--- In qalam@yahoogroups.com, John Cowan <cowan@...> wrote:

> Peter T. Daniels scripsit:
>
> > I don't know what "Unicode combining character" means, but surely

there

> > are some of those that aren't diacritics?
>
> To be sure. In Hebrew alone, we have dagesh/mappiq, which is a

diacritic;

> vowel points, which are vowel points; and accents, which are a

separate

> layer of the text altogether. All are formally combining

characters.

And all can happily count as diacritics when determining whether
pointed Hebrew script is an alphabet or a pointed abjad. However,
these accents aren't relevant to the issue.

> OTOH, the stroke in ?not a combining character (it's not treated
> as a separate character at all, any more than the dot on i), and the
> IPA rhotic hook is not either (it's considered a separate letter).

Sounds like the 'should be' extension would have to be invoked for
rhotic hook if it were relevant. However, if it were relevant, I'm
sure it would have been made a combining character.

The dot on 'i' is not a separate character. It's a bit of decoration
or disambiguation that is supposed to be removed when it gets in the
way of a diacritic, though I've seen fonts where it isn't.

Richard.