Date: 22 Oct 1994 09:38:52 GMT From: bob@bowman.dcs.qmw.ac.uk (Rosenberg) Subject: Re: CHORDED KEYBOARDS Organization: Queen Mary & Westfield College, London, UK Newsgroups: comp.human-factors NNTP-Posting-Host: bowman.dcs.qmw.ac.uk X-Newsreader: TIN [version 1.2 PL2] Mgauf (mgauf@aol.com) wrote: : CTDNews is a monthly newsletter covering RSI in the workplace. For our : Nov. issue, we will be examining the health and safety aspects of split : keyboards vs. standard keyboards. Also we hope to examine chorded boards. : Does anyone know of any research -- published or unpublished -- on : chorded keyboards? I am doing some graduate work in chord keyboards. I can give you a list of articles and an preview from a paper on alternative text input devices, which should be ready for publication in a few months. If you have any questions, comments, or critiques contact me at bob@dcs.qmw.ac.uk ---------------------------- Chord Keyboards Most keyboard alternatives are just modified versions of the standard keyboard. A character is made by pressing one key, or one key and one shift key simultaneously. This layout allows any number of characters, as long as there is room on the keyboard. To reduce the number of keys, consequentially the size of the keyboard, there must be more shift keys. Using this method, each key would have several characters associated with it. A chord keyboard is the logical conclusion to this process. There is a limited number of keys, usually one for each finger. Pressing a key by itself makes one character, like a normal keyboard. Pressing two or more keys simultaneously also makes a character. Pressing combinations of keys in this way is called _chording_. Chord keyboards were first used by the US Post Office in the 1960's for entering numbers for mail sorting. Most early research on chord keyboards concentrated on limited applications. In the 1980's chord keyboards were reevaluated and applied towards a general text keyboard. The major disadvantage to using a chord keyboard for general text entry is the limited number of keys. A one handed chord keyboard has only five keys. This translates to 31 possible combinations. This is enough for all the letters, with room for a few more characters, like space and return. There are numerous solutions to this problem. Below are just some of the possible solutions: Two handed chording A ten key chord keyboard has 1023 possible characters. This more than enough. Most two-handed keyboards do not use all combinations, but uses each hand separately to type faster. This is done by overlap. Overlap is entering a chord with one hand before the other hand is finished with the last chord. This is what is usually done on a standard keyboard. Users of two-handed chord keyboards tend to type 17% faster than users of one-handed keyboards[Go88]. Most chord keyboards are made for two hands. Thumb keys One or more extra keys can be added in reach of the thumb. Sixteen more characters are added per thumb key. More combinations are possible if the thumb keys can be pressed simultaneously. This is the a common solution and is often used in addition to two-handed chording. Sticky shift keys A sticky shift, when pressed once, acts on the next one chord made. When double-pressed (like double-clicking a mouse) it acts on all chords until the shift is hit again. For each shift, the number of possible characters doubles. Multiple state keys Instead of an on/off key like most keyboards, it is possible to have a three or more state key. A three state keyboard uses keys which can be pushed up, down, or not at all. This gives 243 combinations for one hand. Additional finger keys It is possible to have more than one key per finger, such as an extra row, above or below the base row. This is effectively the same as having one multiple state key per finger. By using one or more of these combinations it is possible to create all the same characters that can be made on a standard keyboard. With this problem removed, it is possible to use a chord keyboard as a general text input device. The following is how chord keyboards rate on the text input design criteria: [The text input design criteria are the 7 most important aspects of a text input which affect its usability. I based these criteria on those found in [Po88], with two extra added.] 1. Input speed: beginner: 30-35wpm, novice: 60wpm[Go88], disabled: 12wpm[Ki86] 2. Error rate: 9.88%[So62] or 1%-2%[Go88] 3. Learning rate: normal use chords: 1 hour[Ya93] typing: novice: 20h, expert: 60h[Go88] disabled use chords: 5 hours[Ki86] typing: (novice) 11h 4. Fatigue: Low amounts of finger use. 5. Muscle strain: can rest the wrist on the keyboard. Arm can be held comfortably. Hand is held flat causing shoulder stress. 6. Portability: one-handed is much smaller than a standard keyboard, but it still needs a surface to rest on. 7. User preferences: no negative QWERTY transfer. Impossible to use without some training/practice. Given the above, how does a chord keyboard compare with a standard or alternative one? The expert typists can type faster on a standard keyboard, but, on a chord keyboard, novice typists can type faster with less training. It is not possible to type without training on a chord keyboard, like it is on a standard one. On the other hand, learning to type on a chord keyboard is easier because the chord shapes can have some physical correspondence to the letter being typed[Go88]. For example, holding out only the thumb and little finger makes a Y shape. If the chord for Y is made with the thumb and little finger, it becomes much easier to remember. In addition, learning to type on a chord keyboard does not have any effect on the ability to type on a standard one. A person can switch back and forth without any problem. This is an advantage over keyboards with different layouts. Touch-typing on a standard keyboard is difficult because the large number of keys and movements make it easy to miss. This is alleviated much by most alternative keyboards. With a chord keyboard it is impossible. Each finger uses only one key. since almost no movement is involved, it is impossible to miss, this is especially useful in blind, or otherwise disabled users. This is reflected in the smaller error rates for chord keyboards. Another problem with excessive hand motion is that with most normal keyboards, the hand must constantly move from the keyboard to the mouse. Since a one-handed chord keyboard is not much larger than a mouse, the entire device can be used as a pointing device. This would alleviate much unnecessary motion. ----------------------- [Go88] Daniel Gopher and David Raij. "Typing With a Two-Hand Chord keyboard: Will the QWERTY Become Obsolete?" IEE Transactions on Systems, Man , and Cybernetcis. Vol 18, No 4. July/August 1988: 601-609. [Ki86] Alan Kirschenbaum, Zwi Friedman, and Arie Melnik. "Performance of Disabled People on a Chordic Keyboard" Human Factors. Vol 28, No 2. March 1986: 187-194. [So62] Sorbel, R. "Performance on a Five Finger Chord Keyboard". Journal of Applied Psychology. Vol 46, No 3. 1962: 165-169. [Ya93] Wayne Yacco "The BAT: Infogrip's keyboard good pick for alternative input device". MacWEEK. Vol 6, No 27. 27 July, 1993: 71. [Po88] Kathleen M. Potosnak. "Chapter 21: Keys and Keyboards" Handbook of Human-Computer Interaction. New York, NY: North-Holland, 1988. 475-494.