and vicariously liable for copyright infringement committed by
users of the Napster system. I have been asked to report about
the nature of the Internet community, and about efforts at
regulating behavior on the Internet.
past academic year, I was a fellow of the Wissenschaftskolleg
zu Berlin. In 1983, I received a B.S. in management from the
Wharton School, and a B.A. in economics from the University
of Pennsylvania. I received an M.A. in philosophy from Cam-
bridge University, graduating in 1986. I received a J.D. from
Yale Law School in 1989. After completing law school, I
clerked for Judge Richard Posner of the Seventh Circuit Court
of Appeals, and for Justice Antonin Scalia of the United States
Supreme Court. I was a professor of law at the University of
Chicago Law School from 1991 to 1997, and from 1997 to
2000, the Jack N. and Lillian R. Berkman Professor for Entre-
preneurial Legal Studies at the Harvard Law School.
addition to my recent book, Code and Other Laws of Cyber-
space (Basic Books 1999), I have written over fifteen articles
exploring the relationship between regulation and cyberspace,
and have given many more lectures on the same topic. The fo-
cus of most of this work has been the interplay between tech-
nology and law, and on the use of law to effect changes in
technology and, in particular, internet architectures. In addition
to my scholarly work, I am a regular columnist for The Indus-
try Standard. I have also contributed essays to the Wall Street
Journal, the Washington Post, the LA Times, and the Boston
the last seven years. I taught one of the first “Law of Cyber-
space” classes at an academic law school in 1995 (Yale). Since
then, I have taught five different Internet related classes. I have
regulation, and I have consulted extensively with policy mak-
ers about the regulation of cyberspace.
policy determinations. I testified before Congress on the Child
Online Protection Act. I have met with the Justice Department
and the F.C.C. on matters related to the merger of AT&T and
MediaOne. In 1997, I was asked by Judge Thomas Penfield
Jackson to serve as Special Master in the Justice Department’s
consent decree case against Microsoft Corporation. In February
2000, at the request of Judge Jackson, I filed an amicus brief in
the current case against Microsoft. I filed an amicus brief be-
fore the First Circuit Court of Appeals in a case challenging an
injunction against sites that carried software designed to crack
the program “CyberPatrol.” Microsystems Software, Inc. v.
Scandinavia Online, No. 00-1503 (First Circuit Court of Ap-
peals, 2000). I am currently representing a group of plaintiffs
challenging Congress’ Sonny Bono Copyright Term Extension
Act. Eldred v. Reno, No. 99-5430 (D.C. Circuit Court of Ap-
peals, 1999). My testimony in these cases is available on my
web page,1 and my full c.v. is attached as Exhibit 1.
computers or computer networks per se, nor to evaluate the
strength of the technology at issue from the perspective of a
computer scientist. Although I have written software, and have
studied network design and technology extensively, I have not
been formally trained in computer science or the art of pro-
gramming. My expertise is in understanding the relationship
between law and technology, and, in particular, the way in
which law might affect the technologies of the Internet.
provide evidence about the character of the Internet, and about
the consequences of efforts to regulate the Internet to protect
intellectual property rights. I am not being compensated for this
testimony, beyond expenses incurred in its preparation.
about Napster, in addition to the material provided to me by
attorneys representing Napster. I have used a version of the
Napster technology, as well as a version of the Gnutella tech-
nology. I base my conclusions upon that review, as well as
upon my experience studying Internet culture and the regula-
tion of the Internet over the past seven years.
fectiveness of regulating the Internet through an injunction
banning Napster-like technologies, in light of the values of the
Internet, and the values the Supreme Court has identified for
assessing regulations of the Internet. My analysis is unavoid-
ably a mixture of legal and technological considerations. I take
it as given both that the law has a proper role to play in pro-
tecting copyrighted works against violations of the Copyright
Act’s “exclusive rights,” and that those protections must be se-
cured through means that are consistent with other values that
the Supreme Court has identified as relevant in this context.
