(Cite as: 23 Colum.-VLA J.L. & Arts 137)

                              Columbia-VLA Journal of Law and the Arts

                                           Spring, 1999


                                     Jane C. Ginsburg [FNa1]

                 Copyright © 1999 by the Trustees of Columbia University; Jane C. Ginsburg

In October 1998, Congress passed two major copyright bills, the "Digital Millennium Copyright Act" [DMCA], [FN1]
and the "Sonny Bono Copyright Term Extension Act." [FN2] Moreover, the Senate ratified U.S. accession to the
WIPO Copyright and Performances and Phonograms Treaties. [FN3] The DMCA implements the obligations set forth
in articles 11 and 12 of the WIPO Copyright Treaty [WCT] (and articles 18 and 19 of the WIPO Performances and
Phonograms Treaty [WPPT]) to protect technological measures against circumvention, and to protect "copyright
management information" against removal or alteration that facilitates infringement. The DMCA also includes a
chapter on the liability of online service providers. That chapter's provisions in part codify judicial developments, but
in other respects significantly reduce the liability of service providers for copyright infringements committed by their
subscribers. Another chapter of the DMCA includes detailed amendments to the public performance right in sound
recordings, entitling webcasters and certain other providers of digital radio to a statutory license for the digital
transmission of recorded musical compositions. The Sonny Bono Act extends the term of copyright by twenty years, but
also includes "Fairness in Music Licensing" provisions that introduce limitations on authors' exercise of their
nondramatic musical public performance rights with respect to certain business establishments. Almost all the
provisions examined here are extremely complicated - the fruit of intensive lobbying by a wide range of interest groups
of copyright owners, on the one hand, and, particularly, users, on the other.

                           I. THE DIGITAL MILLENNIUM COPYRIGHT ACT

                               A. WIPO TREATIES IMPLEMENTATION

1. Circumvention of Copyright Protection Systems

Title I of the DMCA addresses "copyright protection systems and copyright management information." These
dispositions respond to new obligations introduced in the 1996 WIPO Copyright and Performances and Phonograms
Treaties [henceforth referred to as the WIPO Copyright Treaties]. While U.S. copyright law generally met or
surpassed most of the WIPO treaties' requirements, in two respects U.S. law required adaptation. Both concerned the
modernization of the Copyright Act to meet the challenges of digital creation, communication, and exploitation.

The first WIPO norm, set forth in article 11 of the WCT and article 18 of the WPPT, labeled "Obligations Concerning
Technological Measures," requires Contracting Parties to afford "adequate legal protection and effective legal
remedies against the circumvention of effective technological measures that are used by authors in connection with the
exercise of their rights under this Treaty or the Berne Convention ...." [FN4] Does the norm cover more than an
obligation to protect authors against persons who directly circumvent technological measures? The text is not limited
to persons who circumvent; moreover, as a practical matter, protection against circumvention is not likely to be
"adequate and effective" if only the direct actors are covered, because in mostcases, "the circumvention" will be
accomplished by means of machines, devices, or computer programs. As a result, Congress determined that it was
necessary to provide for adequate and effective protection both against persons who circumvent (for example by
performing circumvention services), and against the manufacturers, importers and distributors of the devices (etc.)
that are produced and used to circumvent. [FN5]

The technological measures that the WIPO norm protects are those that "restrict acts ... which are not authorized by
the authors concerned or permitted by law." Thus, for example, the WIPO text would permit circumvention of a
technological measure that restricts copying that would qualify as a fair use. In implementing the WIPO Copyright
Treaties, Congress therefore confronted the problem of devising a system of legal protection of technological measures
that would accommodate fair use and other copyright exceptions. At the same time, the drafters were aware that
proper sensitivity to limitations on copyright protection should not result in inadequate or ineffective prohibitions on
circumvention of technological measures that restrict unauthorized acts not permitted by law. The difficulty of the task
of affording meaningful protection while preserving exceptions in part explains the complex and cumbersome nature
of the resulting legislation.

New section 1201 of the copyright act defines three new violations: (a)(1) to circumvent technological protection
measures that control access to copyrighted works; (a)(2) to manufacture, disseminate or offer, etc. devices or
services, etc. that circumvent access controls; and (b) to manufacture, disseminate, or offer, etc. devices or services
etc. that circumvent a technological measure that "effectively protects a right of the copyright owner ...." It is
important to appreciate that these violations are distinct from copyright infringement. The violation occurs with the
prohibited acts; it is not necessary to prove that the dissemination of circumvention devices resulted in specific

a. s 1201(a): Protection of technological measures controlling access to copyrighted works

This subsection sets out a right to prevent circumvention of technology used to control access to a copyrighted work;
the right is articulated separately and treated differently from the circumvention of technology used to protect a "right
of the copyright owner" under Title 17 (for instance, to authorize or prohibit reproduction, creation of derivative
works, distribution, public performance/display - subsection (b) covers these rights, see infra). The separation of access
from rights of copyright owners responds to the different balances struck depending on whether (a) access to the
copyrighted work is offered to the public subject to the copyright owner's price and/or terms, or (b) access having been
lawfully obtained, members of the public now seek to reproduce/adapt/distribute/publicly perform or publicly display
the work (or portions of it).

The DMCA gives the greatest protection to copyright owners' right to control access, since it makes it a violation both
(1) for users to circumvent access controls, and (2) for others to manufacture, disseminate or offer devices or services
that circumvent access controls. As for post-access circumvention, while the law prohibits the manufacture,
dissemination, offering etc. of devices or services, etc. that circumvent technological protection of rights under
copyright (e.g., anticopying codes), the bill does not prohibit users themselves from circumventing these protections.
The contrast indicates that this law tolerates direct end-user circumvention of post-access anticopying measures, to a
far greater extent than it does circumvention of access controls. In the course of its journey through the legislative
process, however, the prohibition on circumvention of access controls attracted an every-widening array of
exceptions. [FN6] Moreover, Congress delayed implementation of the prohibition for a two-year period, [FN7] and
further instructed the Librarian of Congress, in consultation with the Register of Copyrights, to identify particular
classes of works whose users who would be "adversely affected by the prohibition ... in their ability to make
noninfringing uses under this title of a particular class of copyrighted works." [FN8] Do these classes of works include
scientific journal articles and other materials relevant to nonprofit instruction? Unfortunately, neither the statute, nor
the legislative history give sufficient guidance to determining what the classes should comprise, apart from cautioning
that they should be neither too broadly nor too narrowly drawn. During the three-year period following the rulemaking
procedure that identifies the works, users of these works will not be subject to liability for circumventing measures
that control access to the specified classes of works. The law requires that another rulemaking be conducted every
three years. [FN9]

b. s 1201(a)(1): Prohibition on end-user circumvention of access controls

The DMCA distinguishes between access to the work, and use of the work once accessed. In old technology (hardcopy)
terms, the distinction might be between acquiring a copy in the first place, and what one does with the copy thereafter.
The fair use concerns primarily focus on the second stage. That is, it may be fair use to make nonprofit research
photocopies of pages from a lawfully acquired book; it is not fair use to steal the book in order to make the
photocopies. [FN10] To that extent, the notion of "access" appears to resemble the traditional copyright concepts
inherent in the exclusive distribution right. [FN11] The Supreme Court has construed this right to give the author
control over the determination to grant "access" to her work, that is, to disclose and offer it to the public, for
purchase if she chooses. [FN12]

i. What is "access"?

But it also seems that the "access" that section 1201(a) protects goes beyond traditional copyright prerogatives.
Indeed, the text indicates that "access" is distinct from a "right of the copyright owner under this title." [FN13] The
difference becomes apparent if one compares the consequences of protecting a measure controlling "... access to a
work ..." with a measure controlling "... access to a copy of a work ...." The latter corresponds to "access" in the
copyright sense of the right to distribute copies of the work; the former is the new right introduced in the DMCA. The
following example illustrates the difference between "access to a work" and "access to a copy of a work." Suppose I
purchase a CD ROM containing a copyrighted work, such as a video game. Suppose also that to view and play the
game, I must register with the producer, using the modem in my computer. The producer in turn communicates a
password to me. A technological measure included in the CD ROM recognizes my password, and my computer.
Thenceforth, each time I wish to play the game, I must enter my password, and play the game on the same computer.
This means that I cannot use that copy of the game on another computer. It also may mean, at least in theory, that I
cannot communicate my password to a friend or family member to play the game on my computer, since the password
protects access to the work, and my disclosure of the password is an act that circumvents a protective measure that
had limited access to me.

In this scenario, by purchasing the CD ROM, I have acquired lawful access to a copy of the work. Section 101 of the
Copyright Act defines "copies" as "material objects" in which "a work" is fixed. The CD ROM I purchased is a
material object. But I do not access "the work" until I have entered the password (from the correct computer).
Thus, when the law bars circumvention of controls on access to the "work," "access" becomes a repeated operation,
whose controls will be substantially insulated from circumvention under the text of section 1201(a). I would therefore
not be permitted to circumvent the access controls, even to perform acts that are lawful under the Copyright Act, such
as using my copy in another computer or lending it to a friend - acts permitted to the owner of the copy under the
"first sale doctrine" codified in section 109(a) of the Copyright Act, and under section 117, which allows the owner of
a copy of a computer program to use it in a computer. By contrast, had the law barred circumvention of technological
measures controlling access to "a copy" of a work, then once I had lawfully acquired my copy, I should have been able
to use it in a computer, or to circulate that copy, without further prohibitions imposed or reinforced by the Copyright
Act. (If I want to make or transmit additional copies, however, I would most likely be infringing the reproduction or
public performance/display rights, and might also run afoul of post-access technological protections attached to my

How would the distinction between access to "a copy" and access to "a work" apply to online access? The online user
who "accesses" a work is also making a "copy" of the work in either or both RAM and hard drive. [FN14] If the user
lawfully stores the copy of the work, then its subsequent consultation should not be considered "access" to the copy,
because access to a copy would already have been acquired at the time of downloading to a permanent storage
medium, and it is (or should be) permitted to make a further copy within the computer as part of using the
lawfully-acquired copy in conjunction with that computer. [FN15] In adopting an "access to the work" standard,
Congress has placed the user who has lawfully stored a copy of an access-controlled work in the same position as a
user who does not retain the copy, and who must therefore re-connect to the online source to view the work. Each
viewing from the online source is a new "access" to the work. But so are viewings from a downloaded version (or, for
that matter, a free-standing version such as a CD ROM). In each of these circumstances, the user may not (subject to a
variety of exceptions discussed below) circumvent a technological measure that controls the user's ability to apprehend
the work.

What is the justification for this apparent extension of the copyright owner's prerogatives? The WIPO texts require
that signatory nations protect technological measures "that are used by authors in connection with the exercise of
their rights under this Treaty or the Berne Convention ...." Do either the WCT or the Berne Convention provide a right
of access to a "work" in the sense of s 1201(a)? The WCT introduces a right of distribution of copies to the public, but
this right addresses material (rather than electronic) copies; it would not extend beyond conferring a right to
control access to a physical copy, rather than to a work. [FN16] The WCT also synthesizes the Berne Convention's
various provisions on public performance into a general right of "communication to the public," including by digital
delivery. [FN17] This right would appear to cover "access" to a work through online media; it is considerably less
clear that it would also cover subsequent apprehension of a work once the user downloads a copy to the user's storage
media, or once the user acquires a free-standing copy, such as a CD ROM. [FN18]

Since the WCT does not compel signatory countries to protect technological measures that control the breadth of
"access" covered by s 1201(a), Congress has independently determined that this scope of protection is necessary to
afford meaningful protection to copyrighted works in the digital environment. What supports this determination? The
changing economics of exploitation of copyrighted works may supply a justification. In the digital environment,
exploitation of works of authorship will differ from exploitation in the analog world. Professor Jessica Litman has
observed that in the analog world, a book could not "sprout wings and fly back" after a certain number of readings.
[FN19] As a result, the purchasing user paid a price commensurate with unlimited consultations of the book, or
listenings of the sound recording, or viewings of the motion picture (if the film was not available for rental). In the
digital world, by contrast, the price can be set to correspond to the number of times the user wishes to enjoy the work.
It may, for example, be more attractive to a user to purchase a $5.00 DiVX videodisc, that permits three viewings of a
film, than a $25.00 DVD videodisc that allows unlimited viewings. [FN20] But if the copyright owner is to make this
marketing option available, it is necessary to ensure that the user who purchases the $5.00 videodisc cannot then
circumvent the technological protection that limits the user to three viewings, in order to obtain unlimited viewings
without paying the price.

Arguably, the s 1201(b) prohibition on provision of devices to circumvent measures that protect "a right of the
copyright owner" (see discussion infra) could suffice to assure copyright owners that users will not be able to pay
$5.00, but through circumvention obtain $25.00 of value. Section 1201(b), however, does not prohibit direct acts of
circumvention; the technologically adept user thus faces no liability under that section. As to the provision of
circumvention services or devices, it is not clear in this situation that a "right of the copyright owner" would be
implicated. The user's viewing of the film would most likely be a private performance (i.e., there would not be a
"substantial" number of viewers other than the user's family and social acquaintances), [FN21] to which the
copyright owner's rights do not extend. In order to view the film, the user would be making a copy of the contents of
the DiVX onto the user's RAM, but this copying might not be actionable, either. [FN22] As a result, the user would
certainly be depriving the copyright owner of revenue, but that deprivation might not contravene a "right of the
copyright owner," and s 1201(b) might therefore be ineffective. By contrast, if each viewing is an act of "access" to
the work, then, after the third viewing, the DiVX user would be circumventing an access control, and would be in
violation of s 1201(a).

