Incompetance abounds at the InterNIC

Let's for a moment assume you're argument that real (i.e. non-RFC1918)
addresses are a requirement. Let's also assume some new companies wish
to join these private internets which use real addresses.

Just how exactly are those new companies to get real addresses? They
can't get them from IANA. I suppose they could pay an ISP for a block of
addresses and then not actually connect them.

That is exactly how we told them to get addresses in my previous job.
If they wanted to use a non-RFC1918 address (and many did) we required
that they be legitimate addresses. Several connecting customers had
randomly chosen network blocks and we refused those connections, even
if the block hadn't actually been assigned to anyone yet (had one of
those cases, too). Most of these were resolved with NAT and a few
actually learned they needed to go get their own address space.

We also did administer RFC1918 space, and accepted no connections on
any of these addresses that had already be allocated to someone else
based on our allocation records.

Perhaps you've come up with a good and workable reason to deploy IPv6.

I'd love to see IPv6 get rolling, personally. I do know that a _LOT_
still has to be done with various programs to make them IPv6 ready.
I was discussing plans just last night with a friend to set up an IPv6
tunnel between our homes and addressing some technical issues in how
to connect 2 tunneled LANs that were both homed on dynamic addresses.

Ultimately ARIN is an agency for assigning IP addresses for use in the
PUBLIC Internet. If you choose to use those addresses for other
purposes, I don't see that as a problem the folks managing the public
Internet need be concerned with.
Similarly the InterNIC is an agency for handing out namespace for use in
the PUBLIC Internet. If you wish to use their service for private use,
and you pay your money, there's no harm in that. But, you'll be expected
to play by the rules. That means name servers and valid contact info. It
doesn't mean you have to have a working web server. Name servers need
only serve SOA info. That, at least, provides evidence the domain is
valid and in use, and gives the contact information for anyone wishing
to challenge an organization's right to use a particular domain name.

The rule may say there needs to be a name server, but tell my why we
even need such a rule? The "validity" of a domain name is an issue
only between the registrar and registrant.

I do agree with the valid contact info.

As thhis whole thread of discussion really centers around private use of
internetworking technology, I have to wonder how it affects operational
issues (which just happens to be the subject of the mailing list).

I think it got started with domain names without name servers that
speculators were registering, and I mentioned the analogy of private
address usage, which then resulted in a confusion between private
"internal" addresses and private "external" addresses.

It _was_ an operational issue at one time, but it doesn't really resemble
it anymore unless someone wants to fuss about not being able to exercise
his God given right to traceroute to any IP address he pleases.

I think a lot of good points have been put forth by you, Phil, as well as
by Dean.

The problem is that I'm not sure NetSol cares.

I've seen a lot of registrations for *legitimate* companies -- backbone
providers at that -- with obviously bogus phone numbers like (123) 555-1212
and obviously bogus addresses. What is NSI willing to do about it? Nothing.

Of course, the agreement you sign with them says they will revoke domains
with bogus contact info. Yeah. Right.

(cc'd to Mr. Gomes at the InterNIC because I really want to hear why they
refuse to uphold their end of their contract)