Re: [wg-c] Who should vote for new gTLDs

[Kent Crispin <kent@songbird.com>]

On Tue, Jul 27, 1999 at 11:04:13PM -0400, Jonathan Weinberg wrote:
> Kent --
> 
> 	We've got two sets of issues here.  One is politics; the other is policy.
> Both are worth discussing.
> 
> 	As for the politics — I agree with you that a variety of trademark folk,
> in the White and Green Paper comments, urged that seven new TLDs were too
> many.  You may not be able to "document" that large TM interests lobbied
> the USG directly, but I was working on domain-name issues for the U.S.
> government back in 1997 (I was a scholar in residence at the FCC then), and
> I remember that lobbying well.  Lots of folk will push, as hard as they
> can, the position that ICANN should add as few new TLDs as possible, as
> slowly as possible.  So I think that convincing ICANN to authorize a lot of
> new TLDs will be a major challenge.  I don't think we should abandon that
> idea, though, if we think it's good policy.

I am not at all against large number of gTLDs.  I am against
introducing a whole bunch all at once, as our opening shot.

> 	And I think it is good policy.  You argue that it isn't, because the
> addition of a lot of new gTLDs will impose "incalculable" costs on
> trademark owners (or, alternatively, "billions of dollars in legal bills").

There are *many* arguments against it being good policy...this is 
only one.

To be a little clearer, I meant the "incalculable" literally, not as
hyperbole -- we simply can't calculate the magnitude of the cost.  It
certainly won't be zero, but how high it could go is unknown, within
very broad limits.  When one is entering unknown territory, it is
very seldom good policy to rush in with all guns blazing... 

In fact, of course, pushing *any* policy with potential large scale
cost is fundamentally insane, unless there is some strong
compensating gain.  The question is, what is the *gain* of
introducing a whole bunch of gTLDs all at once, vs the *gain* of
introducing gTLDs in a slow, controlled manner?

In general, one could argue that 1) we will get more consumer choice,
quicker; and 2) the names we get will be the ones that people "really
want", because the market is the best mechanism to determine what
people "really want".

As far as 1) is concerned, I agree that 100 names will give more new
instant choices than 7.  In practical terms, however, I don't believe
having the other 93 names immediately available will add that much
social utility, relative to adding names in a more measured way.  In
fact, I think that there is a fair possibility that adding 100 names
will simply strengthen the relative value of .com, and actually
decrease effective competition. 

And I think the advantage you claim for 2) is largely an ideological
illusion -- the situation is much more complicated, and the
advantages and disadvantages much more subtle.  Assuming a DNSO
approval process such as the WG model I described earlier, any name
or names that people "really want" can be proposed, and if they are
indeed appealing, and don't conflict with something, they will likely
be approved.  Therefore, names that people "really want" will get in
the root under either method.  An approval process, however, will
weed out likely losers sooner, which is a real social utility.  (The
competition you describe indeed will have winners, the names that
people "really like".  But the flip side is that if competition is
working correctly there will also be a set of losers, and what do you
do with them: TLD names that are not popular; the registry folds; and
then what? And please be aware that the simple answer is not a good
one.)

You further argue that in addition to lots of new gTLDs we should
have lots of registries, and the competition will result in the usual
benefits of lower cost and innovation.  There are multiple counters
to that, for example: 1) plenty of competition takes place at the
registrar level; 2) the registry function is just a back office
function, and frankly doesn't need much optimization; 3) using the
careful definitions I constructed, competition is more effective at
the *registry operator* level, not the registry level.  Registry
operators are contractors that can be replaced at will by the
registry administrator, etc.

Here are some other arguments as to why it is bad policy.  I have 
mentioned these before, and I apologize for mentioning them again, 
but I don't think you have adequately addressed them:

1) the TLD space is a single name space, and gTLDs are only one
possible inhabitant of that space.  That space must be administered
as a whole, and therefore the gTLD names must be considered in
conjunction with other policy considerations. 

2) sudden introduction of a large number of names with a free-for-all
name selection process seriously hampers structuring the name space


>  Near as I can tell, though, the effects on trademark owners of the
> addition of new gTLDs have been vastly overblown.

I am reminded of the saying that everybody should drive strapped to
the front bumper of their car, so they can fully experience their own
errors in judgment.  :-) It costs you little to say "as near as you can
tell" -- it's all academic and ideals to you, but it is dollars and
cents to the ones who will be affected.

> Bear in mind that we
> will have an ADR procedure for cybersquatting.  We may have a "famous
> marks" procedure as well.

Please note that the same people who argue for large numbers or new 
gTLDs are frequently the same ones who argue against the WIPO DRP.

> I see no reason to believe that the additional
> trademark policing costs attributable to new TLDs will be nearly as
> overwhelming as some of the numbers I've seen thrown around.  And I don't
> think we should deform our DNS policy simply to minimize trademark policing
> costs.

"Deform" is a fairly prejudicial term, don't you think?.  Relative to
what ideal? How do you justify that idea? TM policing costs are part
and parcel of this -- hence the push for ADR, famous marks
exclusions, etc.

> As David Maher said in a post to this list a while back: "The
> Internet is much bigger than the ‘marketplace.' We are serving everyone now
> on the Internet and those who will be.  There are other interests at least
> as important if not even more important than IP."

Of course.  But 1) addressing their concerns about rapid introduction
of gTLDs hardly seems like a bad thing to do, and it costs us very
little; and 2) there are other reasons why the policy is bad.

-- 
Kent Crispin                               "Do good, and you'll be
kent@songbird.com                           lonesome." -- Mark Twain