The Quick Brown Fox Could Soon Be Jumping Over a New Kind of Keyboard
When early versions of the typewriter first appeared on the market 100 years ago, salesmen loved the arrangement of letters on the keyboard because they could write TYPE WRITER without leaving the top row.
At the time, this impressed customers. It hasn’t since.
After a century of typos, back strain and repetitive motion injuries, the standard QWERTY keyboard--so named for the first six letters of the third row--is considered one of the true abominations of modern design. Of the 10 letters ADEHINORST, for example, which make up about 70% of English words, just three are on the middle keys where the fingers normally rest. One of the 10 requires the right index finger to jump to the left, six require the hands to hurdle a row either up or down, and one of the alphabet’s most commonly used letters, A, is struck by the pinky, the weakest finger.
In recent years, the obvious problems with QWERTY have led to a number of radical suggestions for improving the keyboard. Instead of rearranging the keys to minimize hand movement, as proposed unsuccessfully in the past, a number of researchers would junk the conventional keyboard in favor of much smaller arrays in which the typist would play combinations or chords of keys, much like a piano.
One of the most recent and ambitious chording arrangements, developed by Virginia engineer Larry Langley with the help of the Navy, has just eight keys, one for each finger. Each key has two active positions, front and back, which gives the typist--if he or she plays one key from each hand simultaneously--a total of 64 combinations, corresponding to all letters of the alphabet and other necessary keyboard functions.
Chording arrangements have obvious advantages. The fingers don’t have to jump around the keyboard, and the keys themselves are moved back and forth gently, not struck, minimizing muscle strain. Learning the key combinations also turns out not to be appreciably more difficult than mastering conventional touch typing.
Where chording raises questions, however, is over how it compares to QWERTY in typing speed and efficiency. Is it faster to play eight keys than it is 26? Does cutting out the hyperactive hand movements that characterize QWERTY make for a better keyboard?
Consider the the speed at which an expert QWERTY keyboard typist transcribes “The quick brown fox jumps over the lazy dog.”
In stimulus-response experiments, subjects are usually able to press a button in response to a light within 250 milliseconds. Yet according to the graph, the time interval between keystrokes is routinely in the 150 millisecond range. In other words, despite its drawbacks, QWERTY manages to allow typing speeds twice as fast as might be predicted.
How is that possible? The answer lies in two observations, both of which suggest that where keys lie is less important than the skills expert typists master. A good typist, for example, will overlap successive motor functions on the keyboard: When the left index finger is typing the T in THE, the right index finger is already poised above H, and the middle left finger is moving into position above E.
Expert typists also can spit out familiar chunks of text at extraordinary speeds, overcoming the physical obstacles of key placement by short-cutting the mental steps involved. The sequence CH, for example should be more difficult to type than BH, since C is typed by the weaker middle finger and B by the stronger index finger. But typists have so much more practice typing CH--it occurs 600 times more often in the English language--that on average it can be transcribed 50 milliseconds faster.
Chording methods obviously benefit from familiarity as much as QWERTY does. But the possibilities for overlap are curtailed because with only eight keys the likelihood of a single key being used for successive keystrokes is much greater. There is also a cost to typing commands that involve two fingers instead of one. When QWERTY typists capitalize letters, for example, the act of simultaneously hitting the upper case bar doubles the total time taken by the keystroke.
Chording can provide some advantages. The average keystroke latency period in QWERTY ranges from more than 500 milliseconds to more than 100 milliseconds, a reflection of those cases where the haphazard placement of keys does matter and where difficult tasks are assigned to weaker fingers. By assigning commonly used sequences and letters to strong fingers--the index finger, for example, is about three times faster at forming chords than the ring finger--chording systems can avoid some of the QWERTY bottlenecks.
Chording also may lead to fewer errors, because the majority of mistakes in QWERTY come from hitting the wrong key or from transpositional errors, where the H anticipating the T in THE actually comes out ahead--neither of which is likely with chording.
“The success of chording depends on how you measure performance,” said Karl Kroemer, a Virginia Tech engineer working on a chording system. “If you measure just output speed, it’s not as good. But if you measure the health of the operator and accuracy, then things look quite different.”
