So what if Lindows/WINE/whatever actually *does* get to the point where it runs native Winblows applications nearly as well as the original. Would that necessarily make Linux viable on the average desktop?
I tend to think not. After all, that was one of the big draws of OS/2, that it could run Win 3.x apps as well (generally) as Win 3.x. I recall some commentator commenting (as commentators are apt to do) that just doing what the competition does is not enough; you have to do something *different* from the competition. I.e., "Why should I use this to run Windows apps when I can just run them on Windows?" Or from a developer's perspective, "Why should I develop for OS/2 when I can just write for Windows and OS/2 can run them anyway?"
Supposedly better (from the basic user's standpoint) multitasking, etc., wasn't really something to differentiate it enough to make people switch; and admittedly there was a performance hit in many circumstances for the sort of "emulation" OS/2 used.
Myself, I loved OS/2, but eventually realized that with a couple of exceptions, all I was doing with OS/2 was running Windows apps in "emulation", OS/2 development was going nowhere (for many of my basic apps, e.g., SPSS), and Win95 apps wouldn't run on it.
IMO, unless there is some sort of standard file format (XML maybe?), the whole WINE thing is totally moot.