Dissenting Opinions

This page dissuses controversy that has gone on for decades over the value of the Dvorak keymap. Before you read on, I'd like to mention that you don't have to believe either side of the argument. Academicians have to measure a thing before they'll admit it exists. (It's a rule.) If you have tried lately to figure out what is safe to eat, you know how hard it is to make sense of academic studies. But since the Dvorak layout is usually free, you can easily try it and decide for yourself.

[MWB April 5 1999: I just added a Postscript to this page that covers some things I've been wanting to address. I'm sorry I don't have time to work it in properly.]

[MWB Feb 20009: Ten years is old for a to-do item! Truth is, I'm used to the page as it is now, so I'll leave it for a while longer. Any complaints?]

[What are the pros and cons of Dvorak?]


The Fable of the Fable

You might hear comments from time to time about studies showing Dvorak is "no better than QWERTY," or words to that effect. All such comments that I've heard seem to echo an article, "The Fable of the Keys," by S. J. Liebowitz and Stephen E. Margolis, published in the Journal of Law & Economics, vol. XXXIII (April 1990).

Note the word "economics." Liebowitz and Margolis are economists opposed to an "excessive inertia" theory, for which OWERTY is often cited as an example. Rather than try to prove their point with a generally valid argument, they simply attack Dvorak as a dubious replacement for QWERTY. As the article's last footnote explains, there are a number of other possible reasons for Dvorak's failure to replace QWERTY, besides a perceived lack of value. The article ignores those reasons, however, and perpetrates that false perception in a nicely self-fulfilling way.

The argument involves perception in more ways than one. If you read the article carefully, you will find that it seems to claim more than it actually does claim, especially after its implications get paraphrased a few times in conversation. Because their effect is just as powerful, I will address its implications as if they were clearly stated claims:


Claim: Dvorak Is a Fraud, Studies Prove It's No Better

The "Fable" article spews forth a host of references to various studies. In the article's view, studies that present Dvorak in a good light are "suspicious," having been "influenced" by Dvorak and his supporters. Of course, in the article's view, it is not possible that Dvorak and his supporters were influenced by their studies!

The article dwells upon claimed Dvorak benefits that seem outrageous, like the Navy's discovery (in the 1940's) that the cost of switching a group of typists to Dvorak could be recovered in 10.3 days, or conflicting newspaper reports of 180 WPM versus 108 WPM speeds, when a Navy officer claimed that no official speed was released. Against suspicious claims like these, the article contrasts other studies that seem to show Dvorak yields a smaller benefit.

[MWB, 5 April 1999: There's more about this in the Postscript.]

But the article either misses or avoids facts that make the favorable claims seem less outrageous. The "10.3 day" claim doubtless considers that the Navy was already training its QWERTY typists (of whom many were novices) periodically to boost their speed. There was probably little additional cost to train in Dvorak instead of QWERTY, so in the Navy's case the cost might easily be recovered. As for the newspaper reports, someone less subtle than the "Fable" authors might attribute the inconsistencies to typos and Navy information leaks, rather than a Dvorak conspiracy. Certainly a Naval officer might have disavowed the study at some point, since it was classified for a time shortly after it was completed! (For details, see Cassingham's The Dvorak Keyboard.)

The "Fable" article belabors the point that the Navy cannot produce a copy of its study, and that the article's authors had to obtain it from Dvorak International. The authors clearly want us to think this is suspicious, but keep in mind we are speaking of a Government agency in a chaotic time (World War II). By coincidence, some time ago I asked the Navy for information about an airplane they evaluated in the early '40s (the XNL Langley Twin); they could find no record except the plane's name, and eagerly asked me to send them the information I had gathered!

The article also makes much of the fact that the Navy study and some others used questionable practices that might favor Dvorak unfairly. But the article cannot and does not say that these practices, such as averaging certain test scores or throwing out "unfair" preliminary tests, did favor Dvorak, only that they might favor Dvorak. In each case, it seems the original investigators thought (rightly or not) that they were correcting a statistical bias. In hindsight, the "Fable" article second-guesses the investigators and assumes they were biased in favor of Dvorak.