My objective has been to place this particular dispute in the
context of a more general question: How the law, consistent
with these other values, interacts with technology.
ating the application of law to the technologies of the Internet,
as well as a consistent method for evaluating the application of
copyright law to new technologies that might be used for copy-
right infringement. As evinced in Reno v. ACLU, 521 U.S. 844
(1997), before extending a law to the Internet, the Court has
been careful to assess the impact of that regulation on the free
flow of ideas that the decentralized, open network of the Inter-
net creates. Likewise, the Court has been slow to restrict the
development of new technologies, outside of the context of the
Internet, that might be used to infringe copyrights. Here too,
the Court has encouraged extensive and careful factual review
before concluding that a technology can be banned merely be-
cause it can be used in connection with copyright infringement.
stand the interplay between law and technology before banning
a particular technology. It is about that interplay that I offer
in the context of this case. It demonstrates the importance of
considering the practical effect of a regulation upon the tech-
nologies of the Internet before directly applying that regulation
through the use of injunctive power. In my view, a considera-
tion of that effect weighs against a ban on the Napster technol-
ogy. In particular, in my view (1) Napster has a potential for
substantial, non-infringing use; (2) plaintiffs could take reason-
able steps to minimize any harm to them caused by this tech-
nology; and (3) eliminating this particular technology — a step
that no court has ever taken with respect to any Internet tech-
nology — is not likely to have any substantial effect in reduc-
ing any harm that plaintiffs suffer.
ground of certain assumptions. I take it as given that behavior
in cyberspace, as in real space, is regulated by more than law.
Code and Other Laws of Cyberspace 85-100 (Basic Books,
1999) (“Code”). Beyond law, social norms regulate behavior in
cyberspace. So too does the market regulate behavior in cyber-
space. And, of particular importance to this case, the architec-
tureor design of cyberspace regulates behavior in cyberspace.
cyberspace, we must therefore account for the interaction
among these different modalities of regulation. Code, 99. Re-
strictions imposed through law, for example, might be under-
mined by changes in technology; changes in technology, in
turn, can reinforce the regulations sought through law. There-
fore, the question in each case must be the interaction between
the effects of law and technology on a particular behavior be-
example. Cyberspace has made available a great deal of “adult
cult, however, to zone this speech from minors in cyberspace,
primarily because the age of a user in cyberspace is not self-
authenticating. Unlike real space, where the age of individuals
is relatively easy to determine, the age of a user in cyberspace
is not easily verified. This makes it difficult to condition access
upon age, and in turn, difficult for governments to require that
speech deemed “harmful to minors” be kept from minors.
eration of Congress’s first direct regulation of cyberspace, the
Communications Decency Act of 1996. Pub. L 104-104, Title
V, 110 Stat. 56, 133-43. Congress sought to restrict the access
of minors to speech deemed “indecent.” In voiding that aspect
of the CDA, the Supreme Court emphasized the burden such
regulation would impose on those supporting or maintaining
the architectures that constituted cyberspace. Reno v. ACLU,
521 U.S. 844 (1997). It would significantly, and unconstitu-
tionally, burden Internet Service Providers, the Court observed,
to enable code that would facilitate zoning based on age. Id., at
877-79. Alternative, voluntary techniques, the Court suggested,
would equally advance the government’s interest without that
burden on the Internet. Id., at 877; id., at 890 (O’Connor, J.,
law regulating speech on the net in a way that might weaken or
undermine free speech values, the Court considered the inter-
action between that law and the architecture of cyberspace. The
Court was careful to assure that the regulation not interfere
with the Internet’s development. This was true even though the
interest advanced by the CDA was, in the Court’s view, “com-
pelling.” Id., at 869. That interest, however important, did not
relieve the Court of the requirement that it assure that state
regulation not interfere with the free speech potential that the
Internet has produced.
While it is not the first computer network, the Internet is the
first large scale computer network to adopt a particular network
peals as “end-to-end,”2 AT&T v. City of Portland, 2000 U.S.
