From: David Dorman <dorman@in...> - 2007-07-06 16:41:33
It isn't every day that we release a brand-new piece of software, and
this is one we're pretty excited about. We sincerely hope you will
like it, too.
Pazpar2 (pronounced like 'passe-partout', if it please ya) can be
viewed either as a high-performance metasearching middleware or a
Z39.50 client with a webservice interface, depending on your
perspective and needs. It is a fairly compact C program -- a resident
daemon -- that incorporates the best we know how to do in terms of
providing high performance, user-oriented federated searching.
One cool thing it does is search many databases in parallel, and do
it fast, without unduly loading up the user interface.. in fact, we
find it will search more than 100 targets in parallel and 'feel'
faster the more data you throw at it. It retrieves a set of records
from each target, and performs merging, deduplication,
ranking/sorting, and pulls browse facets from them. It does it really
fast, and because this functionality is exposed through a pretty
simple XML webservice, it is really easy to build very cool,
lightning-fast user interfaces on top of it -- interfaces that tend
to make people go 'no WAY is that metasearching'.
It doesn't know anything about data models, so you can handle exotic
data sources if you need to.. you use XSLT to normalize data into an
internal model -- we provide examples for MARC21 and a DC-esque
internal model, and configure ranking, facets, sorting, etc., from
that. Let us know if you try other things.
More details at http://www.indexdata.com/pazpar2/
Isn't Z39.50 a little old-school? Perhaps so, but it's also widely
supported -- probably one of the most vital and viable information
retrieval standards out there today, with new servers popping up all
the time. Most library catalogs and many of the larger, commercial
databases support it. It is trivial to gateway SRU/W to it. You can
use our SimpleServer to turn anything into a Z server. If you need
to do large-scale metasearching across non-standard resources, talk
to us -- we have relationships and tools with traditional metasearch
vendors and can help enable 'database connectors' for use via this tool.
It's GPL -- use it, don't abuse it. We offer the usual services and
support to people, and we encourage anyone contemplating commercial
deployment to work with us; make sure we know your needs, and help us
develop this further and keep it viable for the future. For anyone
else; have at it. We call it beta: Issues are bound to pop up when
new people play with it in different environments.. share your
experiences, your observations, etc. We can offer installation
assistance, training, support, even turnkey development and hosting.
US Marketing Manager, Index Data
52 Whitman Ave.
West Hartford, Connecticut 06107
860-389-1568 or toll free 866-489-1568
INDEX DATA Means Business
for Open Source and Open Standards
- - - - - - - - - - - - - - -