Similarly, in the case of musical compositions and sound recordings, 17 U.S.C. s 1008 permits consumers to make
noncommercial use of "a device or medium for making digital musical recordings ...." While a general purpose
computer does not meet the definition of such a "device," a software program specially designed to make copies of
digital musical recordings might. Under s 1008, then, it might notbe copyright infringement to use the program in
conjunction with a computer to make private copies of sound recordings, either from a free-standing medium such as
a CD, or from a digital network. Circumventing a technological measure in order to engage in this kind of private
copying thus would not violate a "right of the copyright owner." But if one cannot make the private copy without
"accessing" the work, then circumvention would violate s 1201(a). Thus, to the extent that s 1008 applies to copying
off the Internet or from free-standing digital copies onto a personal computer, access protection would trump the
private copying privilege.

As a general matter, one should recognize that, in granting copyright owners a right to prevent circumvention of
technological controls on "access," Congress may in effect have extended copyright to cover "use" of works of
authorship (subject to the many exceptions permitting access-control circumvention). In theory, copyright does not
reach "use"; it prohibits unauthorized reproduction, adaptation, distribution, and pubic performance or display
(communication to the public). Not all "uses" correspond to these acts. But because "access" is a prerequisite to
"use," by controlling the former, the copyright owner may well end up preventing or conditioning the latter.

c. s 1201(a)(2): Prohibition on manufacture, etc. of devices, etc. that circumvent access controls

If users may not directly defeat access controls, it follows that third parties should not enable users to gain
unauthorized access to copyrighted works by providing devices or services (etc.) that are designed to circumvent
access controls. Indeed, the principal targets of the DMCA are the providers of circumvention devices, services, etc. As
a general proposition, the prohibition on providing devices such as "black box" descramblers that enable members of
the public to receive without paying for pay-per-view type transmissions (for example, of music or of audiovisual
works), is (or should be) uncontroversial. [FN23] The question is whether the DMCA's prohibition sweeps too broadly,
and ends up barring the manufacture and dissemination of devices or services that have legitimate uses other than to
circumvent controls on access to copyrighted works. Too broad a prohibition may frustrate whatever legitimate
activities the devices may permit. Equally importantly, too broad a prohibition may frustrate the development of useful
new technologies.

Section 1201(a)(2) does not prohibit the dissemination of any device (etc.) that might be used to defeat an access
control. It does not target general purpose devices (etc.) whose accidental, incidental or unwitting use results in
circumvention. Nor does it bar those devices (etc.) that, while capable of, and even used for, circumvention, are
primarily designed or used for other purposes. The law prohibits the manufacture, etc. of devices, services, etc., only in
the following three circumstances:

(A) The device (etc.) was "primarily designed or produced for the purpose of circumventing a technological measure
that effectively controls access" to a copyrighted work; or

(B) The device (etc.) was not primarily designed to circumvent, but in fact "has only limited commercially significant
purpose or use other than to circumvent ..."; or

(C) The device (etc.) is "marketed" (i.e., advertised or promoted) as a device (etc.) to be used to circumvent access
controls. In this case, the target of the law is the person promoting the circumventing use; it is not the manufacturer
(etc.) of the device (etc.), unless that person acts in concert with the marketer. [FN24]

With respect to (B), s 1201(a)(2) does not define "only limited commercially significant purpose or use," but case law
may help to flesh out what appears to be a fairly intuitive notion. For example, in A&M Records, Inc. v. Abdallah,
[FN25] the defendant supplied "time-loaded," made-to-order blank audiotape cassettes. Because most of the
defendant's clients purchased the tapes to make pirated music cassettes, the court attributed no legal significance to
the few clients who did or might have purchased to record their own musical performances.

But is the "only limited commercially significant purpose or use other than to circumvent" standard too
restrictive? To address this question, it is necessary to envision what legitimate circumventions of access controls the
text might forestall. This universe is probably not very large. The most likely example may be that of a work not
protected by copyright, but which is disseminated with an access control. Use of a "descrambler" to defeat the access
control over this work would be lawful under the bill. But sale of descramblers is probably not therefore legitimated,
since the descrambler device cannot tell the difference between protected and unprotected works, and one may
imagine that by far the greater volume of sales will be to those who wish to access protected works. Thus,
descrambling encrypted public domain works might well not qualify as a "commercially significant purpose or use" of
the device.

This suggests that the DMCA could deprive the public of the technological means to access public domain information.
Closer examination of the text of s 1202(a)(2), however, may permit a happier outcome. The three alternative criteria
may be read, with the help of negative inferences, to authorize the distribution of access-circumvention devices in
certain narrow circumstances. First, s 1201(a)(2)(A) appears to establish a subjective criterion: did the producer of the
device create it for the purpose of defeating access controls on copyrighted works? If the producer responds that she
created the device for the purpose of defeating access controls on public domain works (and the copyright owner
cannot prove otherwise), then subsection (A) would not bar the device's dissemination. That, of course, does not mean
that the device can be freely marketed. Subsection (B) imposes an objective criterion: whatever the device's producer
intended, how is the device in fact being used? If in fact the use of the device to descramble public domain works is
insignificant, then the device violates the DMCA. If, by contrast, the device has a commercially significant use to
access encrypted public domain documents, then subsection (B) would not prohibit its sale.

Under what circumstances would this use be genuinely significant? One might imagine that the problem with the
public domain argument for circumvention of access controls is that, for most manufacturers, sellers and users, it is
probably pretextual. In the name of descrambling a public domain motion picture, a "black box" is distributed to users
who are more likely to descramble all of Disney. [FN26] If, by contrast, the circumventer category were limited to
those for whom public domain access is the sole activity, then the appropriate users and uses would not be prejudiced,
and the public domain tail would not wag the dog of generalized descrambling of protected works. This suggests that
the market for particular descrambling devices or services cannot be the public at large, but might be confined to an
identifiable group whose public domain access claims are more plausible (and who can be trusted not to misuse the
devices to access copyrighted works without authorization). With respect to that market, there would be a
"commercially significant" use of the device or service to access public domain materials. This analysis also
comports with subsection (C), since the device would be "marketed" not as a circumvention device for accessing
copyrighted works, but as a device to access the public domain.

Who would make up this group? Librarians and members of research institutions form an identifiable group who can
credibly claim to employ circumvention devices in order to access public domain documents. But one may be
somewhat skeptical of the further claim that these persons will employ the devices only to access public domain
materials. If their use could be so restricted, then a university's computer services department could lawfully assist a
professor, student, or librarian who seeks to decrypt an unprotected work. Similarly, a library could afford its patrons
access to the decrypted public domain work, either on a walk-in basis, or even by making the work digitally available
to remote users. [FN27]

Perhaps one might meet the requirement, that the use of the device have commercial significance other than to
circumvent access protections on copyrighted works, in some way other than distinguishing by type of user. If the
machine itself could distinguish between copyrighted and public domain elements, and accordingly decrypted only the
access code corresponding to the public domain portion, then its distribution would be legitimate under s 1201(a)(2).
But this will work only if information accompanying the work identifies what is protected, and what is public domain.
Since content- providers may not be eager to furnish this information, the Copyright Office regulations defining
copyright management information [FN28] would need to require that copyright owners who incorporate copyright
management information must encode copies of the works to identify the portions in which copyright is claimed. That
way, if the encrypted document is not protected by copyright, it should not bear the code that will alert the machine to
the work's protected status. The machine could then be freely distributed because it would not descramble codes that
protect copyrighted works.

Problems nonetheless persist. Suppose that information providers, having come to the conclusion that, under certain
circumstances, circumvention devices may be distributed if their use is confined to accessing public domain works,
were careful not merely to encrypt raw public domain documents, but instead packaged the documents with
copyrightable trappings, such as a new introduction, or with minimally original reformatting. [FN29] The
circumvention device now would target a copyrighted work, however scant the covering. Does that mean that the
device, even if needed to access the unprotected work, now fails under subsection (A), because it is "primarily designed
to circumvent" access controls on copyrighted works? If this is correct, then s 1201(a)(2) would not only preclude
spurious invocations of the public domain, it would also bar the distribution of any device intended to decrypt public
domain works that came in a copyrightable package. But were that the ineluctable interpretation of s 1201(a)(2),
then the copyrightable fig leaf a producer affixes to an otherwise unprotectable work would, as a practical matter,
completely obscure the public domain nakedness of the documents. Dog-wagging is no more desirable when waged by
copyright owners than by users.

One might argue that manufacture or distribution of a circumvention device in these circumstances is not primarily
designed to circumvent controls onaccess to copyrighted works, since the primary target is the public domain work.
Similarly, if distribution of the device remains confined to the library and research community, then, arguably, the
commercially significant use would be to access public domain, rather than copyrighted, documents. These arguments
are unconvincing for at least two reasons, however. First, unless the public domain work is accessible separately from
the copyrighted work to which it is joined, the device cannot access the public domain work without circumventing the
protection attached to the copyrighted work. Thus the device is primarily designed to circumvent in a manner
prohibited by subsection (A). Since the s 1201(a)(2) criteria are alternative, rather than cumulative, failure to pass
muster under any of the subsections will doom the device. As a result, even were the producer to propose to limit the
distribution of the device to the library and research community, the device probably should not be manufactured in
the first place. Second, the argument does not afford a means to distinguish the pretextual copyrightable fig leaf from
a substantial authorship contribution. For example, what if the copyrightable work incorporating the public domain
documents is a compilation or collective work containing both copyrightable and public domain elements? Or, more
generally, what if the circumventer claims to produce or use the device in order to access unprotected ideas or facts
within a protected work? If all of these claims also justified access-circumvention, then little would be left of the
prohibition. [FN30] As we will see, some of the specific exceptions to s 1201(a) address some of these situations; by
implication, however, s 1201(a) does not authorize a general exemption for devices designed or used to circumvent in
order to access public domain elements contained within copyrighted works.

How troublesome is this? After all, the copyright law does not oblige publishers to give away copies of the plays of
Shakespeare. But, because "access" in the DMCA means more than access to the copy, the publisher can now charge
not only for acquisition of an electronic edition of Shakespeare (assume no copyrightable value added to the text of the
plays), but for each reading of Much Ado, and readers who have lawfully acquired the copy may not elude those
charges. Do the DMCA's strict controls on access therefore result in overprotection, or are they a necessary
adaptation to the digital world? That is, do traditional categories of rights under copyright fail to respond to the way
works are (or will be) exploited in digital media, so that new rights are needed? "Access" probably will become the
most important right regarding digitally expressed works, and its recognition, whether by the detour of
prohibitions on circumvention of access controls, or by express addition to the list of exclusive rights under copyright,
may be inevitable. But if "access" is a right (express or de facto), it is also necessary to consider what limitations that
right should encounter. As we shall see next, Congress has provided a variety of exceptions permitting users to
circumvent access controls, but the exceptions betray their origins in interest-group lobbying; no coherent vision of
appropriate limits on the new "access" right underlies their articulation.

d. s 1201(d)-(j): Exceptions to protection against circumvention of access controls

i. s 1201(d): Exemption for nonprofit libraries

This provision creates what became known during the legislative process as a "shopping right" for libraries and other
nonprofit educational institutions. It permits these institutions to circumvent "solely in order to make a good faith
determination of whether to acquire a copy of that work for the sole purpose of engaging in conduct permitted under
this title ..." [FN31] but only "with respect to a work when an identical copy of that work is not reasonably available
in another form." [FN32] Moreover, the institution must discard the accessed copy once it has determined whether or
not to acquire a copy (presumably, by paying for it, or through other authorized means). [FN33]

When one works through the application of this provision, its practical import seems extremely slim. Suppose a library
is trying to decide whether or not to acquire a particular work. Suppose further, that this work exists only in
access-protected form (i.e., no hardcopies, and no promotional copies to peruse on the author or publisher's website).
Then suppose that the author or publisher is not willing to make available to the library or educational institution
information sufficient to permit the institution to make the purchasing decision. At that point, the library may defeat
the access code and ascertain the work's contents, but only in order to make the purchasing decision, and not for any
other purpose. One may wonder how often this provision will apply in practice.

ii. s 1201(e): Law enforcement; s 1201(j): Security testing

Section 1201(e) entitles federal, state and local law enforcement officers to defeat access controls in order to conduct
an authorized investigation. The investigation may extend to verifying "information security," defined as "activities
carried out in order to identify and address the vulnerabilities of a government computer, computer system, or
computer network." Similarly, s 1201(j) permits circumvention for purposes of "security testing," defined as "good
faith testing, investigating, or correcting, a security flaw or vulnerability, with the authorization of the owner or
operator of such computer, computer system, or computer network." These provisions are not intended to serve
as licenses to "hack" into a computer system. The persons verifying a system's security must be either law enforcement
officers, or persons authorized by the computer system owner or operator to conduct the test. The exception does not
exempt the technologically adept who unilaterally do owners or operators the "favor" of breaking into their systems in
order to warn them of the systems' vulnerabilities.

