Dvorak is best known today as the inventor of the leading alternative keyboard design, but he was interested in everything that could affect typing productivity, including posture and fatigue. In the text that he wrote with Dealey and two typing teachers, Nellie L. Merrick and Gertrude Catherine Ford, and in workbooks for students, Dvorak actually devoted little space to keyboard reform. He was more concerned with every detail of skill regardless of keyboard design, especially developing a rapid, light touch (“ballistic motion”) and reducing tension. He was a connoisseur of techniques, showing aspiring typists and typing teachers how they could benefit from the “looseness and lightness” of top performers in fields as different as tournament golf and window washing. All of this he supported with citations from the latest in the growing psychological literature of typing behavior. Dvorak’s book, sponsored by the Carnegie Foundation for the Advancement of Teaching, is still impressive for its unpretentious but authoritative manner, and for its belief that typing students could participate in the scientific advancement of their own skills. And Dvorak also fought some dogmas of traditional typing instruction, especially exaggeration of the importance of touch and the “queer bias for petty accuracy.”31
Keyboard reform was still important to Dvorak and Dealey. They were convinced that the universal keyboard was an obstacle to good technique. It was expert typists who first noted the problems of QWERTY. A speed coach interviewed by the Gilbreths had already remarked that “[t]he young ladies think that it is the mechanical construction of the machine that is retarding their speed.” The speed coach went on to note that most of the writing was done with the left hand and that too many words were typed by that hand only Dvorak went on to estimate that overcoming the defects of QWERTY accounted for half to three quarters of instruction time. Work was divided unevenly among the fingers. Overloading the left hand interrupted the ideal rhythm of alternating use of left and right. Too many combinations of two letters (digraphs) were made by adjoining fingers. Eighty-five percent of digraph combinations had to touch the top row, while only 7 percent could be made on the home row.32
August Dvorak, a disciple of the time and motion expert Frank Gilbreth, devised, with William L. Dealey, one of a number of reform keyboards promising new efficiency in writing, as Paul von Jankó’s keyboard did in musical performance. Dvorak’s claims for higher speed proved exaggerated. Comfort was the real advantage of the arrangement. From U.S. Patent 2,040,248, May 12, 1936.
Rejecting previous revised keyboards, Dvorak and Dealey developed and patented a new arrangement. Vowels were relocated to the home row of the left hand, and the consonants most often used were on the right, so that hands had to alternate with each syllable. Loads were adjusted to the capacity of each finger. The largest source of errors, digraphs made by adjoining figures, was cut by 90 percent. Much more work could be done on the home row, cutting “hurdling”—the scurrying of fingers over the keyboard. (Only a hundred common words can be made on the home row of the Universal Keyboard, three thousand with the simplified.) Dvorak and Dealey’s studies in junior and senior high schools suggested substantial gains: an average of 48 words per minute in the second semester, for example, versus a maximum of 28.4 w.p.m. in comparable studies with a Universal Keyboard.33
The Dvorak and Dealey layout, patented in 1936, attracted great interest but appeared at the worst possible time. The typewriter market was glutted in the Depression. Dealers were even buying and wrecking functioning old machines to suppress the used-typewriter trade, but the major brands were solidly built and with proper maintenance could last for decades in normal office use. Schools, the most promising market following the results of the Carnegie-funded experiments, could hardly afford conversion, especially when there was no sign of acceptance in commerce. Wages were falling, and there was a surplus of typists. Dvorak still tried to convince corporations that the new arrangement could reduce their costs by eliminating typing positions. But typists influenced purchases by their employers, and the new arrangement seemed to threaten speedups and layoffs rather than the relaxed high performance that the Dvorak textbook preached. And Dvorak made yet another strategic mistake, linking the simplified arrangement with the Remington so-called noiseless typewriter. Dvorak praised the design for its encouragement of his ideal light staccato touch; like Frank Lloyd Wright’s three-legged chair, it would fail if the user’s body technique was incorrect. But Dvorak overlooked the importance to typing rhythm of the auditory feedback provided by the typebar’s clatter. The Remington also made less crisp impressions, and the mechanism could feel mushy; even Dvorak acknowledged in his text that its feel “disturb[ed] even expert typists” before they got used to it. Receiving a U.S. Navy commission in World War II, Dvorak sponsored studies that appeared to prove that added productivity from the simplified keyboard would quickly repay the costs of retraining, but his recommendations were ignored. All 850,000 typewriters bought by the U.S. government during the war were QWERTY machines.34
