Comments of A. Michael Froomkin
Professor of Law, University of Miami School of Law
Member of WIPO "Panel of Experts" in First WIPO Domain Name Process
froomkin@law.tm
http://www.law.tm
June 4, 2001
Washington, D.C
EXECUTIVE SUMMARY
1. WIPO's proposed expansion of the UDRP is
-
premature;
-
procedurally illegitimate and
-
substantively wrong.
It is premature to propose expanding the UDRP until a host of very substantial
problems with the current system have been solved.
2. It also is wrong to propose 'exclusions' that would remove
words from the namespace on the basis of an arbitrary set of criteria.
If this arbitrary list of factors is adopted, we can expect many others
to follow.
3. In many cases WIPO is proposing solutions to "problems" which it
is unable to prove exist, and which do not exist, or are so minor as to
fail to justify the highly intrusive proposals advanced by WIPO -- especially
when one considers the precedential effect these proposals could have.
4. WIPO is proposing regulations which vastly exceed the current international
consensus of the protection due to intellectual property. In so doing it
seeks to make de facto law in an undemocratic and illegitimate way.
It is striking that the nation whose laws would most frequently be undermined,
overruled, or ignored by the current WIPO proposals is the United States.
WIPO's proposals amount to little more than an attempt to impose European
intellectual property rules on the United States without the consent of
the US Congress and the US political process. To the extent that WIPO's
proposals reflect a central planning concept of the 'optimal' use of some
domain names, or the idea that domain name uses should be curtailed to
serve particular social policies, this reflects an orientation antithetical
to free market values.
WIPO's PROPOSALS ARE PREMATURE
5. It is premature to be talking about expanding the UDRP until we have
had a review of the functioning of the UDRP we already have.
-
The review was due last year, but is only now getting (slowly) under way.
-
There are a large number of legitimate questions about the UDRP which suggest
rather strongly that the UDRP is not a fair and legitimate way of resolving
domain name disputes.
PROBLEMS WITH CURRENT UDRP SYSTEM
6. The RFC proceeds under the quite mistaken assumption that the UDRP is
functioning well. Even if it may be true that "the UDRP has proven itself
to be an effective system for eradicating bad faith cybersquatting in respect
of trademarks" (para 225), WIPO fails to consider the extent to which
this result has been achieved at the expense of the legitimate rights of
non-trademark holders. It is striking that the decisions generally accepted
to be erroneous under the UDRP are almost uniformly drawn from those in
favor of trademark holders and against innocent registrants. The UDRP's
errors are indeed one-sided. Until and unless these are corrected, the
UDRP should not be expanded. Indeed, one might more reasonably ask whether
it should be abolished if these serious inequities cannot be corrected.
7. The UDRP's problems begin with lack of adequate notice for respondents
- the policy fails to require actual notice, or even reasonable attempts
to achieve actual notice. Time for response starts to run when notice is
SENT, not received. Study is required to determine to what extent the high
rate of defaults by respondents is due to lack of timely notice.
8. Some of the arbitral bodies have written rules that are manifestly
unfair to respondents.
-
Example: the NAF "sandbag rule" by which parties (i.e. complainants) can
pay extra and submit a supplemental pleading after the close of regular
pleadings -- including factual allegations to which the respondent then
has no opportunity to reply.
-
Example: the NAF policy by which if the respondent pays for a 3-member
panel, and the complainant then withdraws a complaint without prejudice
before panel action, the respondent's money is not refunded.
9. There is insufficient transparency about how arbitral bodies select
arbitrators for particular cases, and a lack of variety among the arbitrators,
too many of whom are chosen from the corporate trademark bar, and too few
from backgrounds that would tend, on average, to make them more sensitive
to the legitimate rights of indivduals and non-commercial users.
10. Since complainants choose the arbitral body, the current system
creates a financial incentive for arbitral bodies to compete to be 'complainant
friendly'. [How many press releases do arbitral bodies issue saying 'complainant
loses?' compared to 'complainant wins'?] This reasonable suspicion of partiality,
whether or not it actually exists, provides a reason why the impartiality
of the arbitral bodies might reasonably be questioned; as such it taints
the entire system.
11. Equally seriously, the UDRP was sold to the Internet community on
the premise that there would be at least a limited 'parity of appeal'.
In fact, this term was a double misnomer, in that even at its best there
was a subsequent judicial proceeding which was neither an "appeal" (being
as a legal matter a separate and original lawsuit rather than a true appeal
of the mere 'administrative process' in the UDRP) nor one with true parity--and
a recent court decision suggests the law is moving away from even that
rather low 'best'.
-
It was originally foreseen that the UDRP would allow a losing domain name
registrant ten days to file a transfer-blocking action in a court of competent
jurisdiction. By contrast, a losing trademark owner would of course have
had a much longer time to file; there was thus no real 'parity' at all.
