2015-11-23 09:45:49 UTC
"The marks are drawn once for each character. However, emphasis marks
are not drawn for characters that are:
Word separators or that belong to the Unicode separator classes
(Z*). (But note that emphasis marks are drawn for a space that combines
with any combining characters.)
Characters belonging to the Unicode classes for control codes and
unassigned characters (Cc, Cf, Cn)."
presumably these marks should not be displayed for combining characters
thinking a bit further, i wonder what happens for syllables in complex
indic and SE Asian scripts. Grapheme clusters may account for some
associations, but not those where multiple grapheme clusters comprise a
single syllable. Perhaps these emphasis marks are not really intended
to be used with complex scripts - if that's the case, perhaps we should
add a note to that effect.
Note, btw, that there is a possible requirement for this kind of
emphasis in Tibetan (see
http://w3c.github.io/tlreq/#emphasis_and_highlighting), but the mark is
centered across the syllable rather than tied to a single character. Do
we have any thought of catering for that in future versions of the spec?