Re: [dnsext] we need help to make names the same, was draft-yao-dnsext-identical-resolution-02 comment

[Lawrence Conroy <[e-mail]>]

On 16 Feb 2011, at 16:59, Andrew Sullivan wrote:

> On Wed, Feb 16, 2011 at 04:37:13PM +0200, Vaggelis Segredakis wrote:
>> John,
>> 
>> Please do not misunderstand the issue here. We did not ask for a wildcard in
>> DNS, we did not ask for millions of domain names to become one bundle.
>> 
>> We did however ask for the user's choice of registered domain names, the
>> names he thinks represent better the word he chose for registration to act
>> as a bundle.
> 
> As a matter of _protocol_, however, what you asked for devolves to
> "more than two".  Once we're at that point, the size of the set
> doesn't really matter except for the question of where the trade-off
> will be for performance reasons.
> 
> Moreover, you're not the only one asking.  This is certainly not a WG
> devoted to solving the problem of how to make things easier for the
> .gr registry.  Whatever we do has to be a significant improvement for
> many different kinds of uses, or it's not worth changing the protocol,
> even if that means you lose.  Sorry.
> 
> A

To which I reply:
  In principle, universal solutions trump narrow solutions that cover fewer use cases.
BUT ...
  So far, this thread has touched on/drifted towards VERY difficult problems, with
alleged combinatorial explosions and other evils.

Vaggelis seemed to suggest that those allegations might not be important in practice.
Almost all registries will have rules on bundling and mapping. Thus one may not need
to "solve for infinity" to reach a solution that covers many or even most cases.
His comment appeared reasonable (to me :), in rebutting allegations of impossibility.

If I can count the number of "shadows" with my fingers, that seems to ME to be a tad
different from millions of potential domains.

The requirements capture is running apace. It seemed that some were starting to claim
that the (quantum superposed) kitchen sink needed to be in there as well.
We need to know what this is trying to solve, but I'd be surprised if DNSng (or some
LDAP-based oracle) is needed, just to meet the real registry issues already mentioned.

all the best,
  Lawrence



_______________________________________________
dnsext mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/dnsext