And ten days is a ridiculously short time for an unrepresented party to
secure counsel and to file a complaint, especially if the relevant courthouse
is far from home.
-
It now transpires, however, that in the United States, even a losing domain
name registrant who surmounts this hurdle may not be able to secure a judicial
determination of his rights. The U.S. District Court in Massachusetts recently
held, in the Corinthians.com case, that a trademark holder who prevailed
in a UDRP case could prevent the court from ruling on the merits of a domain
name holder's attempt to overturn this outcome via a declaratory judgment
action by the simple stratagem of promising not to bring Lanham Act claims
against the original registrant. Upon this stipulation, the District Judge
dismissed the action for failing to state a claim - a result which (if
upheld by the1st Circuit court of appeals) would allow the UDRP domain
name transfer to go forward automatically without a judicial hearing on
the merits of the parties' claims.
-
This should come as no surprise. I warned against the problem of a potential
lack of a cause of action in my commentary on the report in the first WIPO
process. See http://www.law.miami.edu/~amf/commentary.htm#_1_27
.
PROBLEMS WITH WIPO'S CURRENT PROPOSALS
12. The first WIPO domain name process correctly relied on an important
principle: that WIPO's goal in this non-treaty, non-governmental, non-representative,
process should be to avoid attempting to create some species of new regulations,
unsanctioned by any democratic or legitimate representative process. It
would be a grave error to abandon this wise view. In particular, WIPO is
unwise to suggest in its second process that this bedrock principle could
be allowed to suffer "some adjustment" (paragraph 18). There can be and
should be no "adjustment" to the principles of democracy and legitimacy.
13. In contrast, it is heartening to see that WIPO continues to recognize
the importance of giving proper respect to "agreed rights outside the intellectual
property system" (paragraph 19). It is unfortunate, however, that WIPO
did not in fact give due weight to the importance of the democratic process.
Had it done so, it would not be entertaining suggestions that ICANN impose
preferences for various categories of businesses that conflict with the
delicately balanced intellectual property regime in force in the US, and
no doubt other countries as well.
INNs
14. We begin with the critical fact that WIPO has yet to make out even
a minimal case for the regulation of INNs: "Evidence of actual damage resulting
from the registration and use of INNs is lacking". (Paras 45, 55). Despite
this, WIPO argues for regulation! WIPO's approach amounts to the Alice
in Wonderland principle of "sentence first, verdict later". To which one
can only respond, "If it ain't broke, don't fix it."
15. It would be very unfortunate to embark on a policy of regulating
access to domain names via an exclusion mechanism based on their purported
content or on the consequences of presumed 'misuse' of a name. There are
currently a large number of potential non-commercial legitimate uses of
a domain name of the form 'medicine.tld' where 'medicine' is an INN and
'tld' is a TLD. For example, patients concerned about the side-effects
of a drug might band together to exchange information on a web-based discussion
board. (See, for example, caffeine.com.)
This exercise of communicative and associative freedoms would be completely
legal in the United States and no doubt elsewhere. WIPO should not impose
limits on it.
16. This rule would set a dangerous precedent. It is difficult to understand
why INNs are more deserving of preferential treatment than the names of,
say, major religions or churches, political movements, names of deities,
or a myriad other examples. To overlay the DNS with the regulatory agendas
of various interests would be to limit the simplicity and utility of the
DNS system for no substantial gain, and indeed much future wrangling and
pain.
17. Even if there were evidence of a problem, WIPO's entire approach
to the issue of the registration of domain names carrying character strings
leaves out the simple solution of creating more TLDs: Contrary to what
WIPO implies, registration of a name in one TLD by one party does not prevent
others from registering the same string, so long as they do so in another
TLD. As a result, no domain name registrant need have a monopoly on the
use of an INN (or any other string) in a second-level domain name. To the
extent that WIPO's concerns are that one party may secure an unfair commercial
or other advantage by having an attractive SLD comprising an INN, the proper
responses are
-
The case for why this sector should be exempt from first-come-first serve
rules that apply to everyone else is woefully unproven
-
The solution to this (and other similar problems) is to rapidly expand
the supply of attractive TLDs so that many SLDs in a given string become
available. Indeed, the more SLDs there are in a given string, the less
likely it becomes that anyone would erroneously impute semantic significance
to a given string. This solution also allows market forces rather then
regulation to determine registrations.
NAMES of INTERNATIONAL INTERGOVERNMENTAL ORGANIZATIONS
18. International intergovernmental organizations occupy a privileged position
in the DNS: they have their own TLD, .int. It is the height of special
and self-interested pleading for WIPO, an .int registrant, to demand additional
protections. This is nothing less than a transparent attempt by thin-skinned
international bodies unaccustomed to the rough and tumble of the democratic
process and popular debate to prevent aptly-named criticism sites from
being operated by protest groups. It is ignoble to suggest that these bodies
-- all of whom could and should be accommodated in their own bespoke TLD,
.int -- should have a greater protection against critical web sites than
anyone else.
