federated search - Carrot Clustering Engine

[Grégoire Neuville <gr...@gmail.com>]

Hi,

I've recently posted a message on this list about the ways of
implementing a 'federated search' (a search on multiple indices) whith
Solr. Starting from the answers I got plus informations from the
mailing list archives and the Solr Wiki, I've planned to test three
configurations :

- one sole index for all applications (each one interrogate the index
its own way)
- one index and one schema.xml per application (each index managed by
a Core of the MultiCore system)
- one index per application and a sole schema.xml for them all (each
'shard' also managed individually by a Solr Core)

What do you think of those possibilities ? Have you tested and/or
compared them yourself ?

Secondly, as I am currently developping a portal for federated search,
I'm getting interested in the clustering engine Carrot²
(http://project.carrot2.org) ; as it is described as being able to
read results produced by Solr, I was wondering if anyone of you Solr
users had already used them together ?

Thanks,
-- 
Grégoire Neuville