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

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

Hmm.   After reading Eric's comments, I think I finally
understand the zone-scope problem and the associated
"requirements" issue.   And I have probably been part of the
source of confusion, for which I apologize.   Of course, I could
be wrong; I may still not understand it.

There are, I think, two rather separate "zone scope" problems:

(i) May different zones have different policies about what sorts
of names are permitted to be registered and how they are managed
within those zones?  I think that is (at least most of) the
issue Eric addresses below, certainly it is the issue if we are
getting into debates about what ICANN can or should try to
mandate.

My answer to that one is "yes, of course".  It has always been
that way and it should continue be that way, subject only to
constraints of operational (and interoperational) necessity.
And I would suggest that it is not an IETF problem, except
insofar as we need to provide information and guidelines about
the types of policies that would interfere with successful
global interoperation (see other note about engineering/policy
interactions).

Because I don't think of this as an IETF problem --and because
it seemed completely obvious to me-- it didn't occur to me that
this might be the meaning intended in the statement in the
requirements doc drafts.

(ii) May different zones adopt different ways to resolve or
interpret names, i.e., may they have local variations on the
protocols themselves?  My response to that one would be that the
DNS needs to work globally or it essentially does not work at
all. The combination of caching and cross-tree links makes it
almost impossible to localize a particular protocol variation
and the notion of having TLD-specific code cases in each
resolver is appalling (and something we could never get to work
reliably over time).  So we can't have, e.g., different
name-encoding (ACE or otherwise) strategies which are
interpreted depending on which TLD is in use, etc.  And _that_
is what I've been assuming we were talking about when we've
talked about zone-specific rules.

Was anyone else equally confused?   Does that help with the
confusion?  Is there any chance that we actually agree about
this?

    john


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

>>             ... The DNSO is responsible for determining the
>>             policy aspects of how those identifiers are used.
>> This implies that the IETF should strive to develop
>> policy-neutral engineering solutions.
> 
> Which leads me back to zone-scope, which existed in version 4
> of the Seng/ Wenzel draft as [35], and as [30] in versions 5,
> and 6 of that draft.
> 
> It is a policy to require zone-scope mechanism(s) not exist in
> the IETF work- product, and that policy promotes some local
> policy, possibly .com's to global scope.
> 
> It is a policy to require zone-scope mechanism(s) do exist in
> the IETF work- product, and that policy is neutral w.r.t.
> policies and scope.