This semester, I am enrolled in a course titled "Natural User Interfaces." The course is taught by Professor Doug Bowman, a genius who is much better looking than his bio photo suggests. But enough of this shameless flattery! What is the point of all this blogging?
Each week, he gives the students in this class assignments for their "Design Journal." Naturally (ha!), this journal is not physical, but digital. I am using my (barely) pre-existing HCD blog as my journaling platform, and I will be tagging my class-related updates with the "NUI" tag.
The first assignment is the following question, and my answer follows below.
What is your opinion of current 'natural user interfaces?' What would the ideal NUI be like?
"Natural user interfaces" has no set definition, but almost all descriptions of the concept involve futuristic, or at least bleeding-edge, interfaces. Gestural, touch, and multimodal interfaces are the hot new paradigms in interface design, and people describe them as "natural." It's important to remember that every interface that is currently dominant caught on because it was more natural than those that preceded it. Keyboards replaced punch-cards because of their real-time input and the ability to enter recognizable human glyphs instead of binary code. Monitors replaced paper printouts as the dominant output devices because they could display dynamically changing information in a more natural way. The mouse, when it was first introduced, was a far more natural way to select and interact with items on a monitor than keyboard-based navigation. Now, multitouch interfaces dominate our mobile devices because they allow direct interaction with the objects we see on the screen. Multitouch is certainly more "natural" than using a mouse, but all of these interfaces are on the "natural" continuum -- to say that a touch screen is natural while a mouse is not is to make a rather arbitrary distinction.
With that in mind, I would describe the current state of natural user interfaces to be "we've come a long way, and we have a long way to go." There are a lot of seemingly useful interfaces that seem to be perpetually just out of reach (metaphorically and perhaps literally), like natural language recognition and deviceless gesture control, but they will find their place when the right combination of technologies finally exists to make a truly natural experience. Touch screens existed for more than a decade before the iPhone debuted, but they remained a niche market until the necessary materials, innovations, and computing power were available in a small enough package to make lag-free multitouch practical. I predict that other interface paradigms that are struggling right now will follow a similar trajectory. Maybe the Kinect is the Newton of device-free gesture control: a good idea released to market before all of the necessary pieces were in place to make it useful.
There will never be an "ideal" NUI. That suggests that we will someday achieve interface perfection without the need to improve it any further. On the contrary, we should be seeking the best available NUI for a given problem, with the expectation that it will be made obsolete after a little while by a better one. We will know when we have come up with a truly natural user interface when people respond to it the way they responded to the mouse years ago. It won't be an interface that makes people say "Oh my god, that is so cool!" It will be the interface that makes them say "Oh my god, finally!"