August 16, 2004

User Friendly Gestures
User interfaces: The next generation - Computerworld: "One example is a gesture recognition system developed for the U.S. Department of Defense by Cybernet Systems Corp. in Ann Arbor, Mich. The technology was developed to facilitate silent troop communication during combat. It allows users to stand in front of a camera-mounted monitor and manipulate images, data and application windows by using specific hand movements from a lexicon of roughly 80 gestures recognized by the system. A San Antonio-based TV station is using a commercial version of the product, called GestureStorm, to control computerized visual effects in its weather reports."

Gesture input has received a lot of press lately. You occasionally hear about a specific use of the technology (like in this Dept. of Defense system), with claims of broad future use of gesture input. I think there's still a lot of hype about gesture input, and I think broad usage is unlikely, or at least a long way off - for a couple of simple reasons:

1) Affordances

2) Lack of Standards

Affordance: Affordance is the perceived properties of a thing that determine just how the thing could possibly be used. Gesture input is usually done by moving the hand in a certain way or by moving a mouse in a certain pattern (usually after or while a button is depressed). The problem with gestures is that they aren't obvious. A keyboard or onscreen menu have clear labels - gestures usually don't, and so they place a larger burden on the user to learn, remember and accurately recall the correct gesture for what they want to do. Then they have to perform some small gymnastics - executing the gesture properly so the system can recognize their command. Basically gestures seem to have a lot of the same issues as command line interface input with the added fun of hand gymnastics.

Standards: Who wants to learn different sets of gestures for each application or each platform (PC, cell phone, PDA, etc.)? It probably doesn't make sense to create gestures for low-level functions like copy and paste. After all, how would a gesture be better than a CTRL-C keyboard shortcut? If gestures are used for more complex, higher level functions, then they are likely to be industry, context, user or at least application specific (e.g. "Open a list of all bugs assigned to me" or "Close this order and generate an invoice")

It's possible that standards for gestures at the platform level might be established in the short term (e.g. "Email this file"), but how would they be any better than what we have today?

My prediction is that voice recognition will slowly, but eventually take off in mobile applications. We'll get voice recognition software for composing text messages - only requiring the user to edit the message with a keyboard. Voice menus might also eliminate the need for gesture input in some cases.

Gestures effectively flatten the menu structure - making many or all functions available at once - much like keyboard shortcuts do today. In a similar way (to shortcuts), I think they will mainly be suited for power users - primarily in specialized applications.

No comments: