Add Persian (Farsi) notification support

MMKH said:

This alphabet difference is very short in size... I think, around 20 Bytes at all! and this would be able to take place easily on Instinct's firmware...

You probably do not know how characters are being parsed, so that they can be displayed on a screen like the display of your watch. You need either bitmap or vector images of each single character, and that usually in numerous forms (normal, bold, italic,...) and sizes (at least at the bitmap fonts) - they are commonly called fonts, in this case Unicode fonts. And as I already explained, although the differences of Farsi and Arabic alphabets are minor, they have each own character codes. That also means that standard libraries and fonts usually use independent bitmap (or vector) definitions of both alphabets separately. The size difference would definitely not be 20 bytes, but rather in the order of Megabytes. Perhaps there is place enough for that, but perhaps not.

Um... Yes you're right about bitmaps & vectors... But thay dont have independent unicode characters... 

Take a look at Persian_alphabet (Wikipedia) and Arabic_alphabet (Wikipedia). They both have almost same unicodes...

MMKH said:

Take a look at Persian_alphabet (Wikipedia) and Arabic_alphabet (Wikipedia). They both have almost same unicodes...

OK, I see. But still even a single letter would require much more than 20 bytes. Though, if there is just a few characters differing between Arabic and Farsi, and the rest uses identical Unicodes, it could be indeed perhaps doable without big hindrance.

How many characters are missing on the watch exactly? Did you test it by sending yourself a message containing the entire Farsi alphabet, so that you know which and how many characters are not being displayed correctly?

Yes i tested it...

4 Characters(Which is not available in arabic alphabet) + 2 Characters(Which has an equivalent in arabic) + 10 Digits(3 of them have different bitmaps&codes, 7 others just have different codes but exact the same shape & bitmap) = 16 at all. (Which requires 9 new codes with their bitmaps & 7 with codes only)

trux said:

Did you test it by sending yourself a message containing the entire Farsi alphabet, so that you know which and how many characters are not being displayed correctly

OK, I finally did it myself, and it looks like only 4 characters are missing: پ  چ  ژ  گ  (U+06AF, U+0698, U+0686, U+067E) 

As far as I know, Farsi/Arabic do not have lowercase/uppercase letters, but there may be perhaps some ligatures missing (not sure whether they are used in Farsi at all, in contrary to Arabic). However, you wrote 8 charactes are missing. Are you sure about that? Which are the other 4?

I believe that should be indeed relatively easily doable to add them into the firmware (unlike what I claimed previously). When contacting Garmin, make sure you provide them the exact information which characters are missing, including their true froms and the Unicodes (both shown above), so that they do not need to make the research themselves, like I did. You can also send them the attached screenshots from my Instinct displaying a message with the complete Farsi alphabet:

OK, I did not see your reply, before doing the test myself. So 7 of the Farsi numeric characters have different unicodes despite being identical in shape with the Arabic forms? The question is whether Garmin can easily implement a transcription like that. Replacing the original font with a wider definition containg some more characters may be easy (if it is available and does not need to be tediously encoded, pixel by pixel), but adding a transcription algorithm, if the software does not handle such cases yet, may complicate the task, so they may not be willing doing such change.

I tried that in the past, It did't work. Instinct shows squares with question mark instead of some characters...

trux said:

So 7 of the Farsi numeric characters have different unicodes despite being identical in shape with the Arabic forms? The question is whether Garmin can easily implement a transcription like that.

Yes thats right. Although, They (digits) are not necessary to be added... Because even in Farsi, Arabic & English digits are being widely used... 

trux said:

OK, I finally did it myself, and it looks like only 4 characters are missing: پ  چ  ژ  گ 

In addition to those 4 characters, There are two others which are ی (U+06CC) which is ي (U+064A) in arabic, And ک (U+06A9) which is ك (U+0643) in arabic (Which arabic equivalents are currently supported). In addition to that, these two character's shapes (which is very similar but not equal) really doesn't matter in Farsi (and like what i said about Farsi digits, It's code can be transcripted to Arabic equivalent and then being displayed).

Thanks for your GREAT help trux. So, how to contact Garmin now? Do they read these posts?