App. LEXIS 14383, *23 (9th Cir. June 22, 2000), this design
has enabled extraordinary innovation on the Internet. By re-
quiring that the network remain simple, and by counseling de-
signers to place “intelligence” at the ends, or in the applications
using the network, the Internet has offered innovators a neutral
platform upon which to develop. Innovators realize that if they
conform their applications to the basic Internet protocols, the
Internet will run their application.
for the extraordinary growth that the network has seen. By
keeping network controls simple, and by not optimizing the
network to any single purpose, the network has invited new
protocols and new uses that could not have been originally an-
ticipated. Saltzer/Clark/Reed describe the relationship, for ex-
ample, between this principle of neutrality and the birth of the
World Wide Web. During the 1980s, as the authors describe,
many debated optimizing the net for telephony, which was at
that time, in the minds of many the expected use of Internet
technology. Had the net been optimized on that design, how-
ever, in violation of the principle of end-to-end, it would not
have been possible for Tim Berners-Lee to develop and im-
plement the protocol supporting the World Wide Web.3
to be relevant to the Internet’s design, especially as the Internet
becomes increasingly regulated. In my view, regulations of the
Internet should be consistent with this fundamental architec-
tural design, if the innovation and diversity of views that the
original Internet protected is to be preserved.
P. Reed, and David Clark. See End to End Arguments in System Design,
to-End perfectly into its design. It was because of changes in the 1970s sug-
gested by Vint Cerf and David P. Reed that the network we now recognize as
the Internet conformed to End-to-End.
single network, nor a single, permanent protocol implementing
this end-to-end design. Code, 101-102. The Internet is consti-
tuted instead by many networks, all of which respect a com-
mon set of protocols. These protocols are not static. They in-
stead evolve over time as new functionality gets added to the
basic set. Networks “on” the Internet are those that respect
these protocols. See Pete Loshin, TCP/IP Clearly Explained
19-20 (2d ed. 1997) (listing basic protocols).
“TCP/IP.” Among TCP/IP protocols are basic protocols facili-
tating transfer of packets of data among networks. Code, 101.
But in addition to this basic set, the Internet also includes ap-
plication protocols that enable familiar Internet functions.
These include, among others, a “file transfer protocol” (FTP),
that enables a file to be transferred between two computers; a
“simple mail transport protocol” (SMTP), that enables email to
be transferred; and a “hyper text transfer protocol” (HTTP),
that enables the linking of documents across different ma-
chines. Code, 102.
applications to facilitate an extraordinary array of network ac-
tivity. Though each is relatively simple, like DNA, they to-
gether facilitate an extraordinary diversity of applications and
uses. By remaining committed to the neutrality of end-to-end,
and by permitting a gradual evolution of protocols within the
basic Internet set, the Internet has, through a decentralized
process of development, become the most important communi-
cations network in history.
It supported the earliest applications on the Internet, by facili-
tating the transfer of files between two computers. See Reply
Comment of Andrew Oram, Concerning the Copyright Office’s
Rulemaking on Exemptions from Prohibition on Circumvention
of Technological Measures that Control Access to Copyrighted
Works, March 31, 2000 (available at < http://www.cpsr.org>).
These files could be computer programs (so users can, for ex-
They could be web pages produced by design software on a de-
signer’s computer, which then get transferred to a server where
the web page will be accessed. Or they could be movie files, or
music files, that would be transferred to permit a user to play
that content on his own machine.
protocol and extended its functionality. Napster is just one ex-
ample. Other programs too facilitate the easy location and
transfer of files among network servers. The program “Fetch,”
for example, uses FTP to make it easy for a user to build a list
of sites for downloading files. The Archie system is another
example, building indices of files on Archie servers which can
then be downloaded using an Archie client. Both are examples
of applications that build upon the general file transfer protocol
and thereby extend the functionality of the net.
opers to draw upon in building applications and uses for the
Internet. They are general purpose protocols, though their bias
is to enable the decentralized development of new applications.
Unlike the original telephone network, no single institution or
entity or government decides which applications will run on
the Internet. Any application that complies with the protocols
will run on the Internet.
design is that innovators will develop technologies that will af-
fect legal rights or regulation. Sometimes this effect will be
intended; other times it will be unintended. In both cases,
regulators must decide how best to respond to this change in
the effectiveness of a pre-existing legal protection or regula-
tion. That decision must take account of the interplay between
the technology and law.
the network cannot discriminate among the content of