Certainly in my experience of talking to real people - friends, family, etc. - who are not techie, none of them care about the latest stuff. Most of them are running some several year old version of Windows or OSX and running a mixture of older and newer apps. They have a very reasonable expectation that software just installs and works, and that software they bought 3 or 5 years ago just installs and works. They also have a reasonable expectation that brand new software still works because the platform they are on is not such a moving target that the developers of those apps are able to target something a few years old and reach everyone.
It's really not rocket science. It's *boring* science. It isn't fundamentally incompatible with distributions, it's fundamentally incompatible with "oooh, shiny". It requires extreme restraint, a co-ordinated, standardized approach, and dare I say it, a little intentional "stagnation" in terms of the burn and churn. And also, I'm not saying this has to happen. I'm only saying it has to happen if we care about getting mass adoption of Linux as a desktop and having articles like this one - and a number of other ones recently - end very differently.