Sounds good to me, +1. The main thing will be to get volunteers to do the updates, on approximately the road map schedule.
I think the NLP tasks could be seeded by harvesting our mailing lists for useful nuggets. Of course, another thing could be to get someone to write a book covering these topics. -Marshall On 11/18/2012 3:10 AM, Tommaso Teofili wrote: > Hi all, > > currently I don't think we have a proper roadmap on the addons development > so I'd like to propose one. > > Some of the addons are based on some other libraries (e.g. Lucene, Solr, > Tika, etc.) so the first thing I'd like to do is gradually update those to > newly released major versions; these touches: > - TikaAnnotator : currently at 0.7, migrate to 1.2 and make a release > - Lucas : currently at 2.9.3, migrate to 3.6 / 4.x and maybe make one > release each (because of API / index change from 3.x to 4.x) > - Solrcas : currently at 3.1, migrate to 3.6 / 4.x and maybe make one > release each (because of API / index change from 3.x to 4.x) > > The following addons providing services from external services may need an > update as well in order to get latest features > - OpenCalaisAnnotator > - AlchemyAPIAnnotator > then release each of them > > For all the others I think it's more or less just bugfixing. > > Another thing I'd like to add to our website / documentation is something > like "how to do NLP task X within Apache UIMA"; we currently have a list of > external sources plus the addons page but I think it'd be nice for a user > to know which known open source UIMA enabled options exist for doing e.g. > language detection, tokenization, etc either in the addons package or from > some other sources. > > What do you think? > Regards, > Tommaso >
