[DNSOP] on private use TLDS

[Roy Arends <[e-mail]>]

“ZZ” was used in my presentation as an example. Since this bikeshedding is siphoning attention from the important part of the discussion, I’ll try to re-focus on the question here:

"Is it safe to use ISO3166-1 Alpha-2 code elements from the User Assigned range as top level domains for my own private use?"

1) RFC1591: Selection of 3166 was made with knowledge that ISO has a procedure for determining which entities should be and should not be on that list.
(Note of the word “entities” and “procedure”)

2) From the ISO 3166-1 Alpha-2 standard, on the construction of an alpha-2 code: "42 alpha-2 code elements are not used in the ISO 3166. Users sometimes need to extend or alter the use of country code elements for special purposes.” 

It is my understanding that the ISO3166 Maintenance Agency can not re-assign codes from the User Assigned range. This needs an action from ISO TC46. 

3) Jaap Akkerhuis writes:

> Shane Kerr writes:
> 
>> Certainly I have 
>> made heavy use of .Q* and .X* in my own testing, with the assumption 
>> that these would never be assigned (and yes, there is .TEST but 
>> sometimes you need more than one one TLD).
> 
> Yes, that is a perfectly legit use. 

For those who are not aware, Jaap is a member of the ISO-3166 Maintenance Agency and a Category C liaison for ICANN to TC46/WG2 [1].  Therefore it is likely that he knows these things. Note that the IETF has its own liaison to ISO-TC46 (John Klensin) [2]. Conversely, I (Roy Arends) make no representation for ICANN.

4) The XN— ACE prefix for internationalized domains was chosen by IANA, because "The use of ISO 3166-1 user-assigned elements removes the possibility that the code will duplicate a present or future ccTLD code.” [3]

5) ISO, ICAO, IATA, WIPO, UNLOCODE, UNICODE, Worldbank, Interpol, CABforum, and the IETF  make use of these User Assigned codes as intended by ISO3166. These codes have been assigned for different purposes that only have meaning in the local context of their organization and their constituencies.

6) It is highly unlikely that any of the 42 codes will ever appear in the root zone (especially following Postel’s Law: Be conservative in what you do), as we follow RFC1591. 

When a private space name is chosen, the 42 top-level labels are safe to use, as it is highly unlikely to collide with any top-level domain now or in the future. 

Please see the presentation during DNSOP at IETF106 here: 
https://youtu.be/yUO1J9s8BdE?t=5042

Follow the slides here:
https://datatracker.ietf.org/meeting/106/materials/slides-106-dnsop-sessa-draft-arends-private-use-tld

Warmly,

Roy Arends
Not representing ICANN.

[1] https://www.iso.org/organization/567449.html
[2] https://ietf.org/about/liaisons/
[3] https://psg.com/~randy/lists/iesg/2003/msg01081.html