At 04:36 11/01/2010, Arnt Gulbrandsen wrote:
Shane Kerr writes:
Various top-level domains are reserved by [RFC2606],
including "INVALID". The use of "INVALID" as a codified,
non-existent domain was considered. However:
o INVALID is poorly characterised from a DNS perspective in
[RFC2606]; that is, the specification that INVALID does not exist
as a Top Level Domain (TLD) is imprecise given the various
uses of the term TLD in policy forums;
Hm. Then why doesn't this document supersede 2606's imprecise
specification with a better one?
That is a decent suggestion, not sure how well it will be accepted.
We tried to be much more precise in what sink.arpa is than invalid
specification.
My suggestion would be to say that any NEW uses of ".invalid" as a
non existing name MUST use sink.arpa. As for old uses if possible they
should migrate to sink.arpa. there is no inter operability issue
that I can think of only lag in changing code.
As for ".invalid" its intended use seems to be more of a "documentation"
use than protocol use.
Note:
As for ".invalid" the definition seem to me be more of a document
".invalid" is intended for use in on-line construction of domain
names that are sure to be invalid and which it is obvious at a
glance are invalid.
"Obvious at a glance" implies human to me.
o the contents of the root zone are derived by interaction with many
inter-related policy-making bodies, whereas the
administrative and technical processes relating to the ARPA
zone are much more clearly defined in an IETF context;
That can be put that more clearly: "The IETF doesn't have sufficient
authority over the root zone to publish 2606 and ensure its
continued accuracy." My answer to that is that if so, then most of
2606 is broken, and it's necessary to much fix more than just the
paragraph that defines .invalid.
I prefer not to do 2606 rewrite, I can agree to update 2606 saying
".invalid" SHOULD NOT be used on the wire.
o the use of ARPA for purposes of operational infrastructure (and,
by inference, the explicit non-use of a particular name in ARPA)
is consistent with the purpose of that zone, as described in
[RFC3172].
Ie. if .invalid has to be dumped, the replacement should be in
.arpa. I can accept that. _If_ it has to be dumped.
Maybe .invalid was a bad choice in the first place. But that's water
under the bridge.
Arnt
__
Historical note: RFC2606 was done in hurry to get the names into document
before ICANN had really started to function.
Speculation: Even if the current ICANN is unlikely to allow wild card
in the root zone, that may change ==> ".invalid" suddenly exists.
IETF/IAB can prevent addition of wild card to .arpa.
Action: Next version of sink.arpa should say that .arpa MUST NOT
have a wild card as that will render sink.arpa. useless.
Olafur
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf