On Tue, 8 Mar 2011 23:12:54 -0500, Alan Munn <am...@gmx.com> wrote: > I'm no expert on this, so I'd best stop while I'm ahead. Does makeindex > sort non-English things properly without resorting to special devices in > the index entries? Anyway, there is no problem as far as I know with > hyperref and makeindex; the problems with makeindex are related to its > ability to deal easily with non-English sort orders.
Ok, that makes sense. In our case, sorting doesn't cause a problem, because the index entries I mentioned contained phonemic (linguistic, non-orthographic) representations, and there *is* no "right" way to sort them. (Fortunately for us, the index entries contain English text introducing them, so the sorting works on that, and any remaining ambiguity in sorting affects at most a few entries.) If on the other hand you were creating index entries for words written in an orthography for which there was s "correct" sort order, like a Perso-Arabic or Indic script language, then I would imagine that makeindex cannot know how to sort them. You can't even sort by the Unicode code points, since different languages using the same Unicode blocks can (and often do) have distinct sort orders. All this is explained in xindy's FAQ: http://xindy.sourceforge.net/doc/faq-1.html#ss1.1 which I should have looked at first. <search for "unicode" at xindy site> Hmm, looks like the real problem with xindy is that it doesn't speak Unicode...there was a little discussion about changing it to be UTF-8 aware, but it's not clear that ever got done. So yes, I can see how that would be a problem! Similarly for BibTeX, I guess; we're using it successfully with XeLaTeX, but probably only because our use of Unicode in the citations is minimal, and we've been lucky. So yes, I see your point! Mike Maxwell -------------------------------------------------- Subscriptions, Archive, and List information, etc.: http://tug.org/mailman/listinfo/xetex