Sorry, but I beg to differ.  It's not just that Colemak moves a lot fewer keys than Dvorak.  Since you seem to have missed the points on this made in the FAQ, I repeat them here for you.

Keys remain close to their QWERTY positions

The keys in Colemak are just a key or two away from their QWERTY position (except E/P/Y). This makes it easy to visually 'hunt' for the keys. Moreover, it makes it easy to 'hunt' on QWERTY when you're temporarily forced to use the QWERTY layout.

Only two keys move between hands from QWERTY

The hardest keys to learn are those that move between hands. Colemak only moves 2 keys between hands, compared with 22 that move hands in Dvorak.

The bottom rows remains mostly unchanged

Except of 'K' instead of 'N' the bottom row remains the same. The bottom row is a bit more difficult to learn because of the way the keys are aligned.

No typing tricks required

To achieve a high typing speed on QWERTY, most people employ tricks to avoid same finger typing, such as when typing the ED digraph. On Colemak these kinds of tricks aren't necessary.

Many low frequency keys do not move

Colemak minimizes moving low frequency keys and except ; and : all punctuation keys remain in their QWERTY positions. Low frequency keys take the longest to relearn because you practice them less often.

Memorization isn't linear

Memorizing a 10 digit number takes much more time than memorizing a 5 digit number (more than twice the time). Colemak moves only 17 keys, while Dvorak moves 33 keys.

Avoids confusing key switches
Colemak avoids confusing changes, such as T and Y being switched on Dvorak.

Backspace key on home row

During the learning phase errors occur very often. By placing the Backspace in easy reach reduces the frustation and allows to easily correct errors.

The point is that Colemak keeps some very common ones.  Moving keys is always going to come with compromises.

For a Mac user like myself, you notice that Colemak offers a much easier one handed Cmd-T shortcut for tabbing in Safari. 
Dvorak requires use of both hands.   The all important Cmd-Q for quit and Cmd-W for closing a window have not moved.

Another issue you seemed to have missed that Colemak pays attention to while Dvorak being from a pre-PC era does not,  repeating from the FAQ

Potentially destructive shortcuts:

Some users in Dvorak intend to hit Ctrl+V (Paste) but instead hit accidentally the adjacent Ctrl+W, thus losing everything they've typed. Colemak has been designed so most of the potentially destructive shortcuts (Q - Quit, X - Exit, Ctrl+Z - Undo, Ctrl+X - Cut, Ctrl+W - Close Window), will remain the same as QWERTY.

No, I didn't miss the other points made in the FAQ. Some of them are valid, I don't dispute that. However, all but one of those you highlight refers to the changes made from the Sholes layout. The main assumption made is that the Colemak is superior to Dvorak because it is easier to learn Colemak if you already know the Sholes layout - I'm not disputing this, though I would dearly like to see an objective study of this assertion to prove whether it is true. I would also like to see research into the comparison of learning ease for those completely new to keyboards.

My main concern is the fact that children - some of them very young - are being effectively forced, by peer pressure, into learning an obsolete layout. From the perspective of the totally new user, is the Colemak better than Dvorak? More than this, is it sufficiently better that it is worthwhile further confusing those who are in a position to promote a change from the Sholes standard by bringing Colemak into the discussion?

Your assertion is misleading. It takes a smidgeon over twice as long to memorize a 10-digit number than a 5-digit one. The smidgeon arises from splitting the task into the memorization of two 5-digit numbers.

Moving the backspace key would certainly appear to be a change that would have many benefits to the touch-typist. My own preference here would be to split the - unnecessarily long - space bar into two, and assign the backspace function to the left thumb: but that would require an alteration of the physical keyboard, something that Dr. Dvorak himself recognised as better avoided.

No, I didn't miss that point, it's very valid. I too hit Ctrl+W by accident once when intending to hit Ctrl+V, and I agree that it is a problem. My solution has been to retrain myself NOT to use the Ctrl+C/V/X keys, but to use Shift&Ctrl+Insert & Delete, instead - easily done with one hand (the right) - it comes in handy being able to use the mouse with either hand :)

The problem is by focusing so much on Colemak vs. Dvorak, I think you are really just struggling with the idea that there is one best solution, one best layout.  There isn't.  There are far more variables than constraints and not everyone agrees with what the constraints should be to make an optimum layout.  All that can be really said is that both are much better than Qwerty by many metrics commonly used.  The problem is it is quite clear that many people will argue with and disagree with you forever on whether those metrics are meaningful when you can point to all those touch typists using Qwerty at over 60 WPM, over 80, over 100.  Just go here ( http://www.ryanheise.com/typing-test/ ) and look at the top 10 typists are all Qwerty going over 125 wpm.  After decades of advocacy, academic texts, Navy tests, etc. where is Dvorak layout ?   

I think that it is quite clear that Colemak is not the "half-way house" as you put it.  Colemak shows that you can get an optimal layout that compares favorably with Dvorak without the radical reorganization that Dvorak's layout demands. Part of the reason for that is in such a high dimensional solution space, there are lots nearly the same optimal solutions exist. 

As alternatives in the huge number of optimal solutions, both layouts have good arguments going for them for different reasons.  It's also clear neither is going to ever dislodge Qwerty.  All they can do is be universally available for those that would desire something other than Qwerty can use them. 

If you like Dvorak great.  There is no reason to switch to Colemak.
If someone likes Colemak great.  There is no reason to switch to Dvorak.
and for most touch typists, they simply are not even listening.

