Here’s my wish for the long-term future of Android phones to achieve faster software updates: separate the carrier portion of the OS from the rest of it to allow the “rest of it” to be upgraded/replaced on less restricted timescales.
The problem is that carriers are slow to qualify new OS updates for technical, legal and business reasons of various legitimacy. But the parts of the OS that they really need to care about are fairly small, as CyanogenMod has proved to many people.
How, technically?
1) Virtualization: the user OS (e.g. Jellybean, KitKat) runs in a container above the carrier OS, or they both run above a hypervisor.
2) Hybrid hardware: the carrier OS runs on a low-power, completely separate CPU with it’s own ROM and the user OS talks to it via internal bus
I was going to say the same thing Winfred Byrd. I think that is a good step to get “features” out to android users faster.
I think the fact that any update needs to go through the carrier is ridiculous. I can see vendors needing to tweak Android for their devices, but after that let it loose.
Good example, and I agree with you. I was thinking in terms of bringing new features to old phones where this should be a rare case, but this problem is encountered every time new hardware is used in a device. Carriers become the bottleneck that hardware vendors have to get past.
Just thinking about who might absorb the cost of such a project, I think your first solution is more likely, but even then, I still probably wouldn’t bet on it.
Winfred Byrd – it occurred to me this morning that Google Glass is already following model 2 in a way: it tethers to a nearby phone which provides connectivity. Because of that Glass can rev whenever Google wants it to. I’ll bet that reasoning weighed at least as high as battery life when they architected the hardware.