That said the fact that a new Android release has to pass through two other sets of hands before it can get to an end user sure ain't helping.
It takes money and time to upgrade all their carrier-specific "value-added"(i.e. shit) software on top of the new operating system, and they hesitate to do it, because they've already got it working, and taking no action is much easier and cheaper than upgrading.
Besides there isn't enough customers yelling about it anyway (the ones who really care only buy the phones with known fast OS upgrade cycles, like Nexus).
Still I agree the problem is devices are not getting update and it's not a problem of want as the article hints, but of can't, at least at the user level: Of course carriers and manufacturers hate having to support devices instead of selling new ones!
The problem that Google apologists/Android fanboys forget... is that deploying an Android update is a HUGE cost to manufacturers for customization, testing, and deployment.
Google ships endless new versions without any respect to this, and leaves most of the work, and the cost, on manufacturers, who already had thin margins to begin with. Especially on entry level phones, hardware costs have been a race to the bottom.
These people are afraid to point out that Google has created an OS that's hard to update universally, requires a custom deployment for each and every hardware model, and is shipped without working hardware support for most of the hardware using it or even a modicum of stability even on their own reference platforms.
if the linux kernel driver model wouldn't be so hostile to driver backward compatibility, it'd be trivial for soc provider to have a single device driver delivered with the soc and cut the certification costs by orders of magnitude.
this is something google could fix, btw, creating something like FUSE but for wireless chipset integration and keeping that backward compatible.
Manufacturers can get their drivers merged into the mainline kernel and they won't have a driver backward compatibility issue to worry about.
They choose not to for business reasons (market segmentation/artificial hardware limits via drivers) or because they are being cheap. And because much of the market doesn't know enough to care, it mostly works for them.
In the long run I think it limits the IoT ecosystem from progressing and it's a practice which must stop. If your hardware is not fully supported by the mainline kernel, it may as well not exist.
The argument isn't bad per se, it's just that reality disagrees.
and these people are usually out of touch enthusiast-types who are happy with half-baked releases that require a lot of tweaking, rooting, and replacement launchers and such to be usable. The 5.0 release a year ago was just a nightmare. It was slow and introduced a significant memory leak that I find impossible to believe that Google didn't know about at launch. But Google likes to cheer on its product and have Nexus yearly launches. Tying your physical product to your software one seems a questionable strategy at best, and a boneheaded decision at worst. I'd be happy with a new version every two years with significant refinement done in the meantime.
I have a Nexus and I fear every major update. It just breaks my phone and causes me stress until a .1 or .2 update finally fixes the issues Google introduced. If we're in a position where even Nexus owners are scared of updates, then there's something fundamentally wrong here. Worse, this morally empowers OEMs to say, "Well, google releases unfinished code, so we'll going to skip this update for a while." Honestly, the phone OEMs that went from 4.x to 5.1, skipping 5.0, were on the right side of history here. Every other Google update is half-baked and everyone knows it.
I don't think Google is giving itself enough time to iron out the issues. Slowing down this train is the only solution here that makes sense, imo. Meanwhile my ipad is bulletproof. Sure it gets minor issues now and again, but Apple's QA level is leaps above Android.
Lastly, the mobile revolution, at least in terms of phones, has long matured. Every update is just another denial about how close we are to the innovation grave here. There's just nothing compelling that requires frequent updates. Its time to move into a refinement period, the same way MS took NT4/2000 and made XP/2003 or Apple took a 3-4 year innovation break between Tiger/Leopard and Lion/Mountain Lion.
I switched to iOS on iPhone 5S last year after Lollipop being a shitshow, but iOS 9 has shown me that Apple isn't anything better, because they have to sell new hardware, so gimping older devices on purpose is the name of the game. I don't care about version numbers, if the OS is a stutter show after an update.
All those iOS update bragging would be interesting if Apple would allow downgrades to previous major versions of iOS until next one gets released.
Now a rule of the thumb is find the device/software you like right now and decide about updating it only after research of feedback.
After the signing window was closed, I was stuck with it. I tried ongoing betas until 9.2, but then just had enough and sold it. I refuse to buy into the Apple greed of selling new hardware with shoddy software updates and so far I've read at the time, this wasn't the first time (iOS 7). The majority of non-power users just think "hey phone is too old I have to upgrade" and buy a new device.
I mentioned fear of updating because this was the second time in one year I was burned with an update on two different platforms.
The grass isn't greener on the other side.
Statistically, I don't know what percentage of people want upgrades vs don't, but anecdotally, I feel that most non-technical users don't want upgrades because they've learned how to use the phone and they're comfortable with it.
I wish it was disintermediated from the carriers. I bought a generic android phone from BLU last year for $99 and wanted to update on my own to latest version of android, and it was such a pain to figure out how (I gave up).
Now it's mostly backend stuff as Play Services regroups much more of the user visible stuff. Boring is good too though, just not as exciting
This is becoming more and more important to me as I have ranged from the updating on nightly builds to only updating when I see real performance benefit - and even then I wait till I see more users from my model and service provider commenting on how much the change benefits them.
*edit - of course application updating is an entirely different process and I allow this to update ASAP due to unlimited data plans but updating core os features has become more and more towards the above versus the application update idea.
The problem is the carriers and manufacturers, not the users' wants. They aren't updating enough. And I can totally understand the pain for them, having to QA all those new versions and release them.