[wg-c] Initial Numbers

[Milton Mueller <mueller@syr.edu>]

There has been some discussion about the choice of 6-10 as the
initial number of new TLDs. Some say it is too large for a
"test" and too small for robust competition. The latter
proposition is true, but is not a significant criticism if 6-10
is seen as the first step in a continuing process. The "too
large for a test" argument, however, lacks any supporting logic
that I can discern.

Kevin C's discussion of the topic was designed to create the
impression that the authorization of new TLDs is a step into
uncharted territory, a mysterious process fraught with
horrendous risk and irreversible consequences.

Absurd. We have been adding new TLDs to the root for ten years.
Here are some undisputed facts about the number added since
1994:
1994: 22
1995: 30
1996: 31
1997: 47
1998: 2
It should be noted that no restrictions were placed on the
business or administrative model of any of these TLDs, other
than the broad guidelines contained in RFC 1591.

Technically, adding a new TLD to the root means adding a few
lines of text with the character string and pointers to two name
servers. There are no technical issues whatsoever as long as the
number stays below one million, which it certainly will do.

Let us then move on to the question of whether there are
significant economic and policy risks posed by adding new TLDs.

First, let me concede that in purely administrative terms, there
is a limit to the number of applications that ICANN is able to
review, monitor, and approve. But since ICANN took only
two-three months to receive dozens of applications and accredit
the initial 5 testbed registrars, and has now accredited approx.
100, no one can reasonably contend that ICANN is incapable of
going through a similar process for registries. In this respect,
10 registries introduced over a six month period looks about
right.

What about economic policy considerations? Clearly, there are
uncertainties here but most of them argue for more, rather than
fewer, TLDs. From the standpoint of economic policy, one new
TLD, devoted to some relatively minor market segment such as
.mus, will tell us absolutely nothing about any trademark or
competition policy considerations that might be posed by the
addition of new TLDs. A .mus TLD might be lucky to get 500 new
registrations in its first year. What would we know at the end
of that period? Nothing new. We had similar experiences with the
addition of .int a few years ago.

Suppose we fail to add *any* generic TLD that could compete
directly with dot com, net and org (such as dot web, dot info).
Would we learn anything about the cross-elasticity of demand for
domain names? No. Would we have addressed the single biggest
reason to introduce new TLDs, which is to reduce the scarcity of
good names in the NSI gTLDs? No. Will we learn whether the UDRP
and other existing legal protections for trademark are
sufficient to prevent name speculation? No. Will we learn
whether expansion of the name space reduces the incentive for
name speculation? No.

No, the only effect of such drastically limited "testbed" will
be to prolong the artificial scarcity, which means that the
suppressed demand will build up even more, and make the prospect
of handing out authorizations for new TLDs that might compete
with dot com even more difficult and contentious. I can think of
no better demonstration of the utter bankruptcy of the entire
ICANN endeavor than an inability to confront the com registry
monopoly in the first round of additions.

And suppose we fail to add any for-profit registries, or any
non-profit registries. Will we learn whether for-profit or
non-profits are better, or whether there is any discerable
difference in their behavior? No.

Will the addition of one new, highly restricted TLD allow us to
know whether opening up the name space will encourage innovative
service development? No.

Whatever we do in the initial stage must yield enough variety to
provide some valuable insight into the policy and economic
impact of expanding the name space.

I see no feasible way for the the initial introduction to be
less than 9. We need at least 3 highly generic TLDs such as
.web, .zone, or .nom to compete with NSI, 3 policy-driven or
"chartered" TLDs to see how that works, and at least 3-4
designated for experimental new services or applications, such
as the E-164 proposal. The initial applications should also be
evenly balanced vis a vis for profit and nonprofit. This does
not even begin to address issues of cultural variation -- what
about TLDs in Spanish, German, Korean, etc?

--
m i l t o n   m u e l l e r // m u e l l e r @ s y r . e d u
syracuse university          http://istweb.syr.edu/~mueller/