People with the inclination are already doing that. Spending lots of time coming up with their own solutions or buying other peoples solutions.   Personalized layouts though don't have anything to do with Dvorak's standard layout.  The real point is barriers that Dvorak's layout poses to a Qwerty touch typist to making the transition.  For some like yourself, you got through the extra load on the right pinkie.  For others, it may have caused them to quit in the first week for all you know. Know one really knows.  What we do know is that Dvorak's layout despite being so easily available for many years is still rarely used compared to Qwerty and that's unlikely to change.

Attacking Colemak isn't going to help Dvorak be adopted.

The one thing that we are agreed on is that the Sholes layout is a disaster.

Colemak may or may not be better for those who already know the Sholes layout; it would be interesting to test this empirically.

My intention in starting this thread was not to attack Colemak.

My point is that when it comes to attempting to persuade those who have the power to influence national education policies that there is a significant benefit in considering a move away from QWERTY, it muddies the waters when there is so much confusion over what alternative is 'the best'.

It was one reason I chose to try Colemak rather than Dvorak - I'm left-handed and my left hand is noticeably more dextrous than my right (even after years of practising the piano). My objection to Dvorak is not just that the right hand fingers get a bit more use than the left, but that the right hand fingers have to *move off the home keys* more than the left, so it favours right-hand dextrousness. I don't have any data on that, but it *appears* fairly clear from the layout.

Shai - any chance of modifying https://colemak.com/Compare to show Distance separately for left and right hands? I'd certainly find it interesting.

I had other reasons for preferring Colemak over Dvorak too - mainly that Dvorak's design principle of maximising hand alternation didn't make much sense to me (based on my experience of playing the piano), and that Colemak keeps the Ctrl-Z/X/C/V shortcuts easy to use with the left hand while mousing with the right.

In case you missed it above, the RyanHeise typing test has a replay feature which simulates the finger movement on the different layouts.

"In fact, in my experience, it is relatively trivial to memorise a 4-character sequence. If I were to attempt this, I would therefore split the sequences down further:

1. WLJc
2. 8Rs2
3. MqwC
4. hYM1
5. pK3Q
6. CfyT"

That's the silliest logic I've ever heard.

"I can memorize 2 digits in a split-second. That means I can memorize 4 digits in a half-second. That means I can memorize 8 digits in about a second. That means I can memorize 16 digits in about 2 seconds. That means I can memorize 32 digits in about 4 seconds. That means I can memorize 64 digits in about 8 seconds. That means I can memorize 128 digits in about 16 seconds. That means I can memorize 256 digits in about 32 seconds. That means I can memorize the first 500 digits of pi in under a minute!"

Good luck with that.

I agree with you that in a professional environment, that is in the case of a professional teacher being told you have to be retrained and now teach Dvorak, the investment would seem to be trivial.  That's not the case we find ourselves in 60+ years after the Navy study.  The situation in choosing Colemak or Dvorak is completely voluntary and there are downsides to learning an alternative layout because of the universality of Qwerty hardware and software.  In an environment like we have even small subjective things can make the difference to someone who is going against the flow.  Colemak does not  look so radically different from Qwerty. It get's a lot of gains even with for the touch typist having keys in the same place or a related space to Qwerty.  Some of this subjective stuff can be reassuring.  Maybe it takes someone just a few days less to get to a competent speed and that's not really significant when viewed on the scale of a lifetime but it's just enough to get someone over that hump and on to different path despite going against the flow.  Given the way things are there is a lot to like about Colemak including because of some choices made is cross language international appeal.  Both layouts are rational and as such there are always going to be people that learn either layout in a snap.  For the rest, when it's voluntary, you need everything possible to make it as easy as possible to switch. 

As a disclaimer, I am a lefty also.  Until this thread and watching those movies, I hadn't realized the right handed bias of Dvorak and perhaps that is another reason I did not have a good experience trying to learn it.  In the tech industry they talk a lot about the influence of early adopters.  Maybe lefties already used to living against the flow are ready to be early adopters here and Colemak has by being nearly neutral found a hook not open to the righty Dvorak.  We may be a minority but a receptive minority can be a powerful thing in attempting a paradigm shift.

>About being lopsided: according to my statistics, it's more like 40%/60% for Dvorak, you can try it for yourself on the compare page (which doesn't take into account backspace and the space bar). If you take in to account the backspace key, enter key and the space bar, it becomes more like 33%/67%. See also the ryanheise.com typing test, and compare it visually for yourself. That said, I don't think it matters that much, and I don't think anyone has mentioned it as a reason for switching to Colemak.

Where did you get your statistics? For letter frequency, I use either the Wikipedia page on letter frequency or an analysis of a bunch of my own documents (138 pages worth). I find that my documents are about 5% off from the Wikipedia page, mostly in the mid range of c to b (cumwfgypb according to Wikipedia, ucmygwfpb according to me). But that's not the important part. The important part is, using my data, which was more precise–not as accurate, but more precise, I found that Dvorak is about 46%/54%. Taking into account delete/tab/return etc., for me it's probably about 42%/58%.

About ease of memorization, I find that if a string of digits are longer than 4, I break it up: 5 becomes 2-3, , 6 becomes 3-3, 7 becomes 3-4, 8 becomes 4-4. Longer than 8 becomes increasingly difficult to hold in short term memory. A lot of this is about short term memory length. If it's 7 digits, you can read it once and say it to yourself over and over until you know it. If it's 10 digits, which is only 3 more, you have to keep looking back at what you're reading the digits from to memorize them, because people find it a lot harder to hold 10 digits in their head at once. I'm not saying nobody can, but not many people can.

That is true. But you'll notice that of the top 10, 1 is Colemak and 1 is Dvorak. MAYBE 1% of all people use Dvorak or Colemak, yet they make up 20% of the top 10. That means they are overrepresented in the high scores by 2000%. Plus the world's fastest typist used Dvorak.