The Future of Human to
Computer and Communication Hardware Interfaces

Human to computer interfacing is more and more the big bottleneck in computer use. In light of the fantastic advances of information input and access possibilities with modern computer and communication technology, it seems ludicrous that virtually all computer users are interfacing using a keyboard perhaps designed to slow down typists on the first typewriters, with newer keys patched on to the ends. It's well documented that these keyboards cause frequent worker disabilities. We also use a basically immobile, 2-D screen.

Strangely, few people see other options. The goal of simply building faster and faster computers in smaller spaces suggests building them to compete with the human brain. I like to think more in terms of cooperation between humans and computers. We hear figures about ten to twenty percent of the human brain in use at any one time, on the average. What's the equivalent figure for a computer, while it's turned on? How much could we increase both those figures by developing better interfaces?

Here are several possible improvements, mainly in keyboards, going from the simplest to the most elaborate. Professional typists won't like most of these ideas since they would involve relearning typing. But if the interfaces from keys to electronics and keyboard electronics to computer was more standardized, or if we had appropriate adapters, keyboards could be built according to personal preference, and given a chance to evolve to a new and better forms. One keyboard could be used in several different modes.

Improvements to Conventional Keyboards:

1. If you're learning to touch type, lettered keycaps only encourage looking at the keyboard rather than the screen. If you've already learned, you don't need them. Ideally, they should be blank, like in typing classes. For now, if you're learning, you might want to build a cardboard box with an open front to cover the keyboard, so you can't even look at your fingers. The display of keyboard format, if necessary, should be on a chart near the screen (or on top of the box). Keyboard format can then be changed with a different keyboard configuration file and a different chart, without changing keycaps. Watching the screen makes error correction easy, and also helps teach touch-typing. (As computers, and OCR scanners, are used more, we get away from having to re-type while looking at printed copy.)

2. The Dvorak format was a small improvement, but the QWERTY monopoly was preserved. Anyway Dvorak designed for 1930s (mechanical) typewriters. It's time now for computer-assisted development of a new format for computers.

3. Building a keyboard which is different only in shape or key format can be relatively simple process of electrical soldering rather than designing electronics. Some Technical Stuff

4. Each thumb should control two or three keys, instead of two thumbs sharing the space bar. For "now", the thumb keys should include Shift, Control, Alt, space, return, delete. On my "conventional" keyboard, this was a simple operation of re-mounting and re-wiring keys This is how I finally learned to touch-type. It's much easier to hold Shift, Control or Alt with my thumbs while hitting other keys. I've built several experimental keyboards and the above improvement has been the most useful for me. (But I'm not at my keyboard 9 to 5, at least until I started editing these web pages.) Simply adding more keys to the ends of the keyboard encourages hunt and peck typing. (Question: Many keyboards produce different codes for left and right Shift. Does that mean keys can have different meanings for each? There's a lot I don't know yet about key codes.)

5. The keyboard should be in two sections, curved to fit the bent fingers. At least when my fingers are bent to the keys, the don't all point in the same direction. The keys should probably be in columns in line with the fingers. Then having a numeric keypad would only be a matter of switching key definition files, perhaps with one keystroke, to define some keys as numbers. I've seen one keyboard, I think an IBM brand, that has two sections, as adjustable as possible. But each section is still flat, and instead of one space bar for two thumbs, it has two.

Note that on a conventional keyboard, the 6-T-F-C line has a different angle with the horizontal rows than does the 7-U-J-M line. Why aren't they symmetrical, mirror images for the two hands? Because on a mechanical typewriter the lever arms, of T and V for instance, must be side by side. Isn't it time we broke with tradition, and with the structural requirements of mechanical typewriters?

6. The flexibility of software for completely programmable keycodes, means every key can have programmed definitions for normal, Shifted and Ctrl and Alt. Then one can have keyboard definitions for easy one-handed operation, or independent format. With this, and keys in "vertical" rows, the separate numeric keypad becomes unnecessary.

7. Ideally, the thumbs should operate four or six Shift/Control/Alt type keys (requiring a different keyboard interpreter chip). Each "character" key could be programmed to express as many meanings as there are "shift" keys, plus one for no-shift. On my conventional keyboard, there are 83 keys (not counting cap lock and numeric keypad) for a potential total of 332 key codes, but apparently only about 174 codes are used by the computer. Only thirty-six "character" keys (four for each finger) and six "shift" keys (three for each thumb) would be enough to produce 252 programmable codes. Furthermore, keys under both thumbs, or adjacent keys under one thumb, could be pushed simultaneously. This would produce 36 possible combinations of shift-type keys, for 864 possible codes, if you care to use them all.

Advanced Input Methods:

I predict that personal, technical communication in the future will be centered around electronic hat and gloves.

