Am 29.04.2010 06:15, schrieb Manuel Pégourié-Gonnard:
Amongst them, three have no patterns nor exceptions at all (nohyphenation, arabic, farsi) and one has only two patterns (dumylang). Which leaves us with 4 "real" languages (german-x, ngerman-x, ibycus, mongolianmlc). There are probably arguments for not supporting any of them. However, my proposal is to support them in the following easy way: treat them the good old way, load them at format generation time.
I'm not really closely following the current restructuring efforts, but could you please point out what makes first-class support for patterns from dehyph-exptl (german-x et al.) desirable? How do the timestamped patterns break your scheme for all the other languages?
Note, those experimental pattern sets are relatively large. To me, it doesn't seem right to load them into the format.
Best regards, Stephan Hennig
