Another Transition Plan Proposal

[Gervase Markham]

On 10/12/09 15:05, Cary Karp wrote:
> I do not believe that many TLD registry operators will agree that the
> registration of one element in a pair of labels such as "fuss" and
> "fuß", or "möller" and "møller", should, or even can, be made
> conditional on the registration of the other (bundling).

Although I respect your greater knowledge both of the space and of the 
people, I find that highly surprising, particularly "or even can". What 
is preventing such bundling? Nothing technical, surely. Only registry 
policy, and perhaps a determination to charge a flat price per single 
zone file entry. Of course, you can't make the registration of "fuss" 
conditional on the registration of "fuß" if you charge €19.99 for each. 
But if you charge €19.99 for both, suddenly it becomes entirely 
reasonable to make it conditional, or even to enforce it, and even 
without necessarily bothering to tell the registrant what you've done.

> The only way to know if the holder of a name containing ss would have
> preferred it to be an ß is by asking. Any registry that wishes to avoid
> the potential snags that we're talking about will therefore need to poll
> the holders of all ss names no matter what.

I don't agree. Why can't you just give them all an additional 
registration, for free, of the ß form, for every domain containing the 
string "ss", even those for which a human would say it is inappropriate?

> When doing so they can
> easily offer the holder of, say, fuss.tld privileged opportunity to
> register fuß.tld before the ß is added to the generally available
> repertoire (sunrise).

Would registries really prefer to administer such a scheme rather than 
just run a script over their zone file?

> Beyond that introductory action, the issues attaching to the maintenance
> of a single zone that contains names which differ only by the element of
> the ß/ss pair they include, become no different than those associated
> with ä/a/æ, ö/o/œ, ü/u/ue, é/e/è or any other of the many situations
> where such clustering might be considered.

I entirely agree with that. Which makes me surprised that you think 
there's some question about whether such bundling can be done.

Gerv