Re: overloading cn= in the DIT

Andrew Findlay <> Fri, 24 July 1992 15:54 UTC

Received: from by IETF.NRI.Reston.VA.US id aa04244; 24 Jul 92 11:54 EDT
Received: from NRI.NRI.Reston.Va.US by IETF.NRI.Reston.VA.US id aa04240; 24 Jul 92 11:54 EDT
Received: from by NRI.Reston.VA.US id aa19734; 24 Jul 92 11:54 EDT
Received: from by with local SMTP id <>; Fri, 24 Jul 1992 16:40:02 +0100
Via:; Fri, 24 Jul 1992 16:39:54 +0100
Received: from by with SMTP (PP) id <>; Fri, 24 Jul 1992 16:39:45 +0100
From: Andrew Findlay <>
Message-Id: <>
Subject: Re: overloading cn= in the DIT
To: Andrew Waugh <>
Date: Fri, 24 Jul 92 16:39:37 BST
In-Reply-To: <9207180222.AA03735@squid.mel.dit.CSIRO.AU>; from "Andrew Waugh" at Jul 18, 92 12:22 pm
X-Mailer: ELM [version 2.2-hd PL 10]

>> Now how do I deal with people who do
>> search cn=<string> and find a bunch of DSA and MTA get returned?
>This is the problem with the ****** DUAs, not the standard. The

Absolutely correct. Hit the DUA designers to do the job right.
(I speak as one, and cower as I write :-)

There has been a raging flame-war in the UK lately about the existance
of animal names in the DIT. The same comments apply: DUAs should show
the user what they want *and nothing that would confuse them*.

>user should be asking (in an appropriate GUI way, of course :-)
>'search for a person with the name of George Michaelson'. Never,
>ever, 'search cn="George Michaelson"'.

This is exactly what the (later) Brunel DUAs do. Try `doog' (line
mode) or `xlu' (MIT X) or PC-pages (MS-windows/CONS). They all take
great care to show only `real' people when that is what the user is

>(ps I should make an honourable mention of ITAXA, which at least
>looks at the object class of the returned entries _and_sorts_them_.

It is much better to specify the object class in the search. Then you
never have to process the irrelevant entries at all (and there might
be a lot...)


|      From Andrew Findlay at Brunel University, Uxbridge, UB8 3PH, UK     |
|       +44 895 203066 or +44 895 274000 x2512 |