browse comments: WIPO RFC-1

WIPO RFC-1
mueller@syr.edu
Mon, 24 Aug 1998 10:04:07 -0400

Browse by: [ date ][ subject ][ author ]
Next message: christian.bock@ipi.ch: "WIPO RFC-1"
Previous message: ANG_Lee_Chen@MLAW.gov.sg: "WIPO REQUEST FOR COMMENTS (WIPO RFC - 1) WIPO PRESS RELEASE"

Attachment: http://wipo2.wipo.int/dns_attachments/attach904053846.doc

File notes: Posted on behalf of the submitter

WIPO RFC 1: Comments of Dr. Milton Mueller

These comments are submitted by Dr. Milton L. Mueller, Associate Professor and Director, Graduate Program in Telecommunications and Network Management, Syracuse University School of Information Studies, USA. I have published several reports and scholarly studies on telecommunication policy and Internet governance issues (see references
at end).

Let me begin by applauding WIPO's willingness to conduct an open proceeding and especially its willingness to entertain comments on its
terms of reference. The Internet domain naming process affects the
interests of many parties, most of them outside the traditional purview
of the intellectual property law profession. It would be miraculous if
WIPO was able to define the terms of reference regarding such a complex
issue to everyone's satisfaction the first time around.

I believe that the terms of reference will require substantial
modification if the proceeding is to realize its goals.

1. Documenting the problem.
Perhaps the most disturbing aspect of the RFC is its failure to call for
factual submissions documenting the nature and scope of the problem. The
RFC seems determined to ignore the effectiveness of litigation under
existing trademark law. The RFC language openly assumes that some method
"other than litigation" is required to resolve domain name
disputes. (4.A.Dispute resolution. Section a) If WIPO expects its
proceeding and its conclusions to be taken seriously by the Internet
community, it must be prepared to prove this assertion. It must document
both the existence of a problem and its size and significance.

The RFC's assumption that methods other than litigation are required can
be easily challenged. My empirical study of trademark-domain name
disputes projected, based on statistical evidence, that of the 2 million
or so domain names NSI registered, actual infringement cases constitute
a very small number-about 257 cases in the open, generic TLDs. This
constitutes only 0.013 percent of all NSI-registered domains. The study
proved that litigation has provided an effective and rapid solution to
real trademark infringement problems. (See
http://istweb.syr.edu/~mueller/study.html) When a domain name is used as
part of a pattern of infringing behavior, trademark owners have won
their legal challenges 100% of the time. Trademark owners also have a
perfect record against name speculators. No court in any jurisdiction
has ever upheld the right of a name speculator to hold on to a
trademarked name when challenged by the trademark owner. This has
happened even in cases where the trademark owners' already had several
domain names corresponding to their trademarks, and when the trademark
owners' claim was dubious, as for example British Telecom's claim that
the domain "bt.org" violated its mark.

Most cases of speculative registration of trademarked names occurred in
1996 or 1997, in an environment of legal ambiguity. As one legal
precedent after another favoring trademark owners "right" to domain
names has been delivered, it becomes increasingly clear that no one is
going to make serious money from the resale of trademarked names. Going
forward, it is far from self-evident that name speculation constitutes a
significant problem.

Another fact ignored by the RFC is the existence of commercial services,
such as those offered by NetNames of the UK, that coordinate the efforts
of trademark holders to control or retrieve the domain name character
strings they desire. The NetNames service uses both economic and legal
means available under the status quo to attempt to reserve clients'
names across all TLDs. As an opponent of expansive trademark rights in
domain names, I do not approve of all of Netnames' tactics. The point,
however, is that it represents yet another way in which trademark
holders' concerns about domain names are being addressed.

No medium is ever going to be entirely free of trademark
infringement. The basic issue before us is whether trademark
infringement via domain names is so pervasive and costly, and current
methods of addressing that problem so inadequate, that major forms of
institutional innovation are required. So far, neither WIPO's RFC nor
the trademark community at large have succeeded in making that case.

