Re: Who's problem? (was Re: [idn] Mixed TC/SC (was Re:Layer 2 and "idn identities")

[John C Klensin <klensin@jck.com>]

Eric,

I agree with much of this analysis, disagree with a bit, and
suggest that some qualification or explanation may be in order
for the others.  Let me focus on the latter two categories...

--On Friday, 07 December, 2001 08:28 -0500 Eric Brunner-Williams
in Portland Maine <brunner@nic-naa.net> wrote:

>...

>> > The only entity that can define what is and is not allowed
>> > in domain names (not host names) is ICANN and although they
>> > have doen it before, it doesn't happen often.
> 
> True for those operators that delegate this policy to ICANN.
> This includes gTLD (unrestricted and restricted), the .AU
> ccTLD operator. At the moment, I'm personally not aware of any
> other TLD operator that has delegated this policy to ICANN.

ICANN has three types of authority in this (and several other)
area(s).

	* One is contractual or otherwise formal, and that is
	what you are referring to here: A TLD (or, in principle,
	lower-level domain) agrees to delegate certain decisions
	to them, as you suggest.  The "agreement" may be
	voluntary or thrust upon them; makes little difference.
	
	* The second is moral and/or advisory.  ICANN can make
	recommendations to domains, registrars, or registries as
	to what they ought to do.  Those domains, registrars, or
	registries can follow the advice or not as they deem
	appropriate and useful.  It is worth remembering that
	anything the IETF does in terms of a standard is in this
	same category: we have no authority, other than
	persuasion, to force anyone to conform to a standard or
	to do anything else we recommend.
	
	* And the third is that, in principle, ICANN can remove
	an offensive party/ TLD from the root or redelegate a
	domain against the wishes of the administrative contact
	and relevant legal authorities.   I hope they never try
	this,  at least without very broad public support in
	advance, because I don't want to think about what would
	happen if they did.  But it is possible.

None of this differs from the status of IANA long before ICANN
came into being, except that IANA never resorted to a contract.
The tools were recommendations, education, persuasion, and the
implied threat of involuntary redelegation.

>> > If you want to continue down this path I suggest you submit
>> > a proposal to the DNSO and that it is not a path we wish to
>> > take in this wg.
> 
> Assuming that the scope of the policy sought includes the set
> of operators and their domains for which ICANN, and thus its
> DNSO, are the delegated policy formation bodies.

I am not aware of anything in the DNSO's charter that would
prevent their generating recommendations about best practices
that would apply to all domains, independent of where they could
enforce such recommendations.  Of course, there is a separate
problem, which is that the DNSO rarely demonstrates an ability
to agree on anything more sweeping than when to hold a NC
meeting :-(
 
> It is worth reminding ourselves from time to time, that some
> IANA root TLD operators (other than Verisign) have operational
> practices that differ from the current ICANN gTLD contract.
> ...

I agree completely. But that doesn't seem to have anything to do
with either IETF's scope, ICANN's scope, or whether
recommendations to domains ("registries") are possible.  In
principle, the IETF could make a recommendation in this area
too: we make operational recommendations fairly regularly.  But
such recommendations aren't in the IDN WG's charter, so anyone
wanting to generate such a document should probably start
charter-drafting.   And, without having consulted anyone on the
IESG about this, I'd guess that the threshold for convincing
them that the IETF actually should take on such work would be
quite high.  But, again, to suggest that we could not make such
a recommendation would be going a bit past the facts and history.

>> The responsibility and expertise to define domain and host
>> name formats lies with the IETF.
> 
> Assuming the ICANN MOU, s/IETF/PSO/, and of course we all hope
> that in this context, the ITU, W3C, and ETSI are
> non-contributory. 

Actually no, the IETF, subject to the comments above.  The PSO
would, as I understand both the MOU and current practice, get
involved only under one of two scenarios: (i) if there were a
dispute among the SDOs about the allocation mechanisms for
protocol parameters that could not be settled bilaterally or
(ii) if the DNSO or ASO or ICANN Board were about to generate a
policy or recommendation that might have technical impact.  It
would really be stretching the imagination to believe that the
formats of domain (or host) names were "protocol parameters".
And the PSO is carefully designed to avoid situations in which a
technical matter can be settled by majority vote.

> Again, ICANN itself is binding as the policy
> making body only where there is consent, currently limited to
> com/net/org/info/biz/pro/museum/coop/aero, and AU, again AFAIK.

And IETF itself is binding on nothing.  So?

> I appreciate the opportunity to have observed ICANN process
> first-hand at all three of its meetings this year, and to have
> observed CDNC and JET process as well. It has been educational.

See if you still believe that after a few more years of such
"education"  :-(

    john