Impact of Punycode

Shawn Steele Shawn.Steele at
Thu Mar 25 23:35:17 CET 2010

Yes, DNS doesn't have such a mechanism to extend the protocols.  That wasn't my point, my point was that, regardless of the reasoning, now we have tons of really strange cases that clients (and other kinds of servers) have to try to figure out themselves.

Some DNS servers also have strange cases with IDN.  IDN breaks AD DNS servers, which were using octets as permitted.  Perhaps some folks don’t consider those "real" DNS servers, however they are a scenario some folks now have to reconcile.


-----Original Message-----
From: John C Klensin [mailto:klensin at] 
Sent: Poʻahā, Malaki 25, 2010 3:05 PM
To: Shawn Steele; Vint Cerf
Cc: idna-update at
Subject: RE: Impact of Punycode

-On Thursday, 25 March, 2010 18:33 +0000 Shawn Steele <Shawn.Steele at> wrote:

> I am suggesting that for Email, the EAI approach is less problematic.  
> Yes, e-mail servers will need updated, but there will be less 
> long-term confusion.


Without in any way disagreeing with Patrik's or Vint's comments, let me also note that there is an extremely basic difference between the DNS situation and the email one.  With the latter, we are using a very carefully defined extension mechanism in which the client gets to find out if the server is prepared to handle certain types of strings (in this case, UTF-8 addresses)
and then to announce that such a string is coming.   If the
server doesn't like strings of that sort, it never sees them
(unless the client is very, very, broken).   The DNS permits no
such negotiation or authorization, which completely changes the problem.



More information about the Idna-update mailing list