Re: [wg-c-1] WORK: Question #1 New GTLDs

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

Kent --

	I'm glad we agree more than I thought.  On ICANN's role in the
introduction of new TLDs, I can imagine a continuum of possible solutions.
At one pole might be an approach in which ICANN opens up the root to new
registries without paying any attention at all to the name or content of
the new TLDs.  At the other pole might be a system in which ICANN itself
(perhaps utilizing a WG process) announces the names of all the new TLDs,
the order in which they will be admitted to root, and who will be allowed
to get SLDs in each.  Other solutions might lie in between those two poles
— for example, an approach in which ICANN selects new registries, and those
registries have wide discretion as to what names they will choose, but
subject to an ICANN veto if the name violates certain rules.

	My thought is that we should figure out what policy interests ICANN needs
to vindicate in introducing new gTLDs, and then choose the approach that
vindicates those interests with the *least* degree of central control.
Granting the central decision-maker more policy authority than that, it
seems to me, will yield worse substantive results, run contrary to the
spirit of the Internet, and (incidentally) unnecessarily slow down the
introduction of new gTLDs.

	You wrote that we must introduce initial new gTLDs slowly because "[w]e
don't know how policy mechanisms will respond to the introduction of new
gTLDs."  I just want to make sure I understand — you're talking about
trademark, right?  That is, I gather, the argument here is that we must
introduce new gTLDs slowly so that we can see whether their introduction
leads to trademark problems.  Are there any other policy mechanisms we need
to worry about here?

	On IP rights in TLDs, I can't help but get the feeling that we're arguing
over very little.  You've suggested that to avoid legal complications in
ICANN enforcement of its rules, a TLD registry should be required to waive
any trademark rights in the TLD string.  I've got no problems with such a
waiver, but I've suggested that there are other legal mechanisms that would
achieve the same result (including a contractual provision in which the
registry agrees not to assert any such claimed rights as a defense to
ICANN's enforcement of its rules.  That would also be a waiver, albeit a
more limited one.)  However we resolve this, I'm confident that the problem
can be solved — that this issue, by itself, should not be a bar to getting
any registry online.

