If I end up making and using a "Dvoremak" just because I'm stubborn, I'll let you know how it feels.  :)

If I do, it looks like it would be best to keep Dvorak's placement of both dot, comma and W to keep that bit simple (for me!). Then move the J to where the QWERTY dot is, and the F needs to go somewhere so it'll end up right where Dvorak put it! Sounds like a barrel of laughs.

Hope it's okay with you if people play around. I don't mean to be disrespectful. Some people just have to try things out for themselves.

No, I'm not perfectly happy with Dvorak and I think I'd be willing to try something else. Especially the Norwegian Dvorak which has my left pinky straining to find the not uncommon Ø at the far left of the keyboard, has made me frustrated. But there are other issues that you've mentioned on your pages, too. Also, it'd be interesting to see how fast I can learn this layout and whether my speed picks up. After some months with Dvorak, my speed pretty much stabilized itself around the 60 WPM that I'd get on a good day with QWERTY before the switch - so I wasn't really impressed. The accuracy is better with Dvorak and it feels better though.

I noticed that the H isn't in the home position, so just forget I mentioned it. I do think that you can be more confident about non-English language compliance; it seems to be hardly any worse off for those languages... once you've put the special characters for each language in there. And users from various countries will be able to do that simply, mostly in accordance with their QWERTY national layouts.

Of course, I'm talking about a nonstandard layout for personal use. And since I'd be moving from Dvorak my mileage would be varying from the people moving directly from QWERTY, for instance.

Now, for some random further thoughts:

I'm curious about the home row. Now that I've been thinking about it, the Dvorak home row isn't bad at all: It facilitates many nice rolls for the diphtongs and things like SH or TH, which I really like. However, it may be a good idea to reduce hand switching a bit. The Qwerak layout (by Fingerworks) tries to achieve this by switching the O and S on a Dvorak home row, which may work well for all I know. The Qwerak is described as "experimental". (On a side note, writing "say" on that layout looks hard!)

I don't think that moving a key between hands is as bad as Shai seems to think. It should be easy enough to learn nevertheless.

I understand that Shai moved the S to the middle finger from the ring finger. Is this just because the ring finger is weaker? Well, the R is an important letter too so it can't be just that.

As a pianist, I think that the ring fingers are strong enough, but not flexible and independent enough. So they may well be assigned common characters on their home row, but less so on particularly the lower row.

I think that letter frequencies alone are less important than di- and maybe trigraphs. That's trickier to implement in an analysis maybe? It seems to be what Capewell's been getting at, although I'd like to see him finish his project before I pass further judgement there. To me, it looks like his weighting functions may still need a reworking.

From my eyeballs, the Colemak home row doesn't seem to roll common digraphs so much? Maybe it's just my eyes not seeing straight? Please enlighten me.

My current vision is something like this:
|Q|,|.|U|P|  |F|G|D|Y|;|
|A|S|E|I|L|  |R|H|T|N|O|'|
  |Z|X|C|V|B|  |K|M|W|J|-|

It'd be interesting to hear some criticism on this layout attempt! Preferably the constructive kind of course.

Not too sure about the O-S switch yet. It does have the appeal of QWERTY compliance and Ctrl-S compliance.

Other than that, the placements of U L R D Y were made as I thought best based on letter frequencies and some intuitive digraph considerations. There's room for discussion there.

In a way, it came about as an attempt firstly to keep the necessary QWERTY elements and not mess about with the ones that could be left alone, like Colemak and Capewell do. Then, to use the elements I like from Dvorak. Then, to look towards Colemak and a few others for further guidance while trying to think straight. I fully understand that my attempt is nowhere near as thoroughly researched as Shai's for instance. I hope y'all bear with me.

I haven't considered such things as same finger statistics. I can't manage to implement new layouts into the Java tester, unfortunately.

