"HO" == Heiko Oberdiek writes: HO> But it would be easily possible to have a data structure as HO> superset: % language, patterns file, exceptions file, left- and HO> right-hyphen minima % dialect
HO> Then both language.def and language.dat could be generated easily. probably, it would be good to have the target encoding (such as t1, il2, ly1, eu1) in which to load the patterns, specified in the unified file. (also, versioning, etc). then the same UTF8-encoded pattern file can be used to pre-load patterns in any supported target encoding. Best, v.
