In video games, going all the way back to the Atari 2600, we've seen plenty of new types of input controls tried and failed. Joysticks, rotary controls, gamepads, keyboards, mice, trackballs, touchpads, light guns, robots, power gloves, giant floor-mats with buttons, analog sticks, analog buttons, vibrating controllers, skateboards, drum kits, guitars, microphones, steering wheels, arcade sticks, etc...
Some were great, some were terrible. The "great" tended to be so because they were tailored to the software they were programmed for. Playing Guitar Hero with a guitar controller is good, but playing Call of Duty with it is bad. There are plenty of different types of inputs that exist and will be invented, but some work best for some things and terrible for others.
As you mention, Microsoft Office is designed around a keyboard and mouse. A touchscreen is a hybrid of the two, more portable, but clunkier. An accountant who lives by arrow keys, hot keys, and their numberpad in Excel is going to hate a touchscreen. For MS Office to work on a tablet, it'll have to be re-designed from the ground up, and even then it may not be superior to its desktop counterpart in an office setting.
I don't know what MS wants to do with Win 8. To think all desktops are going to die and become tablets is wrong, just like thinking that T.V. was going to kill radio. Both have their places, but a tablet != desktop.
Maybe you're right that they just want a "quick success", but that strategy hasn't paid off for them recently (Zune vs. iPod, Bing vs. Google, their fragmented mobile efforts vs. iPhone). MS has been in reaction mode for a while, and now they're reacting to the iPad with Metro.
One thing Microsoft does have going for them is the corporate market; Apple hasn't directly targeted that yet, though to think that Apple is not going to go after it shortly is foolish. MS is entrenched in here with Active Directory and Exchange. But cloud efforts are going to shortly give that a run for its money.
Thanks, I agree. To reiterate something that I think you're saying (correct me if I'm wrong), "An accountant ... in Excel is going to hate a touchscreen. For MS Office to work on a tablet, it'll have to be re-designed from the ground up." Ok then, why would Accounting firms switch to Metro?
With multiple competing options (some in the cloud), surely one of them will cater to the old UI that uses the arrow keys.
I think you're correct that with multiple computing options, some cloud-based, you'll see tailoring to specific jobs. Accountants, lawyers, salespeople, programmers, all do slightly different things, and will need different UIs. Also, as contexts switch, so will devices. Lawyer in litigation making a presentation will need a different device than when sitting alone in an office.
Specific to accounting, my anecdotical experience has been that many accountants don't like change unless it makes them more productive. If it's confusing or has a large learning curve, expect lots of complaints and resistance. I upgraded a few with new computers, from Win98 to Vista, and from Office 2000 to Office 2007. The "ribbon" UI was so terrifying to them that they literally unplugged the new computers, plugged the old ones back in, and used them for 9 more months before they finally crapped out.
Any type of manual data entry needs to be done really fast and accurate. Keyboards are better at this than touchscreens in most cases. Touchscreens will have to become more keyboard-like to compete (this is going to begin happening soon, but we're a few years away before it becomes good enough: http://cnettv.cnet.com/senseg-demos-prototype-touch-feedback...).