Since my discovery of it in the 1990s, this article has caused me to ignore literally hundreds of reports from programmers that the keyboard is faster than the mouse because those reports gave no indication that the speeds had been objectively determined.
The most important thing I gained from this article is an awareness of the effect by which planning a sequence of editing keystrokes tends to interrupt tasks like writing prose or writing code more than "purely perceptual" tasks like positioning the mouse cursor tend to do. Highly recommended.
But the experiment does not show that the mouse is faster than the "keyboard". It shows that the mouse is faster than using the cursor (arrow) keys for a task where doing the latter would obviously by extremely slow. For me, "using the keyboard" means typing ":1,$s/|/e/g". Say there are 100 instances of "|" to replace. Which do you think will be faster?
Lets ignore the fact that this is a search and replace since that just makes the contest painfully unfair. Normally when using a good text editor like emacs or vim and you're moving somewhere from far away you do a search. Or if you move manually you start out moving by paragraphs, then by sentences, then by words, then finally by characters when you get to the right word. But using the arrow keys you go the same speed the whole time no matter how far away you are. I understand that Word might not have all those fancy movement modes, but it does let you go faster by using Ctrl-Arrow and I don't see why Tog didn't at least let his test subject do that.
Indeed. I'm at a loss trying to understand why he thought he was demonstrating anything interesting. There would be no harm, except that now and then you hear people claim that it has been proven, by studies, that the mouse is faster than the keyboard. I suppose they are quoting people who are paraphrasing people who have sort of heard of this trivial study.
This article only really applies to the "general population".
When you are the master of your trade (read: touch typist who types on a daily basis e.g. developer), the keyboard is much faster for most tasks. That is why vim is so popular.
I have been on and off a programmer over the last 10 years. and the last 3 have been using vim(only these 3 were on linux), used to be scite for the first 3 years and whatever was available the rest of the time (which was mostly part-time programming, was otherwise studying cog. sci). I have so far been against the mousing faster than keyboarding argument and even now i reach more easier for the keyboard.
But over the last year i have been writing and i can relate to that effect of interference caused by planning keystrokes in my writing(guess still am not enough of a programmer to notice it in writing code).
So i am beginning to understand that the question assumes a simplicity that plain and simple speed in some specific and narrow lab task is indicative of real-world problems and challenges. (Aka, problem analysis, solution space search, choice of solution, planning execution and then execution).Depending on the experimenter and the lab group he's working with, these experiments tend to focus on isolating one of these phenomena/mental processes(sometimes not even clear to the experimenter).
Anyway, at the moment i am convinced that writing is best done in blind manner, ie: no visual input, switched off monitor. I currently use a white on black bg on 750 words, but wish i could find a way to turn off monitor on my laptop and keep typing.
P.S: i found the same thing about QWERTY vs DVORAK Layout. i was a touch-typist(barely average) on QWERRTY too and all the research suggested it's not worth re-learning DVORAK. I started it off as an experiment(1 hr a day of DVORAK for 2 or 3 weeks) find that my writing stamina has improved. I'll bet that there are hand/palm size(area, angle,distance between fingers etc.) variations in the effect of KB layout and that's a statistically significant factor.
>I started it off as an experiment(1 hr a day of DVORAK for 2 or 3 weeks) find that my writing stamina has improved.
I think that anything that causes the person to pay more "perceptual" attention to the details of the physical process of typing will improve "writing stamina" and will reduce the chance of RSI. I tend to think that the reason people's RSI goes away when they start to learn Dvorak is that the need to learn causes them to pay more attention to the physical process of typing (and that a Dvorak user who never learned Qwerty would see approximately the same benefits if he started to learn and to use Qwerty). I am much more confident of my first sentence than my second sentence.
(To be painfully explicit, I tend to think that there are better ways of paying more attention to the physical process of typing than learning Dvorak. E.g., just resolve to pay more attention works pretty well.)
What I consider the important negative effects of overly keyboard-centric editing (and overly keyboard-centric window management) take years to show up, are hard to measure or quantify and have nothing directly to do with typing speed.
Typing speed is important only in that when I realized that people could be so mistaken about typing speed, I wondered what else they could be mistaken about.
The most important thing I gained from this article is an awareness of the effect by which planning a sequence of editing keystrokes tends to interrupt tasks like writing prose or writing code more than "purely perceptual" tasks like positioning the mouse cursor tend to do. Highly recommended.