Even foreign manufacturers failed to challenge the QWERTY arrangement. In France there was strong opposition to the local version, AZERTY. An influential journal, La Revue Dactylographique et Méchanique, established a commission that proposed a layout designed especially for the French language. One company, the delightfully named Manufacture des Armes et des Cycles, adopted it for its own version of a British design in 1909. Unfortunately for its advocates, the French typewriter market remained small by international standards; in 1913, France produced 4,000 machines and imported 28,000, mainly from U.S. makers who had standardized on QWERTY. A leading French psychophysiologist and foe of Taylorism, Jean-Maurice Lahy likewise tried to challenge American-style ten-finger typing, but had limited influence. By the end of the 1930s, the technique of touch typing and its associated technology, the QWERTY typewriter, had spread with only minor variations throughout the world of the Roman alphabet and had influenced even non-Roman national keyboards. Reform proposals from as far away as India had been defeated like alternative musical keyboards, and for similar reasons. Typists formed a community with hard-won motor reflex skills; they were reluctant, in the business world, to trade certain proficiency for uncertain gains, no matter how critical of QWERTY they could be in laboratory settings like the Gilbreths’. While Dvorak continued to fight for his design through the 1960s, he had been defeated twice: not only in the wartime military market but in the early postwar civilian boom. An incisive analyst and accomplished teacher, he lacked the skills of bureaucratic infighting that might have implemented his ideas from above, and the knack for public relations that could have excited enthusiasm from below. There would be one more window for innovative keyboards in the early microcomputer era. But first it is time to examine how the keyboard became a computer device in the first place.35
August Dvorak calculated the data shown here to prove that his key arrangement evened the distribution of work among the fingers. But his data also show that the conventional arrangement was well conceived for the original four-finger style of typing, before typists and typing teachers worked out the five-finger system. Only the middle finger of the right hand appears underutilized. (Courtesy of Scot Ober)
THE BODY ELECTRIC
The electrification of the keyboard began at a surprisingly early date. We have already noted the rise of electrified typewriters in telegraphy In 1902, George Blickensderfer introduced the Blickensderfer Electric, with an interchangeable type element and keyboard control of underscoring, line spacing, carriage return, and tabulation. Hammer force could also be adjusted from the keyboard to produce as many as fifteen carbon copies. The inventor’s only mistake was ignoring a different set of network effects: utilities were still supplying electricity only after dark, and the machine’s power cord was designed to be screwed into a lightbulb socket. Blickensderfer’s death in 1917 left his ideas orphaned by the company’s successors, but Remington and Electromatic (absorbed by IBM) began making commercially successful machines in the 1920s.36
By
the 1930s, IBM was using typewriter-style keyboards for card punching; even early scientific computation used stacks of punched cards, and the keyboard was already an important input device. Yet well into the 1950s, keyboard control was far from universal. Scientists and engineers continued to rely on slide rules for personal computation until inexpensive handheld calculators with keypads appeared in the 1970s. (The ingenious but costly portable mechanical alternatives like the cylindrical Facit used slides and cranks.) The nineteenth and early twentieth century developed wonderful precision instruments for calculating data such as the tides and the areas enclosed by irregular curves. The electrical engineer and scientific administrator Vannevar Bush even built a machine for solving differential equations, applied to problems from atomic structure and ballistics to railroad timetables.37
