Comments on draft-ietf-idnabis-defs-10

[John C Klensin]

--On Monday, August 31, 2009 14:17 -0400 Vint Cerf
<vint at google.com> wrote:

> andrew,
> 
> can it be argued that the only way a domain name containing an
> xn-  label could be formally registered in the DNS with upper
> case ASCII   present would be through violation of the
> IDNA2008 protocol specs? In   a strange way, while one would
> actually find and match such a domain   name because of DNS
> rules, that object, if returned with its upper   case
> components, would fail to convert to a proper U-label.
> Somewhere   in here we might want to say that such an object
> (an A-label with   upper ascii characters in it MUST not be
> registered).
> 
> does that help? we already say something like that by
> definitions I   think.

We say it by saying that nothing but [valid] A-labels can be
registered and that strings that cannot be obtained by
conversion from U-labels are not [valid] A-labels.   ("valid" in
the previous sentence is completely redundant and added only for
emphasis.)

I suspect that, in light of this discussion, we may need to make
that more clear, but that we need do little else.  I don't know
whether it would also be desirable to add specific language
permitting lower-casing of the strings; I'd like to see Lookup
explicitly banned from sending A-labels with upper case
characters off to the DNS.

    john