PERSONAL NAMES
19. WIPO's analysis of the so-called jurisprudence relating to personal
names under the UDRP completely misses a key point. It was a commonplace
among all those who drafted the UDRP that personal names other than those
which were clearly trademarks were excluded from the UDRP. WIPO staff believed
this, as did all member of the WIPO Advisory Panel of Experts on which
I served. Yet, somehow, some arbitrators have managed to find that some
personal names are protected. The conclusion one reasonably draws from
this is not that personal names should therefore receive protection above
that already provided by the legal systems of the world, but rather that
either the UDRP is badly drafted, or the arbitration process is flawed,
or the arbitrators are badly selected, or all three.
20. In particular, panels have been sloppy in their acceptance of allegations
of claims of common law rights in names. It is not an exaggeration to term
this area an abuse of the UDRP.
21. United States law is quite clear at the federal level on the very
limited protection available to personal names, and in particular in the
requirement that a cybersquatter be seeking to sell a domain name in order
for the rights of the named person to be infringed.
22. It follows from the above that, first, WIPO should not propose rules
that contradict or exceed US federal law, and second, that the existing
UDRP process needs to be studied and reformed before it can be expanded.
GEOGRAPHIC DESIGNATIONS
23. The paucity of 'abuses' of the type that are commonly understood to
constitute cybersquatting in relation to geographic designations underscores
the extent to which the regulatory impulse here is not one designed to
combat bad faith (e.g. of someone using nation.net for a web site pretending
to be the official web site of the government of that nation), but rather
a desire to subject the name space in certain TLDs to a centrally directed
concept of optimization. This dirigiste conception is antithetical to the
free-market orientation that produced the Internet. It is particularly
upsetting to see a United Nations body opining that some supra-national
body should be empowered to decide when a web site is using a domain name
in a "worthwhile" way (paragraph 268). This is not a judgment that should
be left to governments, and still less to the unelected mandarins at WIPO,
or to arbitrators selected via a process equally lacking in democratic
legitimacy.
24. Indeed, it is striking that WIPO has been able to find only one
instance of a name in Annex XIII offered for sale.
25. To the extent that there are interests that seek to promote or express
themselves under a geographically identifying domain name but find that
they are blocked by someone else first in time, the solution is clear:
create more TLDs in order to broaden this opportunity. In particular, a
measured, consistent, predictable roll-out of new TLDs addresses the legitimate
'digital divide' issue noted in paragraph 267.
26. It is important to recall that the DNS, and especially the gTLDs,
is a private network. Names are registered by individuals who transact
with private registration bodies. It is antithetical to the free market
system to impose some externally generated requirement of optimum use on
these private arrangements. It is appropriate to police them for mis-use
(e.g. 'passing off'), but this is in all cases a judgment which can only
be made AFTER the registrant has made some use of a domain name -- it can
NEVER be made ex ante. Yet, this is precisely what WIPO seeks to establish.
27. Given that every country has a ccTLD which is fully subject to its
national law and regulation, it is hard to see what legitimate national
interests are served by giving nations or national adminstrative units
extra-territorial rights over registrations in gTLDs. The country's needs
for authoritative and reliable names on which its citizens and other can
justifiably rely should be fully served by the ccTLD.
TRADE NAMES
28. While as an abstract matter the case for protecting trade names in
the DNS shares many similarities with the case for protecting trademarks,
there are practical difficulties which suggest that achieving such protection
without great and undesirable side-effects might be very difficult. Indeed,
one of the chief lessons of the UDRP is that the lightweight rapid online
arbitral system does a poor job of sifting contested facts, and this is
rarely more evident than in the very poor job that panels have often done
with respects to assertions of unregistered (common law) trademark rights.
Lacking the ability and incentive to cross-examine parties or indeed seek
further and better particulars (it takes longer, the rules all but prohibit
it, and it doesn't pay extra), panels have tended to shoot from the hip,
with very uneven results.
29. In formulating any scheme for the protection of trade names, three
considerations should remain paramount. First, the extent to which additional
protection would solve an actual problem as opposed to merely providing
an attractive intellectual symmetry. The case here remains unproven. Second,
given that tradenames are frequently unregistered, by what means one might
avoid replicating and furthering abuses similar to those occasioned by
the assertion of alleged unregistered common law tradenames under the existing
UDRP. And, third, how to avoid breaching the principle of respecting
national law rather than supplanting it..
WHOIS
30. WIPO's discussion of privacy issues in the WHOIS fails to meet the
standard of thoroughness set by much of the rest of the report. In particular,
the report's very cursory treatment of existing privacy law fails to discuss
the relevant privacy laws of major legal systems (most notably the EU member
states and Canada), and the extent to which the current WHOIS policy currently
in use likely breaches them. This issue needs to be addressed before
even opening the issue of whether the WHOIS should be expanded.