First candidate layout of 24 home block letters

[binarybottle]

I just generated a first candidate layout of 24 home block letters for Spanish. For native speakers (alert to @NickG13, @Lobo-Feroz, and @sunaku) please tell me if there are any glaring concerns, or if there is anything awkward or offensive that happens to have arisen in the layout of letters:

    H Y O Z  M L B J
    P I E A  D S N R
    G Q X U  T C V F
    
      7   7  74   3   22  48  10   4
     23  57 109 103   42  61  58  55
     10   7   2  33   36  36   8   7
    
    left: 43336111 (11.72%)  right: 38788861
    Total same-finger bigram frequencies:    106090282428
    Total bigram inward roll frequencies:   6394006706059

[Lobo-Feroz]

Looks awesome! Let's give it a try!

thanks for everything @binarybottle !

[binarybottle]

W above K received the highest score:

Rank Score
1: H Y O Z P I E A G Q X U M L B J D S N R T C V F ' , - " . ? W K 0.6315284924483411
1: J B L M R N S D F V C T Z O Y H A E I P U X Q G ' , - " . ? W K 0.6315284472608467
1: H Y O Z P I E A G Q X U M L B J D S N R T C V F ' , - " . ? K W 0.6315270680131808
1: J B L M R N S D F V C T Z O Y H A E I P U X Q G ' , - " . ? K W 0.631526057420239

So the candidate letters of the Engram-es layout are:

H Y O Z  M L B J W
P I E A  D S N R K
G Q X U  T C V F

[binarybottle]

Thank you very much! Don't get too attached to this initial candidate, As I am crunching through over 40 million more layouts over the course of this week based on this one, and we will find other candidates to compare against.

[NickG13]

This looks awesome!
Just to test it and also practice and learn this layout I've encountered a huge tool, it's called Type Fu.

I will add here two practice files for spanish words, that you can add into the app itself (on custom words section)
Type-Fu
I was using this to test and train AEIOU layout, we could also use it in the long therm for the definitive one, but it's awesome to gain some speed and see how the layout performs after finger travel comfort.
Spanish training.zip we will only need the layout file. I will surely create this in a few hours and upload.
UPDATE: Here is the layout file
Engram-es.zip
It shall look something like this

[Lobo-Feroz]

@NickG13 I've tried your layout and word files and they work nicely. And I have the initial, subjective feeling that this layout will be very easy on the hands.

Also, I propose that we move the typing learning discussion to its own thread.

[ficion]

Would the bigram 'qu' be comfortable to type in this layout?

[binarybottle]

I am waiting for a refined set of bigram frequencies, and will replace this tentative layout with a better one...

[binarybottle]

Please see #7 (comment) for suggested symbols and diacriticals.

[binarybottle]

@NickG13 @Lobo-Feroz @iandoug --

I am running a stability test to see if I can find higher-scoring layouts, but here is the preliminary optimal Spanish layout based on the cleaned-up text corpus:

  { | = ~ +   <  >   ^ & % * } \
  [ 1 2 3 4   5  6   7 8 9 0 ] /

    Z H O X  '(  ")  M D B J W #€$ @`
    P I E A  ,;  .:  T S N R K
    F Y Q U  -_  ?!  C L V G

Next: finalize non-letter characters!

What is the current stance on diacritical trigger key(s)?

[binarybottle]

@iandoug -- What is SFB?

[iandoug]

Same-Finger bigrams

Used a lot on https://www.reddit.com/r/KeyboardLayouts

[binarybottle]

ST is a much more frequent bigram than SD or DS, so it makes sense to avoid ST as a same-finger bigram.

[NickG13]

I'm late to the party! I would consider maybe trying to put C in home row also. I've seen it's very frequent with R and other vocals.
As per UA/AU (not even considering AX or XA) or other combination with near keys with index finger, as I see with ian stadistics it will have a higher SFB with comma and period than E

I have not seen any slang or offensive abbreviation on this layout, neither with PIAE if you finally change it.

[Lobo-Feroz]

Awesome to have a first version!

Do you see anything offensive, slang, abbreviations, or otherwise in this layout?

Nothing with any meaning pops up.

[binarybottle]

I'm not as concerned about same-finger bigrams with an intervening space as I am with the bigram with the preceding letter, as I believe that the cadence is quite different.