Some digraphs/words seem to roll well on this layout. There will always be some troublesome words on any layout - the question is how common they are I guess. I had to think again to get something that worked for LY for instance, and ended up with Colemak's placement of the Y where I first tried the QWERTY one.

The choice to keep the home row closer to Dvorak's came about when I considered rolls. Please prove me wrong with good arguments. Of course, bad things like the awkward placement of I and L (and in my opinion, R to some extent) in Dvorak had to be fixed.

This particular layout would be an easier switch for a Dvorak user I think. At the same time, it'd be easier than Dvorak for a QWERTY user and conserve the important hotkeys. Now, the big question is how well it'd perform eventually.

I'm not too sure about the standard performance tests. Dvorak probably believed too much in hand alternation, and there may have been too much focus on other good ideas such as single letter frequencies and the home row too. All good ideas, but the challenge is getting the mix right and the ideas properly weighted and implemented.

But having to teach each new layout attempt to a group of people and then monitor their long-term performance is grueling. So I guess we're back to thinking and using tests paired with good sense. If you guys don't make me see the error of my ways, I may give this layout a spin soon.

DreymaR, if you're already experienced with Dvorak, I'm a bit curious why you want to switch to a new layout. It serves the need for speed well enough, doesn't it?

By the way, I'm starting to recommend Colemak over Asset even though still I think S ought to have stayed in place. Despite that, I'm learning Asset because probably if I don't (as the designer), no one will.

I wonder the best layout would look like under the constraints that:

- QWASZXCV must stay in place
- only two keys can switch hands compared to Qwerty
- any 3-5 other Qwerty letters must stay in place
- all punctuation stays in place except ;:

... The answer usually depends on the scoring system, I guess ...

The nice thing about 21st century layout design is that unlike Dvorak we have computers that can evaluate a thousand possibilities per second. All that's needed is somebody to write the algorithm.

Your reasons for learning Dvorak were an interesting read. It's good that you're so driven. Some days I don't feel like doing anything....

>I think Shai is right: As the Colemak stands now, you'd be hard pressed indeed to find even 2 keys to move that wouldn't mess up something else enough to ruin the benefit. It seems that solid.

Oh, at this point I agree. Within constraints, I don't think Colemak could be improved by moving two, three, or even four keys. If S is to stay put, the current efficiency of Colemak could not be attained without a substantial overhaul of the layout, because key placements are highly interrelated. That's why I suggested an algorithm would be needed to search for it.

Thank you, sorenk, for your vote of confidence - but as far as I can see Shai has given the matter at least 10 times as much thought as I ever did. And that's a conservative estimate!

Furthermore, I think he got it right. You're doing what I did and throwing out your thoughts in layout form which is good fun for us keyboard "nerds" and may lead to valuable insights on our parts. But I came to the final conclusion that even considering my own private needs I couldn't come anywhere near the optimization level of the Colemak. Hence I'm a Colemak advocate and user now. And a satisfied one I might add, especially as the Dvorak quirks are fading somewhat from my active motor memory.  :)

Shai may take the time to highlight some problems with your layout attempt, as he so patiently did for me. In the meantime, I'll throw out a few thoughts:
- Why thumb Shift at all, really? It makes no sense to me to take Space (used for every word) away from the impossible-to-miss key of the strong thumbs, in favor of the much less frequently used Shift.
- The Enter key is somewhat of a point since it's a long stretch in its current position, and fairly often used. But is it worth it?
- Also, you cannot remap Shift/Enter at the layout system software level. A minor point since Colemak already moves an unmappable key (Backspace).
- Speaking of which: You seem to lose the backspace remapping with your double Enter setup. It's really handy to have on the home row - even more so when learning a new layout!
- Then you have to make way for your dual space keys and in the process hurt other keys' efficiencies. In particular, the very common t and n (# 2 and 6 in English) are now on stretches. That's almost as bad as what QWERTY does! And although the f is fairly uncommon it deserves better than the left pinky stretch position. You'll end up with undesirable AF and FA pinky digraphs there, too (I don't know their frequencies but I suspect they're not uncommon).
- You also end up farther from the QWERTY thus hurting acceptance. Putting the t back may seem to help, except that as mentioned it really deserves a home row position. You move the Ctrl-Q shortcut which will annoy some users. (Yes, you said to disregard compatibility but I'll urge you to reconsider as it is important to many and a genuine advantage to almost all.)
- International compatibility would be about as good as on the Colemak as far as I can see, disregarding the unfortunate key placements mentioned.
- The brackets look neat. I'd put the parenteses there and the square brackets on the shifted keys instead (and then the curly brackets on the AltGr mappings). Fun. The reason Shai hasn't been messing around with punctuation much is that they're rarely used signs and moving them would confuse people without helping efficiency noticeably.

