At the hugely fun animation-themed LA Flash meeting two months ago, I was stunned by the lightning speed with which the presenters jumped around the Flash IDE. They were interface acrobats, as supple as the lemur monkeys swinging on the fake trees at the San Diego Zoo.
Memorized keyboard shortcuts are faster than pointing and clicking, of course. And a series of consecutive keyboard commands works in conjunction with muscle memory: Before an animator even thinks Select All–>Copy–>Delete–>Paste in Place, his fingers have signed, sealed and delivered it.
That got me thinking: can designers take advantage of muscle memory in point-and-click interfaces, especially interfaces built for a lot of back-and-forth, revision, searching (RIAs)?
In the Minority Report-like future, it'll be cake: Tom Cruise's hands will dance on a hologram, and plot-driving relevations will pour forth…but now?
One thing to do is to make sure controls are closely adjacent; I doubt tracks in muscle memory will be laid down if users must mouse over miles and miles of pixels to get to the next button. A tightly packed mass of buttons, on the other hand, would facilitate quick action and muscle-like memories.
A design generated from this sole consideration would place the controls at the center of the screen and relegate content to the periphery, like this:
It's true that the mass of buttons could be placed more conventionally at the left or the bottom…but then, what if pressing the center buttons called forth new submenu buttons? These too would have to remain adjacent to the original mass, and if the original mass is not at the center, the submenus might spill over or crowd the boundaries.
Note that the division of content into this tic-tac-toe pattern does not squeeze content into boxes that are smaller than normal (most pages are already broken up anyway)–and keeps 8 "tabs" open permanently, a strength for a working/searching interface.
1 response so far ↓
I’d like to see the idea extended with the concept of a neural network so that the UI learns the user’s behavior and adapts to it. If it learns that you like to click buttons *a* then *b* then *c*, it will see you click *a* then *b*, and it will enlarge the hit area of *c* to make sure you snag it easily.
Or, you could turn a click sequence into a gesture and use gestural pattern recognition: click *a*, then *b*, then do a quick upwards flick of the mouse to complete the third click of *c*.