iii. s 1201(f): Reverse engineering

This provision addresses decompilation of computer programs to achieve interoperability. It permits persons who have
"lawfully obtained the right to use a copy" of a computer program to defeat the access code (either directly or by
"developing and employing" a circumvention device) "for the sole purpose of identifying and analyzing those elements
of the program that are necessary to achieve interoperability of an independently created computer program with
other programs ...." [FN34] These elements must not have "previously been readily available" to the person engaging
in the decompilation, for example, in an unencrypted version of the computer program, and the acts of identification
and analysis must not infringe the copyright in the computer program. The concern underlying this provision was that
software producers would defeat decompilation by issuing computer programs only in access-protected forms. Because
access protection is distinct from copyright infringement, s 1201 in its original form would have prevented the
circumvention of an access code even if the purpose of the circumvention was to perform an act that courts had ruled
could be fair use. [FN35] On the other hand, reverse engineering should not become a pretext for defeating access
controls in order to acquire computer programs for free, or in order to make infringing copies of the program. Hence,
the prerequisite that the reverse engineer have "lawfully obtained a right to use a copy" of the program, and the
requirement that the decompilation remain within the bounds of what the copyright law permits.

iv. s 1201(g): Encryption research

This provision shares some features of s 1201(f) (reverse engineering) and s 1201(j) (security testing). It defines
encryption research as "activities necessary to identify and analyze flaws and vulnerabilities of encryption technologies
applied to copyrighted works, if these activities are conducted to advance the state of knowledge in the field of
encryption ...," and permits specified users to circumvent (and to "develop and employ" circumvention devices) for
this "sole purpose." [FN36] This provision also imposes detailed limitations on the would-be encryption researcher:
the encrypted copy must be "lawfully obtained"; defeating the access code must be "necessary to conduct such
encryption research"; the researcher must have made a "good faith effort to obtain authorization"; and the "act of
good faith encryption research" must "not constitute infringement under this title or a violation of applicable law
other than this section ...." [FN37] Out of apparent concern that "encryption research" could degenerate into a
pretext for indiscriminate hacking of access controls, this provision further attempts to restrict the class of persons
qualified for the exemption by listing factors to consider: whether the information derived from the research was
disseminated in a manner "reasonably calculated to advance the state of knowledge or development of encryption
technology" or whether instead it "facilitates infringement"; whether the person "is engaged in a legitimate course of
study, is employed, or is appropriately trained or experienced, in the field of encryption technology"; and whether and
when the results of the research are disclosed to the copyright owner. [FN38] As a further caution, Congress
instructed the Register of Copyrights and the Department of Commerce to report to Congress within one year of the
date of enactment concerning the impact of this provision on encryption research and on the protection of copyrighted
works. [FN39]

v. s 1201(h): "Exceptions Regarding Minors"

This provision addresses the concerns of parents who seek to prevent their children from seeing inappropriate material
on the Internet. Congress was persuaded that screening devices that parents would employ to protect their children
might include a "component or part" that circumvents access controls, but that these devices should be available if the
"component or part" is necessary to a device that "does not itself violate the provisions of this title" and has the "sole
purpose to prevent the access of minors to material on the Internet."

How significant is the need for parents to circumvent access controls? One might suppose that if the material is
access-protected, then the children will not see it anyway. But the access control might be easily satisfied. For
example, suppose the [fictitious] website "weird-sex-with-orcas.com" imposes a control requiring the user to state
that his or her age is over 18, and seeking other demographic information as a condition of access. If the child
provides false information about his or her age, and other demographic details, she will gain access to the site. One
may imagine that parents would wish to avoid this result. Hence, the desirability of a device that will scrutinize the
content of access-protected sites (according to criteria that parents or entrepreneurs program into the device), in order
to prevent children from seeing them. Once again, however, it is important that the parental screening device not have
a dual functionality, that is, to circumvent in order to block children's inappropriate access on the one hand, but also
to circumvent in order to facilitate parents' (or children's) unpaid access on the other.

vi. s 1201(i): Protection of personally identifying information

This provision responds to significant privacy concerns raised by users of digital networks. Some website operators
have, without notifying users, engaged in the practice of collecting or disseminating information about the online
activities of persons who contact the websites, for example, by sending "cookies" to the user's hard drive. This
exception therefore permits users to circumvent an access control in order to discover and disable an undisclosed
information-gathering feature. If, by contrast, the information- gatherer provides "conspicuous notice" of its
information collection, and enables the user to restrict the collection or dissemination of personally identifying
information, then the user may not circumvent. [FN40] Similarly, the right to circumvent is limited to identifying and
disabling the undisclosed "cookie" or similar device; it does not entitle the user to "gain access to any work." [FN41]

As this survey demonstrates, s 1201's panoply of exceptions presents something of a grab-bag; it is difficult to discern
from the various exceptions an overall legislative policy regarding the kinds of access control circumventions that
should offset copyright owners' enhanced ability to control the use of their works. From the user's perspective, the
proliferation of narrow exceptions prompts the negative inference that any circumvention not expressly exempted is
prohibited. Far from promoting a flexible, fair use-like approach to the circumvention of access controls, the
overspecification of special exemptions may well make it more difficult to articulate a general user privilege
supporting circumvention of access controls.

Arguably, s 1201(c) accords courts residual authority to expand exceptions to access control. That provision specifies
that nothing in s 1201 affects "defenses to copyright infringement, including fair use, under this title." [FN42] One
might object that a violation of s 1201(a) is not copyright infringement, it is a new violation for which the DMCA
provides distinct remedies. [FN43] Nonetheless, circumvention claims remain copyright- dependent, since s 1201(a)
covers only measures that protect access to copyrighted works. Perhaps, were a court persuaded that the challenged
production, distribution or use of a circumvention device would not under the circumstances lead to copyright
infringement, because--as in the case of a "copyrightable fig leaf" covering a public domain work--whatever access to
a copyrighted work the device permits is incidental and necessary to access an unprotected work. "Accessing" the
copyrighted work will generally involve making a digital copy in temporary memory, but, under these circumstances,
the copying would probably be considered fair use. Perhaps, under appropriate circumstances, the act of accessing
should be deemed fair use as well.

e. s 1201(b): Circumvention of technological protections of (traditional) rights under copyright

The DMCA also addresses technological protection of the post-access rights of reproduction, adaptation, distribution,
public performance or display. Here, the bill addresses only the producers and suppliers of circumvention devices,
services (etc.); the end-user's activities are not at issue. As with section 1201(a), section 1201(b) does not target all
devices (etc.) that are capable of being used for circumvention. Manufacture and distribution of post-access
circumvention devices, services (etc.) are prohibited only if (A) they are "primarily designed" to circumvent; or (B) if
they have "only limited commercially significant" uses other than to circumvent; or (C) if they are "marketed" as
circumvention devices. [FN44]

The prohibition contained in this subsection is not as stringent as that of the subsection concerning access controls.
The post-access devices (etc.) here targeted are those that circumvent "protection afforded by a technological measure
that effectively protects a right of a copyright owner under this title ...." [FN45] The exclusive rights under copyright
set forth in section 106 of the Copyright Act are expressly made "subject to sections 107 through 120," sections that
set forth a variety of exceptions to and limitations on copyright (referred to collectively as "fair uses"). [FN46] If the
circumvention device (etc.) is designed for or can be put to commercially significant fair use, then it is not a violation
of s 1201(b) to sell the device or to offer the circumvention service. Here, as in the case of circumventions of access
controls, however, the device itself probably cannot distinguish between circumventions for fair use purposes, and
circumventions aimed simply at obtaining unauthorized copies. But were the device exculpated simply because it is
capable of being put to fair use, then, as a practical matter, the fair use tail would again wag the copyright
infringement dog.

Nonetheless, this need not mean that there can be no manufacture and distribution of circumvention devices (etc.). The
lawfulness of the manufacture and distribution should turn on the definition of the market for the device or service
(etc.). A copy protection-defeating device addressed to the general public may not be likely to have commercially
significant fair uses; one created for or disseminated to a community of researchers and scholars is a better candidate.
Put another way, so long as university or library personnel employ circumvention devices or services that they
have devised (or that are created at the library or university's behest), and so long as the devices or services are used
to make copies (or adaptations, distributions, public performances or displays) that would qualify as fair uses, there
should be no violation of section 1201(b).

There is at least one major objection to this market-based analysis: it appears to privilege formal fair use
communities, such as universities and libraries, over the general public (although all members of the public can be
library users). All members of the public are entitled to invoke the fair use exception (in appropriate circumstances).
Non-scholarly fair use can include parody and other forms of commentary. [FN47] Under Sony Corp. of America v.
Universal City Studios, Inc., [FN48] some kinds of temporary noncommercial copying for personal convenience may
also be fair use. If one assumes that technological measures will accompany all copies of the works targeted for
parody, or for private copying, then how will the user (who is insufficiently computer-adept to do the circumvention
herself) be able to obtain and exercise the means to make the further copies necessary to carry out her fair use
project? The assumption that copyright owners will only make their works availablein copy-protected form may well
be overstated; nonetheless this is one issue that Congress instructed the Copyright Office to take into account in
preparing its study of the "impact that the prohibition on the circumvention of technological measures applied to
copyrighted works has on criticism, comment, news reporting, teaching, scholarship, or research ...." [FN49]

Underlying this study is a fundamental question about the respective rights of copyright owners and privileges of
users. Should the law reinforce the copyright owner's efforts to prevent unauthorized (and unmonitored) copying, or
should Congress have left technology and the market to develop defenses and counter-ripostes, leading to an
encryption "arms race," [FN50] if need be? Arguably, such an arms race wastes creative resources that might be
better directed toward creating original works of authorship, rather than devices that promote piracy. One might
respond that without descramblers, copyright owners will be able lock up works of authorship, and will charge for
every act of reading, [FN51] let alone quotation and commentary. Circumvention devices therefore would be essential
to preserve our basic civil liberties. Of course, this contention assumes that works will be available only in encrypted
formats. [FN52] That may be more true for some works, such as software, digital sound recordings, and
databases, than for others, such as books and other literary works. But even where works are susceptible to effective
technological protections, copyright owners may not choose to restrict access to every copy. Moreover, copies will
often still be available for anonymous consultation (and limited copying) in places such as public libraries (who, under
the analysis offered earlier, are entitled to circumvent access and anti-copying codes, under appropriate
circumstances). So the debate may in fact reduce to the question whether it is good policy to forestall the general
distribution of circumvention devices whose objective is not to provide unpaid access or copying when none was
otherwise available, but to make unauthorized unpaid access or copying easier. When the public increasingly expects
to acquire works with a click of a mouse, should the law stand back while third-party entrepreneurs endeavor to make
works available without authorization in the most copy-convenient format (and without compensation for resulting

In the world of analog copies, the question did not arise because the work was not amenable to effective copy
protection. Thus, not only was a book unable to "sprout wings and fly back," [FN53] but it could easily be
photocopied; a sound recording could be re-recorded; a television program videotaped, or a videocassette copied
(although the latter depended on whether or not the videocassette was copy-protected). On the other hand,
analog-world users could not copy and redisseminate on the scale that digital media permit. [FN54] Legal protection
of technological measures arguably sends consumers back to an era in which technology was too rudimentary,
cumbersome or expensive to permit private copying to rival significantly the copyright owner's control of markets for
the work.

It may in fact be misleading to condemn (or to celebrate) the DMCA's impact on the "copyright balance" [FN55]
between owners and users. After all, which "balance" do we mean? The one in which technology did not offer much
potential for consumptive copying? Or the one in which technology enabled widespread copying, but did not afford
adequate and reasonable means of preventing or charging for the copying? Or the one in which technology permits
massive copying, but also enables copyright owners to be paid for it? Or one in which technology enables copyright
owners to prevent copying? Which "balance" one selects depends in part on how one views the respective entitlements
of owners and users. From a user's point of view, the owner's "bundle of rights" never extended to all uses of works;
owners now are relying on legal protection of technological measures to secure the airtight coverage the substantive
law never previously afforded. From a copyright owner's point of view, if in the past low technology imposed a
tolerance for widespread copying, that state of affairs should not be confused with a legal right to engage in
widespread convenience copying. Newer technology undermines the factual premise for the tolerance; if the tolerance
is to persist as a prerogative, it requires a legal basis.

