By Jared Chausow
By Katie Toth
By Elizabeth Flock
By Albert Samaha
By Anna Merlan
By Jon Campbell
By Jon Campbell
By Albert Samaha
Efforts to eliminate keyboards generally involve settling for less precise input, while compensating with machine intelligence. One research project by Ericsson Radio Systems of Sweden is a textbook example of the trade-off. Wearing gloves with pressure-sensitive fingertips, the touch typist taps against any surface without reaching for specific key positions. How do you get 26 letters from only 10 fingers? Each finger is responsible for the keys it would normally reach for: a tap of the left index finger, for example, represents either f, r, t, g, b, or v. The computer then does dictionary and grammar checking on the inputted letter possibilities and comes up with a surprisingly accurate reconstitution of what the user must have meant, like an elaborate version of the methods by which Movie Phone knows which movie you've selected using telephone touch-tone spellings. Gloves have a certain cyborg charm, but the intelligence of the system is what pushes it way ahead of the fixed keyboard: it harnesses what computers do well (probability, statistics) to make our job (finger wiggling) a little easier.
Voice-recognition systems also rely on such guessing, especially to generate spellings for something like "Please write to Mrs. Wright right now." Voice-recognition researchers believe the day when keyboards can be tossed aside is fast upon us. But articulation hurdles, such as getting the software to distinguish a "vase of begonias" from a "case of pneumonia," are still tricky. "There's the mechanics of training the system, but then there's really learning how to speak in order to get the best results," explains Nicole Yankelovich of Sun Microsystems. Still, at least the software is beginning to understand something humans do naturally.
A keyboard just doesn't appreciate any of our hard-won individuality. One possible replacement, however, offers a reward for the hours spent in grade school practicing penmanship. The CrossPad, developed by IBM, is another technology simply awaiting its alpha- consumers. It uses a digital pen that writes with normal ink on paper but simultaneously creates a record of the handwriting. After the strokes are uploaded to a PC, handwriting-recognition software translates the scrawl into alphanumerics. "The keyboard forces you to translate your ideas into a representation that the machine understands," says James Landay of the University of California at Berkeley. "It's totally unambiguous." But the pen stroke, he argues, enables a complexity that the keyboard does not allow. Much sophisticated information is contained in a quick squiggle in the margins. "Sometimes I want ambiguity. Maybe not in writing, but for art or design it's really important." This technology captures a handwritten style that people have spent their doodling hours perfecting.
New interfaces can help the digital slip stealthily into social settings where a keyboard seems clunky and imposing. Landay prefers his digital notepad's versatility. "A keyboard can be socially disruptive in a meeting and takes my focus," he says, "whereas a pad is acceptable; it fits how we think of recording information."
Trying out new interfaces might not be everyone's bag. The customizing, learning curve, and frustration with accuracy rates can be a deterrent to the casual user. People who are stressed out might not find wacky guesses amusing. And switching over work habits to another system of devices can be simply an exchange for new and different health problems. "People are experiencing vocal fatigue now from using speech recognition all day," explains Gregg Vanderheiden, director of Trace Research and Development Center, who studies interfaces for people with disabilities and has seen alternate input modes become just as unnatural as typing. "When you're talking to machines you tend to talk continuously and [are] very careful to articulate, so you tense up and you have great strain."