Re: X.500, Naming and the Internet

Einar Stefferud <Stef@nma.com> Thu, 06 February 1992 16:07 UTC

Received: from nri.reston.va.us by NRI.NRI.Reston.VA.US id aa14795; 6 Feb 92 11:07 EST
Received: from bells.cs.ucl.ac.uk by NRI.NRI.Reston.VA.US id aa14791; 6 Feb 92 11:07 EST
Received: from ics.uci.edu by bells.cs.ucl.ac.uk with Internet SMTP id <g.02412-0@bells.cs.ucl.ac.uk>; Thu, 6 Feb 1992 12:20:17 +0000
Received: from nma.com by q2.ics.uci.edu id aa20703; 6 Feb 92 3:51 PST
Received: from odin.nma.com by nma.com id aa09441; 5 Feb 92 22:23 PST
To: osi-ds@cs.ucl.ac.uk
Subject: Re: X.500, Naming and the Internet
In-reply-to: Your message of Wed, 05 Feb 92 19:13:13 -0500. <9202060013.AA01546@spartacus.psi.com>
Reply-to: Stef@nma.com
From: Einar Stefferud <Stef@nma.com>
Date: Wed, 05 Feb 1992 22:19:17 -0800
Message-ID: <7900.697357157@nma.com>
Sender: stef@nma.com

Christian said:

>> uh? What the hell is the French government doing here!

And Wengyik said:

> What the French government is doing here is central to this
discussion.

Which gives me a chance to step in and possibly help Weng out a bit.

Thanks Weng for carrying the load so far!

This sure is a hard subject to sort out.  It took Marshall Rose and I
about six months of rather intense discussion to resolve it for
NADF-175 (RFC1255), and we are now having a very hard time explaining
it to some people around the world, while others seem to catch on
right away.  

Marshall may or may not agree with the details of what I say here, but
he is free to try correct me any time.

Some of these thoughts are going to seem pretty wild, so hang onto
your hats.

First, in X.500 the DN (made up of a sequence of RDN AVAs, in which
some RDNs may have multiple AVAs -- e.g., residentialPerson) is the
conceptual equivalent of a "machine address" in the DIT.  A DN is
constrained to be chosen so that it uniquely designates exactly the
location in the DIT of exactly one entry which abstractly represents
one real world entity.  That is, you can use the DN as a key to locate
its unique entry by using a well defined algorithm to follow the DN,
using one RDN at a time, each in its sequential turn, down the DIT
from the ROOT.  By X.500 definition, no two real entities may have the
same DN.

Now, this is a new wrinkle in the art of computer science, where-in
the real world entry "owner" is required to bring his/her/its own DN
(machine-address), and further that he/she/it had better be certain to
bring a DN (machine-address) that is unique in the DIT, so no one else
will try to locate their entry in the same exact place.

This is where we come to the fork in the road between the REGISTRATION
MODEL and the LISTING MODEL.

In a REGISTRATION MODEL, the DIT operator assigns the RDN values,
arbitrarily, without regard for other uses in other contexts for the
same RDN name values.  This is fine within the DIT, if and only if it
is not a public DIT.

But, if the DIT is for holding public information, now we have to deal
with intellectual property rights with regard to the public use of
names.  Intellectual property rights in general derive from "fair
trade laws" of various kinds in various countries, but generally,
every country and every society has learned the same lessons from long
history, that we cannot allow any party to take over (appropriate) the
use of a name that has come into common public use by another party.
Every country thus has established (long ago) a civil naming structure
for people, places and things (entities) within its jurisdiction.

So, to answer your question: 

	"uh? What the hell is the French government doing here!"

I can only say, it is in here doing its proper job of regulating the
public use of names in France.  Who else might do this for you?

You can try all you wish, but in the end, in France, and in any other
country, you are going to have to come to grips with the national and
local civil naming infrastructure.

You can avoid this by keeping your X.500 DIT private, but it is going
to be hard to provide much of a public directory service with a
private directory that misuses the names of thers in public.

Now, lets take this up close and personal.

How will you (and INRIA) feel when some public DIT builder in France
decides that her pet cat should be given the DN of { c=FR o=INRIA
cn=Christian Huitema } for use in a public directory service (after
all, you say the French Government is busy with other more important
things than to concern itself with the names used by its citizens).

AHA! Your say -- She cannot do that because "I own that name space
within INRIA, and INRIA owns that name space within France!"

And, in response, she quotes your arguments in this discussion thread.
Now what do you do?

My suggestion is that you recant, and initiate a legal action to force
her to vacate her use of that DN for her cat, in your favor.  I expect
that INRIA would also want to do the same.

[Sorry to be the one to point out that you have gotten into a box!]

It is very clear to me that c=FR has a civil naming infrastructure,
and that you will find yourself using it sooner or later, regardless
of how much you would prefer to be free of it.

Actually, as you work your way around this problem, I think you will
soon come to appreciate the fact that all the localities in France are
already named, numbered, and registered, and that all companies and
legal organizations are already named and numbered, so you don't have
to invent it all from scratch.

In c=US, we find that it is wonderful that virtually everything is
already done (e.g., FIPS-55, state and local business registrations),
and all we need to do is use it.  Only a few hundred (most likely)
organizations are going to find that their natural civil name
registration is not adequate, and they have access to the c=US
register that is administered by ANSI to obtain c=US National Standing
for use in X.500.

Nuff for now.  Hang in there Weng!   Best...\Stef