The question is not whether fair use is an integral part of U.S. copyright law; the statute, and perhaps even the
Constitution, [FN56] clearly make it so. Nor is the question whether the copyright owner may employ technology to
preclude unauthorized copying altogether: as to that question, by instructing the Copyright Office to study the impact
of protective measures on the exercise of fair use, Congress has implicitly answered "no." (Although it is worth noting
that the kinds of fair uses about whose future in the digital environment Congress has expressed concern are all of the
"transformative" variety; Congress did not direct the Copyright Office to determine if technological measures unduly
threatened "consumptive" copying.) Rather, the issue is whether a proper copyright balance requires copyright owners
to tolerate some level of unauthorized convenience copying, either because there is a justification independent of
transactions costs (such as the privacy interests of the user), or because, in any event, the law should accommodate
convenience copying lest a broad prohibition on post-access circumvention risk eliminating transformative fair use as

Perhaps, then, Congress should have favored technological neutrality (or anarchy), in which entrepreneurs would
have been entitled to make whatever protection-defeating devices technology allowed, so that copying and anti-
copying technologies could have competed freely. In that case, recourse would have been available only against
end-users (unlikely objects of pursuit), or against purveyors whose devices lacked a substantial noninfringing use.
[FN57] But if convenience copying is, at least in some circumstances, a noninfringing use, then, under the Sony
standard, most post-access circumvention devices would have been legitimated. By outlawing the general distribution
of post-access circumvention devices, Congress has instead adjusted the technological status quo in favor of copyright
owners, and, at least for now, set the copyright "balance" against unauthorized convenience copying.

f. s 1201(k): Anticopying controls and analog devices--a belated legislative response to the "Betamax Case" and home
taping of audiovisual works

The DMCA's provisions on technological protections also include a special chapter on nondigital video recording
devices and technological protections of audiovisual works. This chapter offers a welcome, if belated and incomplete,
response to the Supreme Court's decision in Sony Corp. of America v. Universal City Studios, Inc., [FN58] in which a
majority of the Court rejected liability for contributory infringement of the manufacturers and distributors of
videocassette recorders, on the ground that the recorders were capable of a noninfringing fair use, in that case,
time-shifting of free broadcast television programming. The Court never held that retaining copies of recorded
programs, or recording pay-TV or cable programming was fair use, but in practice, the public engaged in all these
varieties of uncompensated taping, and copyright owners initiated no further suits. Unlike many European countries,
the U.S. never instituted a levy on the media or equipment of videocassette recorders in order to compensate copyright
owners, authors, and performers of audiovisual works. [FN59] Section 1201(k) now addresses the problem of home
videotaping by prohibiting the manufacture and distribution of certain analog videocassette recorders unless the
recorders are equipped with a designated copy control technology. [FN60]

The copy control technology will work in tandem with audiovisual works and transmissions that have been encoded to
prevent or limit consumer copying. Section 1201(k), however, restricts the circumstances under which a copyright
owner or other person may encode the audiovisual work or transmission. These are limited to: transmissions of live
events; transmissions of audiovisual works delivered on a pay-per-view basis; copies of transmissions of live events or
of audiovisual works made available through a subscription to a television channel (be it herzian, cable, or satellite);
"physical medium containing one or more prerecorded audiovisual works"; and copies of transmissions of live events
or of pay-per-view motion picture transmissions, or copies made from a physical medium containing prerecorded
audiovisual works. [FN61] This means, for example, that the producer of the [fictitious] motion picture "It Came
from the Titanic" may distribute encoded copies of videocassettes; if a purchasing or renting consumer attempts to
make an extra copy using a videocassette recorder that conforms to s 1202(k), then the anticopying technology
incorporated in the recorder will produce an extremely degraded quality copy. It also means that if "It Came from the
Titanic" is transmitted on pay-per- view, the producer may encode the transmission; the consumer will be able to view
the film at the time requested, but if she attempts to use a conforming videocassette machine to tape the film as it is
being transmitted, the quality of the resulting copy will be too low to be worth viewing. By contrast, the copyright
owner or broadcaster may not encode programs broadcast over television stations to which access is available without
payment. Thus, consumers may continue to time-shift (or even retain) copies of "free" television programming.

2. Integrity of Copyright Management Information

The second WIPO norm requiring U.S. implementation sets forth "Obligations concerning Rights Management
Information." Contracting parties must provide adequate and effective legal remedies against one who engages in the
knowing removal or alteration of rights management information, "knowing, or ... having reasonable grounds to
know, that it will induce, enable, facilitate or conceal an infringement of any right covered by this Treaty or the Berne
Convention ...." [FN62] The DMCA implements this obligation in new s 1202 by defining a violation of the rights
management information provision in terms of facilitation (etc.) of "infringement."

Section 1202 prohibits: (a) knowingly providing false copyright management information, with the intent to facilitate
or conceal infringement. The provision also prohibits (b) knowingly or intentionally altering or removing copyright
management information, knowing (or having reasonable grounds to know) that the alteration or removal will
facilitate or conceal infringement. Subsection (c) defines copyright management information. It includes: the name of
the author; the name of the copyright owner; and the "[t]erms and conditions for use of the work." [FN63]

The provision is designed to promote the dissemination of copyrighted works by facilitating the grant or license of
rights under copyright (particularly through electronic contracting). Because accurate and reliable information about
the work is essential to its lawful distribution (particularly online), the bill identifies that information and protects it
against falsification, removal or alteration. These are important goals whose achievement will further the interests of
the public and of authors in the digital communication of works of authorship. To a significant extent, s 1202 does
advance these objectives.

There are, however, some respects in which s 1202 ensures the desired reliability and accuracy only imperfectly.
Moreover, the text does not fulfill all U.S. obligations under article 12 of the WIPO treaty. Article 12 requires
Contracting Parties to prohibit, inter alia, unauthorized removal or alteration of electronic rights management
information, when the actor knows or has reasonable grounds to know that the removal or alteration:

will induce, enable, facilitate or conceal an infringement of any right covered by this Treaty or the Berne Convention

(emphasis added). The "right[s] covered by the Berne Convention" are not only those found within the U.S. Copyright
Act. Article 6bis of the Berne Convention declares authors' rights "to claim authorship of the work ...." When the U.S.
adhered to the Berne Convention, Congress announced its position that it was not necessary to incorporate the Article
6bis right of attribution into the body of the Copyright Act, because other provisions of U.S. law, federal and state,
adequately assured that right. [FN64] This means that misidentifying the author of a copyrighted work violates
some legal norms in the U.S., but it is not copyright infringement. Thus, s 1202 falls short of the WIPO treaty
requirement, because section 1202 concerns only copyright management information whose removal or alteration
facilitates or conceals copyright infringement. Removal or alteration of copyright management information identifying
the author of the work would violate the WIPO norm, but since it is not copyright infringement even willfully to
miscredit the author, there would be no violation of s 1202, unless it could be shown that miscrediting authorship
induces infringement. Arguably, improper authorship credit could complicate or defeat title searching and rights
clearance, leading some users to infringe in frustration, but this seems rather attenuated (particularly if other licensing
information is not tampered with). Section 1202 thus inadequately protects authors' Berne-mandated rights of

At least as importantly, it also disserves the general public interest in knowing who is the author of the work. The
DMCA recognizes the importance of and public benefit to authorship credit, since s 1202(c)'s definition of copyright
management information includes "[t]he name of, and other identifying information about, the author of a work."
Section 1202 does not oblige the rights owner to attach copyright management information to distributions of the
work, but if the rights-holder does attach copyright management information, then that information should include the
name of the author. [FN65] So understood, s 1202 expresses a public policy favoring author identification as part of a
reliable system of dissemination (especially electronic distribution) of copyrighted works. Unfortunately, linking the
violation of copyright management information to copyright infringement does not effectively achieve the objective of
ensuring the accuracy and reliability of a key component of copyright management information.

There is another essential public interest, and a Berne Convention norm, that s 1202 does not secure as effectively as it
could. That is the interest in ensuring that the work as distributed is the same as the work was when it was first
publicly released by the author or rights-holder. It is increasingly important, especially with digital media, to be able to
ensure the authenticity of the work. This interest complements the Berne Convention article 6bis right of the author to
ensure the integrity of her work against alterations that damage her reputation. Unauthorized alterations to a work's
contents threaten the credibility both of the document and of the author's reputation. Section 1202 affords authors and
copyright owners some means to protect the content of their work against alteration or falsification, but these means
may not respond sufficiently to the significant threats to the integrity of a work's content that digital media pose.

The statute offers the elements of an argument that the dissemination of an altered work with copyright management
information that described the work before alteration constitutes the provision/dissemination of false or altered
copyright information. This is because copyright management information can be intentionally "altered" in two ways:
directly, by changing the text of the information; and indirectly, by changing the work to which the information
applies, so that the information no longer accurately describes the work. But, even if the information is deemed
"altered," s 1202 further requires that the alteration promote (etc.) infringement before a violation will be found.
Arguably, the alteration does promote infringement, at least if the copyright owner did not consent, because the
alteration may result in the creation of a "derivative work," and the exclusive right to create derivative works
(adaptations, etc.) resides with the copyright owner.

The success of this argument depends in part on whether the alterations will be deemed to produce a derivative work.
Some courts have conditioned the finding of an infringing derivative work upon proof of a substantial authorship
contribution to changes to a preexisting work. [FN66] But, while the alterations to a work may not be so substantial
as to yield a derivative work, they may well be more than sufficient to compromise the work's authenticity. Moreover,
when the author and the copyright owner are not the same person, any alteration undertaken or approved by the
copyright owner, however massive, will not promote copyright infringement, although it certainly can undermine the
integrity and authenticity of the work. The public interest in ensuring the authenticity of digital documents thus may be
disserved by the law's linkage of a violation of copyright management information to copyright infringement.
Similarly, the linkage to copyright infringement may fail to fulfill our obligation to ensure the Berne Convention right
of integrity guaranteed by art. 6bis. Rights outside of copyright, such as those against defamation, may afford claims
against alterations that damage an author's reputation, [FN67] but they do not apply here, since the DMCA defines a
violation of copyright management information solely in terms of facilitation of copyright infringement.

To some extent, the Copyright Office may fill in the gaps in the copyright management information scheme. Section
1202(c)(8)'s definition of copyright management information provides a residual category of "[s]uch other
information as the Register of Copyrights may prescribe by regulation ...." For example, while the statute lists "[t]he
name of, and other identifying information about, the author of a work," [FN68] the statute does not clarify whether
the creator of a work made for hire also should receive authorship credit. [FN69] The Copyright Office Regulations
should interpret copyright management information consistently with the WCT and the Berne Convention art. 6bis
right of attribution, by providing that not only must the statutory "author" receive authorship credit, but the
contributors to a work made for hire should as well. With respect to a work's integrity, Copyright Office regulations
might help ensure authenticity by requiring that the copyright management information include a statement that
the work made available to the public corresponds in content and form (the latter, where relevant) to the work as
created. The regulation might further provide that any subsequent alterations to the work must either be disclosed (if
they are authorized) or are prohibited.


This title of the DMCA adds a new section 512 to the remedies chapter of the Copyright Act. It exempts online service
and access providers from liability for damages for copyright infringement, and significantly reduces the scope of
injunctive relief, if the providers meet the bill's cumulative factors assessing the independence of the providers from
the infringing content they transmit or host. Failure to comply with s 512 does not of itself subject the service provider
to liability for copyright infringement; section 512 offers a safe harbor, but a nonqualifying provider must still be
proved to have infringed, and may still invoke traditional copyright and other defenses. [FN70] Section 512 also
includes strong incentives to the providers to remove allegedly infringing material, once they are notified of the alleged
infringement. Section 512 divides the universe of service providers into three categories: (a) those who merely transmit
third party content "through an automatic technical process without selection of the material by the service provider";
(b) "system caching" of third party content if the "intermediate and temporary storage" of the material "is carried out
through an automatic technical process for the purpose of making the material available to users of the system or
network who ... request access to the material" from the third party that initiated its transmission; and (c) hosts of
third-party content residing on systems or networks at the direction of users. [FN71] To the extent that s 512 exempts
providers who simply transmit or temporarily and automatically cache other persons' communications (including
electronic mail), the provision substantially codifies the "mere conduit" concept set forth in Religious Technology
Center v. Netcom On-Line Communication Services, Inc. [FN72] The court there held that an access provider's
automatic creation of temporary copies as part of the process of transmitting third-party communications did not
constitute "reproduction" in violation of the author's exclusive reproduction right, because the creation of volatile
copies lacked a necessary (albeit interpolated) element of volition on the part of the access provider. Section 512 does
not, however, alleviate the liability of those who originate content on a website.

Finally, it is worth observing that s 512 addresses only liability for copyright infringement. It affords no safe harbor
from liability for hosting or transmitting other kinds of illicit content, such as child pornography. [FN73] Notably,
given the DMCA's distinction between "access" and a "right of a copyright owner," as well as its distinction
between liability for circumvention and liability for infringement, [FN74] it appears that the host of a website that
offers software that "descrambles" protection devices would be subject to normal principles of primary and secondary
liability, including damages, for violating the DMCA's anti-circumvention provisions.

The ensuing analysis will consider the copyright liability regime established for service providers that host third-party
content, as well as the special exceptions created for the benefit of nonprofit educational institutions.

