6 ms·

Typing Errors: The standard typewriter keyboard is Exhibit A... Dvorak not so great either

26 points by nickb 18y ago

ellyagg 18y ago
Based on the distance your fingers have to move to type average words on qwerty and dvorak, it's really impossible that Dvorak wouldn't be loads better.It's sort of like claiming that someone wouldn't be faster at running 50 meters than 100 meters. Of course, where this analogy falls down is that you have to learn a new skill, which means there's an upfront cost and thus makes the whole issue interesting economically. The article claims that the cost of learning the new skill could never be amortized and uses this evidence: > In the first phase of the experiment, 10 government typists were retrained on the Dvorak keyboard. It took well over 25 days of four-hour-a-day training for these typists to catch up to their old QWERTY speeds. (Compare this to the Navy study's results.) When the typists had finally caught up to their old speeds, the second phase of the experiment began. The newly trained Dvorak typists continued training and a group of 10 QWERTY typists (matched in skill to the Dvorak typists) began a parallel program to improve their skills. In this second phase the Dvorak typists progressed less quickly with further Dvorak training than did QWERTY typists training on QWERTY keyboards. Thus Strong concluded that Dvorak training would never be able to amortize its costs. He recommended instead that the government provide further training in the QWERTY keyboard for QWERTY typists. I really can't express how stupid this is. The logical implication the article would have you draw from this study is that qwerty actually may be MORE efficient. To believe this, you'd have to believe that the top end hypothetical speed of qwerty is higher than dvorak. This is literally physically impossible, as I've already pointed out. Moreover, the fastest typist ever used Dvorak. While that's not conclusive, it's suggestive. Folks trying to optimize a skill to the highest level are very good at cutting out all inefficiencies. To understand the results of the study cited there, we need only look to exercise science. First of all, what most lay people don't realize is that a great deal of exercise results for activities like weight training is explained by the exercise's affect on your nervous system, not your muscles. Properly managing your nervous system response is at the heart of continued gains, and not managing it properly leads to overtraining and plateaus. All top flight Olympic athletic coaches spend much of their time thinking about this. As a matter of fact, it is a well known and well documented phenomenon in the literature that if you train at close to maximal capacity on the same training regimen for longer than 3 to 4 weeks, you will stop making gains. You'll see big increases the first 3 weeks, and after that, nothing. This is a nervous system effect. Armed with that knowledge, it becomes clear why the study's results look the way they do: The dvorak trainee were already in a state of CNS fatigue, whereas the qwerty trainees were starting out fresh. Dvorak is clearly better than qwerty. Colemak is even more better. However, I don't think this situation is much of an indictment of the free market. I don't think the bottleneck in most people's productivity now (or ever) was top end typing speed and it would take quite a long time to recover the cost of the switch. Considering that this is THE MOST concrete example ever cited where you might be able to say that government meddling could induce a positive effect, I'm not inclined to agree that it's in general a good idea to have our economic lives centrally manipulated. There are far more examples of net economic losses from government interference.
- pragmatic 18y ago
  The point seemed fairly clear. Retraining on a Dvorak held no real advantage. The fact is that almost all people learn on QWERTY so the necessary retraining isn't worth it. Not to mention the millions/billions of keyboards in existence that would need to be scrapped.
  - tome 18y ago
    Touch typists don't look at the keys anyway, so you don't need to scrap the keyboard. Windows keymap switcher, or the equivalent in your OS of choice allows you to quickly go back to QWERTY on the (very) rare occasions where what you need to do corresponds with the actual symbol on the keys.
- Rabidmonkey1 18y ago
  Hmm... Colemak looks interesting. I've always had a hard time learning to use Dvorak, especially since everywhere else you work it's going to be Qwerty by default. Colemak is also public domain, according to Wikipedia. I might have to give it a try. http://en.wikipedia.org/wiki/Colemak#Colemak http://en.wikipedia.org/wiki/Colemak#Colemak
  - cninja 18y ago
    I actually installed Colemak on my computer yesterday. I played with it for a couple hours before uninstalling it. Most of what I type is programing code, not prose. It dawned on my that most my typing delays come from all of the parens and brackets and other programming symbols. Since all these fancy keyboard layouts have been computationally optimized for prose language, they don't address the primary cause issue. I decided that if I'm going to take the time to learn a new keyboard layout, then it will need to be one that has been optimized for programming.
    - Shamiq 18y ago
      Fix this, please. It's making this thread just about unreadable.
pragmatic 18y ago
This is not just about keyboard layouts. It's about the concept the poor products (VHS, Windows) win over superior products (Beta, Mac) because of a winner take all economy. As you can guess the author shows this to be bunk science.
- davidw 18y ago
  What they are afraid of is the idea of the free market not always and in every case producing the best possible result, so they need to aggressively attempt to demonstrate that 1) either market failure just doesn't exist, as they attempt to show here, or 2) that in no case, ever, will government intervention improve the situation, which is the topic of other libertarian writing. In other words, this is just another episode of "Libertarians vs Socialists": http://news.ycombinator.com/item?id=419503 http://news.ycombinator.com/item?id=419503 For a serious treatment of the economics of digital goods and networked economies, I think Varian and Shapiro are a better read: http://www.squeezedbooks.com/book/show/7/information-rules-a-strategic-guide-to-the-network-economy http://www.squeezedbooks.com/book/show/7/information-rules-a... (Edit: to be fair, it turns out the authors have a more scholarly treatment of the subject matter, which is a better read: http://www.utdallas.edu/~liebowit/harvj/harvard.html http://www.utdallas.edu/~liebowit/harvj/harvard.html Still though, I think it's pretty clear that they're coming from a point of view, and "backfilling" to try and get there)