But what about studies showing Dvorak is worse than QWERTY? There aren't any! The "Fable" article's star performer is a 1956 GSA study conducted by Dr. Earl Strong. (More than one aspersion against Dvorak springs from Dr. Strong.) In short, Dr. Strong recommended against the Dvorak layout, but why?

The "Fable" article uses a different slant, but here's what it tells us: The GSA study put 10 QWERTY typists through a blistering 25-day regimen, training four hours a day in Dvorak to reach their pre-conversion speed. Four hours a day! A well-worn rule of thumb (Dvorak cites studies from the 1920's) is that it's a waste of effort to speed drill for more than two hours daily. After this torture, Strong found that the overtrained Dvorak typists didn't gain from additional training as fast as 10 fresh QWERTY typists did.

(And Cassingham points out that the QWERTY group's error rate increased after the additional training, while the Dvorak group's errors decreased.)

The "Fable" article puts a lot of stock in the total hours of Dvorak training required to regain one's old QWERTY speed, doubting that the Navy Dvorak group could really retrain in about 50 hours when, for instance, Strong's Dvorak group took about 100. The assumption is that all training is equal, but of course that's not true. I think it's an interesting coincidence that only the hours differ; both groups took about one month to regain speed!

At best, Dr. Strong's study is ambiguous. More realistically, it is a gross example of why you should not put blind faith in studies. In any skill, there is usually a burst of progress when you start speed training after a time of routine work. Clearly, the burned-out Dvorak group would have to be exceptional to match (as they nearly did) the progress of the slack QWERTY group. Studies are often ill-conceived, but how could Dr. Strong make such a horrible error in his study's design?

Look at the "Fable" article in another light: what if Strong was biased? The article puts a lot of effort into defaming Dvorak, (giving Strong as a source in one place) but it casually discards (in a footnote) the possibility that Dr. Strong's GSA study--the article's star witness, and possibly the worst setback Dvorak ever had--was itself subject to bias. Again, the article either misses or avoids key facts that cast doubt on Strong's study.

For example, the "Fable" article cries foul over "missing" data (preliminary tests, etc.) in favorable Dvorak studies. But where is the GSA study data? In The Dvorak Keyboard, Cassingham reports Dr. Strong destroyed almost all his data before other researchers could verify the results!

Suspicious enough, but there's more. As evidence of Dvorak's possible bias, the "Fable" article quotes some of Dvorak's enthusiastic comments about overcoming QWERTY's flaws. Cassingham quotes a different sort of statement written by Dr. Strong a few years before he conducted the GSA study:

"I have developed a great deal of material on how to get this increased production on the part of typists on the standard [QWERTY] keyboard. ...I strongly feel that the present keyboard has not been fully exploited, and I am out to exploit it to its very utmost in opposition to the change to new keyboards."

There are other reasons to suspect Strong and his GSA study harbor bias against Dvorak, but I think that quote just about says it all, don't you?


Claim: Dr. Dvorak Was a Mercenary Huckster

The "Fable" article takes a direct swipe at Dvorak's motives; pointedly mentioning that he owned the patent for his keyboard, and, in the same breath, that he received at least $130,000 for "studies." The clear implication is that Dvorak was out for profit and used grant money for his marketing R&D.

First, the grant. Cassingham reports the Carnegie Foundation issued two grants totaling only $14,000, and points out they came after the keyboard design was mostly in place, but let's accept the larger figure and look at what it bought.

Dvorak's "report to the Carnegie Foundation" (Cassingham's quote) was the book Typewriting Behavior, a 500-page textbook on the psychology and physiology of typing. Dr. Dvorak was only one of the book's four credited authors, and it cites well over 200 other researchers. (Probably 300 or 400; I gave up counting halfway through the Index.) The book represents a lot of study and work, and might well have been worth 130 grand, if that's really what it cost to write.

