promyloph / lulua Goto Github PK

Have you tried implementing this layout on 40% keyboard? I find this project to be quiet interesting

Get access to the original standard and verify the layout is correct. I’ve seen different layouts in different publications.

This seems to be a photocopy of the layout, which matches our current layout:

Figure out a good, readable colorscheme for the website and layout SVG’s.

After all what we and the algorithms think could be a good layout may be total nonsense in people’s daily life.

Ideally we’d want to have a number of people learning the new layout while monitoring their progress in terms of hard numbers (typing speed, accuracy, …) and soft indicators (ease of use, comfort, …). A good setup would also include a control group learning the “standard” layout with the same typing tutor under the same conditions. Evaluating the impact of the compose-based Alef-Hamza approach would be very important, since people will feel weird about it at first.

The main challenge here is finding enough participants willing to screw up their typing habits for us (by learning a new layout) without providing any kind incentive. (There is no budget to pay them.)

Windows seems to think lulua uses a different physical key layout:

Arabic (102) displays correctly, so there must be an undocumented(?) switch somewhere:

Document how to use this software, how to create new layouts, that Martin Krzywinski’ carpalx is used.

• Why compose-based?
  • pro: consider hamza optional diacritic, backed by Buckwalter https://doi.org/10.1007/978-1-4020-6046-5_3 (section 3.3)
  • con: could reduce weight of plain alif → more even distribution
  • con: more keys pressed, more combinations
• Why no العامية/dialect?
  • Lack of corpora, but we could scrape social media in the future
• Why European keyboard, not ANSI?
  • Arabic Mac keyboard is uses 102 button layout
  • Better support for 3rd layer (right shift key)
  • But both variants exist for traditional Arabic layout

Making a new keyboard layout is such an ambitious goal, but it's not easy especially for an Arabic Layout. And I respect that!

The hamza mechanic is quite unique and I like it. Though, it may cause issues for the uniformity of the word.
For example, the word "أحمد" (Lulua) is not identical to "أحمد".
There is a difference in the Unicode encoding which may cause huge problems in search, registration forms, etc...

Figure out a way to cross-compile the Windows driver, so we don’t need a Windows VM and the build flow can be fully automated.

Since our audience is the Arabic-speaking/-writing population this would be quite nice.

i genuinely love this layout and consider to use it for daily use. hence, i try to learn touch typing with it in past month.
then i've encounter problem with أ إ ىٔ ىٕ all the ٔ and ٕ addition doesn't showing in the klavaro practice programs.
furthermore i try online typing lesson such monkeytype.com or tybaa.com but the problem keep occurring. every hurf with ٔ addition keep identified as wrong words,
is there any solution for this?

Right now the corpora used are heavily leaning towards MSA (i.e. the news) without diacritics. It would be nice to include literary and scientific Arabic, as well as kid’s books (which usually contain diacritics).

These were automatically generated by lulua-optimize, but are impossible to remember right now.

• Windows
• Android
• macOS: keylayout files have no support for custom modifier keys. State machine is only useful for deadkeys.
• iOS: Someone with access to Apple’s developer program needs to look into that.

Might be possible to use

• Keyman for all platforms (no idea how to implement custom shift states)
• AHK on Windows
• Native driver on Windows, see also Neo2 driver (done)
• ASK on Android (done)
• Ukelele or keylayout XML files on OSX

Support klavaro typing tutor. Needs layout export and a few example texts.