james mckay dot net

Now when I saw what this guy had to say about Colemak, my initial reaction was that he was being a jerk. Four days is nowhere near enough time to come to a reasonable conclusion about whether or not you’re going to get anywhere with an alternative keyboard layout, as even the most diehard fanboy would admit. Colemak actually has a lot going for it — it is easy to learn, and well supported by a vibrant online community, which comes in handy when you’re doing something as off-beat as using a different computer keyboard layout to everyone else.

But you can’t say the same thing about someone who draws exactly the same conclusions after having been at it for several hours a day for four months — by that time you should certainly be able to tell whether it’s going somewhere or whether you’re wasting your time. And in the past week or two, I have done exactly that.

My switch back to qwerty was partly prompted by our recent recruitment drive — as part of the interview process I’ll be wanting to do a little pair programming exercise with potential developers, and this is the kind of situation where an alternative keyboard layout would get in the way. However, much more significantly: I have found that Colemak has failed to meet my expectations.

My top Colemak speed of 71 words per minute may sound pretty impressive, but when you consider that my top qwerty speed on the same test was 90, the picture looks quite different. My typical results for Colemak have stuck stubbornly in the 62-64 range without budging an inch in three months, occasionally even dropping down into the 50s.

I’m sorry, but a net speed loss of 20% must be some new meaning of the word “fast” of which I was not previously aware.

I haven’t noticed any significant difference in comfort or accuracy either. Colemak initially gives the impression of being more disciplined and comfortable, but after four months of it, I was still making just as many typos and mistakes, and when switching back to qwerty, I did not notice any difference in long term comfort whatsoever.

Psychologistst talk about something called “cognitive dissonance.” This is where you get into something at considerable personal expense, then eventually, further down the road, it begins to dawn on you that you may be barking up completely the wrong tree. At this point, what many people do is to start rationalising their decision, and even defending it vigorously — the classic attributes of fanboyism. I sometimes wonder if this is what we see to a certain extent among devotees of alternative keyboard layouts, leading to the advantages of their layouts and the disadvantages of qwerty being exaggerated. They certainly would have you believe that qwerty is a total disaster area. They love to quote statistics about how much less your fingers travel on their layouts, how much more you use the home row, and so on. Frequency usage diagrams are all very well, but to be honest, that’s just theory, and unless you can demonstrate that this translates into a clear and obvious advantage in practice, which outweighs the disadvantages involved in using a non-standard layout, these statistics become no more meaningful than lines of code as a metric of developer productivity.

There have never been any scientific studies that have demonstrated significant advantage to alternative keyboard layouts, and even those that demonstrate relatively minor advantages are disputed. “The Fable of the Keys” by Liebowitz and Margolis is the well known paper here: its bottom line was that there were conflicts of interest behind wartime studies showing an advantage to Dvorak, and while it has seen one or two rebuttals from Dvorak fans, these don’t seem to have been given any serious consideration whatsoever by ergonomics researchers.

To be honest, I think this is why alternative keyboard layouts simply aren’t going to take the world by storm. Colemak is probably about as close as you’re going to get to attaining that goal, and sure, it’s easy to learn, and yes, its lively, friendly online community is fantastic, and yes, it’s maybe better than Dvorak, but its advantages are simply not sufficient to present a convincing case for its widespread adoption.

So sorry to disappoint any of you alternative keyboard fans out there. If you’re already a satisfied Colemak user, don’t let any of this put you off, of course. If you’ve found that it works for you, that’s fine — it’s just that it hasn’t worked out for me as I’d hoped.

Nothing personal…

(Update 4 June 2008: added a note on cognitive dissonance. Hat tip: Joel Spolsky and Jeff Atwood, who discuss the topic in their latest podcast on stackoverflow.com.)

Comments (22) »

Posted at 08:07 in Typing