- twopoint718 18y ago
  It would be interesting to construct an experiment to see what kinds optimal solution a free market can and can't find (though I have no idea how this would be done). In search algorithms you have to worry about merely finding a local maxima rather than a global maximum. In algorithms such as simulated annealing you perturb your solution randomly at the outset, trying to "bump" your solution off of one of those local maxima. What if the free market is prone to getting "stuck" in some locally optimal, globally sub-optimal solution? Depending on initial conditions (price, number of firms in the marketplace, demand, etc.) we could end up with great solutions, but in others with terrible results.
Hexstream 18y ago
What bullshit. There's a common pattern in Dvorak VS Qwerty: Dvorak advocates sometimes rely on history and sometimes on purely technical grounds to prove their claims, while Qwerty advocates always only rely on history because there's simply absolutely no way to prove Qwerty approaches Dvorak when it comes to logic. Here are some facts about Dvorak that are easy to verify and the utility of which is obvious: - Dvorak has the most used keys on the home row. Compare: AOEUIDHTNS VS ASDFGHJKL; - Dvorak optimizes for alternating hands, which obviously helps speed. Ever tried typing "street" or "states" in Qwerty, two common words with common letters that you have to type all with the left hand? On the Dvorak side the worst word is "joke" I think. - Dvorak optimizes for typing a bit more with the right hand instead of the left hand, which makes sense because the right hand is usually the strongest. - Dvorak optimizes for typing from the outside to the inside (try tapping your fingers in rapid succession on the desk from outside to inside, then inside to outside, and you'll see immediately which one is easier and faster). - Dvorak optimizes for typing more with stronger fingers (index and friends) than weaker ones (pinkies and friends). - Dvorak has all the vowels on the left hand on the home row (except Y which is conveniently on top of I), one beside another, which is great for learning. It also makes sense because you use consonants more than vowels, (see right hand argument) and since the most used consonants are on the right home row, it optimizes alternating hands. No need to mention Qwerty makes no such optimisations and the result is correspondly atrocious. See? No need to bullshit with biased reports about history.
- lliiffee 18y ago
  You weaken an otherwise useful post by using words like "bullshit". The problem with "bullshit" isn't that you are cursing. It is a way to dismiss an article with out actually engaging in any of the arguments made by the article. At the end you do hint that you believe the article gives a biased report of history, but you give no specifics, and it comes many paragraphs after "what bullshit".
  - cchooper 18y ago
    I see nothing wrong with beginning the post "What bullshit". It's a concise statement of what he believes. And what's wrong with stating your belief up-front and then providing the argument afterwards?
elviejo 18y ago
If we were designing a keyboard from scratch: What shall we do? How about we find the most popular letters and put those under the strongest fingers. then we favor hand to hand switch, since that is faster. etc. etc. Seems to me that the logical conclusion of such a study, would be a Designed Keyboard (DK). Said GK has to be better than any alternative that came from a) Random arrangement b) being able to type 'typewriter' at sales demonstrations or c) designed to to avoid jams of moving parts. We would call this Not Designed Keyboard (NDK) So we have DK and NDK. Almost by definition DK has to be better than NDK. Now we all know that in order to do a good job we need to use the right tool for the job. If the job is typing which it is for 100,000 of people then the only logical conclusion is that we need to use the better keyboard in this case Designed Keyboard. (DK) Next what if you had an organization with 1,000 secretaries that their whole work was to type documents day in and day out for the next 20 years and you could improve their typing in just 1% using a better keyboard. Would it be economically worth doing? Of course. In conclusion switching to a DK is the smart choice, is the cheaper choice. And the only thing absent from this article in "reason" magazine is precisely that; reasoning.
- s_tec 18y ago
  Our wonderful universe acts in surprising ways. Human intuition is often wrong, and the only way to know something for sure is to test it. The Dvorak keyboard's logical organization seems intuitively faster, but we don't know how the human brain is wired, so we don't really know the limiting factors on typing speed. Anything is possible, and the only way to know the truth is by experiment. That's the scientific method. According to this article, the valid experiments support QWERTY over Dvorak, or at least show little difference. I don't know if the article true without examining the experiments myself, but I wouldn't call the article "unreasonable" for valuing science over fallible human reason.
  - tokenadult 18y ago
    Human intuition is often wrong, and the only way to know something for sure is to test it. Hear. Hear. http://norvig.com/experiment-design.html http://norvig.com/experiment-design.html
elviejo 18y ago
For "real science" on keyboards go to: http://mkweb.bcgsc.ca/carpalx/ http://mkweb.bcgsc.ca/carpalx/ From the site: "The carpalx project attempts to find the best keyboard layout to minimize typing effort for a given set of input documents. These documents may be English text, programming code, or whatever you find yourself typing for hours" Take a look at the studies from those results.
- silentbicycle 18y ago
  Thanks for posting this! I was in the midst of writing something very similar.
tlb 18y ago
The Dvorak keyboard might have some slight advantage for English text, but it's noticeably worse for programming where you need ;=+-,./{}[]() and the numbers quite often. There might be a advantage for a programmer's keyboard, basically QWERTY with common punctuation on a fifth row above the numbers. But my few brief experiments with unorthodox layouts caused no end of hassle when trying to use other people's machines, so I gave up.

