On Mar 12, 2010, at 9:58 AM, Bernd Fondermann wrote: > On Fri, Mar 12, 2010 at 15:39, Grant Ingersoll <[email protected]> wrote: >> I have no problem with you proposing to bring in Nutch's overlap. The fact >> is, the Board doesn't like subprojects anyway and we are likely headed for >> some consolidation/spinning out anyway (see the December Board Minutes). > > In fact, I was waiting for this argument to be made... > > The truth is, umbrella projects didn't go well and the board is only > watching over this, while the ASF membership thinks umbrellas are no > good. > > And as everybody can see now, although there is a large overlap in > Lucene/Solr committers, people talk like there are two different > projects. This is wrong. There is only one project, named Lucene, with > one PMC, and one committership. >
I think that is where we are headed, but it isn't where we are right now (at least at the committership level). The Board will likely be seeing a proposal for Mahout as a TLP next month (we are in the middle of a release cycle so we don't want any distractions at the moment). I think Tika can stand on it's own, too, and the community there should have the discussion. At the same time, I don't want to "kick them out", either, but I would encourage them to at least have the discussion. The Ports of Lucene are a bit tricky in my mind. Both of them are auto-generated for the most part, so they don't require a super amount of work to produce, but they don't really seem to be standalone either other than there isn't much committer overlap. I personally think the status quo works really well there, but again, just my opinion. That leaves Solr and Nutch. The past vote has answered the question for Solr. I guess I'd encourage the Nutch community to have a discussion on it. There isn't much committer overlap there with Lucene or Solr but there is some code overlap. Personally, I think the crawling/plugin stuff could spin out but the core Lucene/analyzers stuff merits a review and a merge. Again, that is up to Nutch to decide. Last I looked at Nutch they were moving to a more modular architecture that focused on crawling and handed off the other stuff to things like Solr and Tika. -Grant
