@grenudi said:
In a few years, those dozens of devices will surely run the mainline kernel and have the daily driver base needed,
I would like to add the following about the mainstream vs. halium topic:
We have to work with an industry that produces devices and chipsets that are not standardised as would be a PC, that are made for Android with closed-source drivers, and are made to be obsolete after only a few years (at least are for sell only for a few years). Also the constantly evolving telephony standards are making devices rapidly obsolete and unusable. For example, some parts of the US are already dropping 4G, making some devices that are only few years old completely unusable. I really hope this logic can change.
But until then we have to work with it and it requires a lot of effort to port a new device to the point where it is fully usable, while the device has only a short period of availability. Given that it takes a lot of time to port the devices the interval between the time when the devices is ported and the time it's removed from market or not compatible with latest telephony standards may be short.
Then in this context the practical difference is that porting a device to mainstream Linux takes a lot more time and effort to achieve the same level of functionalities as porting it with Halium.
I think what would really be good is that the manufacturers start to make things meant to work with mainline linux in the first place: devices and CPUs adapted to phones. And this would certainly happen if we started to have a decent adoption of Mobile Linux and some traction.
But other than that, I wonder whether porting mainline Linux to a constantly evolving hardware base, is not just too much work for the community, that has to be endlessly renewed like the myth of sisyphus.
Therefore I think a strategy can be: first attract a decent userbase with Halium, then use it as a leverage to have manufacturer produce hardware really designed for Linux on smartphone.