Eszett again (was Re: Parsing the issues and finding a middle ground -- another attempt)

Wil Tan dready at
Thu Mar 5 02:35:52 CET 2009

I realized that this is slightly off-topic. If so, please accept my
apologies, and consider it a question for Marcos.

On Wed, Mar 4, 2009 at 3:26 PM, Marcos Sanz/Denic <sanz at> wrote:

> Breaking backwards compatibility is to my eyes the big stigma of IDNA2008.
> So:
> e) If mappings are to be removed from the standard, as we thought they
> were, then we fall back to our pre 2003 position, that is: we would like ß
> to be PVALID (and this is reflected by the current draft situation). Then
> there is no havoc anymore, it is up to us as a registry to deal with
> eszett, and we'll do it the right way.

What would be the right way for a registry to bundle the existing names such
that compatibility with IDN2003 can be maintained?

I'm not sure if denic solicited the "original" domain name (before ToASCII)
at the time of registration. For at least some registries who implemented
IDNs, only the punycode domain name is needed for registration (along with a
language tag.) Therefore, the original domain name, if it contained eszett,
would not be known to the registry.

Would these registries have to bundle all domains that contain sequences of
"ss" by "mapping it back" to eszett? So, the registrant for "grosses.tld"
gets ""großes.tld". For longer sequences of repeated "s", there are also
permutations of mapped and unmapped eszetts so multiple variations will have
to be generated and placed in a bundle. As a contrived example, "ssss" may
get "ßß", "ßss", "ssß", "sßs".

-------------- next part --------------
An HTML attachment was scrubbed...

More information about the Idna-update mailing list