On 6/1/2007 6:27 PM, Jordan Hayes wrote:

> Someone asked me today: why don't we just use Solr for this?
> 
> So I checked out their web page: http://lucene.apache.org/solr ... and 
> it doesn't look to me that this could be all that different, except at 
> the edges.  I like the idea of a built-in query mechanism available as a 
> servlet, but ...
> 
> Anyone here have experience with both Solr and swish-e such that I could 
> simply say "Nah, let's use swish-e instead!" (because of course I 
> already know swish-e ... :-).
> 


I have experience implementing swish-e, and experience thinking about
Solr and looking at others' work.  :-)

IMO, Solr gets you built-in faceting plus support for multi-million
record databases.  It also brings with it some Java requirements on the
server side and a requirement to talk to the search engine through HTTP
queries - not bad stuff if you're working at a certain scale, but
something I'm happy to avoid fiddling with swish-e in a good old Perl
CGI environment.

Also IMO, the Solr documentation consists of a tutorial that you either
grasp or you don't, and there aren't a lot of alternatives.

So, for example, I index about 135k records for electronic theses and
dissertations in swish-e, but at some point in the future we'll index
(and facet) ~10M journal articles with Solr.


-- 
Thomas Dowling
tdowling@ohiolink.edu
_______________________________________________
Users mailing list
Users@lists.swish-e.org
http://lists.swish-e.org/listinfo/users