It's great that other layout makers are taking the sound design principles of Colemak for the better. The ZXCVSTNEIO cluster is identical to Colemak, and looks like it would work for the language. Hopefully more layouts like these are made for all kinds of languages (cough cough I want a custom Spanish Colemak cough cough just because Spanish Dvorak has all of the disadvantages of Dvorak in Spanish).

It mentions Colemak in several places: the FAQ, the learning tips (largely just translated from the equivalent Colemak page), and a page where they explicitly state that it's been inspired by Colemak.

For the layout comparison: probably the Java applet can only handle base characters and no AltGr combinations, so comparing with Colemak is impossible for Lithuanian texts?

Look more closely at the Qwerty image.  It's not US Qwerty, but Lithuanian Qwerty (it has additional letters on the number keys).  They seem to use both an Azerty and a Qwerty derived layout there?

Yeah, that's silly.

Hi Tautrimas!

Ideally, I suppose that rather than insisting on keeping so-and-so key in place (in the case of A and Q) you should assess a 'value' of keeping them there for the purposes of your optimisation. A for instance, is nice to keep in place because it's a home row (common) key and a quite common shortcut - but maybe it could be even more centrally placed too. I'd give it an intermediate 'moving penalty value'. Q is a rare key that's already in a hard-to-reach position and also a common shortcut key; I'd give it a much higher penalty for moving. I don't have any such algorithms but as you'll find on these forums there are several other users who are developing just that kind of thing.

Since Colemak is becoming more and more common, I would personally see the point of keeping a positive value on Colemak likeness in your project too, but maybe that doesn't really matter to you personally.

I see that 9 diacritics is a lot to fit in. As a Norwegian, I only had to cope with the 3 special letters 'æøå'. My solution which I would recommend for any country with that few special signs is:

- The most common of them goes to the VK_102 key (which I have in the middle bottom row, but normally it sits at the bottom left of 102-key boards); in my case that's the 'ø' (I've also put the anglo-french 'œ' there for the English layout).

- The other two go to the [] keys. If you want a more coding-friendly layout, they'll be on AltGr plus []; otherwise, the '[]' and '{}' signs are on AltGr mappings. You can even have those two layouts side by side and switch between them for text or coding, keeping everything else the same so it's easy to learn and remember.

- If you need even more, then the ' \ / keys are given the same treatment. I'd be careful with putting the apostrophe sign on its' AltGr mapping if you're writing any amount of English though, because it's really common. The ` key can also serve for a semi-common character.

- Taking -= for two more signs is certainly an option, especially if you have a solution for NumPad input from the home position. Otherwise, like the apostrophe there's a real danger of getting annoyed with an AltGr-based hyphen far too much. You could of course mix it up, putting a rare diacritic on a dead key or on AltGr mappings of the hyphen and/or apostrophe keys for instance - while using the other sign keys the other way around.

This way, you have from 3 to 7 diacritics to play with and no need to change the main letter block: You've made a national Colemak (or whatever it is you're using for your main block) variant instead of a new layout!

I've no interest in optimizing the main letter block for my language: It turns out that Colemak does a surprisingly good job of most non-English European languages! (Did you see our topic on that from a while back?) And I write so much English anyway that making another Norwegian-only layout to wringe out a few percent more efficiency would be just silly for me... and for nearly all people nowadays I feel. YMMV of course, and maybe Slavic languages for instance have radical enough differences in consonant use to make Colemak really bad - what do I know. For Norwegian it's completely trivial problems such as the 'r' being a little more common in Norwegian and some digraphs being a bit awkward for their frequency.

Your AltGr-based NumPad is nice, by the way. Almost makes me want one for myself, but I'll use another method still.

If I may suggest something: Maybe you could put the ž on the VK_102 next to the Z to make it easy to remember? Unless you introduce a ZXCVB shift (like I heartily recommend) so the VK_102 key is in the middle; then it's a more open question where to put it again.

I'd consider whether any of the diacriticals are very rare (so, e.g., the Ž could sit on a dead key mapping or on the AltGr mapping of the Z key after all), but they may be too common to be on AltGr keys by default I guess. Our Norwegian signs are in the ~1% frequency range, so for me it's perfectly acceptable to have one of them on non-AltGr mapping and the two slightly more rare ones on AltGr mappings. As I said, I write a lot of English anyway too.

In my opinion, even if the 3 4 8 9 keys are a bit nice they're not quite super and are probably best kept they way they were. I didn't care much for standards before but I've started to love them more and more because they make things simpler, easier to remember and less likely to fail (if, say, a somewhat simple-minded US programmer supposes you have a hyphen on that key and you have it on AltGr+some-other-key, you may be in trouble or at least inconvenience using his program; yes, it's his fault but it's still your headache!).

I see the point of the W, H and F keys versus all those other more needed ones. Especially when the Colemak F position is one of the best on the board next to the home row. One possibility would be to make one anglocentric layout and one lietuvacentric one where the F and for instance Š (since it'd then be right above the S) letters switch between being on AltGr and normal mappings? Similarly W and for instance Ą or Č? Then again, you'd make it simple to remember where the keys are and if you type more English or Lithuanian at different points you could easily do the 'microswitch' and get the needed signs more accessible but essentially still in the same positions.

The H key should hold a consonant that minimizes same-finger digraphs with the right index finger. Same-finger digraphs are evil.

What I'd do with the NumPad would be to faithfully reproduce it under the right hand so that anyone used to a real NumPad could just jump right in and use that. But I'd use a special Fn/Extend key (which requires running a program like PKL in the background) instead of the usual AltGr mappings - then again I keep the numbers in their top-row position too and I wouldn't want to rely on a program to reach them all the time because that'd create problems in some situations for sure. I'd also make the +-/* keys like on the NumPad so I'd use the Colemak KM<>NEILUY keys for the numbers 0123456789, respectively. Then the NumPad Enter and + keys go on Colemak O and ; keys (the Enter key is of course only remappable by program, not layout).

If one option is to make a layout that makes it possible to train your Lithuanian typing speed a little higher but makes everything else really inconvenient, you should consider what flexibility and ease of use and remembering may also be worth in the large picture. I for instance, have probably saved more time by putting the arrow keys on my home row (with an Extend key) than on optimal positions four other letter keys - even if Colemak has raised my typing speed a little. It's not very common to be typing miles and miles of pure text anymore, but more often you'll find yourself working in a complex way with constant switches between typing/editing/navigation (which makes editors like Vim and Emacs very cool, hehe).