Studying languages to test my multilingual keyboard layout

Continue or start your personal language log here, including logs for challenge participants
User avatar
Deinonysus
Brown Belt
Posts: 1216
Joined: Tue Sep 13, 2016 6:06 pm
Location: MA, USA
Languages:  
• Native: English
• Advanced: French
• Intermediate: German,
   Spanish, Hebrew
• Beginner: Italian,
   Arabic
x 4620

Re: Studying languages to test my multilingual keyboard layout

Postby Deinonysus » Sun Nov 21, 2021 6:22 am

It's been a while since I posted on this log! I have been using my "balanced" layout for the year and a half or so since I mentioned it in my last post. However, I have recently made a lot of changes to the AltGr keys, and I recently had a big breakthrough the other day. SpanishInput mentioned something about the kana layer in his thread about his own layout, and that gave me the idea that rather than having to press the AltGr or Ctrl+Alt keys every time you want to use a deadkey, what I could do instead was to turn the right Ctrl button into a "kana" toggle to turn the numbers and minor punctuation into deadkeys! That saves a keystroke every time one of the deadkeys is used, and it also adds a lot of extra room for AltGr keys, so I was able to make assorted European quotation marks easier to type. But the most important change is:

I'm able to support Vietnamese with this layout now! With the deadkey toggle on, it takes a maximum of three keystrokes to type any vowel with its tone mark (as opposed to two keystrokes for every vowel with tone mark on the standard Vietnamese layout). However, I think the vowel keys on my layout are better placed and allows typing with fewer trips to the top row, so I think that ultimately it isn't a worse typing experience than the standard Vietnamese layout, and may actually be a little bit better.

Here is a map of the layout, with the main layer in red on the left, the AltGr layer in blue in the middle, and the kana (deadkey) layer on the right (purple). All deadkeys have a blue background.
2021-11-20 Layout.png

Supported languages

I made a pretty major change based on the 2021 Ethnologue data on languages by total number of speakers that was posted to Wikipedia. I now have enough AltGr keys to cover quick typing (2 keystrokes per lowercase letter excluding tone marks, and at most 3 keystrokes per lower case letter with tone marks for Vietnamese and Yoruba) for the 15 most spoken languages in the world that commonly use the Latin script; these languages are all in the top 40 languages with the most speakers. The languages are:

    1. English (1.348 billion speakers)
    2. Spanish (543 million)
    3. French (267 million)
    4. Portuguese (258 million)
    5. Indonesian (excl. Malay) (199 million)
    6. Standard German (135 million)
    7. Turkish (88 million)
    8. Vietnamese ( 77 million)
    9. Hausa (75 million)
    10. Swahili (69 million)
    11. Javanese (68 million)
    12. Italian (68 million)
    13. Nigerian Pidgin (48 million)
    14. Filipino (excl. Tagalog) (45 million)
    15. Yoruba (43 million)
Note that the keyboard is only machine-optimized on the main layer for English, Spanish, French, and German. In 2022 I am hoping to harvest my own letter frequency data for each of these languages and use that data to optimize the layout for all of them, using a genetic algorithm. That should blow my current layout (which uses a much less efficient brute force algorithm) out of the water and be much more competitive with other popular layouts.

In terms of weighting, I will plan on half of the optimization being dedicated to English (a quarter of the total weight) and Spanish and French (another combined quarter), and the other half being dedicated to all of the other languages. In other words, English will count six times, Spanish and French will each count three times, and the other twelve languages will each count once. The extra weight towards English, Spanish and French is because not only is English the international language, but these make up half of the official UN languages, and over half of all countries have at least one of these three languages is an official language. But the lower-weighted languages will have veto power, because the worst score for any individual language will be counted against a candidate layout's total score, meaning that the final layout shouldn't be terrible for any of these 15 languages.