Hope that's helpful. Always nice to meet an enthusiast!

From DreymaR's first post:
>But I feel bad about going back to those horrible punctuation placements from QWERTY.

I wonder if you would be more comfortable with Qwerty's placement of , and . if the bottom row was shifted a bit to the left. I believe keyboard makers should shift the bottom row left by 1/4 the size of a key. That way the keys would be more aligned with your fingers when you move them down from the home row. Some have suggested completely aligning the keys vertically (here's such a keyboard: http://www.typematrix.com/ezr2030/), but I think that a smaller degree of change would make the transition painless for all typists.

Thank you, I think your assessment was astute. Especially regarding the cost of removing two keys from the home row. I was just thinking that typing would be faster if spaces could be entered more efficiently and that maybe index fingers can enter data faster than thumbs. And that thumbs could be better for shifts because it's easier to move your hand around with a thumb held down. But even if the index finguers are faster my idea is still quite impractical. Still,I have been working on alternating my thumbs depending on what key starts the next word. I think it helps a bit.

I also corresponded with Michael Capewell who's doing some impressive work on keyboard optimization too. And he had much the same conclusion. He added that most people would end up using just one of the keys for space and the other key would lose functionality (wasted home row key bad).

I think some one did research on using a foot pedal for a space bar, giving a 15% increase to typing speed. But given the standard keyboard, every key on the home row necessary. I just need faster thumbs. The shift key's lack of mappablity is another good point (I didn't know they posed a challenge for remapping).

At the moment I'm still with QWERTY. Over the past few months I raised my speed from just bellow 40 to just below 60 WPM. A long time ago I tried Dvorak and I liked it because it was comfortable. But I couldn't escape QWERTY. Also I had difficulty difficulty making up for lost combinations, as in typing F-A-C-T instead of FACT. Even though the CT is an awkward reach (on QWERTY) just having a memorized pattern of laying down a set four keys helps. But I came across a tutor, by Capwell, that drills common key combinations, or "cords" as Capewell coins them ( http://www.geocities.com/smozoma/progra … rogram.zip ).

The existence of this tutor provides some incentive for me to learn an alternate layout again. My memory of the comfort of typing Dvorak is another incentive. On the other hand, given my current progress, I think I can achieve close to 100 WPM in QWERTY given another year or two of intensive practice (about 10 hours per week). The advantage to sticking with QWERTY is that (I feel) I've over come 99.99% of my Dvorak motor functions (I still have some occasional error with "comma" and "period" which I attribute to the Dvorak moving the punctuation to the left hand without transposing their order. Dvorak coded "period" to my left ring finger so on QWERTY I end up with a comma using the same finger. This experience gives me reason to side with Colemak theories about relearning a layout.

If I go with Colemak, I'll leave Caps-Lock where it is because I try to avoid the back space as much as possible in mush the way an alcoholic maintains sobriety. I think maybe I should rig a electric-shock pad to the key. My experience is that convenient correction yields convenient mistakes.

