Bad Keyboards

In ZombieTechnologies, Qwerty-based keyboards are described as an "older technology" and that "we'd be better off in the long term converting." (See also comments in the PsychologyOfEverydayThings book.)

The QWERTY example is also suspect. See TheFableOfTheKeys. It argues that the Dvorak advantage is much smaller than most people believe. For most people there's no advantage at all. For example, consider how PairProgramming works despite halving the typing speed. -- DaveHarris

IdeologicalWarrior notes that TheFableOfTheKeys is full of crap, and why.

...but the whole purpose of QWERTY is to slow down typists so that a mechanism doesn't jam. That mechanism isn't used in modern keyboards - but we're stuck with what was established.

[Note: QWERTY was designed before there were typists to slow down. The earliest typewriter models used gravity to reset the letters, so Sholes put common English digraphs as far apart as he could. He also made sure you could spell "typewriter" using just the top row, so this is not to say his design criteria are not suspect. --ChaseAlbert?]

Did you read the reference I cited? The effect of the QWERTY design criteria is to have consecutive keys serviced by alternating hands. That's quite a good criteria for speed, too. -- DaveHarris

hmmm. well, without getting into all the rest of the QWERTY vs. Dvorak points, the above is simply not true. QWERTY does a very bad job of alternation, barely better than alphabetic

I assume DaveHarris's point was that there is no advantage in switching, so it's not one of your ZombieTechnologies because we wouldn't be "better off in the long term converting"

Although there is certainly some debate about it, typewriter historians generally agree that QWERTY was not intended to slow down typing; in fact, the inventor of the commercial device (Christopher Sholes) apparently worked hard to achieve the opposite. While it is true that if you typed too quickly you'd jam the hammers (which was and still is true on all hammer-based manual and electric typewriters), it is not the reason we have QWERTY. Apparently, the reason we have QWERTYUIOP on the top row was to allow otherwise un-trained Remington typewriter salesmen to quickly and elegantly type the brand name of the device, TYPE WRITER, without making it obvious that the layout was kludged to be used as a marketing tool. This layout appeared on Remington's first commercial typewriter. Prior to that point, there were very, very few "typists," regardless of definition. Various Web sites and other sources will confirm this. -- DaveVoorhis

Well that's something but it's very easy to imagine that there were competing keyboard layouts for typewriters. And if there were then the one that slowed down typists enough to not jam keys would be selected for. Were there competing layouts?

There certainly were. The best-known of these was the Ideal layout of 1880 -- which placed the most commonly-used English letters (DHIATENSOR) on the home row -- and, of course, the Dvorak layout of 1932. The Remington QWERTY layout held sway for the simple reasons that Remington Typewriter had considerable marketing clout, having emerged from the Remington Arms company, and it was what most typists -- starting in the late 1880s -- were trained to use. The parallels between this and the popularity of certain programming languages, large software companies, and well-known operating systems is obvious. By the way, Remington Typewriter eventually evolved into what we now call Unisys. See http://unisys.com/about__unisys/history/index.htm. -- DV

This is discussed on TheFableOfTheKeys, but it should be noted that the above source is full of no-good non-information. If you have ever taken typing, you will know it is easiest to hit homerow keys and not particularly easy to hit things below it. Dvorak puts the most common keys in the easiest to hit places; anyone who has ever used it will testify it is infinitely faster for word processing (programming is different, it's pushed the SemiColon off to the side).

The increasingly widespread use of 'net jargon, emoticons and abbreviations is giving us the solution: don't change the keyboards to match our language, change the language to match our keyboards! After all, old technology hangs around for ever, but language is much more malleable and fluid :-) --NatPryce.

Raw typing speed is not the most important skill for many situations involving keyboards. I know that I don't write *good* code at 50 words/minute.

[Are you sure? I write all my code, good and bad, at 50wpm. Usually in 4-second bursts. My point is that typing faster doesn't (I hope) impact the quality of my code, it merely reduces the time it takes to transfer the concepts in my head to the program I'm writing, so I can spend less time bashing keys and more time thinking, and I need lots of time for that :). wpm == bandwidth -- DarrenHobbs]

(I'm happy if I can write 50 good lines (under 500 words) in an hour.) Keyboard layouts may only matter when the bottleneck is pure typing speed. (The only time I find typing to be the bottleneck is in entering large amounts of English text, and continuous speech recognition is good for that.)

I personally switched to Dvorak at a time when I had to write a large amount of english text every day. (on the order of 20-30 pages) And I found that Dvorak substantially eased that task. When I'm doing something different, such as a tool which is primarily mouse-driven and uses the keyboard for shortcuts, the Dvorak in neutral or even a hindrance.

Writers should seriously consider Dvorak. Programmers somewhat less so.

I'm interested in radical replacements for the keyboard, but nothing looks very promising (as of late 1999). The DvorakKeyboard layout may be useful, but it still maintains the conventional physical interface. See the DataHand and KinesisKeyboard pages for useful alternatives to the standard physical layout. -- CliffordAdams (a Kinesis user since 1999)

The best of both worlds for me has been setting Vim to use the Dvorak layout in insert mode only. I tried using as the OS layout for a while, but couldn't retrain my fingers for all the little shortcuts and two- or three-letter commands, particularly in normal-mode Vim. However, Dvorak simply feels better for major typing. On the other hand, that approach significantly slows down the rate at which you regain typing speed.