It was digital processing that wedded the QWERTY keyboard to the computer. Early World War II and postwar digital machines used keyboard-encoded punch cards and tape, first paper and then magnetic, and teletype-style printers. More and more computer functions could be managed by typing rather than by rearranging plugged wires and setting dials. Academic scientists began to use time-sharing systems controlled entirely by teletype. The crucial date for the keyboard, though, came in the later 1960s, when the Multics time-sharing system developed by MIT, Bell Laboratories, and General Electric replaced the clacking teletype with a video display terminal. In 1972 the addressable cursor appeared, and keyboard control sequences were standardized. In the late 1970s, when microchip costs had dropped low enough to make hobbyist microcomputers possible, the obvious control hardware was the keyboard plus monitor. The Altair, which employed switches, was soon overtaken by new consumer brands running the BASIC programming language with a video display terminal and keyboard similar to those of the Digital Equipment Corporation VT100, ubiquitous in industry and academia. With the expansion of personal computing in the 1980s, accelerated by improved Web browsers and Internet service from the mid-1990s, the QWERTY keyboard has become a universal interface.38
But along with this victory of QWERTY has appeared new hope for reform. Electronic keyboards can be reprogrammed with free software for the Dvorak layout; ever since version 3.1, Microsoft Windows has joined the Macintosh in offering built-in Dvorak support. Keys can easily be relabeled, and moderately priced replacement keyboards offer dual labels and push-button toggling between layouts. From the mid-1980s to the early 1990s the future of the Dvorak layout seemed brighter than ever.
One early hope was healthier keyboarding. As I discussed in Why Things Bite Back, the exposure of millions of professional and amateur typists to the new equipment in the 1980s produced an unforeseen epidemic of overuse injuries of the hand, most of them known as carpal tunnel syndrome (CTS). The condition’s origins and treatment are still imperfectly understood, including the parts played by physical anomalies and psychological states of users, by working conditions and employer attitudes, by variables such as the response of keys to fingers, and by use of the mouse. The apparent efficiency of the computer keyboard clearly played a part. The manual typewriter had demanded a healthy elevated position for the wrists and had enforced breaks for throwing the carriage and for removing and inserting paper. Computer keyboards permit dangerously rapid and uninterrupted data entry coupled with awkward wrist placement.39
The Dvorak layout, which reduces finger travel dramatically, at first seemed an ergonomic blessing. People with CTS and related ailments have praised the arrangement. Yet its medical advantages still have not been proven. Without significant numbers of Dvorak users in organizations it is difficult to gather statistics, and voluntary Dvorak adopters probably are different on average from their QWERTY colleagues: perhaps more affluent or health conscious, but also possibly more concerned about CTS and other conditions after previous episodes. Since total reported cases of CTS dropped from 41,000 in 1993 to 26,300 in 1998, most of them industrial, prevention in the office has seemed less urgent. The schools would be a logical place to test the benefits of the Dvorak layout, as computer-related injuries have been growing among young people and there is no doubt that the Dvorak system is much easier to learn. But few schools even appear to offer Dvorak instruction.40
Productivity is another matter. Dvorak’s own writings, reflecting the Gilbreths’ influence, assumed that adults would be able to sustain the advantage of up to 40 percent in favor of his system that he found among students. Efficiency is easier than health to monitor in controlled experiments. And when a new round of studies began early in the personal computer era, their results were surprising. In 1982 the cognitive psychologists Donald A. Norman and Diane Fisher compared computer simulations of QWERTY, Dvorak, and alphabetical keyboards. Norman and Fisher found that although the Dvorak arrangement did indeed save on motion as its advocates had long claimed, gains in speed were modest: the advantage was only about 5 to 10 percent. They found the long-maligned QWERTY keyboard surprisingly rational in its high number of alternating-hand sequences. The Norman studies and others bolstered an influential 1990 rebuttal to Paul David’s analysis by two fellow economists, S. J. Liebowitz and Stephen E. Margolis, who argued that the Dvorak arrangement had lost on its merits.41