I'm still thinking of waiting for Capewell's final layout. So far he's come up with:

     .YWDF JPLUQ
     AERSG BTNIO
     XZCV; KWH,'

But it's still a work in progress. Capewell is trying to keeps some concessions toward QWERTY conversion but with a far more aggressive bent towards optimization. The really interesting bit of his research is analyzing key combinations so that many letter sequences can be entered with a roll of the fingers.

Of course the longer I wait the more I'll have to relearn if I do make the switch. And if I lean to type very fast in QWERTY I'll be able to type fast anywhere at any computer.

On the down side I'll never be able to type as long, as comfortably or as safely as I can with an alternate layout. My stint with Dvorak taught me that much.

Does any one here know the fastest speed achieved on Colemak by a QWERTY convert and what their speed was before the shift?

I've read a couple of reports of people getting close to their original qwerty speed within a matter of weeks, but I haven't come across any reports of passing the original qwerty speed (except for one mention of reaching 85wpm on the tutorial, but that's not the same as free typing).

But this sounded promising.

*(e is the right-home-row-middle finger on Colemak and s is the left-home-row-middle fingure which is just bellow qwerty e (left-top-row-middle finger on QWERTY).

Specifically I'd like to know if anyone who typed 60WPM+ in QWERTY achieved near 100WPM with in a couple years. (I know the world record holder started out in QWERTY, but I don't know what her speed was before transition or how long it took her too exceed 100WPM). Regardless, I'll be happy with +90 WPM. I don't expect to be the next Barbra Blackburn.

On a side, regarding the atavistic type-bar row shift. I like David Piepgrass' idea to encourage manufacturers to restore the rows a quarter key at a time. And one company, Maltron, has already designed a keyboard with straight rows. You can have one now for only a few hundred pounds. I think their "Executive keyboard is rather sexy ( http://www.maltron.com/maltron-kbd-flat.html ), but I'm not about to shell out 600 Pounds. Besides if I learned to type on a straight keyboard I'd feel queer using my laptop.

WARNING, MIND FART IN PROGRESS:

MAWHAWHAWHAW!!! I know what to do with the Caps-Lock key, map it to BackSpace and replace it with "the" key and clobber all the other English layouts for "Total keys."

The father, the mother and their Heather gathered thither the theater.

QWERTY 70, Colemak 70, Dvorak 70. THEmak ***50***

BWAHAWHAW!!!

Of course ye can also mean you depending on how one pronounces the y(j or TH). I think symbols such as þ and ð on retain their sound only because we've forgotten how to mispronounce them.  "Day yough I warg thoun yeh fally av theaD I fir know ofaul"
I think phonetic representations have a place in dictionaries and stenography. But written language is better from a historical standpoint if it expresses meaning independent of sound because it can remain stable across dialects. I think traditional Chinese writing is a perfect example. Many of their written words are used in languages like Japanese that are no closer than Italian and Russian. But the Chinese and Japanese literate can both comprehend the meaning of the same written word.

I think it would be neat to design a phonetic keyboard layout (not like a corded: too imprecise) but a system that you can use to precisely represent any word in your head. I think one advantage to this would be not getting your motor reflexes jumbled between layouts (but then I haven't tried it, so it's just a guess for the guy who thought of the "shift bar" in exchange for the two most prominent keys on the home row: be weary of practicing another man's idle thoughts). You'd have more basic symbols to deal with, but you could take advantage of the similarity between voiced and unvoiced consonants (place both on the same key and use shift as in shift-ð for þ because þ is seldom used). Though I don't think I'd use this system myself (I'm modifying a system of Gregg to suit the same purpose, mostly by using annotation so that I have the choice of writing fast or accurately), I think it's a good idea all the same. A major advantage of writing phonetically is that you don't get slowed down by an unfamiliar word. sn..f.l..p..g..s (ASCII phonetics) is sn..f.l..p..g..s regardless of spelling.

Thank you that helps me, especially as his qwerty speed was about where I'm at.

