Re: Coloured Punctuation and Annotation

From: Richard Wordingham <>
Date: Thu, 6 Apr 2017 05:41:07 +0100

On Thu, 6 Apr 2017 01:11:09 +0100
Michael Everson <> wrote:

> On 5 Apr 2017, at 22:48, Richard Wordingham
> <> wrote:
> > I tried to read it from UTS#51 ‘Unicode Emoji', which is not part
> > of TUS, but I couldn't deduce that a font that enables U+10B99
> > PSALTER PAHLAVI SECTION MARK to have exactly two (as opposed to
> > none or four) red dots is in breach of the guidelines therein.
> Kindly explain how ANY font could do this.

Is this a trick question?

The character consists of 4 dots arranged at the corners of a
diamond. The top and bottom dots are traditionally red. From the
proposal, they may be drawn as circles, but I'm not completely sure
from the wording that this isn't a transcriptional convention. The left
and right dots are in the colour of the accompanying letters.

I haven't done it myself, so I can only give you my interpretation of
the OpenType standard.

An OpenType font not using SVG would use 3 outline glyph
definitions - one for ultimate monochrome rendering, one for the red
dots, and one for the black (or whatever) dots.

The cmap table would map the character to the first glyph.

The COLR table (the 'color table') then maps the
first glyph to a combination of the second and third glyphs. The
second glyph would have its colour specified by an index into the
currently selected colour palette. The third glyph would use 'index'
0xFFFF to specify that it be displayed in the foreground colour.

The palette is defined in the CPAL table (the 'color palette table') . All palettes
(there need only be one) should have a common index value designating

And far as I can see, that's job done - a natural two-tone glyph.

If one set the foreground colour to purple, then the top and bottom
dots will be red and the left and right dots will be purple - exactly
two red dots. Of course, if one selects red as the foreground colour,
one will get four red dots, as with a monochrome font.

> > Are we really going to have to set up Psalter Pahlavi emoji?
> > There's also some encoded Ethiopic punctuation that certainly used
> > to have red dots.
> If you want 10B99 to have different coloured dots (the rings? the
> dots?) the only precedent we have in the UCS is (1) to name a whole
> glyph with a colour like RED APPLE and then to hatch the glyph in
> black and white or (2) use the emoji property.

Chapter and verse for (2), please. I searched and couldn't find it.
To be precise, what says that a 'text presentation' has to be

> > I think the emoji database has overlooked an entire script of emoji
> > - the Egyptian hieroglyphs!
> Put it out of your mind.

A prohibition on fonts delivering appropriately coloured hieroglyphs
seems wrong. The hieroglyphs don't have the emoji property, so
compliance to the standards promulgated by the UTC would rule such a
font out if you and Asmus are correct.

Received on Wed Apr 05 2017 - 23:43:07 CDT

This archive was generated by hypermail 2.2.0 : Wed Apr 05 2017 - 23:43:07 CDT