This guy got it right

For the average user, Linux is actually easier. Power users have problems because they expect Linux to work like Windows, and expect to do a lot of more complex things with it. Average users just click on programs and type letters, so Linux pretty much just works once it's installed.

But either way, the point stands that if Microsoft killed of XP support tomorrow, people would not move to Vista en masse. They would look seriously at alternatives, and keep XP in the interim. On their next machines, they would get a Mac, or businesses would start pushing Linux to the desktops. Vista doesn't do enough to make it worthwhile to 90% of people, and changes enough to make the actual switch a pain in the rear. When you combine a change that requires relearning and work with little to no perceptible gain, you get Vista's position, which is active resistance to change. Win98/ME -> XP was a very perceptible gain. And that's how the actual upgrade went... Win 2000 was mostly business-class. It wasn't on consumer machines en masse. WinXP -> Vista gives very little actual benefit, and is actually MORE of a change to the interface and such than Win98/ME to XP was.

Now, when you've accepted that you're going to have to do a pretty much wholesale change, what do you do? Do you pay $150 for a new OS that every says is buggy, or do you pay $0 for a version of Linux that everyone says is buggy? And if you're a business, do you pay $150*X where X is the number of machines you have, or do you pay $0 and invest in quality personnel to roll it out? Just a couple things to think about I will probably install Vista in a vm eventually, but the few experiences I have had with it do not leave me with a very good taste in my mouth. Hopefully Windows 7 will give the perceptible improvement that Microsoft needs to have people switch. If they're reading this, I'd suggest they implement open standards and make it easy to access things, since they have to compete on quality now, since they can't compete on price