2011/1/19 Jörn Kottmann <[email protected]>: > Hi all, > > as everyone knows OpenNLP needs statistical models. Over at sourceforge > we simply had a model download page and offered the models there for > download (we actually still do that). > > We might come up with a project internal process to test the quality of new > models before we release them. Beside that are there any rules > we have to follow ? E.g. a vote on the incubator mailing list, like we would > do > for a release of OpenNLP itself. > > Any insights to this issues are very welcome, maybe I just need to ask on a > different > mailing list. I think we should start releasing the old models we created > over at sourceforge > at Apache, since the models do not need to be changed for the 1.5.1 release.
It would also appreciate it very much if the binary model could be packaged as (versioned) maven artifiacts, maybe one artifact by language and by type of model. That would make it much easier for third party apps to depend upon opennlp. For instance the clojure community is pushing some of those models on clojars.org. It would be better if they could be made available on the same maven repo as the opennlp binary code IMHO. -- Olivier http://twitter.com/ogrisel - http://github.com/ogrisel
