Today, I am willing to suspend my disbelief, the same way I do whenever I turn on the cable and some movie adapted from a comic book is playing. Because Microsoft's developers would have done a truly Superhero thing if they've been able to shrink Windows down such that it doesn't rely on Intel's brawn.
(Digression: in the mobile age, Intel's x86 CPU increasingly looks like this guy: super-strong, but pointlessly so (why do I need six cores if none of my software can take advantage of them?). ARM CPUs, with their long battery life and efficient power usage, are more like this yoga dude.)
So here are three ways I think this thing could play out when Steven Ballmer steps on the CES stage in January, in increasing order of likelihood:
1) Microsoft shows off a polished-and-optimized version of Windows 7 for ARM tablets that will ship on tablets in 2011.
Advantages: Immediately makes Microsoft a player on tablets, due to Windows' massive ecosystem of software and its army of developers. "This will bring a tidal wave," my colleague at Sybase, Tony Kueh, wrote to me in an e-mail, pointing out all of the .Net apps that will immediately be able to run, and all of the native apps built with Microsoft's popular Visual Studio tool that will only need a simple recompilation.
Disadvantages: It pisses off long-time partner, Intel, which did a great job co-opting netbooks, but has yet to prove its Oak Trail processors are efficient enough to make them relevant for tablets. Also puts it in competition with its largest OEM, HP, which is backing its own WebOS. Also renders Windows Embedded Compact 7, which it had been pre-announcing all of 2010 (though to mostly-deaf ears), totally irrelevant.
Likelihood: 25%. Rumors about Microsoft porting the PC version of Windows to ARM have been circulating as long as those the Apple tablet ones (and look at how that proved true). Where there's smoke, there's probably fire...Also, it's believable that Microsoft's Windows chief, Steven Sinofsky, was able to throw a few of his several thousand developers onto a skunkworks ARM project, especially after the success of Windows 7. Kueh, who was a longtime employee at Microsoft who helped launch Windows Mobile, believes in this scenario (though he says he has no insider knowledge), pointing out that "the NT kernel is designed for multiple architectures -- recall that NT used to support x86 AND MIPS." (MIPS is a type of CPU that was the equivalent of ARM in the 1990s)
2) Haha, did you think we were actually bringing Windows 7 to ARM tablets? We meant Windows Embedded Compact all along.
Advantages: Remember those dot-com-era Pocket PCs? Those were ARM-based devices running an earlier version of Windows Embedded Compact, then called Windows Compact Embedded (the flipping of the last two words, conveniently, means no more puns at WinCE's expense). So making Windows Embedded Compact 7, the latest version, run on ARM tablets is no problem. In fact, as pointed out above, this is the drum Microsoft has been beating all year.
Disadvantages: Did you know that Windows Embedded Compact is basically the same as Windows Phone 7 under the hood? Most of us don't, which testifies to the lack of mindshare Win Embedded Compact has among even techies. Unless Microsoft is going to start investing heavily in marketing Windows Embedded Compact (step #1, change its mouthful of a name), this feels a bit like too much of the same old, same old.
Likelihood: 30%. Initially, my cynical self thought this was the most likely scenario. But I think Microsoft's PR consigliore, Waggener Edstrom, is smart enough to advise against such a bait-and-switch strategy. Hell hath no fury like a reporter who has been tricked.
The great-great-great-grandchild of the OS that ran this 1996-era iPaq could be Microsoft's new Windows OS for ARM tablets. Now aren't you excited?
3) Microsoft shows off a rough version of Windows for ARM tablets that it says will ship in several years when it unifies its PC-tablet roadmap under Windows 8. Meanwhile, it pushes Windows Embedded Compact for ARM for the next several years.
Likelihood: 45%. Microsoft has always loved to pre-announce, so this makes all sorts of sense to me. Microsoft's licensing deal with ARM to make ARM CPUs is only six months old after all. So while I think Microsoft is embracing ARM, I doubt Microsoft is shifting this fast, despite how far the rumors go back - see Mary Jo Foley's coverage. Also, it would be too much of a radical power shift towards Sinofsky and away from Bob Muglia, the senior VP who runs Server & Tools (Exchange, SharePoint, cloud, etc.) and whose division only absorbed the Embedded group 3 months ago.