The "Fable" article would have you believe Dvorak's work was just a huge, expensive advertisement for the new keyboard. It was not. Of the book's 500 pages, only about 60 actually discuss keyboards, and only 15 mention Dvorak's (which he just called the "simplified" keyboard). Indeed, several other attempts to improve on QWERTY are discussed. The rest of the book is dedicated to careful study of behavior, language, learning patterns, fatigue, and other general concepts as they applied to all typists. Dvorak clearly was not hawking his wares. It is also pretty clear that he knew his stuff.

What about Dvorak's patent? The article's authors might like us to think anybody who holds a patent is a fraud. I doubt it. The article doesn't mention that Dvorak sank a good deal of his own money into building Dvorak typewriters (Cassingham). On the face of it, that does show vested interest, but it also shows that Dvorak believed in his invention!

(Rather than combining his layout with a tried design, the typewriter backed by Dvorak used an ill-received "silent" mechanism that made a poor impression on the market.)

The "Fable" article also stresses the Navy study's enthusiastic endorsement of the simplified layout, and concludes that Dvorak must have driven the study because he was the Navy's top typing expert, he had the patent, and anyway Dr. Strong claimed as much. Again, the same footnote that discards Strong's bias also discards opposing reports saying that Dvorak's only participation was to loan the Navy some converted typewriters.

Given conflicting opinions, whom should we believe? Rather than assuming the Wartime Navy was a Dvorak club, it seems more reasonable to think that Commander Dvorak had superiors with a genuine stake in getting a fair evaluation. In that position, an officer might well ban Dvorak from the effort because he was the expert, unless the officer could be certain that Dvorak was earnest and fair. While it is possible that Dvorak's superiors were idiotic sheep, the "Fable" article offers no proof of it.


Claim: Dvorak's Keyboard Failed, So It Must Be No Good

The "Fable" article stresses the economic opportunity that Dvorak's keyboard offered him. With the protection of his patent, didn't he have an unobstructed market and ample opportunity to make his keyboard stick? No. Even if Dvorak was the slick, devious salesman that the article assumes, he might still have found the task too difficult.

The typewriter keyboard (QWERTY) was invented in the 1870's, and the typical manual mechanism was first built around 1900. I've used a typewriter dating to that period, and I assure you they were built to last. By the time Dvorak patented his simplified keyboard in 1932, the typewriter market was saturated with machines so durable that some dealers bought up and destroyed used machines in the hope of selling more new ones (Mavis Beacon Teaches Typing Owner's Manual Version 2.0, 1991).

What's worse, Dvorak introduced his layout in the midst of the Great Depression. This was followed closely by the pressures of Word War II, and then the powerful opposition of Dr. Strong. Those are tough hurdles for any technology to clear.

For a parallel example, plastic composite aircraft construction was actually developed in the late thirties (using wood for the fiber matrix). It made a big impression for a time, and was used in such famous planes as the Spruce Goose and the very successful DeHavilland Mosquito, as well as for parts of many other aircraft. It offered superior performance, but all-metal construction won out and only recently began yeilding significantly to composites made with advanced techniques.

Temporary "failures" like this don't mean that Dvorak (or plastic composite construction) is inferior, but just that it carries a cost that made it inappropriate in the past for most applications. Because computers are so much easier to convert than typewriters, and because so many more people are typing regularly, it is now possible for Dvorak's benefits to outweigh its costs.


Claim: It Was Always Easy To Convert Typewriters

The "Fable" article implies that the installed base of typewriters was no real impediment to Dvorak conversion. It says, "for a time before and after the Second World War typewriter manufacturers offered to convert QWERTY typewriters to Dvorak for a very small fee." This is more sweeping than the footnoted reference, which says in 1961, typewriter service shops could convert typewriters for as little as $5.

