On Wed, Jul 20, 2011 at 05:04:58PM +0100, Arthur Reutenauer wrote:
> Two issues here:
Oh yes, and
3. The patterns for Serbian are actually not used at all in hyph-utf8,
since experiments conducted by Nikola Lečić tended to prove that the
ones for Serbo-Croatian did a much better job of hyphenating Serbian
(for both scripts). The different TeX distributions thus load the
patterns for Serbo-Croatian when asked for Serbian, as can be seen in
loadhyph-sr-cyrl.tex and loadhyph-sr-latn.tex. We still ship the
Serbian patterns in hyph-utf8, as taking them out would only render the
situation even more confusing ("Why are there patterns for Serbian
outside of hyph-utf8, and the package only ships patterns for
Serbo-Croatian?").
Arthur
