RE: [idn] Domain names and ASCII compatibility

[Paul Hoffman / IMC <phoffman@imc.org>]

>I would venture that even after several decades such approches
>will have decoding glitches.  Those WILL lead to complaints
>from users that will see that as errors.

Two things here:
- If our goal is to reduced the decoding glitches, then fully-opaque 
downcasing is probably better than partially-opaque because it will put 
that much more pressure on software that shows users the host names to do 
so correctly.
- Complaints from users who see errors will be of less severity than 
complaints from users who cannot access an Internet resource. That is, if 
they can't send mail or can't use a secure server or can't manage a host 
because of the IDN protocol, those complaints will be louder than those of 
folks who errantly saw the downcased name.

>Please target a transition to pure UTF-8 (with fallback names
>selected by a human, for use during the transition).

If we had specific proposals for how to move every protocol that cannot 
handle UTF-8 today to being UTF-8-friendly, we could look at this option 
and compare it to the other options. But we don't. Making up a name for 
solution for one of the protocols affected without even presenting an 
Internet Draft, much less an RFC agreed on by the experts on that topic, is 
not sufficient.

The WG charter at <http://www.ietf.org/html.charters/idn-charter.html> says:

"A fundamental requirement in this work is to not disturb the current use 
and operation of the domain name system..."

The term "current use" seems pretty clear, and unilaterally breaking the 
implementations of  many fundamental Internet protocols could be considered 
a significant disturbance. Similarly, "operation" seems pretty clear, and 
saying that everyone who wants an internationalized domain name must in 
fact get a second "fallback" name for use for an undisclosed period of time 
could be considered a pretty significant disturbance.


--Paul Hoffman, Director
--Internet Mail Consortium