1. Hosting Third-Party Content: The Application of the Exemption

Section 512(c) sets forth cumulative prerequisites to a hosting service provider's qualification for exemption from
direct or vicarious liability for copyright infringement. The host must "not have actual knowledge that the material or
an activity using the material on the system or network is infringing"; and it must not be "aware of facts or
circumstances from which infringing activity is apparent"; moreover, if the host has the "right and ability to control"
the originator's activity, the host must not "receive a financial benefit directly attributable to the infringing activity."
[FN75] Once the host becomes "aware" of infringing activity, it must act "expeditiously to remove, or disable access
to, the material." [FN76] Most importantly, upon proper notification by the copyright owner, the service provider
must respond "expeditiously to remove, or disable access to" the allegedly infringing material. [FN77] A similar
regime applies to "information location tools," such as search engines that hyperlink to websites. If the provider of the
link to an allegedly infringing site does not know and is not aware of the presence of infringing material on the linked
site, and, upon proper notification from the copyright owner, acts expeditiously to remove the link, the provider will
not be liable for damages, and will be subject to limited injunctive relief. [FN78] These provisions establish a
complicated, but potentially effective system of "notice and take down" whose details we will now explore.

a. Notice and take down ... and put back

Section 512(c)(2) requires that a service provider, in order to benefit from the reduction in liability, designate, and
provide contact information concerning, an agent to receive notification of claimed infringements. Copyright owners
alleging that the service provider is hosting infringements must send the provider's agent a signed, written
communication identifying the work infringed, the material alleged to be infringing and its location; a statement that
the complaining party has a good faith belief that the use is unauthorized; and a sworn statement that the information
contained in the notification is believed to be accurate.

Upon receiving the notification, the provider must "expeditiously" remove or block access to the alleged
infringing material, or else face the full range of liability should the author prevail in an infringement suit. But the
provider who removes or blocks the material must also so notify the subscriber; the subscriber may then send a
"counter notification" (whose contents the law prescribes). [FN79] In that event, the service must send the counter
notification to the person who notified the service of the alleged infringement, and must inform that person that the
service will replace the material in ten business days. The copyright owner must within that time "file[] an action
seeking a court order to restrainthe subscriber from engaging in infringing activity relating to the material on the
service provider's system or network." [FN80] If the copyright owner does not initiate the action, and so inform the
service provider's designated agent, then the service provider must put back the material "not less than 10, nor more
than 14, business days following receipt of the counter notice ...." [FN81]

The purpose of these provisions is to ensure that s 512 does not make it too easy for copyright owners to compel the
removal of allegedly infringing material, without judicial process. Section 512 encourages providers who have
received notice to take down the material immediately, since the text insulates service providers who comply with the
statutory requirements from suit by persons (including the subscriber) disgruntled at the removal of the material from
the server. [FN82] From the point of view of authors and copyright owners, the strong incentives to remove material
may offer an effective means of enforcement, since the author's first goal will often be to get the material taken down
before it can be copied/disseminated further. On the other hand, there is the risk that timorous service providers will
remove material whose posting was not infringing, for example, because the posting constituted fair use. Hence the
opportunity for the posting subscriber to demand that the material be "put back," and a corresponding obligation for
the copyright owner to initiate judicial proceedings if the copyright owner wishes to ensure that the "take down" of
the material remain in effect.

The law does not specify where the copyright owner must file the action. The statute obliges the subscriber who sends
a counter notification to consent to the jurisdiction of the subscriber's local federal district court, or, if the subscriber
is located outside the U.S., to consent to the jurisdiction of the federal district court "in which the service provider
may be found." [FN83] The statute thus facilitates suit in the U.S., but does not explicitly compel the copyright owner
to proceed in a U.S. court. In the absence of a statutory requirement that the action be initiated before a U.S. federal
court, [FN84] one should interpret the text as meaning that the action may be filed in any country that has
jurisdiction over the subscriber. This makes sense, for it may often be the case that, while the service provider is
a U.S. entity, its subscribers are not. Residents of any country in the world from which Internet access may be obtained
can become subscribers of a U.S. service, and occupy space on its server. Moreover, the copyright owner may also be
a foreign resident or entity. Suppose, for example, that Jacquie Holiday, a [fictitious] popular French composer,
discovers that the contents of a CD of her most recent songs have been uploaded by a Belgian subscriber to a website
hosted by the [fictitious] U.S. service provider Pangaea. Where will Holiday bring the claim against the Belgian
subscriber? While the U.S. courts will have personal jurisdiction over the Belgian with respect to a claim arising out of
the Belgian's contacts with the U.S. service, either by s 512's enforced consent, or pursuant to principles of minimum
contacts, [FN85] Holiday may well prefer to bring the action in Belgium, where its execution will be immediate (and
where litigation costs are considerably lower than in the U.S.).

But what if Holiday does not know the identity of the subscriber that originated the upload of the CD, because the
subscriber has chosen to go by a pseudonymous handle, such as "Robin Hood" (denoting the subscriber's attitude
toward copyright owners, on the one hand, and users who "hit" the subscriber's site, on the other)? Suppose also that,
upon learning from Pangaea that Holiday has sent the requisite notification, "Robin Hood" in turn sends a counter
notification? How can Holiday initiate the required action within the ten-day period if "Robin Hood's" identity is
unknown? Or even if "Robin Hood" does not send a notification, so that the material remains off Pangaea's server,
suppose Holiday still wishes to sue "Robin Hood" for infringement, to recover damages for the copies of the CD made
by users who previously contacted "Robin Hood's" website. (Recall that if Pangaea complies with all the requirements
of s 512, no damages may be awarded against Pangaea. [FN86])

Section 512 addresses this problem in two ways. First, a subscriber who sends a counter notification must include his
"name, address, and telephone number, and a statement that the subscriber consents to the jurisdiction of [the]
Federal District Court for the judicial district in which the address is located, or if the subscriber's address is outside of
the United States, for any judicial district in which the service provider may be found ...." [FN87] The service
provider must communicate the counter notification to the copyright owner. As a result, if "Robin Hood" objects to
the blocking of his website, Holiday will know who "Robin Hood" is, and where to find him. Holiday will also know
which U.S. court will be competent to adjudicate a copyright infringement claim against "Robin Hood," should
Holiday decide to proceed in the U.S. (If Holiday sues "Robin Hood" in Belgium, the national courts there will be
competent, because "Robin Hood" is a resident.)

Second, and relevant to subscribers who do not send a counter notification, but whom the copyright owner seeks
to pursue, s 512(h) provides for issuance of a subpoena to compel the service provider to disclose the identity of the
subscriber. The copyright owner must present the prescribed documentation to the clerk of "any United States district
court." [FN88] Upon issuance of the subpoena, the service provider must "expeditiously disclose" the requested
information. Thus, upon complying with the requirements for issuance of a subpoena, Holiday could compel Pangaea
to identify "Robin Hood." While Holiday will be able to acquire the information necessary to initiate an action against
"Robin Hood," Holiday will not be able entirely to avoid U.S. courts (and attendant costs). Even if she ultimately
brings the infringement claim in Belgium, she will still have to go before a U.S. court to obtain the subpoena
compelling Pangaea to disclose who "Robin Hood" is.

What if the host service provider is located off-shore? Suppose that a U.S. copyright owner finds her works posted
without authorization on a website hosted by the [fictitious] U.K. service provider CoolBrittania. Does s 512 apply to
foreign host servers if the websites they host can be received in the U.S. (and/or if their subscribers are from the U.S.)?
If CoolBrittania believes it may be subject to personal jurisdiction in the U.S. (for example because it is "doing
business" by accepting U.S. subscribers and by making the sites it hosts available to U.S. users), it may elect to comply
with s 512's notice and take down procedures, in order to benefit from the consequent limitations on its U.S. liability. If
CoolBrittania is not subject to personal jurisdiction in the U.S. (because it is not deemed to be "doing business," and
because its role in making the infringing website available is considered too passive), then it might elude the
competence of U.S. law. The latter conclusion is troubling, since it suggests that service providers may disregard
notice of infringements, and avoid liability (reduced or otherwise) under U.S. law, by relocating to a server-friendlier
jurisdiction, [FN89] while continuing to accept U.S. subscribers, and to transmit the sites to U.S. users. This in turn
suggests that foreign host servers who welcome U.S. subscribers, or who do not respond to notification of alleged
infringements by blocking U.S. users' access to the allegedly infringing site, are likely to be deemed amenable to suit in
the U.S., and subject to U.S. law, at least with respect to transmissions received in the U.S. [FN90]

2. Remedies Available Against Hosts and Transmitters of Infringing Content

Section 512 bars award of damages against any service provider that complies with the statute's criteria. Section
512(j) sets forth the limited circumstances under which injunctive relief will be available against a complying service
provider. The court may order the host or caching provider to block access to the infringing material, and to
terminate the account of the subscriber who uploaded the infringing material. By contrast, if the service provider
merely transmits the infringing material, the court may order the service to terminate the subscriber, but may not
order the service to block access to the material, with one important exception. When the material resides on a site
located outside the U.S., a court may order the U.S. access provider to block access to that site. [FN91] This
exception is consistent with the overall scheme of the Online Liability Limitation Act, which diminishes liability in
accordance with the degree to which the service provider is implicated in content it hosts or transmits. When the
service merely transmits, Congress has, as a general matter, oriented the copyright owner's pursuit toward the host (as
well as toward the primary infringer). But if the host is off-shore, there is a greater probability that the host will not
comply, or at least, that it will be more difficult and time-consuming to obtain an order against the host from a foreign
jurisdiction (particularly if that jurisdiction is a "copyright haven"). In that case, the only meaningful relief will be
from the U.S. participants in the communication, hence the court's power to order local access providers to block
foreign sites (but not to order local access providers to block local sites; the latter order is properly directed against
the site itself).

How meaningful is the relief available to copyright owners against service providers under s 512? The Act's
inducements to remove allegedly infringing material can ensure a speedy response, but the preclusion of monetary
relief may leave significant remedial gaps. Recall "Robin Hood." Suppose that 10,000 websurfers had "hit" the site
before Pangaea received notice and blocked access. Holiday seeks to recover damages for the 10,000 unauthorized
copies made off Robin Hood's site. To that end, she obtains a subpoena compelling Pangaea to disclose Robin Hood's
identity so that she may sue him for damages. Pangaea complies with the subpoena; Holiday files an action against the
now unmasked "Robin Hood"--only to learn that he is seventeen years old, and judgment-proof. In foreclosing
monetary claims against service providers, Congress has also denied any practical prospect of recovery in damages
when the primary infringer is underage or otherwise unable to satisfy a money judgment. Moreover, this class of
infringers is likely to be significant on the Internet, since one may anticipate that "Internet cowboys" will upload
unauthorized copies, not for commercial gain, but to "liberate the content" of copyrighted works for all to enjoy. Even
if not commercially- motivated, this kind of conduct can have deleterious commercial consequences for copyright

3. Special Exemption for Nonprofit Educational Institutions

In general, a person or entity who posts infringing content is subject to the full range of liability for copyright
infringement. Congress made a special exception, however, for nonprofit educational institutions. Normally, by
application of the doctrine of respondeat superior, professors and employed graduate students who act within the
scope of their employment engage the liability of the university that employs them, if their acts prove to be
infringing. Universities therefore faced the unpleasant prospect that they would be fully liable for damages and subject
to injunctive relief if professors or employed graduate students posted infringing content to their university-hosted
webpages. Universities also, no doubt, anticipated that professors and employed graduate students might well commit
many copyright infringements, however unintentional. Universities therefore sought, and Congress ultimately agreed,
that a faculty member or employee graduate student "be considered to be a person other than the institution," and
that the professor's or graduate student's knowledge or awareness of his or her infringing acts not be attributed to the
university. [FN92]

Severance of the employer-employee link between professors and their universities means that universities will be
treated as service provider "hosts" of third party content, rather than as originators of that content. As a result,
universities that comply with the requirements of s 512 (including notice-and-take-down) will benefit from a significant
reduction in liability. But the DMCA does not insulate universities from all online postings made by their personnel.
Rather, a university will retain the characterization of "employer" (and hence be liable for the professors' acts) unless:
the faculty member's "infringing activities do not involve the provision of online access to instructional materials that
are or were required or recommended, within the preceding 3-year period, for a course taught at the institution by
such faculty member or graduate student"; and the institution "has not, within the preceding 3-year period, received
more than two notifications" from the copyright owners that the postings were infringing; and the institution promotes
its system users' compliance with the copyright law. [FN93] This means, for example, that if, at any time during the
previous three years, the [fictitious] casebook "Copyright for the New Millennium" was an assigned or recommended
textbook at [the fictitious] Columbus University Law School, and if the professor who teaches the copyright course at
Columbus uploads, without authorization, substantial parts of the book to the course's webpage, then the University
will be fully liable for the professor's infringement. The purpose of the prohibition is clear: to avoid the competition
that availability of the book on the course webpage would pose to sale of the work to enrolled students.


