non-IDNA LDH Label ?

John C Klensin klensin at jck.com
Mon Jul 19 18:41:44 CEST 2010



--On Monday, July 19, 2010 16:20 +0000 Shawn Steele
<Shawn.Steele at microsoft.com> wrote:

> I'd expect some "fake" a-labels due to the disparity of
> allowed code points between 2003 and 2008.  Presumably some
> labels will be fake by one or the other.

As a set of concrete examples, any label containing a symbol or
punctuation character (most of which were valid under IDNA2003)
becomes a "Fake A-label" with IDNA2008 because it cannot be
obtained by applying the Punycode algorithm to a valid U-label.
Although it was against ICANN (and other) advice, several zones
have permitted such things to be registered.   

The list would get much longer if zones registered things that
were dubious, but not invalid for lookup, under IDNA2003.  An
example of that would be a label that included characters new to
Unicode 4.0 or 5.0 for which someone has extrapolated Nameprep
handling.  I suspect that has been done but, unlike the symbol
case, I can't immediately point to specific examples.

      john






More information about the Idna-update mailing list