Eszett and IDNAv2 vs IDNA2008

[Erik van der Poel]

On Thu, Mar 12, 2009 at 5:56 AM, Andrew Sullivan <ajs at shinkuro.com> wrote:
> If what you're saying is that IDNA2003 is around forever, then how do
> you deal with John's argument that, if you mean that, IDNAv2 won't
> solve the problem either.

If the rough consensus is that it is too late to solve the problem
(that Eszett and ss cannot resolve to two different IP addresses),
then IDNAv2 wouldn't solve that problem either (since we agreed that
it's too late to solve it). IDNAv2 might solve the other problem, i.e.
how to display Eszett when the resolved name had ss.

In the future, we might discover other mappings in IDNA2003 or IDNAv2
that we don't want, but presumably we'd say that it is too late to
solve those problems too.

So the WG needs to decide whether or not to keep the IDNA2003
mappings. If we decide to keep the mappings, we need to decide whether
they are for transition only or forever. And we need to decide whether
or not to add similar mappings for post-Unicode-3.2 characters.

Either way, we should keep the definition of U-label, so that SMTP
extensions can refer to that and restrict to that, if they wish. This
would probably mean that mappings are optional in the base spec.

Erik