It’s somewhat implicit, but still relatively clear:
[...]
You are right about the information re Arabic round double quotation
marks being essentially within TUS; thanks.
I also had in mind Roozbeh's statement about Persian and Greek (and I
won't check right now), so there remains the question of glyph variation
and locale/language-based glyph repertoires and most common shapes. CLDR
has a lot of information here
http://www.unicode.org/cldr/charts//by_type/patterns.characters.html
and here too
http://www.unicode.org/cldr/charts//by_type/misc.exemplarCharacters.html
(it always takes me a while to locate these charts), but I can't find
information covering Roozbeh's answer there. Canonical glyph shapes
might be presently considered to go beyond CLDR (such data also depends
on the era (how would one indicate that, time period–based language
tags?)), but the parts don't yet add to a comprehensive whole. I would
just like to draw people's attention to this.
Stephan
