Questions about WAIS, WAISGate

Steve Romig (romig@cis.ohio-state.edu)
Sun, 13 Dec 92 03:33:47 -0500


Date: Sun, 13 Dec 92 03:33:47 -0500
From: Steve Romig <romig@cis.ohio-state.edu>
Message-Id: <9212130833.AA13526@brachiosaur.cis.ohio-state.edu>
To: www-talk@nxoc01.cern.ch
Subject: Questions about WAIS, WAISGate

I'm trying to set up a W3/wais server for the Internet RFCs and
related docs.  I created a server script in Perl that takes requests,
looks up the RFCs in our ftp archive, and returns an HTML doc with
links to things that look like other RFC names.  So far, so good, this
works great.

I also wanted to be able to search for stuff with a waisserver, so I
indexed them (they're in /n/archive/2/ftp/pub/rfc or something like
that).  WAIS lookups work fine.  

I also set up a waisgate server.  Lookups work ok, but here's the
problem: the doc IDs are created such that requests for the docs get
passed to waisgate and from there to the waisserver.  Of course,
that's the way things should work.  But that means that they aren't
retrieved from the special rfc server that I created.

I could get around this by converting the RFC texts to html, storing
them somewhere, and indexing that, but I'd rather avoid copying
everything and all that.

What I'd like to do (and will, unless someone has a better answer) is
to have waisgate read a config file that might look something like
this: 

    rfc /n/archive/2/ftp/pub/rfc/*.Z http://rfcserver:port/rfc/*.html

so that document ids from database rfc that match the pattern are
seen, they are translated to the form on the right instead of the
usual wais->www ids.  That way I could have waisgate redirect specific
doc lookups to a special server, and not have to copy the databases.

Is there a better way to do this (short of copying the source files,
preformatting them and leaving them there)?  

--- Steve