1. Nature of the Problem

A last-minute addition to the DMCA inserted extremely detailed provisions amending the already extraordinarily
detailed amendments to sections 114 and 115 of the Copyright Act, set forth in the 1995 Digital Performance Right in
Sound Recordings Act. The 1995 legislation introduced into U.S. copyright law a public performance right in sound
recordings; prior to that, the public performance right benefitted authors of musical compositions when
recordings of their music were performed or transmitted, but the artists and producers of the sound recordings enjoyed
neither control over nor received compensation for those performances. The 1995 amendments to the Copyright Act
established a three-tier system. Nonsubscription digital audio transmissions were treated as analog radio transmissions
had been, that is, recording artists and producers enjoyed neither control nor compensation. Subscription digital
transmissions, however, were subject to a statutory license, provided, inter alia, that the transmitting entity did not
exceed the "sound recording performance complement," limiting the number of songs from a particular sound
recording or by a particular artist that may be played during a three-hour period. Finally, the 1995 amendments
conferred on sound recording copyright owners full control over interactive digital transmissions as well as over
subscription digital transmissions that exceeded the complement. [FN94]

Although the 1995 amendments were titled "digital performance right in sound recordings," the text suggests that
Congress was at least as concerned with protecting sound recording copyright owners' reproduction rights as with
instituting a public performance right. The structure of the amendments addressed a spectrum of digital performances,
from those resembling traditional radio broadcasts, to the "celestial jukebox" model of music on demand. While the
former remained outside the sound recording copyright, the more the content of a digital transmission depended on a
particular recording or artist, or could be known by the user in advance, the more subject it became to the sound
recording copyright. This is because the more advance information the user has about the digital transmission, the
more the transmission facilitates a user's private copying (in perfect digital copies) of the recorded performance, or, at
least, enables the user to substitute listening to the targeted performance for purchasing a copy of it. Hence, the sound
recording copyright owner's right to prohibit interactive digital transmissions or subscription transmissions that
exceeded the "performance complement." When the transmission was by subscription, but remained within the
complement, then the sound recording copyright owner was entitled only to a statutory license fee.

The digital transmissions reached by the 1995 expansion of the sound recording copyright turn out to have omitted a
principal form of Internet exploitation of sound recordings: audio "streaming" or "webcasting" of recorded
performances. These transmissions originate on the Internet, and are generally nonsubscription; the user calls up the
"Internet radio" website, much as she or he would seek the channel of a conventional herzian radio station. The
difference is that the geographic factors that limit the broadcast area of traditional radio simply do not apply to
websites: a webcaster from anywhere in the world can reach a world-wide audience. As original (rather than
retransmitted) transmissions that are neither subscription nor interactive (the listener does not select the content),
webcasts were exempted both from the sound recording copyright owner's right of control, and from the obligation to
secure a statutory or negotiated license under the 1995 amendments. (By contrast, webcasters remained liable to
composers of the underlying music if the recordings were transmitted without a license from the music copyright

It therefore appeared that an increasingly significant form of digital exploitation of sound recordings would not only
elude the copyright owners' control, but would not give rise to a right to remuneration. But a different provision of the
Copyright Act, s 112 on "ephemeral recordings," called into question the webcaster's ability to transmit recorded
performances free of copyright claims. Section 112 permitted a "transmitting organization entitled to transmit to the
public a performance or display of a work ..." to make a copy of the program to be transmitted. The ephemeral
recording exception makes it possible for radio broadcasters (and analogous actors) to engage in the frequent industry
practice of prerecording the transmissions, without having to clear reproduction rights. The ephemeral recording
exception, however, was subject to the condition that the recorded program be "used solely for the transmitting
organization's own transmissions within its local service area ...." [FN95] This condition makes perfect sense when the
transmitting organization is a local herzian radio station; the confines of the broadcast signal constitute the local
service area. The condition makes no sense when the transmitting organization is communicating the program over
the Internet: at the point there is no local service area, the Internet service reaches the whole world.

The lack of fit between the ephemeral recording exception and webcasting led the sound recording copyright owners
to contend that even if the Digital Performance Right in Sound Recordings amendments entitled nonsubscription
webcasters to transmit without paying performance rights, they nonetheless could not transmit prerecorded
programming without clearing reproduction rights as well. This impasse set the stage for negotiations between
copyright owners and digital transmission services, leading to a substantial revision of the performance right regime
established by the 1995 amendments. A detailed examination of the new regime is beyond the scope of this Article, but
a review of its broader outlines is appropriate.

2. How the Amendments Work

The new version of the sound recording public performance right retains the three-tier structure of the 1995
amendments, but substantially narrows the category of transmissions that are wholly exempted from the right. Where
this category had included nonsubscription digital transmissions in general, and hence on its face exempted most
webcasting, this category now is limited to "nonsubscription broadcast transmission[s]." [FN96] The statute defines a
broadcast transmission as "a transmission made by a terrestrial broadcast station licensed as such by the Federal
Communications Commission." [FN97] The 1998 amendments then subject digital transmissions, whether
subscription or nonsubscription, to the statutory license, so long as these transmissions meet the act's exceedingly
detailed eligibility requirements. Ineligible transmissions, whether subscription or not, are subject to the sound
recording copyright owner's full public performance right, including the right to prohibit the transmission.

To simplify, the conditions on eligibility primarily endeavor to ensure that the transmitting organization does not
facilitate a user's substitution of the recorded performance for a purchase of a copy. Hence, for example, the
amendments retain the "sound recording performance complement"; they bar the transmitting entity from advance
announcement of the content of the programs; they oblige the entity to accommodate technological measures imposed
by the sound recording copyright owner to protect the works, and to cooperate with the sound recording copyright
owner to prevent users or other third parties from scanning the transmissions to select particular transmissions.
[FN98] The last requirement addresses the copyright owners' concern that third parties will develop search engines
that will identify and download particular sound recordings from a variety of webcasting sites. Copyright owners
apparently feared that a user might employ such a search engine to find and reproduce, for example, any Beatles song
transmitted by any webcaster, thus enabling the user to avoid purchasing authorized Beatles recordings. [FN99]

With respect to the determination of the statutory royalty rate that webcasters must pay, the 1998 legislation modifies
the criteria of the 1995 amendments. Under both versions, a copyright arbitration royalty panel, subject to review by
the Copyright Office, determines the rates and terms of royalty payments. Under the prior version, however, the panel
set the rate according to its determination of what would afford a "fair return" to the copyright owner, and a "fair
income" to the copyright user. [FN100] The current version instructs the copyright arbitration royalty panel to
"establish rates and terms that most clearly represent the rates and terms that would have been negotiated in the
marketplace between a willing buyer and a willing seller." [FN101] Copyright owners anticipate that this should yield
a higher royalty rate than the 6.5% announced under the 1995 amendments. [FN102]

If the 1995 amendments produced a hybrid public performance/protection of reproduction right regime for sound
recordings, the 1998 amendments more closely approach a full fledged performance right, albeit one in which
statutory licensing may outweigh a right to prohibit. Fear of unauthorized copying still pervades the text, but the
text more fully enables copyright owners to benefit from the public performance right as an independent source of
revenue. This marks an important shift in the economics of sound recordings: before these amendments, sound
recording artists and producers derived a return only from sales of phonorecords. Radio broadcasts of the sound
recordings, albeit uncompensated, in effect advertised the recordings, stimulating sales. In the digital environment,
sales of phonorecords may well persist (including via the Internet), but performances of recorded music, whether by
on-demand interactive services, or by webcasting, are likely to displace acquisition of retention copies. This is because
having the recorded performance is likely to matter less to consumers than hearing it, and the digital environment may
well make it as easy for the user to hear the desired performance via transmission as to play her own copy of it. This
observation applies to many kinds of copyrighted works beyond sound recordings. The more easy it becomes to access
and experience works of authorship by means of digital transmission, the less necessary, and perhaps also the less
desirable, it becomes to possess retention copies. The evolution of the scope of the sound recording copyright thus is
just one manifestation of an overall change in the exploitation of works of authorship, a change that also underlies
Congress' creation of a right to "access" copyrighted works, as discussed earlier.


                                       A. TERM EXTENSION

1. How the New Law Works

Keeping up with the Joneses in the European Union, [FN103] Congress has now extended the term of copyright from
fifty to seventy years following the death of the author. In the case of anonymous and pseudonymous works, and works
made for hire, the copyright term has increased from seventy-five years from publication to ninety five years from
publication. [FN104] Congress has also added twenty years to subsisting copyrights, so that a copyright originally
secured under the 1909 Copyright Act, and extended to seventy-five years from publication under the 1976 Act, will
now endure for ninety-five years from publication.

The "Sonny Bono Copyright Term Extension Act" also adds twenty years to the protection of works that were
unpublished as of the effective date of the 1976 Act, but that are published by the end of 2002; these will now be
protected until the end of 2047 (rather than the end of 2027). By contrast, it appears that any work whose author died
more than fifty years ago, and that was not published by December 31, 2002, will enter the public domain on that date.
This means, for example, that an unpublished letter written by the [fictitious] Mayflower passenger Goodspeak
Brown, who died in Massachusetts Bay Colony in 1650, will enter the public domain at the end of 2002, unless
Brown's heirs publish the letter by December 31, 2002, in which case it will remain protected until the end of 2047.

The primary beneficiaries of the extension are the current right-holders; but these often will not be the authors. When,
in 1976, Congress extended the term of copyright from fifty-six years from publication (if the copyright was renewed
at the end of the initial twenty-eight year term), it also afforded authors the right to terminate prior grants, so that the
authors retrieved their rights during the nineteen-year "extended renewal" period. [FN106] By contrast, the new term
extension entitles authors' heirs to terminate grants with respect to the new twenty-year period only if the authors (or
their heirs) failed to exercise their right to terminate the previous nineteen-year extended renewal. [FN107] This
means that grantees of the nineteen-year extended renewal right receive an extra twenty years without having to
negotiate with the author's heirs; if, for example, the author granted extended renewal rights against a lump sum, then
the heirs will receive nothing, and the grantee will get thirty-nine years of exclusive rights for the price of nineteen.

2. Policy Considerations

Was term extension desirable? The law delays the reception into the public domain not only of future, but of
pre-existing works. Because the public domain is a crucial counterpart to the copyright system--the Constitution
authorizes Congress to secure copyright for "limited Times" [FN108]--Congress should not have revised the 1976
Copyright Act's balance between protected and expired works without compelling reasons for term extension. It is not
clear that those reasons have been demonstrated. The considerations favoring term extension fall into two categories:
promotion of authors' rights, and promotion of U.S. interests in international trade of copyrighted works. While the
latter may afford a convincing justification, the former does not.

Regarding the promotion of authors' rights, it is important to emphasize that the traditional justification of copyright
protection, as an incentive to the creation of works of authorship, simply does not apply to extension of the term of
pre-existing works. The law thus cannot enhance the quantum of creativity from the past, but it can compromise the
creativity of the future, by delaying for twenty years the time at which subsequent authors may freely build on these
works. One might justify the extension to pre- existing works as a means of harmonizing the durations of pre-Sonny
Bono Act and post-Sonny Bono Act works, as ss 303 and 304 of the 1976 Copyright Act equalized the terms of 1909
Act and 1976 Act works. However, the strength of this argument depends on the credibility of a life-plus-seventy
(or ninety-five years from publication) term for new works. Here again, the incentive justification falls short: the
prospect of an additional twenty years of protection commencing fifty years after the author's death seems too remote
to impel additional authorial activity. The 1976 Act term already allowed authors to provide for children and
grandchildren; will the addition of great- grandchildren to the prospective beneficiaries of the author's work likely
inspire the creation of another song or sequel? Arguably, with respect to works made for hire, of which the publisher
or producer is the statutory "author," the extra twenty years will enhance the incentive to invest in the creation of
these works. Similarly, the extra twenty years may encourage publishers to greater investment in the dissemination of
works for which they must acquire rights from the author.

Significantly, with respect to preexisting works, the additional twenty years do not, as we have seen, necessarily
redound to authors' benefit. Because the bill does not vest authors' heirs with the extended term, it raises constitutional
questions. The Constitution authorizes Congress to secure exclusive rights for "limited Times" "to Authors." [FN109]
A bill that not only suspends the temporal limitation on copyright protection, but does not even ensure that authors'
families participate in the extension, may be in tension with the constitutional direction.

The international trade considerations flow from the European Union's adoption of a life-plus-seventy (or seventy
years from publication) term, effective July 1, 1995. Under the "rule of the shorter term" imposed by article 7 of the
Directive, an E.U. country is not obliged to grant national treatment with respect to copyright duration, if the work's
term of protection in the country of origin is shorter. [FN110] As a result, unless the U.S. term of protection was
extended, many U.S. works would not enjoy the full duration of copyright available to domestic works in the European
Union. U.S. term extension might therefore put U.S. works on an equal footing with European works. Since the
balance of trade in copyrighted works between Europe and the U.S. heavily favors the U.S., term extension would
benefit U.S. copyright industries.

This argument requires closer examination. It is important to identify what kinds of works are at issue, and to compare
their U.S. terms with E.U. terms. For example, under the 1976 Act, U.S. works-made-for-hire enjoyed a term of
seventy-five years from publication. [FN111] While most European domestic legislation does not incorporate a
general work-for-hire regime, there are some narrower equivalents, for example collective works, whose term under
article 1.4 of the E.U. Directive is seventy years from publication. As to these works, therefore, U.S. term extension
would add nothing to the duration of protection available in the E.U.

