Searching your web pages
marc at matahari.ckm.ucsf.edu
Tue Feb 20 15:36:53 EST 1996
Z39.50 is a complex, binary, stateful information retrieval protocol best
suited for sessions such as an OPAC search session. It can be strapped onto
most any search engine.
If you take the time to do the original cataloging (as you would any other item
held in your collection) and indexing work on your web site so that each page,
directory or other division has metainformation associated with it, and run a
search engine over it that can take advantage of that indexing, then you might
want to run a Z39.50 compliant server over your content.
The problem with integrating Z39,50 (stateful) and HTTP (stateless) is that the
browsing process can easily be stateful, which is no problem for Z39.50, while
HTTP has a hard time dealing with this. With more simpler protocols, gopher,
nntp and even the stateful ftp, the common applications are simple enough to
incorporate into the web browser.
In most Z39.50 applications, the requirements for interaction are complex
enough to suggest a gateway, where the HTTP server learns how to handle state
and can take advantage of its knowledge of a session and its transactions to
present the data in a session context.
So, if you know that people are going to search your richly-indexed web pages
and require the richness that a full boolean search engine can provide, then
1. find a z39.50 capable search engine to put on top of your data.
2. get an off-the-shelf state package, like Isite from cnidr.org
(ftp.cnidr.org) as a front-end to your communicate with your server and
it up to your HTTP server.
But the task of doing this is not trivial. You will need some programmer
resources, database expertise and a Z39.50 guru or two wouldn't hurt either.
// Marc Salomon - Software Engineer - Innovative Software Systems Group \\
\\ Library and Center for Knowledge Management - UC, San Francisco //
// phone : 415.476.9541 - e-mail : marc at ckm.ucsf.edu - fax: 415.476.4653 \\
More information about the Web4lib