Closed simoncozens closed 1 year ago
Note that subscript –ha in modern Punjabi is not performing the same role as in other Indic scripts: even though it is being triggered by a virama character it is not forming a conjunct; rather, it is a kind of tone marker.
Thinking this through, recently, I am inclined to think that Jelle may be right that this Noto behaviour is a better rendering for what will be, after all, a very rare situation.
Maybe let's not change it then unless there's significant user feedback.
Regarding shared image, I'd say that it looks weird. It must create a half ha underneath la but if someone adds halant after that, it must retain the first half-ha and must not create any half form after that for same consonant.
Just ignore the space between letters. This would be ideal IMO. Your thoughts?
Have we confirmed that this word, or ones involving this kind of sequence, actually exist? I have found ਓਲ੍ਹ੍ਹਗੀਆ — i.e. without the second halanta — in the Sikh scriptures and elsewhere, but not spelled as in this issue.
Practically, as far as I know, you will never encounter this kind of word in Punjabi. In modern Punjabi, even the experts are trying to reduce the usage of half-forms.
I searched for ਓਲ੍ਹ੍ਹਗੀਆ in Google and found a result in GuruGranthDarpan.net for this word. But I noticed something weird there.
They are using half-ha but the shape of the glyph rendered on the screen by the font suggests that they want to use Udaat (U+0A51) there.
So technically, the right spelling would be ਓਲੑਗੀਆ, not ਓਲ੍ਹਗੀਆ. Here, usage of udaat is to indicate high-tone (uccha swar - ਉੱਚਾ ਸੁਰ) of the letter.
One result I found in here which can solve our problem. Seems like font developer used a hack to make typing of Udaat easier as there is no direct key on Windows to type udaat and yaksh. On Mac however, I discovered today that there's a key for Udaat/Yaksha but not is not widely popular among most of the Punjabi users.
Summarising, there is no word in Punjabi or Sikh scriptures which uses double half-ha as of now.
Yes, it looks like some fonts are using a glyph level hack to display halanta+ha+halanta+ha as udaat. Oof! That’s a really bad idea.
Have also found live examples of the similar hack to represent the yakash sign using halanta+ya+halanta+ya in ਮਿਥ੍ਯ੍ਯੰਤ (ਮਿਥੵੰਤ). These include https://www.gurugranthdarpan.net/1354.html but also other sites, indicating that these hacks may have proliferated among users dependent on the hack fonts.
As reported by John in #1, the word ਓਲ੍ਹ੍ਹਗੀ forms a conjunct "la ha-ha" whereas other fonts handle it as "la-ha ha". Is this OK? Do we want to change it?