Re: A long strange trip

[Carl Oppedahl <carl@oppedahl.com>]

At 03:39 AM 12/24/96 -0800, Jim Dueltgen wrote:

>The Draft Spec of the 19th is lousy.  It failed to significantly
>incorporate ideas and concerns raised on this list and others in the last
>few months and succeeded in making the IAHC appear to be in the hip pocket
>of "registrars for existing gTLDs."  I think we pretty much need to start
>from scratch.

[...]

>Q: What is driving the need for additional TLDs?  A: The commercial market.
>Who should decide what those new gTLDs should be?  The IAHC?  No, same
>answer as the first question.  

I am sorry but this view is mistaken.  It ignores what in micro-economic
terms is called "the tragedy of the commons".  I hope and trust that the
majority of participants in this group are familiar with it from first-year
economics in college.  Anyone who does not instantly recognize and
understand that phrase needs to go back and re-read (or, if necessary,
read) their economics textbook.  The commercial market is incapable of
making wise decisions about resources that have social costs that are
distinct from the individual benefits.  

Let's try an analogy in some other area.  Back when 800 numbers were
running out, I suppose "the commercial market" could have gone rummaging
through the list of not-yet-allocated NNX's to find not only 888 but also
others such as 999 and 666.  And "the commercial market" could simply have
appropriated those NNX's and started using them.  (Of course this would
only work if all the local telcos could be convinced to add a few entries
to their routing tables, etc.)  Anyway, next thing you know every telephone
number in the North American Numbering Plan is gone.

Or let's try the area that's proposed.  Suppose "the commercial market"
decides to create one hundred thousand new gTLDs.  One one million.
Suppose everyone who now has a COM domain decides to create their own new
gTLD.  That's right, starting now, my email address will be carl@oppedahl
rather than carl@oppedahl.com .  The TLD lookup table expands from ten
entries to one million.  Everybody from Exxon down to the corner liquor
store decides it would like to have its own top-level domain.

I promise you this would be chaos.  Anarchy.  

Oh, and the trademark battles?  Somebody would register the top-level
domain "kodak" to sell it to Kodak.  People would kick and claw each other
to register top-level domains of sex, realestate, houses, hotels, airfares,
and so on through the fifty thousand or so most commonly used English
words.  And if you think the lawyers are all over the place in fights over
second-level domains, just you wait until there would be similar but
bigger-scaled fights over top-level domains.

>Shared domains are not a good idea for the Internet.  I appreciate the
>idealism of the description of registries, registrars and stewards and I
>understand that it is a system that works for 800 numbers, but when this is
>presented as a preservation of the "public trust..."  Sheesh.  All we have
>done here is come up with an abstraction layer to the problem under the
>guise of protecting the "public trust" which doesn't exist.  And then, who
>watches the watchmen?  This is not a "useful market control" if such a
>thing exists.  If there is ultimately a fixed price resource (in this case
>whatever it costs the registrar to register a name with the steward) and a
>fixed number of shared gTLDs then there are a finite number of ways for
>registrars to differentiate their services and acheive profit.  

It works now with 800 and 888 numbers.  It can work with COM domains.  The
databases are of comparable size, actually.

>And make no
>mistake, if there is no profit to be had there are no new gTLDs.

As I say above, "profit" is not the right way to decide things when the
tragedy of the commons rears its head.

>TLDs should be awarded to any organization on the following objective
>technical criteria:  redundant connectivity to the Internet (either two
>physically distant sites, both geographically and in regards to network
>connection, and/or multi-homing) aggregated to at least 3 MBps, functioning
>root servers and a functioning registry database with a whois interface.

Wrong.  If your criteria were used, there would be hundreds of thousands of
top-level domains in the span of a few weeks, as the Oklahoma land rush
surged.  Any one hardware owner who satisfies your criteria could then
attempt to create TLDs for every word in the dictionary.  It would be chaos.

>The IAHC Draft section on Choosing gTLD Registrars only makes a business
>case for large corporations to participate.  There is very little incentive
>for an entrepreneur to be involved.  Most particularly because of the high
>entrance fee, the unpredictable nature of a lottery, and the uneven playing
>field with regard to NSI until 1998 and quite liekely beyond.  This model
>only makes sense for large utility companies that can throw $20,000 away to
>buy a seat on the Council of Registrars and is certainly no guarantee of
>competence or desire to deliver superior product or service.

I agree the playing field is slanted.  NSF somehow gave NSI a plum, and NSI
is squeezing the plum for all it's worth (and making terrible decisions in
domain name trademark matters as well).  When the NSI contract expires, NSI
has said it intends to keep running COM answering to no one.  The Internet
community will be best served if *everyone* keeps clearly in mind,
continuously from now till 1998, the importance of eliminating NSI's monopoly.

>And finally, the trademark issue.  Registrars, registries, ISPs, IAHCs,
>IANA's and anyone else who participates in assigning a mnemonic to an IP
>address should refrain from making any statement about copyright.
>Copyright, like every other piece of information on the net should be the
>responsibility of the data's owner, not the carrier or any administrative
>party.  Any disputes that arise should be handled by entities outside of
>the administrative process of assigning names to IP numbers.  Resolutions
>should not involve the registries except to cancel one registration and
>start another.

Trademark, you mean, not copyright.  Oddly enough, assuming I understand
your recommendation on trademark issues, you seem to agree with the IAHC,
as well as with the International Trademark Association
<http://plaza.interport.net/inta/intaprop.htm> and almost everyone else
(except NSI) who has expressed a view on the matter.