Other minor improvements

Some other changes I've made are:
  • I swapped ? with / so that ? could be written with one keystroke, as it is used much more often in writing text.
  • Assorted European quotation marks can now be typed with only two keystrokes each; they were previously on a deadkey due to space issues and thus required at least 3-4 keystrokes each.
  • ¿ ¡ º and ª were previously on the ISO key based on the mistaken belief that most of the world uses the ISO layout (which has an extra key). However, it turns out that the ISO format is mainly limited to Europe, with the Americas using the ANSI layout. So I can't put vital Spanish functionality on the ISO key! I moved those symbols to easy-to-reach keys on the right hand, and I dedicated the ISO key to symbols that are important for German, since most German speakers use ISO layout keyboards.

Language Learning

I am finishing up with Eastern Armenian for my Every script in the world challenge (which I suppose is the opposite of this project which is dedicated just to the Latin script), and now I need to decide where to go next. I have two hobbies that have been competing with language learning lately: chess and competitive typing. But, my recent work on my layout has given me more energy for language learning. It's tempting to abandon Arabic and go straight into one of these languages, but many of them have some significant Arabic influence, so the more Arabic vocabulary I learn, the easier it will be for me to study many of these languages. So maybe I'll keep going with Arabic for a few more months and then start in on Turkish or something.

PS

Also, whoops! It looks like Min Nan (including Hokkien, including Taiwanese) often uses the Latin script too! It looks like the only character it uses that isn't already represented on my layout is a superscript ⁿ, so I swapped the deadkeys around so that it can be produced with just two keystrokes using the deadkey layer. I'll have to look into it more but it's after 1 am so time to get some sleep!
You do not have the required permissions to view the files attached to this post.
3 x
/daɪ.nə.ˈnaɪ.səs/

User avatar
SpanishInput
Yellow Belt
Posts: 97
Joined: Sun Sep 26, 2021 3:11 pm
Location: Ecuador
Languages: Spanish (N), English (C2), Mandarin (HSK 5)
x 469

Re: Studying languages to test my multilingual keyboard layout

Postby SpanishInput » Sun Nov 21, 2021 11:04 am

Deinonysus wrote:Here is a map of the layout

Hi, Deinonysus! Looks like you've switched from MSKLC to KbdEdit. Please make sure to make backups. KbdEdit is cool, but it happened to me once that my file became corrupted, and I had to start from scratch or from a previous backup (can't remember).

Deinonysus wrote:I swapped ? with / so that ? could be written with one keystroke, as it is used much more often in writing text.

I really wonder what they were thinking when they designed it this way. But the Brazilian layout is even worse: ? can only be typed with AltGr.

BTW, besides the right Ctrl key, another key that IMHO could be used for shift states is Caps Lock. This way you don't need to press Ctrl + Alt when you want to type special characters on the right side.

Deinonysus wrote:¿ ¡ º and ª were previously on the ISO key based on the mistaken belief that most of the world uses the ISO layout (which has an extra key). However, it turns out that the ISO format is mainly limited to Europe, with the Americas using the ANSI layout.

China, Taiwan, Korea all use ANSI keyboards. Here in Ecuador we use mainly the European Spanish layout (ISO) and the Latam Spanish layout (also ISO), but you can find computers being sold with American ANSI keyboards and also with non-standard keyboards, including those with the horrible "J shaped" Return key and the tiny Backspace key.

BTW, in case you're considering a future Mac port: On the Mac side, the grave key moves to the extra ISO key when using an Apple keyboard, and the space where the grave key was becomes an extra key with other characters. This does not happen if you only have a Windows ISO keyboard attached to your Mac. :roll: On the good side, macOS allows you to have more than one code point in the end result of a dead key chain, something that Windows does not support. This means, for example, you could produce g̃ with the tilde dead key. Unfortunately, I haven't found a macOS keyboard design software with the same level of deep editing that KbdEdit allows, so I haven't found a way to use the Kana modifier on the Mac side. This would prevent conflicts with Google Docs. Google seems to be oblivious to the existence of non-US layouts: Gdocs and Gsheets use several Ctrl + Alt shortcuts on Windows, and several Option + Shift shortcuts on macOS. IMHO it's a good idea to test drive your layout with Gdocs, as it's so popular with language learners during online lessons.
2 x

