[tex-hyphen] hyphenation different in xelatex?

Arthur Reutenauer arthur.reutenauer at normalesup.org
Wed Sep 2 00:36:24 CEST 2015


	Hi Karl,

> When run with latex or pdflatex, the (terminal) output is Quell-co-de.
> 
> When run with xelatex or lualatex, the output is Quell-code.
> (The US English hyphenation is also Quell-code.)

  That's a consequence of using the new patterns for UTF-8 engines;
quoting from loadhyph-de-1901:

	\ifx\secondarg\empty
	    % Unicode-aware engine (such as XeTeX or LuaTeX) only sees a single (2-byte) argument
	    \message{UTF-8 German hyphenation patterns (traditional orthography)}
	    \input hyph-de-1901.tex
	\else
	    % 8-bit engine (such as TeX or pdfTeX)
	    \message{EC German hyphenation patterns (traditional orthography)}
	    % Kept for the sake of backward compatibility, but newer and better patterns by WL are available.
	    \input dehypht.tex
	\fi

  (And similar code for [de-1996]).  The UTF-8 engines use the newer
patterns prepared by the German hyphenation pattern group (reachable at
https://lists.dante.de/mailman/listinfo/trennmuster), while the non-UTF-8
ones use the legacy patterns that are more or less frozen.  This is a
decision Mojca and I took back in 2008 and it has been the case ever
since.

> Why?  A difference in allowed hyphenations between engines seems
> absolutely wrong.

  Well, it comes from the format rather than the engines, really.  Since
XeTeX was still very new in TeX Live in 2008, and LuaTeX hadn't been
there at all, and the LaTeX team wouldn't make any decision about what
to put in the format, it seemed reasonable to make them behave
differently.  I'm pretty sure we discussed that back then.

> P.S. Ironically enough, this came up in the first place because Herbert
> Voss told me that Quell-code is preferable in German, too!  Oh well.

  This should be communicated to the German pattern group; I'm
pretty sure they'll have opinions on this.

	Best,

		Arthur


More information about the tex-hyphen mailing list