[XeTeX] Strange hyphenation with polyglossia in French

[Khaled Hosny]

On Wed, Oct 20, 2010 at 09:44:36AM +0100, Philip Taylor (Webmaster, Ret'd) wrote:
> 
> 
> Khaled Hosny wrote:
> 
> >Unicode is full of "compatibility with legacy encodings" non-sense, IMO
> >it should just be ignored. AFAIK, comma forms were added in Unicode
> >3.0.0 and that more than 10 years now, if we continue to support the old
> >broken practice it will never vanish.
> 
> Nor will it vanish just because the TeX community no longer support it.
> Unicode is, in many ways, a complete and utter mess [1], but if XeTeX and
> LuaTeX are to be based on it, then they must be based on it and not
> on some variant that we feel is preferable (for whatever reason).

But, AFAIK, what we are talking about here is a mere recommendation not
a compliance requirement. Also Unicode is actually a set of standards
and neither XeTeX nor LuaTeX implement all of them.

> >Anyway, I were actually concerned
> >about fonts that have a loca(lised) feature to map cedilla forms to
> >comma forms a practice that is making no one any favour.
> 
> If you are transcribing an extant document, and the author used
> the comma form, then the transcription should also use the comma
> form, otherwise it ceases to be a transcription and becomes
> something else.

The "loca" variant is causing just the opposite of this, if you want
cedilla form with Romanian you will instead get the comma form, and in
many applications it is impossible to turn off such a feature which can
become a real pain.

Regards,
 Khaled

-- 
 Khaled Hosny
 Arabic localiser and member of Arabeyes.org team
 Free font developer