With respect to other kinds of U.S. works made for hire, particularly audiovisual works, U.S. term extension may well
increase the duration of protection for these works in the E.U., but it may not necessarily result in affording equal
terms of copyright for U.S. and European films. This is because, in many E.U. countries, films are considered joint
works, whose term is calculated at seventy years from the death of the last surviving designated co-author.
[FN112] In many instances, this term may well exceed ninety-five years from publication. It would therefore be a
mistake to think that U.S. term extension will always give U.S. films protection co-terminous with E.U. films. (Term
equivalence might be achieved were past and present U.S. audiovisual works no longer considered works made for
hire--a proposal that may well be worth contemplating, but that was not urged by the copyright-owner proponents of
term extension.)

In sum, there may be strong internation 13058al trade reasons to favor U.S. term extension, but, as a matter of
copyright law, the extension may have been undesirable. In addition to the criticisms raised above, one may be
concerned that term extension will result in an overall weakening of copyright protection. This outcome may seem
paradoxical, since the idea of term extension is to reinforce copyright protection. In fact, however, one may fear that
defendants will contend, and courts will agree, that copyright now endures for so long a time, that the subsisting (and
excessive) period of exclusivity must be tempered by more vigorous exceptions to copyright protection. The cost of
delaying the public domain therefore may be to compromise the scope and force of copyright coverage throughout the
term of protection.

3. Judicial Challenge to Term Extension

On January 12, 1999, the Eldridge press, an online publisher of public domain works, filed a complaint seeking to
enjoin the Attorney General from enforcing the term extension act, on the grounds that the extension violates the
constitutional provision for "limited Times," particularly with respect to already-created works, and that it effects an
unconstitutional transfer of property from public to private hands. [FN113] Putting aside the important issues of
standing, and of whether, in the absence of any attempt to enforce the act's criminal sanctions, the issue is ripe for
adjudication, the two grounds, as articulated in the complaint, underscore the dubious copyright policy basis for term
extension. Nonetheless, the constitutional objections, while troubling, probably will not prevail.

The complaint's first argument, that extending the term of pre- existing works is "retroactive," and violates the
"limited Times" principle, proves too much. Technically, this term extension is not "retroactive," because the Sonny
Bono Act retrieves no works already in the public domain. It simply retards the date on which currently protected
works would have fallen into the public domain. More significantly, the complaint acknowledges the many times
Congress has previously extended the copyrights in pre-existing works (and even restored copyrights in certain public
domain foreign works), but fails to afford a basis for distinguishing those term extensions from the current one. The
conclusion would appear to be that every term extension since the 1790 Act's fourteen-year, once renewable, term
violates "limited Times." It seems unlikely that a court would follow the logic that leads to such a result.

Nonetheless, one can distinguish at least the most recent term extensions, set forth in s 104A restoring certain
foreign copyrights, and in the 1976 Copyright Act (as well as in the annual term extensions during the 15 years leading
up to the 1976 Act's enactment). [FN114] Assuming, arguendo, that term extension in general may challenge the
"limited Times" restriction, these extensions not only improved the lot of copyright owners, but also implemented U.S.
treaty obligations. That is, s 104A's admittedly retroactive restoration provisions implement the Berne Convention
article 18 obligation to new adherents to protect the copyrights in works of Berne Union members that are still
protected in their countries of origin, but that fell into the public domain in the adhering country before the expiration
of the Berne minimum term. The 1976 Copyright Act's life-plus-fifty term (or 75 years from publication) for new works
met the Berne Convention minimum term, where the 1909 Act's 28- year, once renewable, term did not. [FN115]
Extension of 1909 Act copyrights to 75 years from publication afforded equivalent protection to old- and new-Act
works, and similarly facilitated compliance with Berne standards, thus enabling the U.S. to join the Berne Convention
in 1989. By contrast, no current or impending treaty obligation underlies the Sonny Bono term extension to life-
plus-70 (or 95 from publication). Thus, unlike prior recent extensions, the constitutional treaty power [FN116] does
not bolster Congress' prerogatives under the Copyright Clause. This does not mean that the current extension fails to
pass constitutional muster, but it does suggest that even if the current extension is unconstitutional, its predecessors
are not.

The complaint's second argument for unconstitutionality advances the "public trust" doctrine: the legislature may not
take property that belongs to all, and vest it in private hands. [FN117] This argument, too, however, proves too
much. It appears to imply a static concept of the public domain that any enlargement of copyright, whether in duration
or in scope, would invade. That is, any time Congress broadens the reach of the federal copyright laws, for example,
by extending the public performance right to digital transmissions of sound recordings, including sound recordings
made prior to enactment, Congress has put into private hands control over conduct in which anyone previously could
have engaged freely (in both senses of the word: without constraint, and without paying for it). This in effect equates
Congress' prior (back to 1790?) definition of copyrightable subject matter and protected rights with the constitutional
boundaries of the public domain. It would mean that Congress cannot choose to exercise less than the full scope of
constitutional copyright authority, because once Congress has set the initial terms of copyright coverage, it has also
defined the public domain, and any subsequent reduction in the scope of the public domain would deprive the public of
common property.

Once again, the argument needs a limiting principle. Treaty obligations may afford a partial answer, since the Berne
Convention imposes substantive minimum standards as to subject matter and rights protected. [FN118] But,
more importantly, if the "public trust" doctrine is a viable rule, it is necessary to explain why duration is different from
other aspects of copyright, which Congress may, and often does, enlarge. [FN119] Specifically, why does this term
extension offend the public trust doctrine more than previous extensions of the terms of pre-existing works?

The constitutional Copyright Clause may provide a fuller answer. The clause's formula, "[t]o promote the Progress of
Science ... by securing for limited Times to Authors ... the exclusive Right to their ... Writings ..." [FN120] indicates
that the copyright incentive should be substantial and meaningful ("exclusive Right"), and that it goes to "Authors."
With respect to prior expansions of copyright subject matter or scope, one might contend that the growing gap
between copyright coverage and exploitations of works of authorship diminished the exclusivity of the copyright, and
hence reduced its value as an incentive. Enlargement of copyright subject matter or scope thus adjusted the incentive
to correspond more closely to marketplace and technological realities. For example, in the 1976 Act, Congress defined
"to perform" a work publicly to cover cable retransmissions (and, as it would turn out, digital retransmissions as well),
thus expanding a public performance right whose 1909 Act scope the Supreme Court had previously interpreted as
excluding cable retransmissions. [FN121]

With respect to term extensions, the 1976 Act met the "to Authors" requirement, at least in part, because it afforded
authors (or their heirs) the nonassignable right to terminate the 19-year extended renewal term. [FN122] In the case
of the Sonny Bono Act, by contrast, the benefits "to Authors" are considerably more attenuated, since the law does not
accord those authors (or heirs) who exercised the 19-year termination right even the opportunity to recapture the
extra 20 years.

Thus, where term extension neither corrects diminished incentives to create, nor substantially benefits authors, one
might contend that insufficient public benefit offsets the private gain to copyright owners. Ultimately, however, I
suspect the resolution will rely on the notion that the constitutional "limited Times" restriction imposes a substantive,
and not merely a formal, ceiling on the length of the copyright term. But in that case, the "public trust" doctrine adds
nothing to the argument, and one should return to the main question of how many years are too many for
constitutional tolerance. That is not a question I propose to resolve here, but I believe that any constitutional challenge
to the Sonny Bono extension (should it survive standing and ripeness objections) must offer the court a basis for
identifying when, as a constitutional matter, copyright duration becomes unlimited, without calling into question
previous congressional modifications of copyright term, subject matter, and scope.

                             B. "FAIRNESS IN MUSIC LICENSING"

1. Expanded Exemption from the Nondramatic Music Public Performance Right

Finally, in an act that attracted the criticism of the Copyright Office, the Patent and Trademark Office, the U.S. Trade
Representative, and the Secretary of Commerce, [FN123] Congress attached to the "Sonny Bono Copyright Term
Extension Act" new provisions, cynically titled "Fairness in Music Licensing," that limit authors' and publishers'
exercise of their nondramatic music performing rights, with respect to bars, restaurants, and other small businesses.
These provisions also concern the royalty rate-setting procedures of the music performing rights societies, and afford
the small businesses enhanced opportunities to challenge these rates. This Article will address only the expanded
exception to the public performance right.

The 1976 Act included a special exception, set forth in s 110(5), for the "communication of a transmission embodying
a performance or display of a work by the public reception of the transmission on a single receiving apparatus of a
kind commonly used in private homes ...." [FN124] The legislative history to this provision made clear that the
exemption was limited to small businesses who turn on a "home style" radio or television so that customers could hear
or see the radio or television broadcasts. [FN125] (By contrast, there is no exemption for small businesses who
originate the performance, for example, by playing a cassette or a CD on a home style stereo.) The exemption was not
intended to benefit stores with large surface areas, nor those with sophisticated and extensive stereo systems. Hence,
for example, courts interpreting s 110(5) rejected its application to a retail establishment whose surface area exceeded
1000 feet, and which was communicating radio programming by means of six speakers recessed in the ceiling.

Amended s 110(5)(B) expands the exception's coverage to:

. any establishment with less than 2000 gross square feet of space, and to

. any food service or drinking establishment that has less than 3750 gross square feet of space.

. If the establishment is 2000 gross square feet or more, or if the food service or drinking establishment is 3750 gross square feet or more, the establishment will nonetheless be exempt, if:

. the performance of the radio transmission incorporating nondramatic musical compositions is "communicated by
means of a total of not more than 6 loudspeakers, of which not more than 4 loudspeakers are located in any 1 room or
adjoining outdoor space ...," or

. the performance of a television transmission incorporating nondramatic musical compositions is "communicated by
means of a total of not more than 4 audiovisual devices, of which not more than one audiovisual device is located in
any 1 room, and no such audiovisual device has a diagonal screen size greater than 55 inches, and any audio portion
of the performance or display is communicated by means of a total of not more than 6 loudspeakers, of which not
more than 4 loudspeakers are located in any 1 room or adjoining outdoor space ...." [FN127]

As this summary shows, Congress has wandered far from the original confines of the "home style" exemption. If the
business establishment occupies less than 2000 square feet, or the restaurant occupies less than 3750 square feet, then
there is no limit as to the nature of the stereo equipment that the music user may employ. The limits on the number and
placement of speakers apply only to larger surface-area establishments. But even there, it is no longer necessary that
the receiving apparatus be "of a kind commonly used in private homes"; application of the exemption turns on the
number of speakers, not their nature. Thus, for example, an extremely large restaurant, over 3750 square feet, with
three seating areas, may place two professional quality (and size) speakers in each area, in order to communicate
broadcasts of music radio programming, and yet incur no obligation to pay performance royalties to the composers.

2. Conformity of New s 110(5)(B) with Berne Convention and TRIPs Obligations

The Copyright Office and several branches of the Administration vigorously opposed adoption of an expanded
performance rights exception. Letters from the Copyright Office, the U.S. Trade Representative, the Patent and
Trademark Office, and the Department of Commerce, [FN128] all expressed concern that a broader exemption would
compromise U.S. compliance with its international obligations under the Berne Convention and the TRIPs accord.
[FN129] The PTO letter pointed out that the Irish Performing Rights Organization had already requested that the
European Commission investigate the consistency of s 110(5), in its relatively restrained "home style" form, with
Berne Convention norms. A much expanded exemption will be sure to provoke challenges from U.S. trading partners,
these letters cautioned.

The Copyright Office letter emphasizes the fundamental incompatibility of an expanded s 110(5) exemption with Berne
Convention norms. Register Marybeth Peters cited the WIPO Guide to the Berne Convention's discussion of the
problem. The Guide, in 1978, already addressed the case when:

the work which has been broadcast is publicly communicated e.g., by loudspeaker, or otherwise, to the public. The
case is becoming more common. In places where people gather (cafes, restaurants, tea-rooms, hotels, large shops,
trains, aircraft, etc.) the practice is growing of providing broadcast programmes.... The question is whether the license
given by the author to the broadcasting station covers, in addition, all the use made of the broadcast which may or
may not be for commercial ends.... The Convention's answer is "no." [FN130]

Because the license to the radio station does not cover subsequent communications of the transmission, the business
establishments engaged in those communications should acquire their own licenses from the composers, particularly
when the communications enhance the establishment's commercial appeal. Otherwise, a significant exploitation of
nondramatic musical works goes uncompensated. Given the terms of new s 110(5)(B), it is no longer possible to
rationalize the provision as exempting "incidental uses" of music; now the music may be very much in the foreground
of the business establishment's operations, yet remain uncompensated. Under these circumstances, the U.S. may well
expect that the government of a TRIPs member state, whose composers are deprived of a portion of their performance
rights royalties in the U.S. due to this provision, will initiate a WTO dispute resolution procedure to challenge the
conformity of the provision to Berne Convention and TRIPs norms. One suspects that U.S. performance rights
organizations would welcome such a challenge.


