[tex-hyphen] Add a new language support (+hyphenation)

Reinhard Kotucha reinhard.kotucha at web.de
Thu Jun 12 00:39:10 CEST 2014

On 2014-06-11 at 20:27:59 +0100, Arthur Reutenauer wrote:

 >   Unfortunately not, adding a new language is a highly specialised
 > area of TeX, and we're very few people working on language support
 > generally.  But you're welcome to ask any questions here, of
 > course.
 >   As a side note, since your patterns are for Pāli in Latin
 > transliteration, I'd use the code pi-latin, instead of just pi, to
 > tag the language.

It's true that adding a new language is a highly specialised area of
TeX.  On the other hand not many people are able to create hyphenation
patterns themselves at all but making them available to TeX is not
overly difficult.  The only problem I see is that modern TeX
distributions provide update facilities which overwrite your files if
you aren't careful.

IMO the best solution is to upload new patterns to CTAN.  The main
requirement is a proper license, of course.

Development of hyphenation patterns is easier then.  Determine the
location of the language.dat file being used

  kpsewhich language.dat

and copy it to the appropriate directory in TEXMFHOME or TEXMFLOCAL.

If package updates provide new patterns, these new patterns will not
be used.  But if you intend to upload your patterns to CTAN anyway,
this is only a temporary problem.  I assume that while you are
developing patterns for Pāli you are not interested in new patterns
for, say, African languages provided by TeX Live, at least not for the
time being.  You certainly want to remove the stuff from
TEXMFHOME/TEXMFLOCAL when your patterns are provided by TeX Live.

A file like language-local.dat only makes sense for stuff you don't
want/can share with others.

It would be nice to have patterns for pi-latin on CTAN, at least if
they are correct and the license is appropriate.  Just to prevent
people from re-inventing the wheel again and again.

After all, the brute-force method described above works on all systems
but only temporarily.  In the long term you either have to put your
stuff into a dedicated file like language-local.dat (this doesn't work
everywhere) or you have to upload it to CTAN.  IMO the latter solution
is preferred.

Krzysztof and Arthur, is it possible to prevent people from
re-inventing the wheel again?  I.e., push the patters to CTAN?


Reinhard Kotucha                            Phone: +49-511-3373112
Marschnerstr. 25
D-30167 Hannover                    mailto:reinhard.kotucha at web.de

More information about the tex-hyphen mailing list