non-IDNA LDH Label ?

[Nicolas Williams]

On Sun, Jul 18, 2010 at 05:42:12PM -0700, =JeffH wrote:
> Is it correct to define the notion of "non-IDNA LDH Labels" as the
> union of "NR-LDH labels" and "Fake A-labels" ?

I believe the answer is "no" because one should not register fake
A-labels.  However, this is a "soft" no because there's no strong
requirement to not allow registration of fake A-labels.

In the long run I'd expect that there will be apps and/or libraries that
will produce warnings, and even errors when fed fake A-labels, even if
that's not today either required nor recommented, and even if it were
explicitly not allowed.  Which means: we're all best off not allowing
the registration of fake A-labels in any DNS zones, and renaming any
where they exist.

> In other words, do "Fake A-labels" exist in the wild (whether or not
> their creation was purposeful or inadvertant) ?

It's possible.  It's always been accepted that the IDNA prefix (xn--)
could collide with existing labels.  It's just not very likely.

Nico
--