> Dave, is that what you're doing?
To date, we've just brought the results together side-by-side.
But there are some folks here who are interested in seeing a single result set, so I'm gearing up to experiment with that -- hence my musings earlier on score comparisons.
For my money, NC State does a really nice job bringing results together (was going to mention that even before Steve's email), without some of the accompanying problems, like combining facets, as Molly mentioned.
A single, combined result set would have to be significantly better than that to justify all the extra work -- and I'm not sure it would be. I guess I'll find out.
--Dave
==================
David Walker
Library Web Services Manager
California State University
http://xerxes.calstate.edu
________________________________________
From: Jonathan Rochkind [rochkind_at_jhu.edu]
Sent: Monday, March 28, 2011 9:56 AM
To: Next generation catalogs for libraries
Cc: Walker, David
Subject: Re: [NGC4LIB] The next generation of discovery tools (new LJ article)
On 3/28/2011 12:42 PM, Walker, David wrote:
> I know VUFind also has code for Summon, and maybe also Blacklight? I know of a few other smaller projects doing this as well. You may get more responses to this question on the code4lib listserv.
Last I heard/saw, VuFind only provided Summon results in a seperate tab,
it didn't try to merge Summon results into the result set with local
Solrresults.
Has this changed?
Blacklighthas nothing related to this, out of the box.
I am still very interested in people's actual real world experience with
merging relevance ranked result sets from two different systems
together. And encourage anyone that's done that succesfully to write it
up somewhere, blog, code4lib journal, anywhere. Dave, is that what
you're doing?
My worry is that it'll end up with all the terribleness of our existing
federated broadcast search solutions (such as Metalib), since that's
essentially what it is.
Received on Tue Mar 29 2011 - 15:24:12 EDT