Even if that price claim is correct for some typewriters, it could scarcely be typical. If replaceable-type-ball electrics (like the IBM Selectric) were available that early, then $5 is a reasonable low-end claim (corresponding, you may note, to the most expensive machines). But almost all of the typewriters that I saw in the 1960's had a separate type hammer for each key. All the hammers are shaped to sit at a different angle, so converting such a machine requires disassembling it and replacing them all. That can be a fair amount of work.

In the '40s, the War Department estimated the cost at $25 per machine (Cassingham). That was a lot of money in those days; my Dad says it would pay for a month's rent, a week's groceries, or even a pretty good typewriter! On top of that, Dr. Strong was apparently opposing Dvorak by about this time (The quote I mention earlier was from '49), so anyone who paid for the conversion faced the prospect of paying again if they decided to switch back! (For the War Department, that was $20 million ticket each way.)

By contrast, most of today's computer users can convert their computers for free with just a few mouse clicks. To claim it was this easy before is simply ridiculous. And even now, some users might still find it difficult to convert all the systems they need to use.


Claim: QWERTY Was Designed Ergonomically

The "Fable" article admits that the motivations for the Dvorak keyboard are valid, and names three: home-row focus, balanced use of the hands, and alternation between hands. Although the article apparently concedes QWERTY's shortcomings on the first two points, it tries to show that QWERTY's designer, C. L. Sholes, accidentally hit on the third technique by "moving apart" letter sequences that jammed frequently.

A quick test demonstrates clearly that QWERTY doesn't come close to Dvorak at alternating hands, and it barely surpasses an alphabetic layout. What's wrong with the "Fable" article's logic?

The key is the phrase "moving apart." To cure a troublesome sequence, Sholes had to separate the letters' type hammers (a.k.a. "type bars"), not the keys. For "modern" (post-1900) typewriters it amounts to the same thing; the hammers are laid in an arc in front of the platen, so adjacent key levers operate adjacent hammers. But the Sholes' typewriters (and other pre-1900 designs) arranged the hammers in a full circle (the "type basket") under the platen. On these machines, adjacent key levers worked hammers in the front and back half of the type basket in alternation.

So for Sholes' machine, separating letters in the type basket has a very indirect effect on the keyboard. True enough, at either side of the keyboard Sholes might have needed to separate problem letter pairs by some distance (though not necessarily to the other hand). Near the middle of the keyboard, however, a one position move sends the letter's hammer clear to the other side of the basket!

Given this complex interplay, it's no wonder Sholes came up with such a confusing arrangement of keys. Certainly his options were limited by the mechanics of the situation. He might have managed at the same time to promote alternation between hands, but why would he try to promote fast keystrokes when fighting to avoid jams? He had no good motive to do so. At the time, touch-typing hadn't even been invented.

(On the other hand, if Sholes felt so inclined, it would still be to his advantage to force awkward motions for sequences that weren't easy to break up in the type basket. The usual history claims he did, and the "Fable" article fails to prove that he didn't.)

If you don't believe me, how about Sholes himself? He clearly thought his first try was inadequate (Cassingham). After faster mechanisms and touch typing were established, Sholes took out a patent on a revised layout that was very different from QWERTY:

 X P M C H R T N S D G K
  J B W F L A E I O U Y
 Q V . ,   ' ! ? - ; _

(I'm guessing a bit at the punctuation. Sholes' diagram isn't clear.)

Notice that Sholes' revised keyboard anticipates Dvorak by putting all the vowels under one hand, an obvious move if you want to alternate hands as much as possible when touch typing English text. Hmmmm. Why didn't he think of that before? ;^)


Claim: The QWERTY Standard was Established by Competition

The "Fable" article claims that QWERTY emerged supreme from among numerous competing keyboard layouts. Certainly there were typing machines before Sholes, but few, if any, had keyboards! The prevalent type had a selector dial (for picking a letter) and a lever to make the impression. This type continued to be popular for several years after Sholes' design was introduced by Remington (Cassingham).

