[XeTeX] OpenOffice hyphenation

Jonathan Kew jonathan_kew at sil.org
Sat Nov 11 15:40:08 CET 2006


On 11 Nov 2006, at 1:49 am, Adam Twardoch wrote:

> At
> http://lingucomponent.openoffice.org/
> quite an extensive collection of spelling and hyphenation modules is
> available for OpenOffice. The OO hyphenation dictionaries are based on
> TeX patters "with small corrections":
> http://lingucomponent.openoffice.org/hyphenator.html
>
> I think it would be extremely useful if XeTeX accepted those OO
> hyphenation patters as they are, the main advantage being, that  
> they are
> well organized and easily available.


I looked at the OO patterns earlier this year, and wrote a "wrapper"  
macro to enable xetex to read the files as distributed, but abandoned  
this for the time being because (a) I noticed a number of problems  
with the files, where it appears that they were not correctly  
converted or encoded; and (b) I am reluctant to separate the language  
configuration for xelatex from that used for standard latex.

Instead, I have concentrated on ensuring that the files in TeX Live  
are configured so that they can be loaded correctly under xetex as  
well as tex, which should eliminate the recurring problems with  
format creation depending on enabled languages. This new setup should  
be part of TL2006.

Looking at the current OO package of language support files, I think  
a good deal of clean-up has been done, though I think there may still  
be a few issues of encoding/file format. (Yes, I know: when I can  
find a little time, I should file a bug report.) So issue (a) could  
be addressed.

Issue (b) is harder: I agree that it would be very good to share  
patterns between the TeX and OpenOffice worlds, as the algorithms  
used are similar enough that I believe this could work. I'd like to  
see *all* TeX engines (not just xetex) sharing a common set of  
pattern files with OO, with the use of appropriate macros to load  
them correctly according to the encodings in use. (This may be a  
little trickier for the old 8-bit systems, but probably still  
feasible.) Reaching this point will take a significant amount of  
effort and coordination, however.

JK



More information about the XeTeX mailing list