Missing Kharoshthi conjuncts

[dscorbett]

Font

NotoSansKharoshthi-Regular.ttf

Where the font came from, and when

Site: https://noto-website-2.storage.googleapis.com/pkgs/NotoSansKharoshthi-unhinted.zip
Date: 2018-04-09

Font version

Version 1.03 uh

Issue

Noto Sans Kharoshthi includes only those conjuncts mentioned in the Unicode proposal, but those were just examples, not the entire conjunct repertoire. https://github.com/googlei18n/noto-fonts/issues/420#issuecomment-349215405 has more information. Also, the subscript form of a coda consonant, which appears in this font in dhik, should not be formed when followed by another consonant, as in dhikṣa.

Character data

𐨐𐨿𐨩𐨢𐨁𐨐𐨿𐨮
U+10A10 KHAROSHTHI LETTER KA
U+10A3F KHAROSHTHI VIRAMA
U+10A29 KHAROSHTHI LETTER YA
U+10A22 KHAROSHTHI LETTER DHA
U+10A01 KHAROSHTHI VOWEL SIGN I
U+10A10 KHAROSHTHI LETTER KA
U+10A3F KHAROSHTHI VIRAMA
U+10A2E KHAROSHTHI LETTER SSA

Screenshot

[baums]

I would be interested in using Noto for Kharoshthi, but this and related issues make it impossible. Please let me know if I can help resolving them by providing expert advice (with somebody else doing the technical part). I am one of the authors of the original Kharoshthi Unicode proposal and work with Kharoshthi material on a daily basis in my job.

[SKing-2003]

This issue appears to be resolved.

[baums]

This issue appears to be resolved.

The image looks promising, especially the 𐨐𐨿𐨮, but there are many character combinations that need to be catered to. I downloaded the most recent version of the font from https://fonts.google.com/noto/specimen/Noto+Sans+Kharoshthi and, at a quick glance, at least two things are still not supported: - preconsonantal r, as in LETTER RA + VIRAMA + LETTER JA (𐨪𐨿𐨗) and - consonant with virāma and no following consonant, as in LETTER KA + LETTER KA + VIRAMA (𐨐𐨐𐨿). (screenshot attached). Please reopen the bug.

[baums]

Here the screenshot:

[simoncozens]

@baums Thank you for offering your help with this!

We would need to get a preconsonantal ra glyph drawn and added to the font in order to get that working. But what should it look like? Andrew Glass's paper mentions five different forms for this character; which would be most appropriate for the style of this font?

To support final virama, would we need to draw subscript forms of all consonants? Or are there a limited number of virama forms attested? (I notice we have dhik encoded as a ligature already.)

[baums]

I talked with Andrew, and he provided the chart below. This illustrates all the combinations of simple base consonants with vowels, and then all the consonant conjuncts attested in our complete corpus on Gandhari.org (using the vowel a, i.e., no vowel sign, for illustration purposes, even though many of these are attested with other vowel signs, and a font should support all vowels signs attached to all consonants and consonant conjuncts). The dark grey cell backgrounds are only for orientation. The chart uses the Segoe UI Historic font.

Yes, it would be good if all base consonants at least (and ideally all consonant conjuncts as well) were rendered in subscript form when followed by a virama but with no other consonant after the virama. Even though only KA + VIRAMA is attested at word end in the corpus, the other combinations will occur as intermediate stages in Kharoshthi input on computers.

I hope this helps. Having a complete implementation of Kharoshthi in the Noto font family would be exciting!

[baums]

PS. The upcoming Emacs 29 will contain a Kharoshthi shaper and input method that I wrote. The latter is, to my knowledge, the only existing complete input method for Kharoshthi. You may find it useful for testing purposes.