When I switched to a grid-style keyboard, the SL on the pinky really bothered me.  So i swapped the C and L keys (and for the hell of it, I also swapped the I and U, but that made almost no difference that I could feel) and it felt noticeably better.

On normal keyboards, however, the SL wasn't a problem since I use non-standard fingering.

Otherwise, typing on dvorak felt very nice. I would not have switched to my current layout, but I was making too many 2-letter-swap errors - I guess my left and right hands are not very well coordinated. :)

EDIT: Oh, almost forgot.... I also switched the G and M keys based on my timing-data.

It's using different fingers to increase typing speed. It lowers same-finger keypresses and keeps me typing in constant motion. Comfort increases since you're not locked into typing using unnatural motions.

The downside is that the complexity in learning the layout increases slightly (some phrases would be typed using alternate fingering, while other phrases would use standard fingering). Hence, a slightly higher level of concentration is needed until you get used to it. The end result is that accuracy and speed is improved (at least that was the case in my situation.)

For instance, on standard dvorak, in order to type S-L-I-P, the following standard fingering would be used: right pinky-right pinky-left index-left index.

The fingering I used was: right pinky-right middle-left index-left middle.

I also used other optimisations, but those were just minor "tricks." (like "sliding" fingers between rows, or rocking the finger tips to hit 2 keys simultaneously, etc.)

I employ the same type of tricks on my current layout with words like "know", "ledge", "only", "damn", "queue", etc.

HTH :)

It looks like both Dvorak and Colemak assume that the qwerty-G position is easier to reach than the qwerty-E position.  My subjective experience over a long time reaching for these keys doesn't bear this out.  I find stretching a finger sideways, even on the home row, to be more effort than moving a finger directly up by one cell, especially on the left hand.   Maybe the weighting system shouldn't automatically assign everything in the home row a higher ease value.

I find that to be true. I think part of the reason for this is that the middle finger is long, and it almost wants to stretch out to the key above it.

Some layouts that take advantage of this are Capewell's Layout, Arensito, and MTGAP's Layout (my layout) which can be found at the alternative layouts page.

I won't deny that.

Efficiency is a very subjective/touchy term. Some people equate efficiency with minimising finger-distance. Others equate it with fastest speed attainable. Yet others incorporate a combination of comfort and abstract qualities like finger-rolls, same-key presses, row jumping to come up with an 'effort algorithm'.

It all depends on what kind of a typist you are. If you like the high hand alternation and you're better coordinated than I am, then dvorak will probably be better for you than colemak. It's just that good. This is without the 3 simple modifications (swapping U and I, C and L, Caps and Bksp).

With the modifications, it'll be almost indistinguishable on paper.... but you'll definitely feel the difference when you type.

Here's a thought - why not try them both?... :)

In my experience, the most benefit comes from learning to type properly, and using a good mechanical keyboard (http://ergocanada.com/ergo/keyboards/me … tches.html)

Swapping U and I doesn't help much English typing. I do think Dvorak had good reason to put them where thI've swapped them, but I type mostly Italian, where vowel digraphs involving "I" are way more common.

What's so bad about L?

BTW, have a look at Capewell-Dvorak.

I think it boils down to this:
- you are using a standard keyboard ->
      - you think hand alternation is good (I do) -> go Dvorak
      - you think finger rolls are good -> go Colemak
- you are using a contoured keyboard with more keys for your thumbs -> go with Maltron or Arensito.

That's it.

I wish the ANSI Dvorak layout were more similar to that developed by Dvorak on typewriters, instead of being a knock-off of ANSI Qwerty. Here is an image of an original Dvorak typewriter:

Notice placement of numbers and digits.

EDIT: added reference to switches comparison.

IME switching U and I trades speed/quality to get a lower distance.  We should do some tests and see how distance correlates to typing speed.  This would be incredibly easy to do because you could simply take many basic words in a typing test, record the speed at which they were typed and compare this to the "distance travelled" for typing that word.  I have not done this, but I would really like to see how much "distance" is "worth" and how it seems to work.

I don't think he was looking for correlations between different typists using different layouts, but for correlations in data from individual typists - eg. do you type 5-letter words faster if they involve less distance or not?

But there are more factors than just distance.  Anyone will be slower typing "htkin" than "think" although they involve the same distance.  The former is not a word and the brain needs more time to parse it letter by letter, while the latter is for many people a common word, typed in almost a single movement.

Many people learning about alternative keyboard layouts initially put too much focus on speed, while more experienced Colemak users often find the comfort factors far more important.  I think we should focus less on "speed" and more on purely ergonomical aspects when advocating Colemak (or Dvorak).

For optimizing keyboards i'm wondering what's the correlation between letters and brains? Perhaps vowels or certain letters have more preference for one side of the brain (and hand) which would help make keyboards more intuitive.
Considering hand alternating sometimes confusing, distance isn't always the issue, and learning the layout is an issue, has mental comfort been explored as much as physical comfort?