Comments on draft-ietf-idnabis-defs-10

John C Klensin klensin at jck.com
Tue Sep 1 21:58:57 CEST 2009



--On Tuesday, September 01, 2009 12:08 -0400 Andrew Sullivan
<ajs at shinkuro.com> wrote:

> On Tue, Sep 01, 2009 at 11:10:27AM -0400, Vint Cerf wrote:
>> andrew, the problem with that last point is that the two
>> labels will   match in DNS but produce different U-label on
>> conversion. I think that   is not a good outcome.
>> downcasing would solve that wouldn't it?
> 
> I am really reluctant to add a normative downcasing step
> anywhere, because it's a very late change to the protocol.  I
> can't think of anything it would break, but I didn't think of
> this problem before, either, and mostly nobody else did
> either.  I'd have grave concerns about adding another
> normative step to the protocol during WGLC, and then sending
> the result on to the IESG.

Agreed.
 
> The important thing we have exposed in this discussion is that
> there are _several_ possible matches in the DNS to a given
> valid U-label, but exactly one of those candidate DNS matches
> is itself a valid A-label.  That suggests that saying, "making
> sure you downcase before converting the U-label is probably a
> good idea," may be enough -- roughly what John suggested, I
> think, but I'm not totally sure.  We first need to resolve the
> discussion of uppercase/lowercase form from Punycode.

What I'm suggesting, refined by reading a lot of additional
messages and getting a little sleep, is either the above or the
nearly equivalent "make sure that the string is already in lower
case before converting to a U-label".   Neither results in a
change to the protocol; they are just recommendations of an
extra validation and precaution against nonsense.

   john



More information about the Idna-update mailing list