Re: Widely distributed searching

Paul Barker <P.Barker@cs.ucl.ac.uk> Thu, 19 November 1992 20:41 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa03210; 19 Nov 92 15:41 EST
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa03205; 19 Nov 92 15:41 EST
Received: from haig.cs.ucl.ac.uk by CNRI.Reston.VA.US id aa14532; 19 Nov 92 15:41 EST
Received: from bells.cs.ucl.ac.uk by haig.cs.ucl.ac.uk with local SMTP id <g.02884-0@haig.cs.ucl.ac.uk>; Thu, 19 Nov 1992 19:30:27 +0000
Received: from sol.cs.ucl.ac.uk by bells.cs.ucl.ac.uk with local SMTP id <g.08598-0@bells.cs.ucl.ac.uk>; Thu, 19 Nov 1992 19:30:15 +0000
To: osi-ds@cs.ucl.ac.uk, directory-group@jnt.ac.uk
Subject: Re: Widely distributed searching
In-reply-to: Your message of "Wed, 18 Nov 92 14:12:05 GMT."
Date: Thu, 19 Nov 1992 19:30:15 +0000
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Paul Barker <P.Barker@cs.ucl.ac.uk>
Message-ID: <9211191541.aa14532@CNRI.Reston.VA.US>

Thanks to those of you who passed on comments or raised interesting points.
I'd likely to briefly address some of the issues raised.

Load: to be honest I don't have a lot of sympathy for DSAs.  With a few
honorable exceptions most people running DSAs have not been swamped with
queries.  Taken to an extreme, one could "defend" DSAs by providing DUAs of
such limited usefulness that users were discouraged from using X.500.  There 
must be better ways than this of denying service to avoid congestion.  

I don't believe in practice that users will want to "power search" very
often.  Even if it is surprisingly fast, it is still a lot slower than if
one can provide an organisation name match.  In the short term, of course
people will play.  Good for them, as I believe the more the Directory offers,
the more positively they will feel about the Directory.

Filters: there is room for experimentation / improvement here.  The filters
used when power searching are appreciably tighter than those used in DE's
normal search mode.  If anything they are now too restrictive, but I don't
want to loosen them up until I can analyse the effect of so-doing.

Paul