I love the concepts in SixthSense, and also in this one:
TED Blog: The Buzz: Ramachandra Budihal augments reality
Really, I think the SS folks made a serious tactical mistake in basing their project off the projector. People are private, they don't want that information projected! Furthermore, if I want to pull up a webpage, why should I have to look for a "screenable" surface to project my browser on? Maybe I am outside in a field and there is none! But the gestural interface is GREAT!
I see the future as a combination of the two projects, with a healthy dose of local networking thrown in. Use a HUD style display affixed to glasses for information display. Use the camera to constantly evaluate the environment around you, and also your hand positions. Use gestural interface for anything that is needed. You could even use a camera pointed back at the eye with some eye tracking software to keep track of where the user is focusing their vision, and then of course integrate that into the UI.
Local networking comes into play for collaborating with nearby users. Think grand scale technology that is eventually miniaturized to be something that EVERYONE carries. Just like everyone has a cell phone, if you assume nearly everyone will have one of these devices, then you can assume integration with them. So I could "look" at a nearby person, make a gesture or give them "the eye" so to speak, and they would be instantly be allowed to see what I wanted them to see. So if I am looking at a map and trying to give them directions, that view would immediately be shared with them.
Here's a few examples of tasks and how they would be accomplished. Note that one drawback of my idea is that I think it would require a lot more video processing than the current systems, but not so much that its out of reach. When I talk about using a HUD display, I envision it monitoring the world around and formatting the information to display correctly on that. So things would not always be formatted in flat 2D as we are accustomed to seeing it, but rather in 3D, like they were being "projected" virtually onto nearby surfaces. This is the greatest benefits of the HUD system however, because for information that is better served in the traditional 2D environment, you can always go back to that, seamlessly switching between 2D/3D display of information, even doing both simultaneously for different sets of information being displayed at the same time.
1) Dialing your phone: simply look at your hand, arm, or any other surface and initiate the gesture (with eyes or fingers) that would begin that process. A virtual number pad would appear over your hand, just like in the SS video. /the only difference is that the pad would be displayed onto your eyeglass HUD display, so only you would see the number pad on your hand, no one else. Use your fingers to dial normally.
2) Look up ticket information: Simply look at the ticket. The information you need will be displayed on the HUD overlapping the ticket image, or next to it.
3) Shopping: pick up any item, focus your eyes on the barcode to "scan" it. Voila, information.
4) Books: same as in the SS demo, look at the book and its information will pop up. IF a cover search fails, you can always glance at the barcode on the back. Camera will keep track of the book, so further information can be displayed as you begin to read into it.
5) Newspapers: cool idea adding a video screen to a newspaper! BUT why add the screen when you're wearing it? a gesture around an article could have the PDA searching for video relating to that article and popping it up on the screen next to the article.
Just a few things off the top of my head, I think you see what I'm thinking......