Chris Harrison, a PhD student in Human-Computer Interaction at Carnegie Mellon University, is full of interesting interface ideas. One of his latest projects is called OmniTouch, whose prototype design uses a shoulder-worn, depth-sensing camera/projector to create interactive "touchscreens" on anything from a nearby wall to your own forearm.
The technology is amazing: OmniTouch’s screenless interface can distinguish between a touch and a "click" (i.e., a command to do something), auto-detect the size of the interface surface (e.g., it will project a short strip onto your arm, but a large rectangle onto the wall), and even recognize the orientation of the image in 3-D space (if you tilt your screen-hand toward yourself, it will consider it "private"; a more flat orientation will be deemed "public"). But oh, that rig—are normal people really going to be walking around someday with shoulder mounted screen-guns like the Predator?
Obviously, OmniTouch is a proof-of-concept and Harrison realizes that the physical setup has to be miniaturized and refined to viably offer "Wearable Multitouch Interaction Everywhere." And the idea of turning any tactile surface into a personal computing device is exciting. But if mainstream computer-interface design is naturally tending toward the transparent and invisible—so intuitive that it quite literally disappears—schemes like OmniTouch still seem complicated and fussy. Any surface you want to "screen-ize" has to be oriented correctly toward a lens with a single perspective, the surface itself has to have appropriate attributes for receiving the projected image, and the image itself has to be clearly visible over a very wide range of lighting conditions. And what happens to the whole idea of public space when every douchebag on the street is splashing their email and Facebook clients onto the nearest wall or Starbucks counter? (You thought unwanted cellphone chatter was bad—what if you caught some perv on the subway reading their morning headlines off the back of your rump?)
Maybe advances in microlensing technology and computational photography will allow entire flexible surfaces to function as cheap, disposable camera/projectors, so worrying about whether your OmniTouch-like device is "pointed" correctly will be moot. But that’s a long way off, and this kind of sci-fi-ized interaction seems more suited to specialty applications (warfighting? EMTs?) than mainstream adoption. That said, Harrison and his cohorts are the ones who will be building the future of human-computer interaction, not armchair observers like me—so whatever he’s working on is worth taking seriously.