On 8 Mar 2009, at 03:57, Måns Nilsson wrote:
I mean, that for the (previously alleged to be more "interesting wrt
IDN")
ccTLD part of the name space, ISO 3166, with its (at least actual if
not
formal) limitation to two characters from A-Z, should "fix" the
issue of
IDNA in TLD labels, leaving the gTLDs to break...
If only it could be that simple. There already are some IDN labels in
the root for test purposes: .xn--zckzah for instance is the Katakana
for "test" (テスト). For ccTLDs the direction that ICANN appears to
be heading probably means IDN labels for each country representing the
name of that country in the languages/scripts used there. The ISO 3166
list won't matter in that context. A government or ccTLD is likely to
go to ICANN and say "Here are the IDN strings that represent our
country name in our national languages/scripts. Put them in the
root." What happens for the existing gTLDs and sTLDs is anyone's
guess....
ICANN is looking to the IETF to come up with a deterministic way of
checking these IDN labels and provide confirmation that they don't do
any damage: like break the installed base or somehow destabilise the
DNS. Once the IETF has devised this process, ICANN will apply it to
any of the zillions of new TLDs they want to create.
I'm with Patrik on this one: those proposing new TLD labels should be
required to prove that they do no harm. However that's impractical
because it means proving a negative. [OTOH it would be a very
effective brake on the TLD free-for-all that ICANN seems to want.]
This approach is unlikely to be acceptable to ICANN. They'd like to
have a deterministic (automated?) method to validate applications for
new TLDs.
A possible compromise solution might be to come up with an algorithm
which filters out the obviously broken or dangerous TLD labels.
Assuming such a thing can be devised of course.... Then if a proposed
label passes that threshold, it goes out for wider consultation to
allow the world to object on technical grounds that the introduction
of the new TLD would break something. Which begs the question: what
happens when operational problems arise after the new TLD is introduced?
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