More direct competition arose after the Remington came out. Again, the "Fable" article would have us believe that this competition challenged the QWERTY layout. The article discusses contests pitting the Remington (with QWERTY) against other machines, naming one radically different challenger called the Caligraph.

But how is the Caligraph different? I have a picture (engraving) of it showing a huge keyboard with roughly twice as many keys as the Remington. It certainly looks different. But on close examination you can actually make out the letters on the keycaps. With one or two guesses, the following is the Caligraph layout. (In the picture, capital keys are labeled white on black; lowercase black on white.)

  V W 2 3 4 5 6 7 8 9 J K
   R T E ( $ q & z ) U G H
  A S w t r e y u i o I O
   D F a s d f g h c k N L
  B C j x v b n l m p M P
   Q X : ; ' ? " . , - Y Z

This keyboard's most obvious feature is lowercase typing. (Remington's machine typed only capitals until they added a Shift key later on.) But as radical as it seems, the Caligraph's lowercase layout still bears a striking resemblance to QWERTY. This was apparently Remington's stiffest competition, yet at its heart it is clearly influenced by the established QWERTY layout.

Even ignoring the similarities, I don't see how competition from this design makes QWERTY seem better for touch-typing; except, of course, in that QWERTY has fewer keys!


Claim: Market Economics Prove Dvorak Is Inferior

The core implication of the "Fable" article amounts to circular reasoning. It claims the "excessive inertia" theory is bunk because a popular example of that theory (QWERTY) has only been faced with inadequate challengers (Dvorak). Yet throughout the article runs the implicit claim that Dvorak must be inadequate because in all this time it hasn't replaced QWERTY. ("Appropriate inertia," I suppose.)

There's no proof that Dvorak is genuinely inferior (the article states as much plainly in its last footnote), so the article boosts Dr. Strong's flawed and ambiguous GSA study and strives to implicate Dvorak and his supporters as frauds. But ultimately the article's best and only argument is that Dvorak is no good because, in the past, it failed in the free market.

Yet there are other possible reasons for Dvorak to have failed in the past, in spite of its genuine superiority and, yes, regardless of economic theory. (That last "Fable" footnote admits this also.) I've already mentioned the typewriter glut, the depression, the war, and Strong's apparent malice. More recently, there's the "Fable" article itself, and others like it, that continue to obscure Dvorak's obvious benefits with academic fog. And no matter what the economists say, the Dvorak layout has simply not been a ready option for most people--until now.

It comes down to whom you believe: Dvorak and his followers, or Strong and the economists who cling to him. Perhaps the best argument that Dr. Dvorak was earnest was that he did fail at first. If his primary interest was profit, it's likely he could have done a better job selling his layout.

Instead of marketing, however, Dvorak strove to better the lot of all typists, regardless of the layout they use. Typewriting Behavior's broad approach shows this, as do Dvorak's layouts for one-handed typists. Dvorak's "simplified keyboard" was just one facet of his work--one that, finally, is ready at hand for almost anyone to try, and to accept or not as they choose. Now that's a free market.


Postscript

From ???@??? Mon Mar 15 02:18:58 1999
To: Damian Haslam
From: [Marcus Brooks' old email address]
Subject: Re: dvorak vs. qwerty

At 7:59 PM -0600 3/14/99, Damian Haslam wrote:
>I imagine you've already seen this (or something similar), but on the
>off-chance you haven't, do have a look at this link:
>
> http://www.reason.com/9606/Fe.QWERTY.html [2009: now links to http://www.reason.com/news/show/29944.html]
>
>I'd be interested to hear what you have to say regarding the matter :)
>
>cheers - damian

I'd heard about it but hadn't found it online. I do have some comments about the authors' 1990 article at [gave old address, now http://mwbrooks.com -- MWB]. It covers most of their points (which they've been warming over periodically for ten years or more).

Mind you, I don't really give a damn about QWERTY's relevance to free-market theory. I'm not really in this to convert QWERTY typists, although I've helped a lot and they've all thanked me for it. I'm really in this because people are teaching their children to type on a keyboard that could very well torture them for life. These people might have good reasons, given circumstances, but the relative cost of retraining isn't one of them.

