is among a set of new technologies that present the opportunity
for a radical change in the nature of Internet search technolo-
gies. As the Internet itself advances, and as the World Wide
Web adopts more sophisticated designs (including dynamic
pages), traditional search technologies become less useful.
Search engines that rely upon “bots” to spider the web and in-
dex its content cannot index these dynamic pages. Better tech-
nologies for facilitating searching are therefore increasingly
Napster, Gnutella. Unlike Napster (which relies upon a central
server to coordinate its name space), Gnutella enables simulta-
neous searching on literally thousands of machines, according
to the search protocols set by those machines, without the ne-
cessity of coordinating those searches at a central server. The
protocol lets “sites query one another in a chain.” Andy Oram,
The Value of Gnutella and Freenet, Webreview.com (May 12,
2000). This is, as network architects have noted, a revolution-
ary advance in searching technology. Id.Thus, just as with
Napster, a user of Gnutella can simply enter the name of an
artist, or a particular kind of content, and the system will search
all cooperating sites to locate that content, and facilitate its
far less controllable, technology than Napster. As technologist
Andy Oram describes it,
inadequacy of existing search tools to work in an era when
Web sites increasingly depend on database queries and dy-
namically-generated temporary URLs. Many sites have their
own sophisticated searches, but you have to visit the site and
enter the string manually—or study the site's HTML form and
write a customized LWP script—in any case, you have to nar-
row your search to that single site. …
change information, but says nothing about what each site
does with the information. A site can plug the user's search
string into a database query or perform any other processing it
thus become the union of all searches provided by all sites. A
merger of the most advanced technologies available (standard
formats like XML for data exchange, database-driven content
provision, and distributed computing) could take the Internet
to new levels.9
Gnutella, it is clear that much of its current use is similar to the
current use of Napster. Both are presently enabling the location
and easy transfer of MP3 files, many of which (indeed, the
majority of which probably) are copyrighted material. But both
Napster and Gnutella have potential uses that extend far be-
yond this single potentially troubling use.1 0Both, that is, could
facilitate a much more efficient technology for peer-to-peer
relations about content generally.
abled by Napster and by Gnutella plainly has many substantial,
noninfringing uses. My own interest in the technology arose
because of a discussion with a colleague at Harvard, Professor
Jonathan Zittrain, about using Napster or Gnutella to enable
students in our respective classes to share notes of material
while teaching parallel courses. This is just one potential non-
infringing use of this technology. Other uses include libraries
that could share information in a much more decentralized
way, see Daniel Chudnov, docster: instant document delivery,
< http://www.oss4lib.org/read-ings/docster.php>; antique shops
that might facilitate the exchange of wanted but obscure items;
and auction sites that might enable global “flea markets” from
inventories indexed on individual computers. In each case, the
technology would enable a better peer-to-peer connection
among Internet users, for uses that are not now fully imagined
technology is simply an extension of the fundamentally decen-
deemed to be an infringing use. See supra¶ 46.
“one to many” communication of broadcasting, or the “many to
one” communication of voting, or polls, but the “many to
many” communication unique to this medium.
plainly endangered copyright interests. There can be no doubt
that Napster could be used for making copies of copyrighted
material for resale (“piracy”) and—depending upon how the
doctrine of fair use is applied, and upon how the Audio Home
Recording Act is interpreted, see supra¶ 46—that it has fa-
cilitated the sharing of content in a way the law may prohibit.
While I have no evidence about the proportion of material ex-
changed that is exchanged in violation of copyright laws, I
would not be surprised if that amount was as high as the high-
est estimate that the Supreme Court imagined occurred in the
context of VCRs.
ogy based on its current use, or to ban a technology based on
its initial use, even if significant violations of copyright were
enabled. If that had been the test, then many of the early Inter-
net technologies would have been banned. Likewise would the
VCR have been banned, and possibly even the Xerox machine.
Instead, as the Supreme Court has indicated both in Sonyand in
Reno, important constitutional values counsel a much more
balanced and informed response. An early use of an Internet
technology often has very little relation to its ultimate use.
driving the early development of many stages of communica-
tion technology has been pornography: 1-900 services, the
ability to transfer photographs across networks, real-time chat,
streaming video, interactive video—in each case, the early use
of the technology was the distribution of porn. But early use
notwithstanding, as each technology matured, other uses be-
came significant. To infer from the early use some conclusion
about ultimate use would, in these cases at least, have been a
vant for evaluating a technology that might enable copyright
infringement on the Internet. In this part, I sketch those tests
and how the facts as I have described them might apply. My
purpose is not to argue for the relevance or validity of these
particular tests. Whether they are ultimately relevant is of
course a matter for the Court.
substantial noninfringing uses. Sony, 464 U.S., at 442. Second,
what steps can individuals take on their own to protect against
the relevant harm. Reno, 521 U.S., at 877. And third, what is
the likelihood that regulation will actually be effective against
the alleged harm. 521 U.S., at 877 n.43.
