tag:blogger.com,1999:blog-3380785587681963686.post1123215259025536756..comments2014-03-21T09:03:42.727+00:00Comments on Yet another uninformative blog title: Multiwords GSoC project braindumpjimreganhttp://www.blogger.com/profile/11166627131782365307noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-3380785587681963686.post-10418304825313538492014-03-21T09:03:42.727+00:002014-03-21T09:03:42.727+00:00Awesome post, especially the footnotes =P (though ...Awesome post, especially the footnotes =P (though I had to read it in w3m to see the code examples, hehe)<br /><br />But as our applicants say: Sir, I have a doubt. With the second-pass-over-FST method you're proposing, what would the final analyser and bidix _binaries_ be mapping? Am I right in thinking that the end result compiled analyser would output ^panna młoda/panna młoda<n><f><sg><nom>$ for input "panna młoda", and the bidix would match that input as a simple (<b/>) multiword? (That'd be ideal in my opinion, but I may have missed something.)Anonymousnoreply@blogger.com