It’s not any one platform that is too complicated, it’s that none of it was standardized. So once you have support for one phone completely done, the next model is already released 6 months ago and you have to start almost from scratch again.
Pixel was one exception to this, because Google would release and document all the modifications needed to run Android. Unfortunately they stopped doing this as well.
Contrast this to the x86 PC and laptop market and everything basic, like how to discover hardware, how to boot is all a documented standard. Even though on PC, you still have to deal with drivers for specific hardware.
Another reason why PC is much easier for Linux is that much of the hardware is shared with servers and for servers, Linux is absolutely a first class operating system, which all but some extremely niche hardware manufacturers fully support.
You need way, way better sleep handling. To get decent battery life everything needs to be able to go to sleep really fast but also be able to listen for signals from specific devices like GSM modem and wake up immediately. Without it it’s not really usable. Desktop PCs didn’t have any sleep functionality for a very long time and even now they mostly just disable everything and wait for a button press. Sleep/wake-up cycle can last couple of seconds without issues.
Mobile hardware also has more devices. I don’t have GPS, GSM, accelerometer or finger print reader in my laptop. When Linux was developed they also didn’t have cameras or bluetooth. A lot of this additional devices are not easily available like PC parts and require closed source drivers and firmware.
To make a usable mobile phone you need to figure out all if it at once. You can’t really release a phone without GPS or GSM and expect people to use it as a daily driver. With PC you can live without the camera or build in WiFi. I remember using USB dongles for WiFi and simply not having a working camera in my Linux laptops and I was fine with it.
What makes mobile hardware more complicated than desktop hardware?
It’s not any one platform that is too complicated, it’s that none of it was standardized. So once you have support for one phone completely done, the next model is already released 6 months ago and you have to start almost from scratch again.
Pixel was one exception to this, because Google would release and document all the modifications needed to run Android. Unfortunately they stopped doing this as well.
Contrast this to the x86 PC and laptop market and everything basic, like how to discover hardware, how to boot is all a documented standard. Even though on PC, you still have to deal with drivers for specific hardware.
Another reason why PC is much easier for Linux is that much of the hardware is shared with servers and for servers, Linux is absolutely a first class operating system, which all but some extremely niche hardware manufacturers fully support.
You need way, way better sleep handling. To get decent battery life everything needs to be able to go to sleep really fast but also be able to listen for signals from specific devices like GSM modem and wake up immediately. Without it it’s not really usable. Desktop PCs didn’t have any sleep functionality for a very long time and even now they mostly just disable everything and wait for a button press. Sleep/wake-up cycle can last couple of seconds without issues.
Mobile hardware also has more devices. I don’t have GPS, GSM, accelerometer or finger print reader in my laptop. When Linux was developed they also didn’t have cameras or bluetooth. A lot of this additional devices are not easily available like PC parts and require closed source drivers and firmware.
To make a usable mobile phone you need to figure out all if it at once. You can’t really release a phone without GPS or GSM and expect people to use it as a daily driver. With PC you can live without the camera or build in WiFi. I remember using USB dongles for WiFi and simply not having a working camera in my Linux laptops and I was fine with it.