The critics of the Dvorak layout have a point. Typists’ minds are able to manage the additional 37 percent finger travel of the QWERTY keyboard without a corresponding loss of speed. Differentials range from a mere 2.6 percent for Dvorak, to 11 percent. A 1980 Japanese study suggested 15 to 25 percent faster performance in timed writing and 25 to 50 percent faster production of letters, reports, and tables. Scot Ober, a professor of business education rather than a cognitive psychologist, believes that training methods account for some of the difference. Dvorak was a master teacher and developer of instructional materials. To be truly fair, an experimenter would have to provide equally expert textbooks and workbooks, which are not commercially available for the Dvorak program. Even in Seattle’s Keytime typing instruction school, known for the innovative methods of its founder, Linda Lewis, the great majority of students choose the QWERTY option. There are also large professional groups, notably computer programmers, for whom neither QWERTY nor Dvorak is an efficient arrangement, especially now that the Internet has revised the importance of letters (like W) and signs (like @).42
Innovative keyboard designs have not fared much better than new key patterns. As early as the 1920s, a German industrial health specialist recommended that keyboards be split in the center and tilted like the roof of a tent, to keep the hands in a more relaxed, natural position. In the days of mechanical typewriters this was a prohibitively expensive concept, but the rise of the computer terminal in the 1960s made it feasible. The American ergonomist Karl Kroemer revived it in the 1970s. Since then, a dozen or more manufacturers, including IBM and Microsoft, have offered such designs, but they form a tiny part of the market. The barrier is no longer the technology of manufacture but the investment of users in their technique. Our muscle memories are accurate to within a millimeter, and a new arrangement can be discouraging at first. A few brands, such as Gold-touch, can be adjusted very gradually, but they are rarely offered as original equipment and are sold mainly over the Web and by small independent distributors. One of these, Neal Taslitz, is an ergonomic activist dismayed by the obstacles that small manufacturers face in the marketplace, especially in access to national retail chains. He finds that corporate buyers of computing equipment are so cost-conscious that keyboards are one of the first items on which makers—even prestigious top-tier companies— economize. Thus IBM first spun off and then discontinued using its award-winning buckling-spring technology, famous for a distinctive clicky feel. Most keyboards today are almost disposable. The remaining high-quality, high-price designs seem to be bought by a few affluent individuals and overuse-injury sufferers: the very clientele attracted to the Dvorak key arrangement, though the Dvorak layout and new keyboard contours oddly are seldom combined in the same product. Whatever the reasons, the keyboard is virtually the onl
y part of a computer that has stagnated or declined in performance.43
KEYBOARDERS ALL
The response of some computer industry leaders to the keyboard’s problems has been the dream of its abolition. When the Macintosh graphic user interface arrived in the 1980s, the mouse at first took over some of the commands of the keyboard, but many users soon clamored for function keys and keyboard shortcuts. In the late 1980s and early 1990s visionaries at Apple and elsewhere called for a new kind of computer that would recognize handwriting on a portable tablet. While the movement ended in a wave of failures, notoriously of the original Apple Newton, pen-based computing never actually died, and both Apple and Microsoft are doing their best to revive it. Yet the smaller the portable device, the more likely it appears that someone will find a way to plug a portable QWERTY keyboard into it. Several are available for the Palm Pilot personal digital assistant (PDA). Even graphically oriented Apple included a detachable keyboard with the last version of its Newton MessagePad—as well as a Unix-based text command option in its latest Macintosh operating system. And PDAs without plug-in keyboards are routinely connected to conventional computers for keyboard-entered data.44
Voice control was as exciting in the late 1990s as handwriting recognition had been ten years earlier, and with similar results: a wave of marketing and financial troubles that has failed to shake underlying optimism about the technology’s future. While voice recognition software can promote its own overuse injuries, it can now work with natural phrasing and infer spelling from context. But it is unlikely to eliminate the keyboard, because it will still make errors (or users will still fail to enunciate properly), and editing copy orally is even slower and more tedious than correcting it with a keyboard. Optical character recognition data also need checking and editing. Typing will probably be further reduced in familiar applications in the future, but it will also be extended to new tasks.
Our Own Devices: How Technology Remakes Humanity Page 28