On Mar 19, 2007, at 3:04 PM, Michael K. Bergman wrote:
> This is a *very* interesting thread and a wonderful site. I've poked
> around quite a bit on the Peel Library site and have not seen a
> description of how the site is built and with what components. Do you
> have a reference link?
I happen to know the developers of the site, Peter Binkley and Tricia
Williams. We've met at various library-related meetings. They have
posted on the solr-user list about this site and were an early
adopter of facets even before they were built into Solr, using a hack
I had created for Collex.
> But, as an outsider plus the general lack of informative
> material on the Apache site, I most often come away more confused than
> educated.
No doubt its hard to navigate the guts. But projects like Lucene are
key components in at least the Java-based semantic web engines
(notably Kowari and Sesame, and Longwell too).
> Can either of you point me to some links or provide a basic dump on
> Apache projects (Solr, Cocoon, SolrForrest?, ???, ???) and their
> relation to SemWeb applications? And, Erik, do you know anything of
> Carrot2 and how it might relate as well?
*whew*
Carrot2 is a clustering engine. I don't have experience with it, so
cannot comment there, other than to know it exists and has a great
reputation.
Solr is Lucene made easy to use from any environment, not just Java,
and value adds a lot on top of Lucene with caching, filters, index
reader/searcher management, document update management, and much
more. Building indices is key to navigating lots of data, and Solr
can rock and roll on indexing provided a domain is mapped into a
document/field/term structure. So I think Lucene and Solr are very
valuable components to the semweb tool chain. Cocoon is used by the
Peel Library site for being the rendering side of things, conversing
with Solr and formulating the response. It can tap into Solr very
easily, with XML or JSON depending on which part of Cocoon is gonna
digest the data, coming back from Solr. SolrForrest - don't know
much about, though the publish pipeline sending content right into
the search engine makes a lot of sense.
Erik
_______________________________________________
General mailing list
[email protected]
http://simile.mit.edu/mailman/listinfo/general
