Re: [wg-c] There is no "consensus"

[Jonathan Weinberg <weinberg@mail.msen.com>]

	Three thoughts:

	1. A couple of recent postings seem to assume that we are choosing between
authorizing 1) commercial/proprietary TLDs; 2) shared/nonprofit ones; or 3)
some of each.  That's wrong, though, because the issue of whether TLDs must
be {shared, unshared} and {for-profit, nonprofit) are separate ones.  ICANN
could well decide that: [1] all new TLDs must be shared, because it
benefits the consumer to have a choice of registrars; but [2] all new TLDs
need *not* be non-profit, b/c it is for-profit registries that will be best
motivated to offer good customer service, keep prices low, and otherwise do
good things by virtue of competing for business with other TLD registries.
Indeed, in a system with a substantial (but not huge) number of new TLDs,
at least some of which are general-purpose, it seems to me that that would
work pretty well.

	2. We've been caught up in an argument between the folks who argue that
ICANN should pick the new TLD strings and then issue RFPs for folks to
operate registries associated with those strings, and the folks who argue
that ICANN should select new registry administrators and then let *them*
pick the TLDs they think the market will support.  Javier urges that the
former is the better approach, because that way it is "the Internet
Community," rather than crass entrepreneurs, who will be doing the
selection.  I'm skeptical about this, since it seems to me that it won't be
the "Internet Community" doing the selecting — it'll be a few people,
starting with ourselves, who will be acting in the *name* of the Internet
community but surely will not be representative of it.  The ultimate
decision will be ICANN's, and I don't especially think ICANN is shaping up
as an especially good embodiment of the "Internet community."  Indeed, if
we want the "Internet community" to make the decision as to what names to
pick, the best way to accomplish that would be to get a whole lot of TLD
names into the root and then let each and every person registering a new
domain decide which ones he or she would patronize.  Some new ones would
thrive; others would stagnate; and it would be the Internet community that
decided.  Such an approach could easily accommodate both for-profit and
non-profit registries.

	Kent points out that only central planning — not the decentralized
decisionmaking I've described — can *design* the structure of the name
space.  I think this is an important point.  As I've suggested before,
though, a crucial part of the success of the Internet has been the extent
to which its planners pushed control away from the center to the edges.
Such experience as I've had looking at (and engaging in) government
planning causes me to doubt whether ICANN can really design the name space
better than can the marketplace choices of Internet users.

	I'm especially concerned about ICANN picking all of the new TLDs because
it seems to me that this is the approach that centralizes the greatest
degree of decisionmaking authority at the top.  I think it's important to
expand the name space -- but as ICANN takes its first few halting steps, I
don't think this is the time to give it any more decisionmaking power than
it has to have.

	3. A few people have urged that we need to keep the number of new gTLDs
small (say, 6 or 7) because trademark interests will lobby hard against a
larger number.  To the extent that this thinking is based on "practical
political reality," as opposed to the view that it would in fact be bad
policy to add more than a small number of new gTLDs, it may be misplaced.
Fact is, whatever number this WG may come up with, there will be folks from
the trademark community lobbying ICANN to cut it in half, because that's
where they see their interest.  Let's make a recommendation, if we can,
based on our vision of good policy; I can guarantee that "practical
political reality" will still get its due later in the process.

Jon


Jon Weinberg
Professor of Law, Wayne State University
weinberg@msen.com