Since we invented computers, we've tried to find the best way to interact with them. First came the keyboard. Then the Xerox PARC invented the mouse. And the average consumer has been stuck there for a long time. We have touch screens that are more intuitive but require large control areas. Apple Macs have speech recognition that enables us to talk to the computer but, even though it works, the recognition engine needs training and is not perfect. There are also things for which speech is not the most efficient way to communicate with a machine. PDAs and Tablet PCs tried to re-invent the touch screen by giving us a stylus so that we could point to a precise area of the screen. We added more buttons and scroll wheels to the mouse. Apple even added a squeeze to its Mighty Mouse! All in an attempt to make our interaction with the computer more intuitive.
But the problem with the technology so far is that the input is always sequential: you type one letter at a time, click on one icon at a time, speak one word at a time, etc. Compare this to the way we, as human beings, interact with our world: we can manipulate several objects at the same time such as turn up the volume of the stereo while browsing for the next CD to play. We can't do that with our computer. Or rather, we couldn't until now.
Some smart people at New York University are currently working on something that could transform the way we interact with computers. They call it Multi-Touch Interaction (the page is long to load as it has en embedded video). Take the time to watch the video or at least look at the images taken out of it. This thing is a tactile screen on steroids: not only can it detect contact but it can simultaneously and independently detect contact from multiple fingers, including multiple users, and act accordingly. Have a look in particular at the demo applications with the map and the image gallery. Now, this is a much nicer way to interact with the machine: stretch, move, rotate, flip in the same way you would do with pieces of paper.
The number of applications for that sort of interface could be huge: everything that uses a mouse today would benefit from being able to just directly touch what you are interacting with and interact with several objects at the same time. The immediate application that comes to mind is software design: UML would be fun again! Now, this is a scary thought.
Anyway, I want one of those!