> > The only thing that I recall having said about "smaller" gTLDs is
> >
> > * monopoly gTLDs are probably acceptable where the number of
> > probably registrants is too small to make running a shared
> > registry economically viable
>
> The CORE SRS is designed to handle any number of TLDs; the incremental
> cost of adding a TLD is zero; the incremental cost of running a
> registry for a very lightly used TLD is zero. Therefore, there is no
> such case as you describe -- for *any* new TLD, regardless of the number
> of probable registrants, it will be cheaper to run it in the SRS than it
> would be to create a tiny new monopoly registry.
You assume that there is one and only one right set of policies for
registries. I assume that any set of policies created out of thin
air, based just on theorizing, are likely to be faulty. That is, I
assume that CORE's policies will be flawed. Furthermore, I assume
that variety and competition are necessary for improvement in policies.
This does not strike me as a hot issue. The other approach was tried
and failed utterly in Eastern Europe, the CIS, etc.
In any case, I know perfectly well that at least in Europe a
no-competition policy will not be acceptable. There will be competition
in gTLDs -- if there are any new gTLDs at all.
There will be zero or several competing gTLD registries in our future.
I have already said that in my opinion most of these should follow the
shared registry model that Nominet exemplifies.
Given all this, it seems utterly pointless to lecture me on
the importance of the no competing registries/no monopoly
registries dogma.
Talk about how to conduct experiment with monopoly registries. Talk
about how to set things up so that if a monopoly registry goes bad
it is possible to recover from it. This would make much more sense
then just crying anathema! unbeliever! from the rooftops.
-- Jim Dixon VBCnet GB Ltd http://www.vbc.net tel +44 117 929 1316 fax +44 117 927 2015