Takri conjuncts missing, "post-form ha" not attested

[nik9hil]

Font

NotoSansTakri-Regular.otf

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/tree/master/phaseIII_only/unhinted/otf/NotoSansTakriNotoSansTakri-Regular.otf

Date: 2020-10-09

Font Version

Version 2.001

Issue

A particular occlusion arises for TAKRI LETTER NA+TAKRI SIGN VIRAMA+TAKRI LETTER HA+TAKRI VOWEL SIGN II, where the vowel sign & HA overlap each other. To correct it, bring some gap between the VOWEL SIGN II & TAKRI LETTER HA (which is half over here). An example of a word using this is on the second line of the image, meaning 'lowly'.

Character data

Image 1
U+1169D TAKRI LETTER NA 𑚝
U+116B6 TAKRI SIGN VIRAMA 𑚶
U+116A9 TAKRI LETTER HA 𑚩
U+116AF TAKRI VOWEL SIGN II 𑚯

Screenshot

[nizarsq]

Seems to me it is a harfbuzz issue.
Chrome:

Safari:

[nik9hil]

First of all, the VOWEL SIGN II should ligate with the main consonant.

Now the main thing is, I don't know who are responsible for making this font, (please don't feel offended by this I don't want to sound rude) but I have gone through good number of Takri manuscripts, & I have to say that the 'half HA' I have seen is only with the TAKRI LETTER MA, never with any other consonant. I know I raised this issue yesterday, but I would request you wait for some time because the half HA is simply just wrong with other consonant according to me, as it just occurs with one single consonant. I am pretty sure that the team has strictly referred the Unicode Paper & no real & actual official or proper manuscripts, because there are few more issues (largely with ligating) in the current version of the font.

I will be happy sharing my opinion on further correcting the font of my ancestor's script; but I have been a bit busy lately, so I will take time to properly formulate my proposal.

[nik9hil]

After going through samples, here is what I could find:

TAKRI LETTER MA+TAKRI LETTER E+TAKRI LETTER HA+ "TAKRI LETTER LA+ TAKRI SIGN VIRAMA+ TAKRI LETTER HA+ TAKRI VOWEL SIGN E"

TAKRI LETTER TA+TAKRI LETTER I+"TAKRI LETTER NA+TAKRI SIGN VIRAMA+TAKRI LETTER HA+TAKRI VOWEL SIGN AA"

TAKRI LETTER BA+TAKRI SIGN VIRAMA+TAKRI LETTER RA+ "TAKRI LETTER MA+ TAKRI SIGN VIRAMA+ TAKRI LETTER HA+ TAKRI VOWEL SIGN AA"

The Unicode pdf cites only 1 source i.e. for TAKRI LETTER MA+ TAKRI SIGN VIRAMA+ TAKRI LETTER HA but, makes a glyph which is very different from the actual one. The current font builds upon that & makes a more wrong glyph.

Unicode PDF (observe the difference between above & this glyph):

Noto Font (observe the difference between above & this glyph):

I hope this helps understand what I want to point out.

I know I should be creating another issue to state this, but you can also observe that consonant+virama+Takri letter ra/ya ideally ligates with the consonant. The current version builds upon the Unicode pdf reference which in turn considered handwritten documents only. This script was prevalent in early times, so to know the exact & proper way of writing, manuscripts should be referred.

[simoncozens]

Thanks for this. I can see that this "post form ha" is based on a misunderstanding. We should be using precomposed conjuncts for consonant+virama+HA and consonant+virama+RA/YA. That's going to take a lot of drawing, and we will need either commission people to do that or solicit development from the community on an open source basis.