Tom, I abandoned my full text indexing projects once Rob Newson put the hammer down on couchdb-lucene. The code is still floating around but is hugely lacking compared to couchdb-lucene. I've since always recommended every use that because of all that it does.
As to the JVM, I feel your pain. Though in reality the only reasonable complaint is its memory usage. Chatting briefly with Rob it seems that there might be tricks to reducing the usage in general if you're worried about that. Perhaps a wiki page of "Managing memory usage" might be helpful to future users? HTH, Paul Davis On Sun, Apr 11, 2010 at 12:42 PM, Tom Sante <[email protected]> wrote: > Ok good, Couchdb Lucene it is than, used it before does a great job. > I was just curious because I never noticed there was a python option out > there eliminating the need for using java. > > On 11/04/10 18:34, Sebastian Cohnen wrote: >> >> I would go with couchdb-lucene. Robert Newson is doing a awesome job on >> pushing this thing forward and IMHO it's the best thing we have for FTI with >> couchdb... >> >> On 11.04.2010, at 18:31, Tom Sante wrote: >> >>> Hey, >>> >>> Is hypercouch still maintained somewhere? or is it best to just use >>> Couchdb Lucene instead? >>> >>> Tom >>> >> > >
