In the first place, I'm not generalizing my requirements. I'm considering the millions of people who inhabit "Dilbert World," i.e. those folks who aren't coders or engineers, but they use computers eight or more hours a day in their work. They are apparently invisible to many Apple (and specifically iPad) users, but they constitute the lion's share of the folks in the world sitting in front of computer screens.
You claim that tablets will ultimately replace PC's as tablets connect to multiple screens, enable multiple resizable windows, utilize keyboards, and in general increase their connectivity. Hate to break it to you, but once all that happens, what makes a tablet a tablet no longer exists. It's simply a smaller form factor cpu. And since most workers sitting at their desks have little use for a touch screen interface, a "tablet" isn't a "tablet" anymore. (Note that touch screen interfaces have enjoyed extremely limited acceptance in most office settings. That's because it's far more efficient to control a computer with your hands on a keyboard.)
It's certainly true that tablets are rapidly increasing their year to year sales, especially in terms of percentage growth. But that doesn't mean PC sales are declining or are likely to. It simply means that one is comparing a product category with a huge installed base to a product category with an initially tiny installed base.
It's also true that desktop systems are rapidly being replaced by laptops. But that's not because most workers spend their time wandering about the halls computing, much less jet setting about from meeting to meeting. It's because laptops consume less energy and take up less space in ever shrinking cubicles into which most computer users are stuffed. And without a 30 lb desktop cpu to find a spot for, the potential for an additional monitor increases. That's a very good reason to replace a desktop with a laptop system (with equivalent power.) It's not a reason to replace either a desktop or a laptop system with a tablet that has a 10" screen, limited interfacing capabilities, and lacking in the horsepower to provide reliable multitasking.
-Those Dilbertians could use an iPad 9x% of the time given what they spend their time on. I know, I spent years and years and years in such environments in another professional life.
-Saying that an iPad or other tablet connected to all manner of other devices doesn't in any way blunt the argument that tablets will overtake and replace PCs, sorry. Your straw man about the the touch interface is simply that (and fwiw a touch interface is far more efficient for may tasks and useful for many more people).
-PC sales are ALREADY shrinking in the developed world year over year. Check the facts. So while yes the tablet market is new (or, well, at least the iPad is) and the PC market is old and entrenched, there is no reason to believe that the PC unit sales trend seen in the developed world won't continue in general, AND will be seen as well in the developing world as the better price, energy consumption, portability, increasing capabilities, and ease of use of tablets helps them find their way into more such markets around the world (there are many many nations on earth where it is almost impossible, logistically or financially, to buy anything other than a pos desktop or laptop).
-You can't have it one way re: laptop space/energy/efficiency/portability and another way for tablets. They both achieve the same net result (or can, with continued evolution) once connected to additional "stuff." And, your final straw man regarding lack of sufficient computing power and/or connectivity options in tablets is, once again, not only just that, but misses the mark:
iOS, the latest and greatest apps, hardware add-ons, and the ipad2, are already far more robust, flexible, powerful and capable of serving a broad array of use cases (and in fact in many specialized niches are already more capable than PC equivalents - see the recent batch of iPad-based audio mixing consoles) than their predecessors of only a year ago. Similarly and by extension, there's absolutely no reason to think that a year from now, the number of things "you just can't do" or can't do well, on an iPad 3, or 4, and with it's surrounding ecosystem, will continue to diminish and asymptotically approach zero. Contrast that with the PC ecosystem which is essentially static* and the argument is only strengthened.
Said another way, increased performance/capabilities in tablets is something baked into my projections of the future.
So any debate involving an analysis of trends which attempts to make projections therefrom but that looks at the two things being weighed as equally static is flawed, especially in this case. PCs are PCs, and basically have been since Windows 95. Tablets on the other hand have been and continue to make leaps and bounds in their evolution.
*As someone else posted, Windows8 tablets may hold the key to turning this debate on its head, or at least complicating it, depending on whether you lump these into tablet sales or PC sales. It remains to be seen how wonderful these things are to use in practice (for my part, I'm not a fan of WP7 and am not overly confident in MSFT's ability to not screw up the experience in W8 on tablets), but in general I would call them tablets and point to them as an example of the evolution of tablets becoming increasingly powerful and capable as has already happened from iPad 1 to 2.