Proposed Charter for the IDNAbis Working Group

Martin Duerst duerst at
Thu Mar 27 11:47:14 CET 2008

At 01:17 08/03/27, Paul Hoffman wrote:
>At 7:08 AM -0400 3/26/08, Vint Cerf wrote:

>>particular, IDNs continue to use the "xn--" prefix and the same
>>ASCII-compatible encoding, and the bidirectional algorithm
>>follows the same basic design.
>This is not true at all. The bidi algorithm in IDNA2003 was all about single labels standing on their own; the proposed changes in IDNA200x is about making full names present better. We can remove the last clause completely: it is already covered in the bulleted list earlier.

I'm not sure I agree with Paul here. The goal in both cases was to
try and make sure there was a reasonable chance that users could
do a good job at parsing/keyboarding/... bidi domain names. We thought
that we could achieve the "make full names present better" goal with
restrictions on single labels, but we were not thorough enough.
But the basic approach, in particular the fact that mixing RTL and
LTR characters in the same label is forbidden, stays the same,
and most practical domain names (in particuar those containing
only letters) stay valid in IDNAbis if they are valid in IDNA2003
and vice versa.

I remember that there was a big push for keeping the specs restricted
to single labels, but I think that was not the main driver for the
details in the current algorithm.

Regards,    Martin.

#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#       mailto:duerst at     

More information about the Idna-update mailing list