I already asked about the lessons and he said he hasn't created more and doesn't plan to (not his top priority) but it's a start (10 most used letters at least for english). But that doesn't stop us from making more. I've been trying to find an ordered list of letters that make up English syllables but most of everything I've come across is junk. I think such a list would help with making lessons. But then I'm probably to lazy and impatient to make lessons anyway. A similar technique that I used before is just too type poetry syllable by syllable (I came across this technique as a way of studying poems while practicing typing but it does pretty much the same thing as Capewell's tutor.

As for asking Capewell to add Colemak to his site (he already has eight layouts and would have to drop a layout or further modify the program to accommodate it). I'd be more curious to know why the person who installed the script on this site only posted three (they used the same source code created by Maxwell and modified by Capewell, except the Colemak version leaves keymaps 4 through 8 as "/* unused */").

The script is fairly easy to modify with cut and paste (it helps that there's a built in batch for compiling). My greatest difficulties were figuring out where my java directory was and how to actually run the script (I kept thinking the "comparepage.html" was just some random info link).

Note: I'll be happy to email my cut and paste version of the script to anyone who wants it.

According to my comparison. Colemak does a good job on or above the home row (but not as good as Klausser. And Dvorak seems to follow Klausser as the second contender for avoiding the bottom row...unless your Chaucer, then Colemak beats Dvorak but not Klausser.)  And Capewell clobbers Colemak for avoiding using one finger twice in a row (Colemak does well here but Capewell does much more better). For evening finger load only Arensito and Dvorak seem to do anything about relieving the right index finger. I think Arensito is by far the best for programmers and mathematicians because it provides easy access to AltGr. And contains a shifted layout for full access to special characters and numbers.

Right now I don't know which to choose, an annoying dilemma because I just received my new keyboard (the TypeMatrix EZR-2030 Unlabeled xD thanks  Qwertie). And  I LOVE it (haven't even typed on it yet). But it's truly unlabeled , no logos on the front. It's like the perfect faceless watch. Sorry, Childhood trauma: Once when  I wash young... ... ... I saw a silver handed annalog watch with three symmetrical markings (a line for noon/midnight and a dot for 4 and 8) on a lovely blue solar cell. And that dream of a watch, it would of been perfect if NOT for fat thucking logo smack across it's  otherwise immaculate face of perfect form and function. (Ok, I lied, I never was young, but you know what I mean).

Anyway,  my dilemma is that  I'm itching to test how if facilitates learning a new layout (I think the slight shift of keys will aid reconditioning my motor memory). But I can't decide on a layout. If I was going to learn a new layout for a type bar shifted keyboard I'd go with  Colemack if for  no other reason than  not having to relearn the upper right punctuation ( - = { } and \ ) But since I must relearn the positions anyway I figure I may as well go with an entirely new layout. That and it bugs me that Colemak chose not to include Klausser, Arensito and Capewll on their comparison chart. (See were much better than the other layouts designed a centry ago!)

I'm even considering Dvorak with ',' and '.' transposed ( ' . , instead of ' , . inplace of qwe) because Dvorak does a very good job of figure balancing.

I'm surprised only a paragraph was confusing. I babble a bit much when I try to figure stuff out.

You made a good point about the clarity of the Colemak comparison. But I hope Shai adds a link for more complete comparisons.

Regarding the confusion I created: I think Capewell is still working on his layouts. It is the lessons for the Typing Tutor he designed as an aside that he's not working on.

The main difficulty I've had with finding a layout is finger loading. I don't like to use even my index finger too much (over 16%). Dvorak does the best job in this respect. I also like to avoid using the same finger twice in a row for different keys (such as lo and um on qwerty). Dvorak does better than qwerty here, but not as good as Colemak. And Capewell seams to do the best at avoiding "same finger." Capewell also does well finger balance. But it doesn't use the home row enough.

