either containerisation or virtualisation. Thinking android uses containerisation in their security model these days that does provide a level of tricky.

https://halium.org/ Yes halium does use containerisation to a point. Android drivers are differently done to what Linux Desktop software expects as well reason for needing halium to use android kernels. Of course these abstractions cost.

Android on Linux Desktop or Linux Desktop on Android both require abstraction that abstraction costs. Depending on the abstractions used there are different levels of maintenance cost. Virtual machine is the cheapest on maintenance but most expensive on hardware requirements. Containerisation is fairly cheap on hardware requirements very expensive on maintenance cost to keep current with upstream. Please note these relative costs don't really change if you are containerisation/vm of LInux desktop solution inside Android or Android inside a Linux Desktop solution.

[Quackdoc]

ok, it may be a bit daft question- but why not write it as a some (deep) modification of open-source android, so it could run android app as well as those written especially with kplasma in mind? something like blackberry os did.

if it would base on android, it would be installable on a greater variety of phones. imagine if it would be somehow available as a 'version' of lineageos

the reason why Android ROMs can work on such a plethora of devices, is because specific people maintain those devices. Android is actually incredibly inefficient when it comes to mobile devices, GSI helped a fair chunk, but a lot of phones need proprietary junk to work properly. Halium project might interest you though.

But it's just better (in my opinion anyways,) in the long run to run vanilla Linux under the hood instead of Android

[szymon_g]

ok, it may be a bit daft question- but why not write it as a some (deep) modification of open-source android, so it could run android app as well as those written especially with kplasma in mind? something like blackberry os did.

if it would base on android, it would be installable on a greater variety of phones. imagine if it would be somehow available as a 'version' of lineageos

[stiiixy]

PinePhone aside, which seems to be the dev device of choice to get an actual working phone with real apps now, I'm excited for the project's future.

It has been a long time coming and it's very much appreciated the effort expended to get this far.

And Manjaro have been working with it as well (my favouritest full-blown desktop 😃) so, bonus!

[oiaohm]

yeah it is a fairly big issue that needs to be tackled though, just because there is no good solution now thankfully doesn't mean there won't be one in the future.

The big problem is all the possible solutions to the problem need more processing power than the pinephone like devices will have. To tackle the problem you need users and developers. You will not get either without the hardware platform.

So its a problem that is a future problem to fully solve when we have more powerful phones that can run the Linux desktop OS taken to phone..

[Quackdoc]

Anbox issues I will give it has quite a few. Particularly that it stuck on Android 7.1.1 and we are up to Android 11 now. KVM path you need more cpu cores and more memory than a pinephone is going to have.

Android-x86 has the limitations of being short on resources to stay with current versions of Android as well.

With the pinephone hardware limitations the best option if you need Android applications is in fact dual boot run Android and what ever OS you want for desktop as two different OS installs.

Once you get to 8G of ram and 8 full cores you have enough to be able to VM well.

Remember more Linux desktop users are likely to use a Windows application than a Android one. Until we get a larger population of Linux desktop based mobile phones out there with more processor and ram power integrated support for Android applications is not going be there because there just is not the resources in the devices to-do it.

Once we have the phones out there with the hardware there will be more users wanting the android application support so hopefully more developers for projects like anbox.

The hard bits of the hardware and the base OS rough edges they need to be fixed before Android application support. Once that is fixed there are quite a few different options from commercial products to virtual machine options to containers for running Android applications.

https://www.omgubuntu.co.uk/2020/07/glodroid-android-for-pinephone-allwinner

Different groups are working in the means to run Android as primary OS on pinephone items this is why dual boot is option for this generation of pinephone hardware.

It a wacky one right. Not supporting Android at this time in the OS is not a deal breaker. Its really simple to forgot pinephone and items like it are built from the ground up to be running more than 1 OS. That means if you need android support you can just run Android once you have done what you need in android reboot and go into plasma mobile or some other Linux desktop based OS for mobile. Yes that other OS can allow you todo things that you cannot do well under Android as well.

Things get hard and expensive in hardware requirements when you wish to run both at the same time. This was also true when we only had 2-4 cores in desktop computers.

yeah it is a fairly big issue that needs to be tackled though, just because there is no good solution now thankfully doesn't mean there won't be one in the future.

[oiaohm]

unfortunately anhox isn't most stable of programs. even doing a clean install on recommended setup I've had nothing but issues with it. it's hardly in a place where I could recommend someone using it for business, hell I would recommend someone use KVM and an Android x86 virtual machine first and that's not without its issues

Anbox issues I will give it has quite a few. Particularly that it stuck on Android 7.1.1 and we are up to Android 11 now. KVM path you need more cpu cores and more memory than a pinephone is going to have.

Android-x86 has the limitations of being short on resources to stay with current versions of Android as well.

With the pinephone hardware limitations the best option if you need Android applications is in fact dual boot run Android and what ever OS you want for desktop as two different OS installs.

Once you get to 8G of ram and 8 full cores you have enough to be able to VM well.

