On Mon, May 30, 2022 at 9:45 AM Jan Kučera <jan.kucera@matfyz.cz> wrote:

Dear all,

I am doing a review of Indic scripts rendering mechanisms, and I would be interested if anyone had issues that they had run into during their work. Examples I am aware of include Vedic signs not working on anusvara, nuktas not working on vowels and similar.

There are these issues with the characters in both the unicode devanagari extensions block and the unicode vedic extensions block and Microsoft Word. If you apply a vedic accent to a character in the devanagari extensions block (which is necessary for Taittiriya texts), or to the vedic extensions block then Microsoft Word puts in its dotted circle character. This doesn't happen in LibreOffice or I think Adobe publishing software.

MS Word has another problem with Vedic accents. The Vedic characters 1CE9, 1CEA, 1CEB, and 1CEC are always combined with a following anusvara dot or with candrabindu. I've seen this in a printed editions and see item 2 of this Unicode committee report.

https://www.unicode.org/L2/L2008/08216-cmt-vedic.pdf

But when you put in a following anusvara dot or candrabindu to these characters, MS Word puts in the dotted circle.

I've been told a few years ago that this is an issue Unicode has not resolved. See this response from a devanagari font designer to me.

This issue is one that Unicode has not resolved. Yes, the proposal
documents all indicate presence of the dot, but the characters are
encoded in the Vedic block, and occur in scripts other that Devanagari
(notably in Bengali). So there is an open question about what character
should be used for the dot, since including the Devanagari anusvara
character in a Bengali font for use with Vedic block characters is not
an ideal solution. My preference would be for encoding of an explicit
dot for this purpose within the Vedic block, so it can be generic and
not script-specific, although I understand that some thought has been
given to encoding versions of these characters with the dot explicit.

Harry Spier