Hello, Petr, I'm not sure if the list accepts non-member posts, could you please repost it?
On Sat, Jun 07, 2008 at 06:12:19PM +0200, Petr Sojka wrote: > -- when alphabet does not have more than 255 chars, > opatgen's patterns are "isomorphic" to patgen's. > > -- opatgen is necessary when one needs more than 255 chars, > thus is not needed for German. Absolutely correct. The patterns generated are the same (except a bug in patgen in one character long patterns, which don't make sense in most cases anyway). Moreover, if you can use patgen, do so: - patgen is significantly faster - the data structures are tightly coupled with semantics, i.e., a set of terrible hacks - it doesn't solve unicode, opatgen has its own routines to parse UTF8, at the time of writing, stable libraries were not available - the design of opatgen is quite unsuitable for the purpose; I tried to make it really general which wasn't a good idea - opatgen is not very well coded (neither is patgen, but is works better) If you need more characters than patgen is capable of (which is definitly not the case for German), try to compile opatgen with gcc-2.95. Yours, D.A. -- The great tragedy of Science---the slaying of a beautiful hypothesis by an ugly fact -- Thomas Huxley
