New Input Devices

Let's face it the input devices available to us are pretty boring. Whether it's keyboard plus mouse on the PC or gamepad on the videogame, there's a need for something more. Unfortunately, none of the ideas that I have seen in these many years has been worthwhile.

Brainwave Headband

The absolute worst idea was the "brainwave headband" (also known as the Atari MindLink). Atari experimented with this idea in the early 80s. The concept was simple: Control the computer with your thoughts! A headband with simple detectors would detect your brainwaves and pass the results on to the computer. Such a device was built and actually worked after a fashion; you really could play a videogame using the device. However, the device picked up muscle control neuron activity just as easily as it picked up brain activity, and so it was easier to control it with head muscles than with brainwaves. During intense gameplay, players tended to scrunch their eyebrows around to control the headband, leading to excruciating headaches afterward. The product was never marketed.

Exercise Cycle

A close second in overall stupidity was the "Atari exercise cycle." The idea here was to combat the rather tawdry image of videogames by building a videogame into an exercise cycle, providing good, healthy exercise while you played your videogame. Two design strategies were explored: meandering and fleeing. The meandering strategy offered a screen of pleasant, beautiful bicycle paths among trees. As you cycled along, the scenery changed. This design strategy failed for two reasons:

• First, the sensory input was always much inferior to the real thing.

• Second, it wasn't much fun.

The second design strategy suffered from a completely different problem. What we imagined here was a Tyrannosaurus Rex chasing the player, whose video screen functioned rather like a rear-view mirror. The player had to pedal faster to escape the monster. This was all good fun, and seemed a viable product, until somebody asked what would happen should one of our customers suffer a heart attack while playing. Oops back to the drawing board.

Nowadays, with better graphics capabilities, products using the first strategy have finally come to market.

Force Feedback

One early idea that did impress me was the "force feedback joystick." This little doodad was a joystick with powerful electromagnets built in that allowed the joystick to fight back against you. The demonstration game that the research team put together for this device was quite impressive. It was nothing more than a fishing game. The joystick acted rather like a fishing pole. The rules were simple: the fish would tug at the line, causing the joystick to jerk in one direction. If the player allowed it to travel all the way to the end of its travel in that direction, then the fish broke the line and got away. The fish was programmed to wait random intervals and then jerk in random directions. The kinesthetic aspect of the experience was uncannily realistic; feeling that stick jerk around in your hands really created the impression of a living thing fighting back.

A variation on this is the "force feedback mouse," a mouse with a device that provides extra friction under software control. This could be used to create a tactile sense of boundaries (e.g., the user feels a slight bump when moving out of a window). This simple addition to existing mouse technology provided a useful enhancement, and there were several unsuccessful attempts to market such a device.

The killer problem, of course, was enticing enough software developers to support the device. Even without this problem (suppose, say, that Microsoft builds support for the device into Windows), there are still serious problems. First, the resistance must be tactile, not just visual. It would be simple to build mouse software that slows the cursor down slightly on encountering bumps, but this would provide mere visual feedback; the power of the idea comes from the kinesthetic feeling of the mouse in the user's hand. Unfortunately, mechanical mice are rapidly being replaced by optical mice, which don't offer any possibilities for tactile feedback.

Eyeball Tracker

A particularly exotic input is the "eyeball tracker." This little doodad watches the eye of the user and figures out where it's looking. The algorithm is surprisingly simple: You need merely compare the iris with the cornea; the offset of their centers gives you an accurate indicator of the direction in which the eye is looking. The problem is made difficult, however, by the rapid motion of the eye, which jumps around wildly in compiling a complete internal representation of the scene. With fast enough processing, this problem can be overcome. But even then, there remains the much more subjective task of determining the user's intentions from his eyeballs. Humans are capable of furtive glances, blank stares, sidelong glances, nervous peeks, and all manner of other visual behaviors; what is the poor computer to make of all this activity?

Microphone

The one alternative input device for which I have high hopes is the microphone. The hardware is cheap, the software for voice recognition is well-developed, the machine cycles required are now available with the current crop of CPUs, and most computers now have microphone input jacks. It is a simple matter to integrate voice commands into existing software; indeed, for a transitional period, voice commands could merely duplicate existing keyboard-driven commands.

The power of voice commands lies in two factors:

• First, voice is a completely different input channel. The user can use it while continuing to use the hands for mouse, keyboard, or game controller.

• Second, voice is familiar to users, and we often use voice in conjunction with our hands.

The objection is often raised that voice input is subject to external noise emanating from other persons nearby. I give this objection little credence. Most computers are used privately; loud nearby conversations would be just as distracting to a user working on a spreadsheet as to the computer trying to hear and understand that user. Moreover, the use of head-mounted, noise-canceling microphones dramatically limits the ability of others to interfere. At this moment, thousands of drudges crowded together in telephone solicitation sweatshops are carrying out thousands of simultaneous conversations in close proximity without noise problems. They could just as well be talking to computers with just as little interference.

Specialized Input Devices

There remains a number of interesting input devices custom-designed for particular games. The most impressive of these is the set of pads provided with Dance Dance Revolution; they permit the player to dance with the game, with the player's dance steps becoming inputs to the game. This was a brilliant use of specialized hardware to extend the expressive reach of a game far beyond the traditional. Such specialized input devices are expensive and impose considerable risk on the company placing them on the market; nevertheless, as Dance Dance Revolution shows, such an approach can produce successful products.