tributory infringement raises is not whether a substantial por-
tion of its existing use is “fair” or legitimate, but “merely”
whether the technology is “… capable of substantial nonin-
fringing uses.” Sony, 464 U.S., at 442. On that standard it is
clear that Napster, and Gnutella, plainly make possible a kind
of sharing of information (1) that the Internet has not to date
fully enabled, (2) that would vastly improve the functionality
of the Internet, and (3) that would not even arguably violate
any copyright law.
sic, the sharing of copyrighted music that had been authorized
for sharing (for example, for purposes of sampling), and the
sharing of other non-copyrighted content on the net in a peer-
to-peer manner. Given the wide range of interaction that the
Internet has created, the potential for non-infringing use is vast
— indeed, far greater than the potential non-infringing use of
steps plaintiffs might take to avoid the harm they allege. This
factor too tends, in my view, in Napster’s favor. There are
plenty of technologies that copyright holders might use to bet-
ter enable the control of their content on the Internet. Some of
these technologies would make it harder for music to be
“ripped” from a CD and copied on the Internet. This is the ef-
fect, for example, of the Cactus Data Shield, produced by Mid-
bartech. See < http://www.midbartech.com>. Other technolo-
gies would make it easier to track illegal copying. This is the
effect, for example, of watermarking technologies. Other tech-
nologies would make it possible for sites to identify more eas-
ily properly copyrighted material.
self-protect their content. They would make it easier to selec-
tively attack users of the Napster system who are engaging in
illegal behavior, without attacking the Napster technology it-
self. As in Reno, if implemented, they would eliminate much of
the need for state control of the technology.
fringement that might be accomplished using the Napster tech-
nology. They would not, for example, affect music already re-
leased. This infringement may well be significant, and is obvi-
ously relevant to the facts reviewed by this Court. It might be
significant as well, however, that the recording industry has
been aware of these technologies for more than a decade, yet
has not taken steps effectively to implement them. See OTA,
Copyright and Home Copying: Technology Challenges the
Law 52-61 (October 1989).
tilts in Napster’s favor. This test asks whether the remedy
would be effective. In my view, eliminating the Napster tech-
nology would not substantially change the behavior that plain-
tiffs complain about. This is the implication of the existence of
and the difficulty that any court would have in regulating this
peer-to-peer searching capability, and is operated in a far more
decentralized manner. Because of this architecture, there would
be no way, under the present architecture of the net, for a court
to stop the deployment of Gnutella without essentially shutting
down a substantial portion of the Internet. Gnutella is simply
an application that runs on the net; there is no central server for
this application; links are made in a chain that itself is not con-
sistent or easily tracked.
Napster is likely ultimately to be ineffective. There is no bar-
rier to limit consumers switching from Napster to Gnutella.
Like Napster, Gnutella technologies are free, and there are
many different Gnutella clients freely available. The effect of
banning this one technology, Napster, would therefore be as ef-
fective as banning one of the 50,000 newsgroups on the
USENET. As in that case, the ban would simply move the con-
duct into a different channel.
regulation and technology, this fact about Gnutella would ren-
der, in my view, an injunction against Napster technologies in-
effective. This fact may in turn be relevant to the Court’s ulti-
mate resolution of the case.
has flourished because new technologies can easily be adapted
to the existing protocol base. This neutrality has produced ex-
traordinary innovation in the past; it will produce more inno-
vation in the future.
tant legal values, such as copyright. An appropriate and effec-
tive response to such threats is hard to determine in the ab-
stract. A court, in my view, can properly consider the interac-
tion between its rule and the behavior regulated. This interac-
than the doctrine of any one area of law.
for substantial non-infringing use before, on the basis of exist-
ing law, it bans that technology, then as Napster does have the
potential for substantial non-infringing uses, it would follow
that it should not, through injunction, be banned. This conclu-
sion is only strengthened when, as in this case, the effect of
banning the technology on the harm alleged is so speculative.
under laws of the United States of
America, that the foregoing is true
in Berlin, Germany.
Stanford Law School
559 Nathan Abbott Way
Stanford, CA 94305
M.A. Philosophy, 1986; Honors First
B.A. Economics, B.S. Management
Professor of Law, 2000.
Jack N. and Lillian R. Berkman Pro-
fessor for Entrepreneurial Legal
Studies, 1998; Professor of Law, 1997-
ter term, 1997.
Assistant Professor of Law, 1991-
1995; Professor of Law, 1995-97. Co-
Director, Center for the Study of Con-
stitutionalism in Eastern Europe.
Courses taught or currently teaching:
Constitutional Law I (federalism,
separation of powers, judicial review);
Constitutional Law II (free speech);
Contracts; The Law of Cyberspace
(Seminar); The Public Good (Semi-
nar); Comparative Constitutional Law
(Seminar); Legal Theory Workshop
(Seminar); Fidelity Theory: Theories
of Originalism (Seminar); Antitrust.
ture and Community, University
Harvard University, Cambridge,
Visiting Professor of Law, Spring
term, 1995. The Law of Cyberspace;