8. Let's glorify the principle of #7, pushing keys in combinations, and forget about the conventional keyboard. Go on to chord keyboards - a few keys pushed in many simultaneous combinations, as on a court recorder's typewriter, or "Stenotype". Ten keys for the ten digits give a maximum of two to the tenth power, or 1024 possible, programmable meanings, "without lifting a finger". The electronics is probably simpler than for conventional keyboards, mechanical hardware certainly.

9. But don't stay long with "conventional" chord keyboards. Go on to the "Electric Glove" (my "invention" as far as I know, before the "DataGlove" of virtual reality, which is intended for a different purpose). For "typing", this would eventually have one three- position switch at each major knuckle and two at the wrist - 3 to the 12th power meanings possible on one hand, 3 to the 24th on two hands. They would all be programmable, mostly as common words, phrases, computer commands etc., as far as the software, your fingers and your memory will support. Not all one-hand combinations are easy to use, unless you start using such a glove as a young child. But with a total of over 282 billion possible meanings (three to the 24th power) who needs them all? Communication might eventually be limited only by the speed of thought, and translation between thought and words. Where will the creative mind go with that? A person with no voice and just one hand (producing only about 531,000 meanings) who learned the operation from childhood, could probably transmit written or "spoken" information much faster than normal speaking.

Display Screens:

10. Build a hat with a with a built-in screen to match the compact input. (We're starting to see this now, but I'm still a little ahead in design.), It should include screens, earphones, microphone and perhaps video camera built into a hat (perhaps not much larger than the caps professional baseball players wear) and connected as a wholistic terminal for multi-band radio, CB, cellular and/or cordless phone, audio and video recorders, TV and computer. Instead of screens in front of the eyes, semi-silvered mirrors in front of the eyes would reflect views of small screens on top or side of the head. Turn screen brightness down to see the surroundings through the mirrors, or raise the mirrors on a swinging visor for direct eye-to-eye, human-to-human contact. A separate screen for each eye would make 3-D displays possible. One or both views could switch between computer, lab instrument readouts, or TV. We're seeing this kind of display now for TV and VR. How much longer for general computer use?

After all this talking, I've found people who are doing the "wearable (ultra-portable) computing" bit, including most of the above. But the outlook for seeing it on the mass market isn't real hopeful.
http://lcs.www.media.mit.edu/projects/wearables/

Another improvement I'd like is screen which could scroll smoothly, rather than line by line. One could then set the scroll speed and read continuously as it scrolled. It would have to scroll evenly in lines per second, rather than bytes per second. Bytes per second would speed it up when short lines come on the screen, though you're reading elsewhere. I've seen one, expensive-looking, monitor that did at least some of this. It wouldn't require a graphic interface, moving up one row of pixels at a time. Minor monitor electronics could take care of most of it. Improvements such as these would make reading "electronic text" more attractive, rather than printing it out.

Since the invention of the transistor, continued miniaturization of circuits has meant that the size of most electronic devices is determined by the interface components. Other circuitry gets more and more complex but takes up less and less space. So it makes sense to have multiple uses for interface devices, as well as much of the other circuitry, combined into one totally portable device.

This is all within the ability of present technology. The main question is how big and heavy will the hat have to be for how much automation and integration of functions, and how soon? Hat and gloves would contain mainly direct interface devices, screen, earphone and microphone, with as much support circuitry as possible being worn in perhaps a belt pack.

One day we might have a holographic screen built into a pair of glasses. One doesn't have to focus the eyes on a holographic screen, and virtual images can potentially be displayed at any apparent distance. This is probably a little while off, when we can create microscopic, electronically controlled pixels. (Today, virtual reality glasses are becoming available, which apparently do the same thing.)

Other Improvements:

11. Condense smart terminal (or perhaps whole computer) electronics to fit in the hat. "Superchips" are being developed, but "superblocks" (several layers of circuitry in one "crystal") would be much more compact, and possibly simpler. Heat dissipation becomes a problem here. This might be overcome with built-in electronic cooling based on the Peltier effect, the reverse of the thermocouple effect, which happens to be quite compatible with semiconductor technology. This would make it possible to put the cooling near the sources of heat, which are also generally the areas of heat sensitivity. This might become unnecessary with the imminent development of ambient temperature superconductor circuits, but it would be especially useful with near-ambient superconductors. Perhaps someday the small screen in the hat will simply be one surface of the one-piece, integrated computer.

12. With these improvements, solar-powered computers might soon become practical. Solar cells might cover the top of the hat.

Must we wait for the desktop and laptop market to be saturated before we make such changes?

Why are so many people afraid to get too connected to computers?

Updated: 97/11/05 (rational date notation)


Send me your thoughts.
Dan Robinson, danrob@efn.org, Eugene, Oregon
My home page: http://www.efn.org/~danrob/