As for the Navy study, the article you link has more of the authors' interpretation than I had seen before. I recently obtained a copy of the Navy study and read, for example, the rationale behind averaging the initial scores of the first-time QWERTY typists. I haven't had a chance to roll any of it into my web page yet.

Before you buy into the Margolis and Liebowitz slant on the Navy study, consider this: The "start score" for the Navy's Dvorak group was actually their QWERTY "end score" after an average of 419 hours of training! If the Navy had averaged the Dvorak group's starting scores, they would have had to do it on QWERTY before switching keyboards. (Otherwise Dvorak would have posted _truly_ incredible gains.) Averaging four days' QWERTY typing would have made no significant difference to the baseline because the Dvorak group were already experienced QWERTY typists.

As for averaging the QWERTY group's last four test scores, M&L don't mention that the Navy compensated for the averaging by computing the gains based on _four days less_ than the actual training period. Given that, the only way averaging could hurt the QWERTY group's net gain is if their scores were _decreasing_ at the end of the study.

It seems Margolis and Liebowitz did a very poor job of reading the Navy study. For another example, they howl at the Navy's claim that the cost of retraining was amortized 10 days after retraining. They can't even read their own writing: in one place M&L say 10 days after the end of training; in another they say 10 days after the _start_ of training (their emphasis).

In point of fact, the figure was based on the _end_ of training, and the numbers are all there in the report. Here are the averages per typist. You can add it up and see where the Navy got it wrong. I sure can't tell.

Cost of typist's hours in retraining class:         $59.62
Cost of machine conversion:                          20.00
Cost of instructor:                                  13.69
Cost of decreased efficiency early in training:      12.07 
Total cost of retraining:                           105.38

Value of increased production late in training:      52.80
Value of increased production per day after training: 5.12

Number of days after retraining to amortize net cost: 10.3

Again, this is for the group's "average" typist. The study also gives the ranges, but I won't bother typing them. Considering the wartime economy and the sort of gains realized, it all looks pretty reasonable to me.

[MWB, from a later note: It occurs to me that the $12.07 figure might look suspicious. I didn't think to mention that the Navy had the luxury of assigning their subjects to mostly non-typing work until they had at least regained their old typing speed. So the "cost of decreased efficiency" is much lower than it would be if there's simply nothing productive you can do without typing.

Before you suspect the Navy of cooking the data, note that doing as little typing as possible outside of retraining time is an important factor in any retraining effort. The Navy simply bowed to a fact of life and saved a little money in the process.

So if someone types constantly and can't afford to lose a few weeks' work, I try to talk them out of switching, at least until they can take some time off. Come to think of it, I need to state that a little more plainly on my training tips page.]

Thanks for the URL. This will help my page when I have a chance to update it.

BTW: I also have found more info about other "competing" keyboards. One significant one that failed (the Blick "Scientific") almost certainly did so because it had a very slow (but nifty) mechanism. It was offered with a QWERTY layout, so if folks had really wanted it the keyboard was no obstacle. Since it died out, they must not have wanted the mechanism. The other keyboard (Hammond's "Ideal") also had a nifty mechanism, but the keyboard was an uncomfortable-looking thing like a semicircular piano. Hammond pretty quickly offered QWERTY too, and their machine survived into the 80's as the respected Varityper typesetter. The less-significant Crandall was incredibly expensive and looked a lot like the Hammond Ideal, with different keytops. "Full keyboard" machines like the later Calligraphs and the Yost all used the QWERTY layout, but were still considered "different" keyboards because they had twice as many keys (usually two full QWERTY sets).

Gotta go. I'm supposed to be working. Ta,
Marcus

Back to Introducing the Dvorak Keyboard.


Last update: 5 April 1999
Original page established: 8 December 1996
Marcus Brooks:  HTML Home  Weblog