…

________________________________ From: Ian Douglas ***@***.***> Sent: Wednesday, August 18, 2021 3:48:08 PM To: binarybottle/engram-es ***@***.***> Cc: Arno Klein ***@***.***>; Mention ***@***.***> Subject: Re: [binarybottle/engram-es] First candidate layout of 24 home block letters (#9) [EXTERNAL EMAIL] DO NOT CLICK links or attachments unless you recognize the sender and know the content is safe. Same-Finger bigrams Used a lot on https://www.reddit.com/r/KeyboardLayouts<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.reddit.com%2Fr%2FKeyboardLayouts&data=04%7C01%7CArno.Klein%40childmind.org%7C68e20a001ca042b305a208d962811aea%7C7e9101c66f8942a180ba1af9e680cc3a%7C1%7C0%7C637649128939771840%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=xuOneEWjnQK8hGQrULpoDJ98%2FboO%2FTrCGH1a16lOjAU%3D&reserved=0> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fbinarybottle%2Fengram-es%2Fdiscussions%2F9%23discussioncomment-1202882&data=04%7C01%7CArno.Klein%40childmind.org%7C68e20a001ca042b305a208d962811aea%7C7e9101c66f8942a180ba1af9e680cc3a%7C1%7C0%7C637649128939771840%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=s2tqy6pQ1jRjBNIgIdzzcf1ipH1%2FMjzjD9A15kVkTaA%3D&reserved=0>, or unsubscribe<https://nam04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAABZX5GZQUX7ZPLSCJI4KSLT5QE7RANCNFSM5BJAFSIA&data=04%7C01%7CArno.Klein%40childmind.org%7C68e20a001ca042b305a208d962811aea%7C7e9101c66f8942a180ba1af9e680cc3a%7C1%7C0%7C637649128939781797%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=hDcQ52o2e%2F3Rfjo7CbPabqfw9%2F58GaFNRJEwt3S4uTU%3D&reserved=0>.

[NickG13]

I would also joint both " frequencies as the same character, since in spanish we use it as the same thing

[iandoug]

I would also joint both " frequencies as the same character, since in spanish we use it as the same thing

They are separate now as requested... :-)

[iandoug]

Thank you, @iandoug -- this is exactly what I needed!

That was the point of my original bigram file.

Also so that you could see which capitals are common, and put them on appropriate hand ... distance to Shift is not the same Left/Right ANSI/ISO ...

cheers, Ian

[Lobo-Feroz]

I think some degree of SFB will be unavoidable.

As per Ian's numbers we see that most words end in o, a, e. This is because theres letters denote masculine, femenine, neutral. "esta, esto, este" (this, this, this). And the most ending is in 's' because it's the plural (estos - these). And 'r' shows up a lot because in spanish many verbs end in 'r'.

So the fact of the matter is that the most frequent letters are also the most frequent word ending letters, so I don't think this is avoidable except by not having popular letters in the home row index finger position (which doesn't make much sense), or moving period and comma elsewhere, which doesn't make much sense either, since center columns home row seem a good placement for the most used punctuation signs.

In an ergo keyboard maybe we could move them to the thumb or something, with spaces, the three of them going after the ending letter of a word. But for ISO/ANSI I don't know which could be the best compromise.

Another option could be including the main punctuation signs in the optimization algorithm, but that would yield a keyboard where the signs could be anywhere, different from the engram signature of having them in the center columns.

[binarybottle]

Thank you all. I'm looking at the punctuation/diacritical frequencies now to guide placement. The frequencies above suggest that the comma should probably go on the right.

[Lobo-Feroz]

Made a test layout for type-fu.com with the newest @binarybottle 's layout, and added the comma-period key and the accute accent key to test it.

Here's the type fu layout:

Engram-ES - V0 - 20200820-03.tfl.zip

I've tested it with @NickG13 's spanish dictionary and so far seems easy on the fingers. The home row is very easy to memorize.

[iandoug]

I will need to modify KLA to handle dead keys ... it has no concept of them. Will also need .json and UI changes to add them.

[Lobo-Feroz]

I will need to modify KLA to handle dead keys ... it has no concept of them. Will also need .json and UI changes to add them.

I think that if you want to have international support, this would be the way to go. Even if all new keyboards used a different approach for diacriticals, you would need to support all the existing qwerty-es, -fr, -de etc. even if only for comparison purposes.

I don't know if I could help you with that. I imagine that must be more of a legacy code base issue.

[NickG13]

I don't know any analyzer that has support for dead keys, so that would be surely a milestone in this category

[iandoug]

Adding support is non-trivial... starting with the UI. :-)

[NickG13]

Yeah, that makes it even more a milestone!