[twg-tds] [gaulle@idris.fr: a TDS question]

[Bernard Gaulle]

Karl Berry writes:
 > given an original .:foo:bar, and
 > if this potential language searching is enabled, then it could either
 > search:
 > 
 >   ./language
 >   foo/language
 >   bar/language
 >   .
 >   foo
 >   bar
 > 
 > Or it could search:
 > 
 >   ./language
 >   .
 >   foo/language
 >   foo
 >   bar/language
 >   bar
 
Good question! As language search should precede any other
i guess that the first solution is the good one.

 >      > Finally, in the current TeX Live texmf tree, it seems that
 >      > language-specific files are scattered everywhere ... ?
 > 
 >     it-is an argument pro or con?
 > 
 > As Walter said, it means I don't understand how we can implement the
 > idea in practice, since the files aren't organized that way.  So I guess
 > it's con.

no, it's pro, numerous files have today linguistic attributes in
their name starting with hyphenation files such as ushyph.tex.
That organization is historical but not the best. We could start
gathering all linguistic input files in a same area and thus standard
file names such as hyphen.file could be used for all languages
staying more DEK compliant, more efficient and greatly open to
new multilingual features.

  --bg