Re: root knowledge

Steve Hardcastle-Kille <S.Kille@cs.ucl.ac.uk> Tue, 12 May 1992 14:54 UTC

Received: from nri.nri.reston.va.us by ietf.NRI.Reston.VA.US id aa23652; 12 May 92 10:54 EDT
Received: from nri.reston.va.us by NRI.Reston.VA.US id aa12175; 12 May 92 11:00 EDT
Received: from bells.cs.ucl.ac.uk by NRI.Reston.VA.US id aa12161; 12 May 92 10:59 EDT
Received: from glenlivet.cs.ucl.ac.uk by bells.cs.ucl.ac.uk with local SMTP id <g.25391-0@bells.cs.ucl.ac.uk>; Tue, 12 May 1992 13:20:15 +0100
To: pays@faugeres.inria.fr
cc: "(Paul-Andre.PAYS)" <Paul-Andre.Pays@faugeres.inria.fr>, osi-ds@cs.ucl.ac.uk
Subject: Re: root knowledge
Phone: +44-71-380-7294
In-reply-to: Your message of Thu, 07 May 92 21:26:40 +0100. <705266800.6537.0@faugeres.inria.fr>
Date: Tue, 12 May 92 13:19:46 +0100
Message-ID: <1036.705673186@UK.AC.UCL.CS>
From: Steve Hardcastle-Kille <S.Kille@cs.ucl.ac.uk>

Paul,

Let me step back from this a little:

The key issue:  X.500(88) is insufficient.   The main reasons
   a) there is no replication model
   b) there is no mechanism for representing knowledge in the directory

To operate a service, these problems need to be addressed.   (RFC 1275
explains this in more detail).   

At some stage, use of the 1992 specifications will be the right
solution.   I suspect that there will need to be additional
specification and agreements in order to make this deployable, but
this remains an open question for now.   


When we specced RFC 1276, there was a perceived need for an interim
soultion.  This was based on, but not identical to, the QUIPU
solution, with solid input from other implementors.  I think that
there is much merit in basing standards on things which have been
shown to work.  The RFC 1276 specification is independent of QUIPU,
and fully self standing.


I am loath (very loath) to spend effort on developing ANOTHER interim
solution.   If I am going to spend effort on this area, it would be on
protyping the 92 stuff.   My feeling is that what you propose will be
insufficient to make things work reasonably, although it clearly
remains to be tried out.   


I would strongly recommend that you work on RFC 1276.  If you do not
like the replication, you could at least support the knowledge
representation.  Once you do this, as Colin notes, you can extract and
replicate knowledge by use of standard directory operations.


Steve