Colemak does a good job with the home row, but keeping the quote key in place does hurt it a bit (according my own writing Colemak only gives me 67.7 percent on the home row but that includes 1.9% numbers and symbols). Colemak also uses the bottom row a bit much (Capewell is the worst regarding the bottom row with 16%) and places a lot of work on the right index finger (20% with what I write).

So I decided to design my own optimum layout.

Why try to find the best keyboard when tomorrow there will be a better one. Because it's the best I can do now. And to me, that's worth something. The question is comparable to why try to find the best way to educate my child if tomorrow there will be a better way. Just beat the kid with a rod, he'll learn, it's worked for centuries!

Along the way I think I found the best technique for optimizing keyboard with current technology: make it an online game. Each player starts with an identical board. And the players take turns moving keys (like chess pieces) to achieve a certain goal of optimization. After each round the computer evaluates the board. The first player to reach the goal wins. Or if it a draw (both players decide to pass or have moved every key on the board once) the player with the highest stat wins. Then the board is stored in a data base. In addition the game could ask contestants what keyboard layout they use and then record key timings as they speak to establish optimal board positions.

In short I realised human intuition can do a good job choosing keys. And, unlike Dvorak, I had the advantage of seeing what keys go well together basted on computer optimizations (comparing the maps of different optimized boards) and I could test them by running them through a program (the java comparison applet, I'd have preferred to use Kiwi but I couldn't get it to work...Kiwi kept saying it couldn't save a file).

Anyway, I'm happy with what I developed so, I most likely wont type much for a while.

I intend to cover if it with a labeled skin until I teach myself all the shortcuts. But once I know it yes. I don't know about the black model (they make black skins) or the hinge. But I talked to a customer rep who said they're working on adding a customizable on-board map for their next model.

So here what I came up with.

~123 45 67 890{}
KCD MF VW U>J_+
OST NH LR EAIY
"GP BZ X: <Q?

I tried to avoid moving pairs of keys a cross hand to opposit loactions without transposing them (as in <> on Dvorak which is one combo that really messed with me, I think it would have been better if Devorak did ><). But mostly I just went full tilt towards optimization. I'm very happy with the result. I couldn't have done it with out the Arensito, Klausler (two layouts), Capwell (three layouts), Colemak and Dvorak layouts to guide me.

I also used Capewells java tool for analyzing text and red through the list untill I got to Q (you'd be amazed how many combinations of characters come before Q and even before k). That helped me know which keys I'd want to be able to hit easily together.

The results are (in order of Dvorak, Ostn Reai and Colemak):
Total keys: 92093 (same for all, I couldn't work the "the" key into the compairison script ;)
Finger used (left to right):
08  08  14  14  -  -  16  12  12  11
08  10  13  17  -  -  13  15  10  10
08  08  10  16  -  -  20  15  11  09

Same finger (this can verry a lot from content such as Chausser versus Mark Twain:
2.390%
1.399%
1.865%

Home Row, Top Row, Bottom Row:
65.31%   23.48%   9.477% 
70.04%   21.53%   6.697% 
67.68%   17.25%   13.13%

Same Hand:
22.06%
34.21%
32.52%

Wants a version of the Maxwell's comparison applet modified by Capewell that compares Dvorak, Qwerty, Ostn Raei, Colemak, Klausler and three versions of Capewell, let me know and I'll e-mail a zip.

And if you have any suggestions for further improvement please let me know.

Well Qwerty, I'm going to miss you. And I won't be typing much anytime soon. I'm sooo glad there's always pen and paper.

