IDNs in the root

[Simon Josefsson]

Harald Tveit Alvestrand <harald at alvestrand.no> writes:

> Paul Hoffman skrev:
>> At 4:55 AM +0100 1/23/09, Harald Tveit Alvestrand wrote:
>>   
>>> For one thing, that makes "IDNAv2" likely to be finished well after IDNs have been introduced in the root.
>>>     
>>
>> Serious question: why is this at all relevant? Is there any IDNs that
>> are meant to be entered in the root that will have a different
>> encoding under IDNA2003 than under IDNA2008 or contain characters
>> that make them unrenderable in IDNA2003?
>>
>>   
> There are many strings for which registration could be requested under 
> IDNA2003 that cannot be registered under IDNA2008 (the symbols being the 
> most prominent example). ICANN is aghast at the idea of having to allow 
> something to be registered in the root and then having to take it out 
> when the rules change.

The ICANN can specify a list of acceptable code points, can they not?

Given how most if not all TLDs have implemented IDNA2003, this is
already common wisdom of how to introduce IDNs in a zone.  I would be
surprised if such a list isn't discussed already.

For stability, I would suggest ICANN limits the permitted code points to
those where the encoding under IDNA2003 and the current IDNA2008
proposal are identical for all strings containing the code points.  One
could also test for compatibility with IDNAv2, although given Paul's
design approach I'm not sure that will reduce the set of permitted code
points.

If a good list of acceptable code points is developed by ICANN, I share
Paul's question why this is relevant for TLD IDNs.

> I haven't seen anyone claiming that they want to register a TLD in 
> Dhivehi or Yiddish (the two BIDI cases where it matters that 2003 is 
> more restrictive than 2008). But there are people arguing for 
> registration of a TLD with a ZWNJ in it.

There seems to be security problems with such a zone, if you consider
IDN strings in TLS certificates and Kerberos realms etc.

> With the relatively restrictive rules that ICANN has put in place, it's 
> not likely that any problems will be caused - but one reason for ICANN 
> having to specify the restrictive rules in full rather than saying 
> "stuff legal under 2008 is OK" and adding some short list of 
> restrictions is that 2008 is not finished.

I believe they need similar rules anyway.  Saying that any string legal
under IDNA2008 would be OK to register as a TLD IDN seems like a good
way to damage the Internet name space.

/Simon