[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
RE: Unique identifiers for LDAP attributes
I think the basic issue again comes down to discipline.. The OSI and OID
stuff was considered all too hard and long winded, registration, formality,
etc as opposed to quick fix - cut the code, see it works stuff..
X.500 defined 15 or so OCs and about 40 attributes as its base information
set... with OIDs we all know that.
So where we (in the directory world) build systems for online customers we
go through the formal schema design processes and where required, OIDs
registration under our national bodies. Just to ensure these orgs who spend
a lot of money on these systems - we are always concened that commercial
work does not get undermined by "an experimental document" out of
somewhere..
We cannot change history - but would should learn that the issue with
directories is not one of protocols - but one of information standardisation
and identification. Simply because commincation protocols can go through
gateways, but transforming information - is a real pain (synrax, semantic,
naming, etc.
If there are no engineering or doctrinal rules applied - how is it a
standards process?
XML will hit this issue .. but is anyone concerned? XML - a simple
Ebusiness solution.. seen that before eh!
using OIDs means that the author and the standards organisation under which
they are registered are taking some care - it isnt perfect but it does mean
existing systems should not be compromised.
regards alan
-----Original Message-----
From: Steven Legg [mailto:steven.legg@adacel.com.au]
Sent: Monday, July 17, 2000 1:33 PM
To: 'Mark Wahl'
Cc: ietf-ldapext@netscape.com
Subject: RE: Unique identifiers for LDAP attributes
Mark,
> -----Original Message-----
> From: wahl@austin.innosoft.com [mailto:wahl@austin.innosoft.com]On
> Behalf Of Mark Wahl
> Sent: Saturday, 15 July 2000 1:31
> To: d.w.chadwick@salford.ac.uk
> Cc: ietf-ldapext@netscape.com; Ramsay, Ron
> Subject: Re: Unique identifiers for LDAP attributes
>
>
>
> Its RFC 2252 section 4.2:
>
> > Schema developers MUST NOT create attribute definitions
> whose names
> > conflict with attributes defined for use with LDAP in existing
> > standards-track RFCs.
The problem is not that schema designers are using standard defined
attribute names in new definitions but rather that schema
designers are unwittingly using the same names in new definitions
where those new definitions are different (different OIDs or different
syntaxes or different equality matching, etc). And there is no
guarantee that an unused attribute name I choose today isn't going to be
commandeered by some future standard.
Regards,
Steven
Adacel Technologies
>
> Mark Wahl, Directory Architect, Service Provider/Infrastructure
> Sun Microsystems, Inc. iPlanet Alliance
>
>