Note: The "Ostn Reai Keymap" is registered as Public Domain with Creative Commons.
<!--Creative Commons License--><a rel="license" href="https://creativecommons.org/licenses/publicdomain/"><img alt="Creative Commons License" style="border-width: 0" src="http://i.creativecommons.org/l/publicdomain/88x31.png"/></a><br/>This work is licensed under a <a rel="license" href="https://creativecommons.org/licenses/publicdomain/">Creative Commons Public Domain License</a>.<!--/Creative Commons License--><!-- <rdf:RDF xmlns="http://web.resource.org/cc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:rdfs="http://www.w3.org/2000/01/rdf-schema#">
    <Work rdf:about="">
        <license rdf:resource="https://creativecommons.org/licenses/publicdomain/" />
    <dc:title>Ostn Raei Keymap</dc:title>
    <dc:rights><Agent><dc:title>Soren Kyale</dc:title></Agent></dc:rights>
    </Work>
    <License rdf:about="https://creativecommons.org/licenses/publicdomain/"><permits rdf:resource="http://web.resource.org/cc/Reproduction"/><permits rdf:resource="http://web.resource.org/cc/Distribution"/><rdfs:subClassOf rdf:resource="http://web.resource.org/cc/PublicDomain"/><permits rdf:resource="http://web.resource.org/cc/DerivativeWorks"/></License></rdf:RDF> -->

~ 1 2 3 4 5   6 7 8 9 0 { } 
  K C D M F   V W U > J _ + 
  O S T N H   L R E A I Y 
  " G P B Z   X : < Q ?

Heh. Fascinating boardie. Wonder what Shai has to comment on that one, apart from the obvious (it'll be hell and hell revisited to learn and doesn't have any shortcut keys conserved)?

- I'd call it "OstnReai" if I were you. A space in there is unsexy (looks like a foreigner's name?), and drawn together it's more chic. Like the Macintosh fad, if you wish. Or OSTNREAI, all-caps style like several others choose.
- The placement of <> must be considering programming? I first thought they were on one finger, but it actually looks nice.
- Where does the "/" go, still considering the coding thing? Could have one on the extra key of the 102-key boards (bottom left)?
- (Put a "</" digraph on AltGr+"<", and people may cheer!)
- Grats on beating Capewell at the "stat game"!
- However: Did you run a sufficiently representative and comprehensive syllabus through it?
- It's interesting to see how the contestants fared. To my eye, they all did "well enough" and the differences were small.
- Seems you managed to underwork your right index finger though! Are you playing those crazy clicking games like Insaniquarium; is that why your mouse finger is so tired?  :)  Actually a point there somewhere - that finger gets a lot of mouse workout.
- You have a "the" key from what I gather. If you only use your layout - ever - that'll be great. But it may be frustrating to do without it when you're roaming!

In sum, looks like a right solid and pretty impressive piece of figgering if you're hell-bent on moving absolutely everything around. I know I'm not. Learning Dvorak took me far too long which was frustrating; I firmly believe that he moved far too many keys.

It's been good to see that learning Colemak is faster because I can capitalize on already knowing so many of the keys from QWERTY. As I said it when thinking about layouts myself: "If QWERTY got it right (enough), don't fix it!" The extra minor benefits you may eventually reap at the other end of a protracted and grueling relearning process (according to my fear) could be worth it to you, but I'm not convinced.

You don't create a layout for one specific text. You first make a big corpus (minimum 8MB) of text from many different sources.
Colemak moves the following: FPGJLUY;:RSTDNEIOK
Your layout moves: []{}KCDFBVWUJX-_=+OTNHLREAIY'"GMPZ;:.,<>
In fact it moves even more keys than Dvorak.

I think that the index finger is not overloaded on Colemak. It is the strongest and the fastest finger, and therefore it can do a lot more work in comparison to other fingers, although you might disagree.

I also don't like the KO/OK digraph on the little finger nor the CS/SC digraph on the ring finger. I also don't think it makes sense to put the ;: on the strongest finger.

Also, it's a bit strange to have the Q stuck there between punctuation symbols.

Overall, it's not a bad keyboard layout, but it's not easy to learn, ZXCV aren't there, and it isn't ergonomically better than Colemak IMO.

It seems I've been uninformed as to the full extent of my favourite god Loki's offspring. Søren, you're hilarious. Keep it up!  :D