F Wolff wrote:

>> For Yoruba, we're talking about a maximum of two diacritics, a sub-dot
>> for
>> additional characters and an acute or grave to mark tone.

> Exactly. Is a combining diacritic for tone on top of a sub-dot vowel
> more complicated to support than a combining diacritic involving
> stacking?

No. It's technically very slightly less complicated in terms of the font
layout tables and developmentally about on par. It is implemented either
as an above anchor attachment on the /edotbelow/ glyph or as an above
anchor attachment on the /e/ with another, below anchor attachment for
the dotbelow.

Note that during some forms of normalisation, text may be fully
decomposed to letter plus combining mark sequences. Conversely, during
display, a layout engine may recompose via the font cmap table to
precomposed glyphs if they are in the font. Then, if I built the font,
there is a pretty good chance that these precomposed glyphs will be
contextually re-decomposed for mark positioning; this last stage because
it is more efficient for me to define anchor attachment on individual
base letters only rather than on all the precomposed diacritic glyphs as
well.


John Hudson

_______________________________________________
A12n-collaboration mailing list
A12n-collaboration@...
http://lists.kabissa.org/mailman/listinfo/a12n-collaboration