Remember more Linux desktop users are likely to use a Windows application than a Android one. Until we get a larger population of Linux desktop based mobile phones out there with more processor and ram power integrated support for Android applications is not going be there because there just is not the resources in the devices to-do it.

Once we have the phones out there with the hardware there will be more users wanting the android application support so hopefully more developers for projects like anbox.

The hard bits of the hardware and the base OS rough edges they need to be fixed before Android application support. Once that is fixed there are quite a few different options from commercial products to virtual machine options to containers for running Android applications.

https://www.omgubuntu.co.uk/2020/07/glodroid-android-for-pinephone-allwinner

Different groups are working in the means to run Android as primary OS on pinephone items this is why dual boot is option for this generation of pinephone hardware.

It a wacky one right. Not supporting Android at this time in the OS is not a deal breaker. Its really simple to forgot pinephone and items like it are built from the ground up to be running more than 1 OS. That means if you need android support you can just run Android once you have done what you need in android reboot and go into plasma mobile or some other Linux desktop based OS for mobile. Yes that other OS can allow you todo things that you cannot do well under Android as well.

Things get hard and expensive in hardware requirements when you wish to run both at the same time. This was also true when we only had 2-4 cores in desktop computers.

[Quackdoc]

https://anbox.io/

Most of the OS that install in the pinephone are at core standard Linux distributions this includes KDE Plasma. This does mean if you do all the steps to install anbox you can have anbox to run android applications. This is not without its problems

The problems
1) google play you have to side load that if you want it. Of course pinephone and anbox are not google approved devices so particular Android applications will not work and will not install by google play because Digital Rights Management crud fails. Like if all the applications you use come from fdroid for example you are in business. Google play is a little more dicey.

2) RAM this is a big one. Pinephones have between 2 to 3G of ram. This turns out not to be a lot when you are running two userspaces with two userspaces background services. I know with raspberry pis using anbox this problems goes away when you have 8G of ram.

3) CPU pinephone is Cortex-A53 X 4 so you are not playing with the biggest processor here and now you are splitting it.

4) Battery life. This is the two userspaces problem. You are running more stuff you eat more battery.

Notice 3 of these problems are hardware linked and maybe overcome by using next generation of device. 1 is google themselves.


Its really simple to forgot android has a Linux kernel at core with a few extra drivers. That why anbox is able to work they are able to add the extra drivers and place the open source parts of Android in a container.

Linux Desktop OSs coming a mobile OS is in quite a good place to take on Android.
The hard bits:
1) getting good enough hardware that it can.
2) getting the rough edges on power and so on fixed up for good in Linux Desktop OS solutions..

Remember all the Linux Desktop OSs can solve the Android compatibility problem in a single project they don't have to-do it individually.

unfortunately anhox isn't most stable of programs. even doing a clean install on recommended setup I've had nothing but issues with it. it's hardly in a place where I could recommend someone using it for business, hell I would recommend someone use KVM and an Android x86 virtual machine first and that's not without its issues

[oiaohm]

I have a lot of equipment that needs ether Android or IOS apps for me to be able to work. I wish the devs the best but I couldn't consider this unless I could run those apps even if I have to side load them and they will never be ported to this OS. Breaking into this market is a Hurculean effort becuase of the absolute stranglehold the two big players have over both the app makers and the telecom companies.

https://anbox.io/

Most of the OS that install in the pinephone are at core standard Linux distributions this includes KDE Plasma. This does mean if you do all the steps to install anbox you can have anbox to run android applications. This is not without its problems

The problems
1) google play you have to side load that if you want it. Of course pinephone and anbox are not google approved devices so particular Android applications will not work and will not install by google play because Digital Rights Management crud fails. Like if all the applications you use come from fdroid for example you are in business. Google play is a little more dicey.

2) RAM this is a big one. Pinephones have between 2 to 3G of ram. This turns out not to be a lot when you are running two userspaces with two userspaces background services. I know with raspberry pis using anbox this problems goes away when you have 8G of ram.

3) CPU pinephone is Cortex-A53 X 4 so you are not playing with the biggest processor here and now you are splitting it.

4) Battery life. This is the two userspaces problem. You are running more stuff you eat more battery.

Notice 3 of these problems are hardware linked and maybe overcome by using next generation of device. 1 is google themselves.


Its really simple to forgot android has a Linux kernel at core with a few extra drivers. That why anbox is able to work they are able to add the extra drivers and place the open source parts of Android in a container.

Linux Desktop OSs coming a mobile OS is in quite a good place to take on Android.
The hard bits:
1) getting good enough hardware that it can.
2) getting the rough edges on power and so on fixed up for good in Linux Desktop OS solutions..

Remember all the Linux Desktop OSs can solve the Android compatibility problem in a single project they don't have to-do it individually.

[Corodius]

But I would love to see one of these distros make a x86 oriented port.

KDE Neon has a plasma mobile x86 version specifically designed for testing and tablets etc. Might be worth a look?