26 points by nickb 18y ago

Based on the distance your fingers have to move to type average words on qwerty and dvorak, it's really impossible that Dvorak wouldn't be loads better.It's sort of like claiming that someone wouldn't be faster at running 50 meters than 100 meters. Of course, where this analogy falls down is that you have to learn a new skill, which means there's an upfront cost and thus makes the whole issue interesting economically. The article claims that the cost of learning the new skill could never be amortized and uses this evidence: > In the first phase of the experiment, 10 government typists were retrained on the Dvorak keyboard. It took well over 25 days of four-hour-a-day training for these typists to catch up to their old QWERTY speeds. (Compare this to the Navy study's results.) When the typists had finally caught up to their old speeds, the second phase of the experiment began. The newly trained Dvorak typists continued training and a group of 10 QWERTY typists (matched in skill to the Dvorak typists) began a parallel program to improve their skills. In this second phase the Dvorak typists progressed less quickly with further Dvorak training than did QWERTY typists training on QWERTY keyboards. Thus Strong concluded that Dvorak training would never be able to amortize its costs. He recommended instead that the government provide further training in the QWERTY keyboard for QWERTY typists. I really can't express how stupid this is. The logical implication the article would have you draw from this study is that qwerty actually may be MORE efficient. To believe this, you'd have to believe that the top end hypothetical speed of qwerty is higher than dvorak. This is literally physically impossible, as I've already pointed out. Moreover, the fastest typist ever used Dvorak. While that's not conclusive, it's suggestive. Folks trying to optimize a skill to the highest level are very good at cutting out all inefficiencies. To understand the results of the study cited there, we need only look to exercise science. First of all, what most lay people don't realize is that a great deal of exercise results for activities like weight training is explained by the exercise's affect on your nervous system, not your muscles. Properly managing your nervous system response is at the heart of continued gains, and not managing it properly leads to overtraining and plateaus. All top flight Olympic athletic coaches spend much of their time thinking about this. As a matter of fact, it is a well known and well documented phenomenon in the literature that if you train at close to maximal capacity on the same training regimen for longer than 3 to 4 weeks, you will stop making gains. You'll see big increases the first 3 weeks, and after that, nothing. This is a nervous system effect. Armed with that knowledge, it becomes clear why the study's results look the way they do: The dvorak trainee were already in a state of CNS fatigue, whereas the qwerty trainees were starting out fresh. Dvorak is clearly better than qwerty. Colemak is even more better. However, I don't think this situation is much of an indictment of the free market. I don't think the bottleneck in most people's productivity now (or ever) was top end typing speed and it would take quite a long time to recover the cost of the switch. Considering that this is THE MOST concrete example ever cited where you might be able to say that government meddling could induce a positive effect, I'm not inclined to agree that it's in general a good idea to have our economic lives centrally manipulated. There are far more examples of net economic losses from government interference.

