TINA DAM- Blog on IDNs- Paris ICANN meeting.

[bwhhisc]

http://blog.icann.org/?p=334

June 26th, 2008 by Tina Dam-
One of the main IDN questions asked by end-users in the last few months, and that have been discussed during the ICANN Paris meeting in the recent week is as follows:

“If I have registered <domainname>.tld, then how will you ensure that I am also the registrant of <domainname>.<idn-tld>, for all languages.”

The question shows that there is an expectation that:

(i) there is a way to translate the .tld into other languages. Having done that with .test I can assure you it is quite a challenge to find a word that is an adequate translation for all users in a community. Often there is more than one way to express the word “test” in various languages. Some existing TLDs might be easier to translate than others, but common for them is that they could be represented several ways within one language.

(ii) that the registry operator for the .tld will apply to become the operator for such new TLD(s), and that if they doso and are successful in their application, that they will implement .tld with an aliasing functionality where registrants under .tld automatically becomes registrants of the same domain names under .idn-tld

On the gTLD side of things:

a. the GNSO policy for introduction of new gTLDs states that there is not precedence for becoming an operator of an IDN TLD. In other words, just because you are operating a TLD today it does not mean that you automatically become the operator for any translated version of that TLD (being IDN or ASCII, but mostly discussed in relation to IDNs).

b. in the process for introduction of new gTLDs there are various objection procedures available. While they are not implemented completely yet you might imagine that the .tld registry operator might object to someone else applying for the IDN version the .tld.

c. The policy does not provide a global direction to registration policy regulations, such as for example whether or not new TLDs should be aliased to an existing TLD.

On the ccTLD side, the situation is similar:

a. the IDNC WG final report does not talk about this specific topic.

b. based on community discussions during the last few months, it could be anticipated some IDN ccTLD will be operated as aliased versions of the existing ccTLDs, and others will not. The decision is usually referred to a difference in opinion on whether there should be IP protection or more competition and choice.

On the technical side of things:

a. aliasing have often been connected to the concept of DNAME. DNAME have been initially tested, and indications are that it will not be useful to provide the aliasing functionality. ICANN is looking into the opportunity for having more tests done on this topic.

b. without a standard way of implementing aliasing the concern is that aliasing will be implemented in many different ways leaving users confused and a need to further education than currently is needed, which could be avoided.

In Summary: There is no guarantee to the registrants, it depends on whether existing registry operators will apply for the IDN version of the their TLDs; that the application is granted; and that they will implement aliasing as their registration policy, which we currently do not have a technical standard for, and which the policies are not providing global direction upon.

END QUOTE

[Jay]

Yes, bad news for those hoping for .com aliasing. But no surprise for those who have been closely following ICANN/GNSO's statements on this for the last six months.

[bwhhisc]

Originally Posted by Jay

Yes, bad news for those hoping for .com aliasing. But no surprise for those who have been closely following ICANN/GNSO's statements on this for the last six months.

Not so sure its all bad, just no firm roadmap or answers. Looks like each registrar will have to apply for the equivalent idn gtld, or cctld, and sort out the aliasing or mirroring of the .idn. She indicates that some may not choose to do so. I don't think that will be the case seeing as Verisign authored DNAME, and JPRS has put a proposal for aliasing out probably less than a month ago. On the positive, ICANN again states that the existing registries like .com, .jp etc are protected for their equivalents in various foreign languages.

QUOTE: "in the process for introduction of new gTLDs there are various objection procedures available. While they are not implemented completely yet you might imagine that the .tld registry operator might object to someone else applying for the IDN version the .tld."

[domainguru]

Originally Posted by bwhhisc

http://blog.icann.org/?p=334

a. the GNSO policy for introduction of new gTLDs states that there is not precedence for becoming an operator of an IDN TLD. In other words, just because you are operating a TLD today it does not mean that you automatically become the operator for any translated version of that TLD (being IDN or ASCII, but mostly discussed in relation to IDNs).

b. in the process for introduction of new gTLDs there are various objection procedures available. While they are not implemented completely yet you might imagine that the .tld registry operator might object to someone else applying for the IDN version the .tld.

Great find Bill - finally a decent "summary" from ICANN. (a) and (b) are the most important for me - they pretty clearly show that nobody has a cat in hell's chance of wrestling .com away from VeriSign in any major languages ....

As for aliasing, yeah, seems to me ICANN need to sort through lots of issues before they commit to any "aliasing standards" - so don't expect any new versions of .com in the next two years, that's for sure.

[Rubber Duck]

Verisign introduced IDN without the introduction of ICANN standards.

I think it is quite likely that they will introduce aliasing without formalised ICANN standards. It is not like they cannot easily change from one aliasing method to another. This isn't even really an end user issue.

Yes, I agree. If Verisign can reasonably demonstate that a Unicode form is used to express "dot com" in a foreign language, then it already has their name on it.

[domainguru]

Originally Posted by Rubber Duck

Verisign introduced IDN without the introduction of ICANN standards.

I think it is quite likely that they will introduce aliasing without formalised ICANN standards. It is not like they cannot easily change from one aliasing method to another. This isn't even really an end user issue.

Yes, I agree. If Verisign can reasonably demonstate that a Unicode form is used to express "dot com" in a foreign language, then it already has their name on it.

yeah but idns at the second level could never be said to have a chance of destabilizing the internet, whereas TLD aliases implemented badly do have some chance of messing up the DNS. I think Tina's answer was good, but still more clarity needed to explain why DNAME apparently doesn't work as an aliasing mechanism. Who decided that, when, and why?

[bwhhisc]

Originally Posted by domainguru

. I think Tina's answer was good, but still more clarity needed to explain why DNAME apparently doesn't work as an aliasing mechanism. Who decided that, when, and why?

I asked her that straight up on an earlier ICANN Blog but so far have received no answer. Seems a bit odd that if DNAME was not even an option, the registries would be advised and JPRS would not have authored this document last month (May 08')

https://www.centr.org/main/4329-CTR/...hmentData/data

[Rubber Duck]

Originally Posted by bwhhisc

I asked her that straight up on an earlier ICANN Blog but so far have received no answer. Seems a bit odd that if DNAME was not even an option, the registries would be advised and JPRS would not have authored this document last month (May 08')

https://www.centr.org/main/4329-CTR/...hmentData/data

Yes, but DNAMES can be implemented at the Registry servers, not just the ICANN Root.

Whilst she is twittering on about standardisation of approach others will be implementing DNAMEs or some other form of aliasing.

China has already done at hack at ISP level. This would be much cleaner for them at Registry Level.

If China does it then what is stop JPRS or indeed anyone else?

Of course none of them can do any aliasing until such time at the IDN TLDs go into the Root, which I believe will be end of 2008.