Multi-touch demo at http://cs.nyu.edu/~jhan/ftirtouch/index.html is cool. However these are just first steps on the mountain of inventions that are possible with multi-touch.
<aloril> hmm.. 3D multi gesture ... <aloril> 2 high resolution cameras and .. ;-) <PirateHead> and then suddenly using your cellphone looks like casting spells in old fantasy fic flicks
Fully utilized potential of even 2D only pressure sensitive multi-gestures that can detect different fingers might look like somebody is weaving spell(s). Probably most of time fingers would touch screen with varying strengths.
Here is one quick idea: Use middle finger to enter text using something like http://www.strout.net/info/ideas/hexinput.html . Use index finger to change mode to text entry/ search (file/google/wiki)/ cut/ paste/ move/ etc... A bit like vim, except both modes (and more modes) are available at the same time. Thumb could be used for some additional modes. For example change entry mode from letters to most common syllables or to most common words or depending on mode to most common movements, etc.. This of course still doesn't fully utilize potential of even 3 fingers.
Above is just one quick idea of what is possible.
With full 60cm*60cm (about 24") pressure sensitive multi-touch screen that could recognize different fingers uniquely you could probably input faster utilizing all 10 fingers in spell weaving than any traditional keyboard + mouse can do. However learning full potential might require multi year full time training to become competent wizard in spell weaving ;-)
You might want to also utilize some 3D potential too, tough full 3D might be quite tiring.
It seems that Neo1973 v1 is single touch only. Maybe its still possible to have limited fake multi touch. Software could recognize that average point suddenly jumped and correctly deduce where second finger touched/gestured. More might be possible by using some 'black magic'.
A bit releated idea in paper world: Shorthand