pragmatic 18y ago
The point seemed fairly clear. Retraining on a Dvorak held no real advantage. The fact is that almost all people learn on QWERTY so the necessary retraining isn't worth it. Not to mention the millions/billions of keyboards in existence that would need to be scrapped.
- tome 18y ago
  Touch typists don't look at the keys anyway, so you don't need to scrap the keyboard. Windows keymap switcher, or the equivalent in your OS of choice allows you to quickly go back to QWERTY on the (very) rare occasions where what you need to do corresponds with the actual symbol on the keys.
Rabidmonkey1 18y ago
Hmm... Colemak looks interesting. I've always had a hard time learning to use Dvorak, especially since everywhere else you work it's going to be Qwerty by default. Colemak is also public domain, according to Wikipedia. I might have to give it a try. http://en.wikipedia.org/wiki/Colemak#Colemak http://en.wikipedia.org/wiki/Colemak#Colemak
- cninja 18y ago
  I actually installed Colemak on my computer yesterday. I played with it for a couple hours before uninstalling it. Most of what I type is programing code, not prose. It dawned on my that most my typing delays come from all of the parens and brackets and other programming symbols. Since all these fancy keyboard layouts have been computationally optimized for prose language, they don't address the primary cause issue. I decided that if I'm going to take the time to learn a new keyboard layout, then it will need to be one that has been optimized for programming.
  - Shamiq 18y ago
    Fix this, please. It's making this thread just about unreadable.

pragmatic 18y ago

This is not just about keyboard layouts. It's about the concept the poor products (VHS, Windows) win over superior products (Beta, Mac) because of a winner take all economy. As you can guess the author shows this to be bunk science.

davidw 18y ago
What they are afraid of is the idea of the free market not always and in every case producing the best possible result, so they need to aggressively attempt to demonstrate that 1) either market failure just doesn't exist, as they attempt to show here, or 2) that in no case, ever, will government intervention improve the situation, which is the topic of other libertarian writing. In other words, this is just another episode of "Libertarians vs Socialists": http://news.ycombinator.com/item?id=419503 http://news.ycombinator.com/item?id=419503 For a serious treatment of the economics of digital goods and networked economies, I think Varian and Shapiro are a better read: http://www.squeezedbooks.com/book/show/7/information-rules-a-strategic-guide-to-the-network-economy http://www.squeezedbooks.com/book/show/7/information-rules-a... (Edit: to be fair, it turns out the authors have a more scholarly treatment of the subject matter, which is a better read: http://www.utdallas.edu/~liebowit/harvj/harvard.html http://www.utdallas.edu/~liebowit/harvj/harvard.html Still though, I think it's pretty clear that they're coming from a point of view, and "backfilling" to try and get there)
twopoint718 18y ago
It would be interesting to construct an experiment to see what kinds optimal solution a free market can and can't find (though I have no idea how this would be done). In search algorithms you have to worry about merely finding a local maxima rather than a global maximum. In algorithms such as simulated annealing you perturb your solution randomly at the outset, trying to "bump" your solution off of one of those local maxima. What if the free market is prone to getting "stuck" in some locally optimal, globally sub-optimal solution? Depending on initial conditions (price, number of firms in the marketplace, demand, etc.) we could end up with great solutions, but in others with terrible results.

