Reply to Reply to still on omittedORAddressComponent

Kevin Jordan <Kevin.E.Jordan@cdc.com> Fri, 28 April 1995 13:31 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa02547; 28 Apr 95 9:31 EDT
Received: from CNRI.Reston.VA.US by IETF.CNRI.Reston.VA.US id aa02543; 28 Apr 95 9:31 EDT
Received: from mercury91.udev.cdc.com by CNRI.Reston.VA.US id aa05854; 28 Apr 95 9:31 EDT
Received: by mercury.udev.cdc.com; Fri, 28 Apr 95 08:16:31 -0500
X-From: kej@mercury.udev.cdc.com Fri Apr 28 08:16 CDT 1995
Received: from localhost by mercury.udev.cdc.com; Fri, 28 Apr 95 08:16:25 -0500
To: Peter Cowen <p.cowen@nexor.co.uk>
cc: mhs-ds@mercury.udev.cdc.com
Subject: Reply to Reply to still on omittedORAddressComponent
In-reply-to: Your message of "Fri, 28 Apr 95 11:08:07 BST" <5636.799063687@nexor.co.uk>
Date: Fri, 28 Apr 1995 08:16:23 -0500
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Kevin Jordan <Kevin.E.Jordan@cdc.com>
Message-Id: <2fa0eaa9671d002@mercury.udev.cdc.com>

> P.S.  The OID for at-associated-or-addr has changed from previous
> versions (was at 1 now at 6) presumably because its changed syntax
> from mhs-or-address to distinguished name,
> 
> Why then couldn't the oids oc-rfc822-to-x400-mapping and
> oc_x400-to-rfc822-mapping be changed ?
> The object classes they represent have changed with the dropping of
> the nonAuthoritativeAssociatedORAddress and
> nonAuthoritativeAssociatedDomain attributes and with the change in the
> syntax and OID of the associatedORAddress attribute. 
> 
> This make it a pain to upgrade. Every DSA that has a definition of the
> object classes has to be changed to show the new one. You can't run
> the two concurrently, as they have the same OID. 

Since the OID of associatedORAddress has changed, it is therefore a new
attribute.  So, the object class lost one old attribute and gained a new
one.  We accommodated this in our implementation by updating the object
class definition simply to include the new associatedORAddress attribute.
To avoid the external name conflict, we actually named this
associatedX400Address.  This allows the implementation to be backward
compatible with its older revisions, and it also allows it to interoperate
with up-to-date revisions.  It can also receive slave updates from up-to-date
implementations, and it can send slave updates as long as the replicates
don't include obsolete attributes.  As there really wasn't any mapping
information in the Open Tree previously, this isn't much of an issue.

Now that the documents are RFC's, any future revisions to object class and
attribute definitions absolutely must be accomplished by assigning new OID's
instead of "recycling" existing ones.