RE: [wg-c] Initial Numbers

["Roeland M.J. Meyer" <rmeyer@mhsc.com>]

[To the list:
I apologize for my last post yesterday, it was supposed to be delayed until
0100 this morning. I can only blame that failure on Outlook2K. This should
be my only post today. - RMJM]

> Behalf Of Kent Crispin
> Sent: Wednesday, December 15, 1999 7:30 PM
>
> On Wed, Dec 15, 1999 at 07:05:16PM -0800, William X. Walsh wrote:
> [...]
> > > the problem with permitting a free-for-all, with no
> constraints, is that
> > > that is what you usually get.
> >
> > And what is funny, is that no one is supporting a total
> free-for-all with no constraints,
> > but your arguments in opposition to them is always from the
> premise that they are.
>
> Could you list for us the constraints that you support?

http://www.dnso.net/library/dnso-tld.mhsc-position.shtml

> Would you be
> willing to hazard a guess as to the constraints Milton supports?

http://www.dnso.net/library/dnso-tld.mhsc-position.shtml

The process that I envision would take some serious effort to add each TLD.
There is a lot of validation, cross-checking, and testing involved. I expect
two to three months of work per TLD, maybe four. It is also largely a manual
process with some parts inherently non-automatable, as they deal with
meat-world processes. At this point the process is self-constraining and
de-scales directly with (un)available resources.

Please note that the full spectrum of services, that I require a registry to
provide, is all value-add stuff and NOT fluff for the sake of loading the
registry. The this also increases the vetting load (therefore increasing
schedule run-time) and radically reduces the rate at which new TLDs can be
added. It also adds a higher uniform level of services that will make the
Internet more user-friendly, in the long-run.

> I asked earlier if you might tell us what your view of the "purpose of
> the testbed" might be.  I still would like to see some discussion
> of that...

I agree with Karl, for new TLDs, a test-bed is unnecessary. However,
multi-part root-system support needs some work, mainly in confidence
testing. I was surprised at Peter Deutch's experiment with adding the COM
zone as TLDs. Not so much at the results, but that someone actually went and
did it. It was an obvious, and elegant, way to prove the point, I wish I'd
thought of it <g>.