Hexstream 18y ago

What bullshit. There's a common pattern in Dvorak VS Qwerty: Dvorak advocates sometimes rely on history and sometimes on purely technical grounds to prove their claims, while Qwerty advocates always only rely on history because there's simply absolutely no way to prove Qwerty approaches Dvorak when it comes to logic. Here are some facts about Dvorak that are easy to verify and the utility of which is obvious: - Dvorak has the most used keys on the home row. Compare: AOEUIDHTNS VS ASDFGHJKL; - Dvorak optimizes for alternating hands, which obviously helps speed. Ever tried typing "street" or "states" in Qwerty, two common words with common letters that you have to type all with the left hand? On the Dvorak side the worst word is "joke" I think. - Dvorak optimizes for typing a bit more with the right hand instead of the left hand, which makes sense because the right hand is usually the strongest. - Dvorak optimizes for typing from the outside to the inside (try tapping your fingers in rapid succession on the desk from outside to inside, then inside to outside, and you'll see immediately which one is easier and faster). - Dvorak optimizes for typing more with stronger fingers (index and friends) than weaker ones (pinkies and friends). - Dvorak has all the vowels on the left hand on the home row (except Y which is conveniently on top of I), one beside another, which is great for learning. It also makes sense because you use consonants more than vowels, (see right hand argument) and since the most used consonants are on the right home row, it optimizes alternating hands. No need to mention Qwerty makes no such optimisations and the result is correspondly atrocious. See? No need to bullshit with biased reports about history.

lliiffee 18y ago
You weaken an otherwise useful post by using words like "bullshit". The problem with "bullshit" isn't that you are cursing. It is a way to dismiss an article with out actually engaging in any of the arguments made by the article. At the end you do hint that you believe the article gives a biased report of history, but you give no specifics, and it comes many paragraphs after "what bullshit".
- cchooper 18y ago
  I see nothing wrong with beginning the post "What bullshit". It's a concise statement of what he believes. And what's wrong with stating your belief up-front and then providing the argument afterwards?

elviejo 18y ago

If we were designing a keyboard from scratch: What shall we do? How about we find the most popular letters and put those under the strongest fingers. then we favor hand to hand switch, since that is faster. etc. etc. Seems to me that the logical conclusion of such a study, would be a Designed Keyboard (DK). Said GK has to be better than any alternative that came from a) Random arrangement b) being able to type 'typewriter' at sales demonstrations or c) designed to to avoid jams of moving parts. We would call this Not Designed Keyboard (NDK) So we have DK and NDK. Almost by definition DK has to be better than NDK. Now we all know that in order to do a good job we need to use the right tool for the job. If the job is typing which it is for 100,000 of people then the only logical conclusion is that we need to use the better keyboard in this case Designed Keyboard. (DK) Next what if you had an organization with 1,000 secretaries that their whole work was to type documents day in and day out for the next 20 years and you could improve their typing in just 1% using a better keyboard. Would it be economically worth doing? Of course. In conclusion switching to a DK is the smart choice, is the cheaper choice. And the only thing absent from this article in "reason" magazine is precisely that; reasoning.

s_tec 18y ago
Our wonderful universe acts in surprising ways. Human intuition is often wrong, and the only way to know something for sure is to test it. The Dvorak keyboard's logical organization seems intuitively faster, but we don't know how the human brain is wired, so we don't really know the limiting factors on typing speed. Anything is possible, and the only way to know the truth is by experiment. That's the scientific method. According to this article, the valid experiments support QWERTY over Dvorak, or at least show little difference. I don't know if the article true without examining the experiments myself, but I wouldn't call the article "unreasonable" for valuing science over fallible human reason.
- tokenadult 18y ago
  Human intuition is often wrong, and the only way to know something for sure is to test it. Hear. Hear. http://norvig.com/experiment-design.html http://norvig.com/experiment-design.html

elviejo 18y ago

For "real science" on keyboards go to: http://mkweb.bcgsc.ca/carpalx/ http://mkweb.bcgsc.ca/carpalx/ From the site: "The carpalx project attempts to find the best keyboard layout to minimize typing effort for a given set of input documents. These documents may be English text, programming code, or whatever you find yourself typing for hours" Take a look at the studies from those results.

silentbicycle 18y ago
Thanks for posting this! I was in the midst of writing something very similar.

tlb 18y ago

The Dvorak keyboard might have some slight advantage for English text, but it's noticeably worse for programming where you need ;=+-,./{}[]() and the numbers quite often. There might be a advantage for a programmer's keyboard, basically QWERTY with common punctuation on a fifth row above the numbers. But my few brief experiments with unorthodox layouts caused no end of hassle when trying to use other people's machines, so I gave up.