Hi Eli,
I have attached a new patch with this email containing 5 input methods.
3 of them for Urdu and 1 each for Pashto and Sindhi. Two of Urdu input
methods correspond to the two keyboards at https://www.branah.com/urdu
and the Pashto and Sindhi input methods follow keyboard layouts from the
same page.

The remaining input method roughly corresponds to the one I sent
before. The changes are:
1) Translations are customizable.

2) Numerals can be Arabic or Roman depending on a customizable variable.
This is because both are routinely used in Urdu text.

3) There are a bunch of prefixes which are customizable. Some of these
prefixes are for inputting characters from Balochi, Brahui, Hindko,
Pashto, Saraiki and Sindhi. After including Punjabi which can be
inputted using the same characters as Urdu (along with maybe one
character from Saraiki) this covers all characters from all languages
spoken by more than 1% of people in Pakistan according to Wikipedia
https://en.wikipedia.org/wiki/Languages_of_Pakistan .

In (not too distant) future I would also like to add keyboard layouts
for Balochi and Saraiki. I cannot speak these languages but I can find
some keyboard layouts on internet, e.g.
https://keymanweb.com/#skr-arab,Keyboard_saraiki
https://keymanweb.com/#bal,Keyboard_balochi_phonetic

Are there any copyright issues to be aware of in using the layouts
from web?

Another question: For the customize types, I have opted for character as
the type of a translation. I think it is the most convenient choice for
someone wanting to customize. However it is not most general type quail
allows. Is there any opinion on what the type should be?

Lastly customizing these variables has the potential to shadow some
translations. For this I would like to add a check to the setter
functions that checks for that and raises a warning if such a shadowing
occurs. Is there a standard customize way of raising such a warning?

Please let me know of any changes to be made.

Thanks,
Rahguzar