The point, I think, is to have something called "Windows 7" that runs on netbooks. Microsoft doesn't want to be caught in a position where their flagship product doesn't run on the next big computing platform, letting some competitor become the de facto standard. This was the point of licensing XP for netbooks (a suboptimal solution), and the point of porting Windows to the OLPC XO.
The concerns over application compatibility are moot. If people actually cared about that, they wouldn't be switching to Linux machines.
Running Windows Mobile is not a bad idea, provided they rebrand it "Windows 7 Mobile" and polish it to the point where it looks as much like a "real" operating system as Linux does.
Wouldn't the purpose of porting to ARM be to run on smart-phones and other handheld devices.
Seems to me most netbooks are running on the Atom processor which is x86 based.
The comments that the author makes regarding peripherals and legacy applications shouldn't be as big an issue, as you're not dealing with a 'this should run on a pc' mentality, but rather 'will this run on a mobile device'.
Nobody looks at their iphone and wonders why they can't run photoshop. Yet, it is an ARM port of OSX.
He addresses that point in the article. Microsoft already have the equivalent of iPhone OS X, it's Windows CE for embedded devices and which is also the underpinnings of Windows Mobile for PDAs and Smartphones.
It doesn't run standard Windows apps but neither would an ARM version of Windows 7.
The concerns over application compatibility are moot. If people actually cared about that, they wouldn't be switching to Linux machines.
Running Windows Mobile is not a bad idea, provided they rebrand it "Windows 7 Mobile" and polish it to the point where it looks as much like a "real" operating system as Linux does.