Several good points made; definitely worth the read. I completely agree with him on the topics of eliminating user confusion by disallowing people even the opportunity of failing to run legacy applications, but my bottom-line-point has been the same as Gruber’s from the get-go:
[If MS ships the legacy desktop on ARM] too many developers will think, Why rewrite and redesign my entire app when I can just recompile it for the classic Windows desktop on ARM?
Microsoft needs people to make great Metro apps to catch up with the battery life, stability and perceived speed of the iPad1, and they’re not going to get that if people can just punt and make legacy desktop applications instead.
The “perceived speed” of the iPad is not without merit. My girlfriend has more than once commented that things on the iPad just “feel faster,” despite owning a 12 month old AMD quad-core system. ↩