Typing speed isn't directly of great importance, but I think it matters more than you might expect. (I'm discovering this right now because one of my wrists is giving me pain and I'm having to type one-handed.) The point isn't the sustained speed but the ability to type fluently, using the computer as an extension of your brain. I find it much harder to write good code, or good English when getting it from my mind to the screen is laborious. This probably isn't an argument for switching keyboards (typing at, say, 80wpm on QWERTY isn't a major problem), but speed does matter even when it isn't the limiting factor. I reckon I'm at about 20wpm now; I can't produce good code at a sustained pace of 20wpm unless it's simple and familiar stuff, but it still slows me way down. -- GarethMcCaughan

''To put less strain on your wrists, my personal experience seems to suggest you try making a change to a certain part of your typing habits (no matter what keyboard you're using), or getting a split-type keyboard. I'm talking about the keys on a "standard" keyboard being arranged in straight rows so in order to touch type (the way it is usually taught) you have to keep both your wrists in an unnaturally bent position (since you'll inevitably be sitting somehow in between your own elbows and your keyboard isn't even as wide as your shoulders). That forced constant angle in the wrists is one long known source of harmful strain. I don't have a split keyboard but (and because) I've instead taught myself to touch-type "pianistically", that is, I never use any given finger for any given key just because some touch-typing standard says so, but I use whatever finger is next to a key at any given moment, adjusting the position of my hands at will as it feels practical (this is rather automatic once you get used to it). This allows me to have my hands free-floating above the keyboard (resting my wrists on the desktop edge works, too) without any particular respect for the orientation of the keyboard rows, so there's no angle in my wrists (or at least by far not as much as if I were sticking to the alleged inevitabilities of "standard" touch-typing). While I've never really stuck to the touch-typing manuals (I learned my 50 wpm on the road), I once gave them a try and I sure noticed the difference in strain. It felt immense.''

Don't think of the DvorakKeyboard as a way to type faster. Think of it as a way to put less stress and strain on your body. One side effect of putting the most used keys on the home row is that you simply don't have to move as much to type. I actually used to have wrist pain sometimes using QWERTY. I have never had wrist pain since switching. Also, my neck, back, and shoulders feel better.

And, Dvorak is also optimized for alternating hands, just like QWERTY. All of the vowels are on the home row under the left hand, while commonly used consonants are on the home row under the right hand. Note that this arrangement also slightly favors the right hand, which, for most of us at least, is stronger.

Many of these Dvorak critiques are published in free market economics journals, where it is critical that the author demonstrate that the free market produces the best solutions to problems. The QWERTY keyboard is an oft-cited failure of the market, and so Dvorak superiority is a favored target for debunking.

The study cited above suggests that the gains in productivity do not warrant the cost of retraining, but nowhere does it say that Dvorak is not a better layout, were retraining not a factor. I imagine that ergonomics were not considered in that study, either. -- RobertChurch

I suspect that the "alternate hands" argument might be fallacious - I can't back this up, however. When I'm typing, my muscle memory appears to stem from somewhere near my wrists. By the time that my left hand has signalled something to my right hand (probably via my brain), a lot of time has elapsed. However, if I simply queue up a bunch of impulses to one hand, they arrive closely together, and I go faster. Of course, this could all be bull - direct me to the correct neurophysiological processes involved in muscle memory, etc. I find that this is particularly aggravated by the ergonomic split keyboards - I have difficulty keeping my left and right hands in step.

The idea with alternating hands is that while typing a key on the right hand, the left hand can be moving in position to type the next key. People who know how to type definitely use this to advantage. You can see it happen if you watch a good typist. -- RobertChurch

Here's one data point. I hacked a script together to go through a file and gather home row and alternation statistics for QWERTY vs Dvorak layouts. Running on my sent and received email and a bunch of UNIX man pages, I get: home row--46% Dvorak, 25% QWERTY; alternation--38% Dvorak, 30% QWERTY. So on that test, QWERTY loses in both instances. That said, I still use QWERTY because my fingers have been trained by years of Emacs use, and it's now almost impossible for me to change.

Rather than count distance or finger movement, the best imperical evidence would be contests it seems to me. What keyboards do the record-setters use?

According to http://en.wikipedia.org/wiki/Typewriter#Typing_speed_records_and_speed_contests , Barbara Blackburn (the current (2005) world speed record holder) uses Dvorak. What keyboards did previous record-setters use ?

I was told that the 'real' reason Dvorak flopped was that it was introduced on Remington's "silent" typewriter, which bombed in the market, since typists need to hear the hammers hit. Is there any truth to this? -- BenjaminGeiger (6 Mar 2003)

Flat Keyboards

I just got a new "flatish" keyboard to use with my desktop. It looks and feels more like a laptop keyboard, but full-sized. Is this a trend? A style fad? So far I don't like it, but will give it some more time before I go back to the traditional style. I have to move my hand up higher to press the perimeter keys such as Ctrl, Alt, etc. If there is some benefit to the flatness that compensates for this extra lift movement, I haven't noticed it yet. Anybody else try one?

See ErgonomicKeyboard & ErgonomicKeyboardIdea for some really "out of the box" thinking. ;->

CategoryKeyboard