Andrew Sullivan ajs at
Wed Aug 13 19:40:51 CEST 2014

On Wed, Aug 13, 2014 at 05:27:33PM +0000, Shawn Steele wrote:

> really a problem.  (Changing the behavior is).  There are numerous

Changing the behaviour _later_ is going to be a bigger problem.  And
that, I think, is one of the points that got all of this recent
discussion started.  The experience with IDNA2003 was that it needed
some changes.  We made those changes, with full IETF consensus, in
IDNA2008, and yet we heard in response that a whole bunch of things
couldn't possibly change given what was already on the Internet.  In
some ways, UTS #46 says that: it doesn't provide a way to move from
IDNA2003 to IDNA 2008, but instead offers a different state that
doesn't really permit a number of the code points that IDNA2008 was
specifically designed to permit.

Best regards,


Andrew Sullivan
ajs at

More information about the Idna-update mailing list