http://www.informationweek.com/blog/main/archives/2008/02/mi...
As for crashing, depends how you define "fault". If a program causes Windows to crash, where it wouldn't have without that program, it could be said to be the program's fault. I tend to agree with you, since the point of Windows is to be a platform on which programs run, and found myself on that side of the argument against a number of PalmOS fans not that long ago.
You could easily say either party (or both) were at fault, but Windows will be the one more damaged by it. Microsoft would tend to agree too, which is why they spent so much effort making XP resilient to that sort of thing by SP2.
If a program causes Windows to crash, where it wouldn't have without that program, it could be said to be the program's fault.
This is such an irritating point of view. In a previous job we had a boss like this, who said our code must have bugs in it because the JVM was crashing with a segfault.
Which is not to say that Microsoft didn’t make some mistakes in launching the OS. They clearly made a few. For one, they gave OEMs too much leeway in deciding what they could or could not slap “Vista Capable” stickers on. And even though it wasn’t entirely their fault that manufacturers screwed this up, it reflects badly on Microsoft. It’s just like when a program crashes your PC. It might not have been Windows’s fault, but you get pissed at it for not preventing it, because that’s its job.
Microsoft told the OEMs exactly what they could stamp Vista capable. It is entirely Microsoft's fault.
If a program can crash the OS, it is the OS's fault. (regular user mode program that is, third party drivers are a different story).