User avatar
Deinonysus
Brown Belt
Posts: 1216
Joined: Tue Sep 13, 2016 6:06 pm
Location: MA, USA
Languages:  
• Native: English
• Advanced: French
• Intermediate: German,
   Spanish, Hebrew
• Beginner: Italian,
   Arabic
x 4620

Re: Studying languages to test my multilingual keyboard layout

Postby Deinonysus » Tue Nov 30, 2021 12:00 pm

SpanishInput wrote:
Deinonysus wrote:Here is a map of the layout

Hi, Deinonysus! Looks like you've switched from MSKLC to KbdEdit. Please make sure to make backups. KbdEdit is cool, but it happened to me once that my file became corrupted, and I had to start from scratch or from a previous backup (can't remember).

Deinonysus wrote:I swapped ? with / so that ? could be written with one keystroke, as it is used much more often in writing text.

I really wonder what they were thinking when they designed it this way. But the Brazilian layout is even worse: ? can only be typed with AltGr.

BTW, besides the right Ctrl key, another key that IMHO could be used for shift states is Caps Lock. This way you don't need to press Ctrl + Alt when you want to type special characters on the right side.

Deinonysus wrote:¿ ¡ º and ª were previously on the ISO key based on the mistaken belief that most of the world uses the ISO layout (which has an extra key). However, it turns out that the ISO format is mainly limited to Europe, with the Americas using the ANSI layout.

China, Taiwan, Korea all use ANSI keyboards. Here in Ecuador we use mainly the European Spanish layout (ISO) and the Latam Spanish layout (also ISO), but you can find computers being sold with American ANSI keyboards and also with non-standard keyboards, including those with the horrible "J shaped" Return key and the tiny Backspace key.

BTW, in case you're considering a future Mac port: On the Mac side, the grave key moves to the extra ISO key when using an Apple keyboard, and the space where the grave key was becomes an extra key with other characters. This does not happen if you only have a Windows ISO keyboard attached to your Mac. :roll: On the good side, macOS allows you to have more than one code point in the end result of a dead key chain, something that Windows does not support. This means, for example, you could produce g̃ with the tilde dead key. Unfortunately, I haven't found a macOS keyboard design software with the same level of deep editing that KbdEdit allows, so I haven't found a way to use the Kana modifier on the Mac side. This would prevent conflicts with Google Docs. Google seems to be oblivious to the existence of non-US layouts: Gdocs and Gsheets use several Ctrl + Alt shortcuts on Windows, and several Option + Shift shortcuts on macOS. IMHO it's a good idea to test drive your layout with Gdocs, as it's so popular with language learners during online lessons.

Thank you, that is some great feedback!

You have a good eye for sticker maps! I wasn't aware of that issue with KbdEdit but I'll make sure I stay vigilant about backups.

I might be in the minority here but I actually really like Caps Lock. I used Colemak without it for a while and I missed it a lot. It's very annoying to type more than two consecutive capital letters without it. So I have kept it in all of my custom layouts.

I didn't realize that ISO keyboards are common in Latin America, that's good to know. But I think it's still a good idea to keep the Spanish punctuation and ordinal symbols off of the ISO key considering how important Spanish is in the US where ISO keyboards are hard to find.

I wish deadkeys worked like that on Windows! My current solution on my layout is that you can create a combining diacritic marker by pressing the corresponding deadkey and then spacebar. But this is obnoxious because you need to know in advance which precomposed characters are available in Unicode. For instance, you press 6 then e to get ẽ (with the deadkey layer toggled on, of course). But to type g̃, which is not available in Unicode, you need to type g, 6, and then spacebar.

The issue with Google Docs is obnoxious but I still think it's best to use the Alt-Gr layer for its intended purpose. Notes can be written in another program such as a text editor and then pasted into Google Docs, until Google Docs makes changes to become compatible with international keyboard layouts. Many programs have issues with international compatibility. For example, MS word has a random problem typing Icelandic. For whatever reason, it has a weird bug where the lowercase letter þ cannot be directly typed; you need to have to have to have it on your clipboard and paste it in whenever you need it.
0 x
/daɪ.nə.ˈnaɪ.səs/


Return to “Language logs”

Who is online

Users browsing this forum: MapleLeaf and 2 guests