2. Uniformity.
The RFC language seems to assume that uniformity and centralization of
dispute resolution are desirable, and that the only problem is to work
out a method of defining and achieving it. I wish to question both the
desirability and the practicality of a developing a uniform approach to
the resolution of disputes between trademark holders and domain name
owners.

Uniformity is not desirable. A uniform approach is advocated by large
corporations with significant brand and trademark holdings because it
reduces the cost of challenging domain names. Under a regime of
centralized uniformity, trademark holders can allocate their policing
resources to one venue, operate according to one set of rules, and
intervene quickly, even before a domain name goes into effect. Such
uniformity and speed might be desirable if the entire world had already
agreed upon a single set of criteria for resolving domain name-trademark
interactions. It has not come to such an agreement, however. This fact
is incontrovertible. Some systems of law privilege freedom of expression
over commercial intellectual property while others do not even recognize
a right to free expression. Some jurisdictions may recognize common law
rights to domain name use while others have no common law tradition at
all. If WIPO wants uniformity, it must begin with the harmonization of
law across nations, and work from there to link administrative
processes such as domain name registration to the accepted law. The
process cannot move in the opposite direction.

In the absence of any real uniformity in the legal principles used to
resolve disputes, the only accomplishment of mandating uniformity at the
registry level will be to make it exceptionally easy for trademark
holders to challenge domain names. And this is precisely what many
Internet users fear about a centralized, uniform process: trademark
holders will use it to assert property rights in character strings that
resemble or conform to their trademarks, regardless of whether any
actual infringement is taking place. In other words, it sets up a
process that is intrinsically biased toward parties with a trademark. A
uniform dispute resolution procedure imposed at the point of
registration acts as a kind of "prior restraint." Domain name
registrants are presumed guilty of trademark infringement until proven
innocent. From our experience with the infamous NSI dispute resolution
policy, we know that any such bias can be and will be exploited by
aggressive trademark lawyers. Trademark owners who wish to assert broad
claims to character strings can delay the registration process or impose
other administrative and documentation costs upon registrants. This is
unacceptable to any advocate of a broad, open Internet.

Another important fact ignored by the RFC is that administrative dispute
resolution policies can cause conflict and litigation rather than
preventing it. There is little serious dispute about the fact that NSI's
policy, by privileging trademark rights over prior registration,
encouraged trademark owners to challenge names even when concurrent use
would have been justified. Furthermore, the ensuing disputes between
registrants and NSI often resulted in litigation.

Uniformity is not a practical option at this time. Quite apart from the
normative issue, the feasibility of a uniform approach can be
questioned. The ability to impose uniform methods for resolving domain
name disputes depends entirely upon the economic model that is
ultimately adopted for the management of the name space.

If proprietary, competing TLDs are selected as the economic model, the
dispute resolution procedure in any given TLD will be determined by
competitive differentiation in the market. Some TLD managers may bend
over backwards to accommodate trademark owners, in an attempt to develop
good will and attract high-margin businesses. Others may choose to
maximize the ease and volume of registration, and wash their hands of
all trademark issues. On the other hand, if the open, nondiscriminatory
name space advocated by parties such as Pgmedia prevails, the very
operation of the root would become more decentralized, and proposals for
uniform and centralized vetting of names would become impractical.

The existence of country code TLDs poses a particularly thorny practical
problem for any advocate of uniformity. There are over 200 ccTLDs. They
can be run by anyone from Islamic fundamentalists to subcontracted
private entrepreneurs who purchased the right from a country that was
solely concerned with making money out of the deal. Many governments
consider the policies of ccTLDs to be a matter of national
sovereignty. Insofar as ccTLDs are officially designated bodies of
nations that are signatories to the WIPO treaties, they may choose to
conform to a uniform procedure WIPO develops. But there is nothing
requiring them to do so. This becomes even more problematical because a
growing number of ccTLDs are starting to "act like" commercial generic
TLDs.

Finally, regardless of which economic model is adopted for registries,
the IFWP meetings have consistently rejected calls for a centralized,
uniform process. In other words, the clear consensus of the Internet
community is in favor of hetergeneity in dispute resolution.

In short, the issue of uniformity is probably moot.