Jon


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



At 09:32 AM 7/12/99 -0700, Kent Crispin wrote:
>On Mon, Jul 12, 1999 at 11:15:56AM -0400, Jonathan Weinberg wrote:
>> 	Kent writes that: [1] if we introduce new general-purpose TLDs, then, at
>> least some of the time, firms will simply re-register their existing .com
>> names in the new TLDs;
>
>Actually, I was making exactly the reverse point, in support of your 
>position...
>
>> [2] we need more central planning of the name space,
>> so as to prevent new registries from picking unacceptable TLDs such as .usa
>> or unused ISO-3166 names; [3] "the first few TLDs *must* be introduced in a
>> very very controlled fashion"; and [4] registries' claimed trademark rights
>> in their TLDs will vastly complicate ICANN's ability to enforce its rules.
>> (I apologize for the paraphrasing, but I find it really hard to read
>> interlined comments at the third level and beyond; I've copied Kent's
>> entire message below).
>> 
>> I agree that, some of the time, firms will re-register their existing .com
>> names in new general-purpose TLDs.  NSI, as Kent points out, is encouraging
>> people to do just that in .net and .org. 
>
>No -- I was pointing out exactly the reverse -- that many (probably
>most) companies do *not* register in all three of .com, .net, and
>.org.  New gTLDs will almost certainly continue this trend.  I was 
>agreeing with your point, as I think you will see if you carefully 
>read what I wrote.
>
>> 	I also agree that ICANN should be able to impose *some* limitations on
>> possible TLD strings.
>
>Which implies, therefore, that ICANN must have a procedure for 
>approval of TLD strings.
>
>>  Someone could do a lot of mischief will .usgov, say.
>>  Existing trademark laws are more than sufficient to prevent unauthorized
>> folks from seeking to run .ibm, but it makes sense to me that ICANN could
>> have a role with respect to TLDs that are more generally deceptive, and
>> thus raise consumer protection issues.  I can also understand why ICANN
>> might seek to impose limitations on .fuck, say, or .godhatesfags (though
>> Lord knows that sort of thing has been a headache for NSI).  It's a long
>> way, though, from there to the view that the name space must unfold
>> according to a central plan.
>
>As far as I know, no one at this point is suggesting a central plan,
>and I'm not sure where you got that idea.  Instead, what has been
>proposed is a *process* whereby new gTLD names are justified prior to
>inclusion in the root.  The process I have advocated in the past is
>that a proponent of a gTLD name should make the case in the
>appropriate public forum (the DNSO, now).  That is, TLD *names* must 
>go through a public approval process.
>
>This means, therefore, that a registry can't say "I have a secret 
>TLD name that I don't want to reveal because it will impact my 
>business plan, but you can examine my infrastructure for technical 
>correctness." 
>
>In other words, ICANN approves *names* as well as registries.
>
>> We can address these concerns in a much more
>> limited way, simply by promulgating a few rules describing categories of
>> off-limits TLDs, and designating a decisionmaker whose authority is limited
>> to the question whether a proposed name violates those rules.
>
>That means that the designated decisionmaker would have to know that
>(hypothetically) "bax" is a deadly religious insult in some language. 
>A public process won't catch *all* such problems, either, but an
>extended period of public review and debate over a name is certainly
>going to catch a lot more than any single decision maker would. 
>
>Perhaps more important, though: If there is anything that the DNS
>wars have made absolutely clear, it is that public review and debate
>is the nature of this game. 
>
>> it means.  Is this to say that new TLDs should be introduced more slowly
>> than operational considerations would otherwise dictate?  
>
>Depends on what you mean by "operational considerations".  It is a 
>general character of operations that when you introduce something 
>new and untested, you do it in a controlled and careful way.
>
>> Why, exactly,
>> should that be so?  Because otherwise INTA will scuttle everything?  (Is
>> that, in fact, true?)  Or is the argument (as Kevin urges) that the first
>> new TLD should be one with only limited appeal, to minimize operational
>> challenges?  Or is the point something else?
>
>The point is something else, much more basic:  The fundamental 
>operational uncertaintity in this matter is not technical, but 
>policy.  We don't know how policy mechanisms will respond to the 
>introduction of new gTLDs.  It is this uncertainty that dictates 
>slow introduction.
>
>> 	Finally, as for "independent intellectual property claims on the TLD
>> name": Kent raises the issue of ICANN's enforcement powers if a TLD
>> registry refuses to follow ICANN rules.  As Chris points out, though, the
>> complete answer to this question lies in the contract between ICANN and the
>> registry operator.  That contract should provide for an agreed-upon
>> procedure to resolve such disputes, and the registry should agree in the
>> contract that the outcome of that procedure trumps any asserted TM rights.
>> If there's more to this problem, I don't see it (and I've been a lawyer for
>> a long time).
>
>That's the "argument from authority", right?:-)
>
>Interestingly enough, these contract clauses you describe are an
>example precisely the kind of "signing over of rights" I was
>describing.  But there is legal reality, and there is practical
>reality... 
>
>Suppose IBM proposes a .ibm TLD, and we say "Of course.  You just 
>sign this contract that says that if there is a dispute over the 
>TLD, you agree that the result of the dispute resolution trumps your 
>IP rights in .ibm."  IBM says "Of course."
>
>Now, let us compare the legal budgets of ICANN and IBM...
>
>In my opinion it would be smart to avoid, to the extent possible, all
>such legal conflicts, *especially* in the beginning.  The best way I
>can think of to do this is to just require that TLDs be the moral
>equivalent of public domain (I realize that "public domain" has
>specific legal meaning -- what I would prefer is something like the
>GNU "copyleft", but I don't know the appropriate legal terminology.)
>
>-- 
>Kent Crispin                               "Do good, and you'll be
>kent@songbird.com                           lonesome." -- Mark Twain
>
>