Impact of Punycode

Vint Cerf vint at
Thu Mar 25 19:28:28 CET 2010

shawn, if you have a better strategy maybe it is timely to start another
BOF. In the past attempts to do IDN, the client-only approach has been
adopted because of the difficulty of changing all the servers and resolvers.
If enough simplification would result from such a mandate, perhaps it is at
least worth looking at fully. My recollection is that some people tried for
a new Class of DNS record other than IN because it could have different
properties (that would have to be known to the servers and the resolvers, of
course). That approach did not gain consensus.


On Thu, Mar 25, 2010 at 2:18 PM, Shawn Steele <Shawn.Steele at>wrote:

> > Nothing about IDNA200x should have any effect of any sort on any DNS
> server
> > that "handled UTF-8".  DNS labels are not restricted to ASCII, and never
> have
> > been: they're octets.  The LDH rule is not a "DNS restriction".  It's an
> effect of
> > the way the hostname syntax got used inside the DNS.
> IDN doesn't impact "DNS".  It impacts people using domain names.  Now the
> APPLICATIONS have to have domain knowledge about a protocol that means very
> little to the application.  For example, if you make an http request, the
> xn-- name can get into the http request.  Certainly the IDNAxxxx docs say
> nothing about http requests.  What's a web server to do if it gets a UTF-8
> request?  A Punycode request?
> In making things "simple" for DNS, we made things very hard for the rest of
> the system.
> Sort of way off topic for an IDNA list, and I'm not suggesting that we
> abandon IDNA, but I'd MUCH rather see the EAI approach than experience the
> same problems with other protocols.
> - Shawn
> _______________________________________________
> Idna-update mailing list
> Idna-update at
-------------- next part --------------
An HTML attachment was scrubbed...

More information about the Idna-update mailing list