3. Famous names in Generic TLDs, addition of new Generic TLDs. Your
terms of reference separated these two items but I am lumping them
together. In fact, both are extensions of a more fundamental issue that
the RFC fails to identify and define properly. That issue is the
interaction between the supply and demand for names, on the one hand,
and the way in which Internet users actually make use of domain names,
on the other. WIPO examines expansion of the namespace almost entirely
from the standpoint of its effect on intellectual property rights. This
is understandable and perhaps even appropriate given its mandate. But
trademark rights are not the only issue to consider in the expansion of
the namespace.

Most of the trademark owners' panic over the alleged threat domain
naming poses to their marks stems from two sources: 1) failure to
understand the nature of the name space as an economic resource, and 2)
fundamental misconceptions about how much influence domain names will
have on users once the name space is enlarged. I cannot, in this
context, prove or even provide much analytical support for that
assertion; to do so would take this comment far beyond its appropriate
bounds. Let me simply suggest, instead, that the terms of reference
under your sections B and C incorporate additional items requesting
comments and factual evidence regarding the economics of the name space
and user behavior regarding names.

Specifically, I propose to add to the terms of reference the following
items under B (Process for the Protection of Famous Marks in Generic
Top-Level Domains):
i. Whether the mere possession of a famous name in any TLD is sufficient
to attract a significant number of users to a site;
ii. Whether the ability of famous-name character strings to attract
users to a site is a permanent feature of Internet domain name use, or
merely a temporary feature of the inexperience of users, the immaturity
of the medium, and the artificial restrictions on the name space
currently in force;
iii. What is the estimated economic value of the domain names that would
be pre-empted by granting famous name holders special rights to all
character strings similar to their names across all TLDs?
iv. What impact would granting pre-emptive rights to current famous name
holders have on developing countries, where many allegedly famous names
are not famous to the local population, and whose economies have not had
the chance to develop sufficiently to form their own distinct brands and
internationally famous names?

I propose to add to the terms of reference the following items under C
(Addition of New Generic Top Level Domains and Related Intellectual
Property Rights):
i. Whether the sale of domain names under .com for prices in excess of
USD 1,000,000 and their routine sale for prices around USD 2,000
indicates the presence of an artificial restriction on the supply of
domain names that could be easily remedied via expansion of the name
space;
ii. Whether trademark holders, by attempting to restrict the expansion
of the name space, are attempting to vest their existing gTLD domains,
especially those under .com, with a premium value worth millions of
dollars as Internet commerce expands;
iii. Whether expansion of the name space will encourage innovation in
the supply of Internet-related services
iv. Whether there are any historical precedents in any other media of
communication for restricting the supply of a critical resource in order
to make it easier to police intellectual property rights, and if so,
what these precedents tell us about the effectiveness of such policies
and their contribution to social welfare?
v. Whether an expanded name space will discourage users from attempting
to use the DNS as a search engine or index, and encourage them to employ
more advanced navigation techniques?

4. Timetable
The timetable proposed by RFC 1 is far too rapid. The terms of reference
need to be significantly revised. There is, in fact, no hurry, because
nothing that comes out of this proceeding can be implemented until the
new non-profit corporation is formed and it begins to make
decisions. This will proceed fairly slowly, over the course of two
years. The entire timetable proposed in RFC 1 should be pushed back at
least by six months. More generally, the process must monitor the
progress of the new corporation and especially its policy decisions
regarding the economic model for registries.

Conclusion
WIPO must alter its terms of reference to document the nature and scope
of the problem and the purported inadequacy of existing remedies. WIPO
must ask parties to comment on the potential biases and costs of a
centralized, uniform process, and the potential benefits of the kind of
competitive pressure that might be created by competing registries with
different approaches to dispute resolution. The sections pertaining to
the addition of new gTLDs must take into account the nature of the name
space as an economic resource. It must be based on a realistic appraisal
of the economic costs and benefits of various policies and their impact
on user behavior over the long term. I have suggested several new terms
to solicit comment on these issues. Finally, WIPO should lengthen its
time frame and be mindful of how its process synchronizes with the
creation of a new non-profit corporation to administer the IP address
and domain name space.

 -- Posted automatically from Process Web site