Re: freeWAIS Multiformat type and WWW front end

davidc@erin.gov.au (David Crossley)

Mail folder: WWW Talk Apr 94-present
Next message: Stan Letovsky: "Re: Toward Closure on HTML "
Previous message: Martijn Koster: "to RFC or not (was: Re: Toward Closure on HTML )"
Maybe in reply to: Marc Andreessen: "freeWAIS Multiformat type and WWW front end"

Errors-To: listmaster@www0.cern.ch
Date: Tue, 5 Apr 1994 11:44:51 --100
Message-id: <9404050946.AA08984@erin.gov.au>
Errors-To: listmaster@www0.cern.ch
Reply-To: davidc@erin.gov.au
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: davidc@erin.gov.au (David Crossley)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: Re: freeWAIS Multiformat type and WWW front end
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
Content-Length: 1534

Marc Andreesen writes on Mon, 4 Apr 94:

> David Crossley writes:
> > We do not store our metadata in text files but in an RDBMS.  So we
> > do not have any text files to index.  As far as I can see the only
> > solution is to run a cron job every few days to generate metadata
> > text files from the RDBMS for each dataset and index these.
> > 
> > Another solution may be to create a text file for each dataset of
> > likely keywords, index these and then somehow intercept the query
> > with either a WAIS server script or a WWW server script.  Does
> > anyone know if this would be possible?

> The latter should be possible if you're going through a sufficiently
> smart and customized WAIS/WWW gateway.  In other words, the gateway
> has to be smart enough to map WAIS search hits to some representation
> of files in the RDBMS; a parallel WWW script would then need to be
> able to reach inside the database upon receiving a query for such a
> representation.  I think that's how it would work...

So as not to assume that the WAIS request is always coming from a WWW client
I will probably do all this with a WAIS gateway and WAIS server scripts. 
I hope that I can get it smart enough to then regenerate metadata files when 
a hit is made. Then these could be picked up by the Multiformat Opportunity
of the client. These metadata files would then be regularly re-indexed 
rather than do the keywords caper.

cheers
David

---
David Crossley
ERIN - Environmental Resources Information Network - Australia
davidc@erin.gov.au