The Digital Millennium Copyright Act and the Sonny Bono Copyright Term Extension Act reflect, in their excess
pagination and lack of overall vision of copyright policy, many of the ills of the modern copyright legislative process.
As Professor Jessica Litman has written, copyright laws tend to be made by the affected industries; whoever is at the
negotiating table will get a piece of the expanding pie, or an offsetting limitation on rights. [FN131] Typically,
Congress simply ratifies the results of the negotiation. The greater the number of interested parties represented, the
longer and more incoherent the statute may become. In the case of the DMCA, the ranks of the represented groups
swelled beyond the usual industrial-strength copyright contenders to include telecommunications industries, online
service providers and other Internet entrepreneurs, as well as educational and research institutions. The entry of
some of these user groups might have helped redirect attention to the bigger picture. In fact, however broad the
parties' initial aspirations, the process appears to have submerged or co-opted its participants. Their contribution
seems primarily to have resulted in securing certain extremely specific exceptions, rather than in articulating general
limiting principles. Those exceptions may now seem like so many steaks thrown to Cerberus: with the guard dog
distracted, the gates of Hell fly open, releasing Titans, hundred- armed monsters, and other over-reaching,
access-wielding copyright owners to plague peaceful (and not-so-peaceful) copyright users. [FN132]

Contentious metaphors [FN133] aside, both the DMCA and the Sonny Bono Act offer much to dislike (whether one is
a copyright owner or a user), of which prolixity is not the least offense. Nonetheless, the DMCA represents an
important adaptation of copyright law to digital communications. How the copyright balance between owners and
users will evolve remains to be seen, and will depend in part on how much flexibility subsists between the lengthy lines.

[FNa1]. Morton L. Janklow Professor of Literary and Artistic Property Law, Columbia University School of Law.
Thanks to Professor Jessica Litman, to Shira Perlmutter, Esq., Associate Register of Copyrights for Policy and
International Affairs, to Bruce Joseph, Esq., and to Andrew Thompson, Columbia Law School Class of 1999. This
Article is adapted from Jane C. Ginsburg, News From the U.S., 179 REVUE INTERNATIONALE DU DROIT
D'AUTEUR [R.I.D.A.] 143 (1999).

[FN1]. Pub. L. No. 105-304, 112 Stat. 2860 (1998) (adding ss 512 and 1201-05 to the Copyright Act of 1976).

[FN2]. Pub. L. No. 105-298, 112 Stat. 2827 (1998).

[FN3]. See S. TREATY DOC. NO. 105-17 (1997).

[FN4]. WCT, art. 11; WPPT, art. 18.

[FN5]. The European Commission, in article 6 of its Proposal for a European Parliament and Council Directive on
the Harmonization of Certain Aspects of Copyright and Related Rights in the Information Society, COM(97)628 final,
1998 O.J. (C 108) 6, also set forth prohibitions against the manufacture and distribution of devices "which have only
limited commercially significant purpose or use other than circumvention ...."

[FN6]. See 17 U.S.C. s 1201(d)-(j).

[FN7]. Id. s 1201(a)(1)(A).

[FN8]. Id. s 1201(a)(1)(C).

[FN9]. See id.

[FN10]. Cf. Harper & Row, Publishers, Inc. v. Nation Enters., 471 U.S. 539 (1985) (rejecting fair use defense
because, inter alia, defendant copied from a "purloined manuscript").

[FN11]. See 17 U.S.C. s 106(3).

[FN12]. See Harper & Row, 471 U.S. 539.

[FN13]. Compare 17 U.S.C. s 1201(a) and s 1201(b). See also id. s 1201(f), (g) (permitting circumvention for
purposes of reverse engineering and encryption research if, inter alia, the user "lawfully obtained the right to use a
copy ...").

[FN14]. In U.S. copyright law, "RAM copying" is considered to effect a reproduction. See, e.g., MAI Sys. Corp. v.
Peak Computer, Inc., 991 F.2d 511 (9th Cir. 1993). The DMCA appears to confirm this interpretation in its Title III,
which introduces an amendment to s 117 permitting the owner or lessee of a computer to "make or authorize the
making of a copy of a computer program if such copy is made solely by virtue of the activation of a machine that
lawfully contains an authorized copy of the computer program, for purposes only of maintenance or repair of that
machine ...." 17 U.S.C. s 117(c).

[FN15]. Cf. id. s 117 (owner of a copy of a computer program may make additional copies "as an essential step in the
utilization of the computer program in conjunction with a machine ...").

[FN16]. See WCT, art. 6 and accompanying Agreed Statement.

[FN17]. See id. art. 8 ("[A]uthors of literary or artistic works shall enjoy the exclusive right of authorizing any
communication to the public of their works, by wire or wireless means, including the making available to the public of
their works in such a way that members of the public may access these works from a place and at a time individually
chosen by them.").

[FN18]. The European Parliament and the European Commission, however, have also proposed to implement WCT,
art. 11 by protecting access controls. See Legislative resolution embodying Parliament's opinion on the Proposal for a
European Parliament and Council Directive on the Harmonization of Certain Aspects of Copyright and Related Rights
in the Information Society (COM(97)628 - C4-0079/98 - 97/0359(COD) (Codecision procedure: first reading) (Jan. 28,
1999). Amendment 31 to article 6 of the Draft Directive defines "effective technological measures" to mean "any
device, product or component incorporated into a process, device or product that in the ordinary course of its
operation controls access to a work or otherwise prohibits or inhibits the infringement of any copyright or rights
related to copyright ...." The Commission has reworded the definition, but has adopted the proposition that
access-control measures are covered. See art. 6.2 of the modified proposed Directive
<http://europa.eu.int/eurlex/en/com/dat/1997/en_597PC0628.html> (visited May 17, 1999).

[FN19]. Jessica Litman, Reforming Information Law in Copyright's Image, 22 U. DAYTON L. REV. 587, 601 (1997).

[FN20]. Joel Brinkley, New Recorded Film Format Draws Criticism from Rivals, N.Y. TIMES, Oct. 12, 1998, at C5.

[FN21]. See 17 U.S.C. s 101 (definition of to "perform" "publicly").

[FN22]. Section 117 permits owners of copies of computer programs to make additional copies in conjunction with
the use of the copy in the machine; RAM copying in the case of the DiVX would be excused if one interpreted s 117 to
extend not only to computer programs, but also to any digitally expressed work. It is not clear that the copy owner's
RAM copying privilege does or should extend so far. In any event, s 117 applies only to owners of copies; licensees
apparently do not enjoy the RAM copying prerogative. If licensees can be excluded, the copyright owner can avoid s
117 by leasing, rather than selling the DiVX discs.

[FN23]. Cf. 47 U.S.C. ss 553, 605 (1998) (prohibiting sale or distribution of devices affording unauthorized access to
cable or satellite signals).

[FN24]. See 17 U.S.C. s 1201(a)(2)(A)-(C).

[FN25]. 948 F. Supp. 1449 (C.D. Cal. 1996).

[FN26]. Note that the Communications Act prohibitions on "black boxes" do not exonerate manufacturers and
distributors whose customers promise to descramble only public domain motion pictures carried by cable or satellite

[FN27]. If information providers did not encrypt public domain documents, then manufacture and distribution of
access-circumvention devices would fail under all the subsections of s 1201(a)(2).

[FN28]. See 17 U.S.C. s 1202(c)(8), discussed infra.

[FN29]. See, e.g., Maljack Prods., Inc. v. UAV Corp., 964 F. Supp. 1416 (C.D. Cal. 1997), aff'd on other grounds sub
nom. Batjac Prods. Inc. v. GoodTimes Home Video Corp., 160 F.3d 1223 (9th Cir. 1998) (panning and scanning).

[FN30]. The Copyright Office might, however, require as part of copyright management information that copyright
owners distinguish copyrightable from public domain components; circumvention devices that decrypted only the
public domain components could legitimately be produced and distributed under s 1201(a)(2). This marking
requirement would be feasible only where the public domain elements are not the objects of copyrightable added
value, such as editorial alterations or enhancements.

[FN31]. 17 U.S.C. s 1201(d)(1).

[FN32]. Id. s 1201(d)(2).

[FN33]. See id. s 1201(d)(1)(A).

[FN34]. Id. s 1201(f)(1)-(2).

[FN35]. See, e.g., Sega Enters. Ltd. v. Accolade, Inc., 977 F.2d 1510 (9th Cir. 1992); Atari Games Corp. v. Nintendo
of Am. Inc., 975 F.2d 832 (Fed. Cir. 1992) (both holding reverse engineering to be fair use when defendant had no
other means to study the program's functionality, and the new program resulting from this study did not infringe the
decompiled program).

Had the drafters adopted language making clear that "access" concerned access to a copy of the work, it would not
have been necessary to create this exception, since the exception itself requires that the reverse engineer have
acquired the right to use a copy of the program.

Query whether the reverse engineering exception set forth in s 1201(f) is narrower than the fair use privilege
articulated by the courts. Section 1201(f)(3) states that the "information acquired" as a result of reverse engineering
may be communicated to others if the reverse engineer "provides such information or means solely for the purpose of
enabling interoperability ...." While the judicial decisions holding reverse engineering to be fair use arose in the
context of interoperability, reverse engineering for other purposes, such as to "gain access to the unprotected ideas
and functional concepts contained in object code," when these cannot be examined without reverse engineering, Sega,
977 F.2d at 1525, might also be fair use.

[FN36]. 17 U.S.C. s 1201(g)(1)(A) & (4).

[FN37]. Id. s 1201(g)(2).

[FN38]. Id. s 1201(g)(3).

[FN39]. Id. s 1201(g)(5).

[FN40]. Id. s 1201(i)(1)(B).

[FN41]. Id. s 1201(i)(1)(C).

[FN42]. Id. s 1201(c).

[FN43]. See id. s 1203. Remedies for copyright infringement are set forth at ss 502-512.

[FN44]. See 17 U.S.C. s 1201(b)(1)(A)-(C).

[FN45]. Id. (emphasis added).

[FN46]. See generally Quality King Distribs., Inc. v. L'Anza Research Int'l, Inc., 523 U.S. 135 (1998) (holding
exclusive rights in s 106 are inherently limited by ss 107-121).

[FN47]. See, e.g., Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994).

[FN48]. 464 U.S. 417 (1984).

[FN49]. 17 U.S.C. s 1201(a)(1)(C)(iii).

[FN50]. Trotter Hardy, Property (and Copyright) in Cyberspace, 1996 U. CHI. LEGAL F. 217, 251.

[FN51]. See, e.g., Richard Stallman, The Right to Read, COMM. ACM, Feb. 1997, at 85 (apocalyptic prediction of
future of copyright in a copy-protected world).

[FN52]. This contention may also imply that prior to the DMCA, every work whether distributed in analog or digital
copies, was made available without restrictions on access or copying. In fact, many documents (frequently composed
of noncopyrightable information) have been the objects of limited distribution and site licenses or shrink-wrap licenses
that contractually limit the library's or user's ability to dispose of the information disclosed in the document. See, e.g.,
ProCD, Inc. v. Zeidenberg, 86 F.3d 1447 (7th Cir. 1996) (upholding shrinkwrap license provision restricting
commercial redissemination of telephone directory information contained in noncopyrightable CD ROM compilation of
white pages phone books).

[FN53]. See Litman, supra note 19.

[FN54]. See Hardy, supra note 50, at 224-25 (discussing "state of the art" limitations on unauthorized copying and

[FN55]. See, e.g., Pamela Samuelson, Good News and Bad News on the Intellectual Property Front, COMM. ACM,
Mar. 1999, at 19.

[FN56]. See, e.g., Rosemont Enters., Inc. v. Random House, Inc., 366 F.2d 303 (2d Cir. 1966); Pierre N. Leval,
Toward a Fair Use Standard, 103 HARV. L. REV. 1105, 1110 (1990); L. Ray Patterson, Free Speech, Copyright, and
Fair Use, 40 VAND. L. REV. 1, 3 (1987); but see Tom W. Bell, Fair Use vs. Fared Use: The Impact of Automated
Rights Management on Copyright's Fair Use Doctrine, 76 N.C. L. REV. 557, 585 n.131 (1998) (arguing the contrary
interpretation is more persuasive).

[FN57]. See Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 442 (1984) (the producer or distributor
of a device used to copy protected works incur no liability for contributory infringement if the device is "widely used
for legitimate, unobjectionable purposes. Indeed, it need merely be capable of substantial noninfringing uses.").

[FN58]. Id.

[FN59]. By contrast, Congress did institute a blank tape and equipment levy with respect to digital audio recording
devices. See 17 U.S.C. ss 1001- 10.

[FN60]. Id. s 1201(k)(1)(A) (18 months from DMCA's effective date, all targeted devices must be equipped with
"automatic gain control"); s 1201(k)(1)(B) (as of DMCA's effective date, any extant devices that currently conform to
automatic gain control or "four-line colorstripe copy control technology" must not modify devices so that they no
longer so conform, and any targeted device manufactured or sold as of DMCA's effective date must conform to
four-line colorstripe technology).

[FN61]. Id. s 1201(k)(2).

[FN62]. WCT, art. 12.

[FN63]. 17 U.S.C. s 1202(c).

[FN64]. See Report accompanying the Berne Convention Implementation Act of 1988, H.R. REP. NO. 100-609, at
37-39 (1988).

[FN65]. The definition of copyright management information appears to give the rights owner the option of including
some, but not all, of the listed elements, see s 1202(c) ("the term 'copyright management information' means any of the
following information ....") (emphasis added). But the definition also empowers the Copyright Office to "prescribe"
other information by regulation, see s 1202(c)(8); the Copyright Office thus might make inclusion of the author's name
a mandatory ingredient of copyright management information.

[FN66]. See, e.g., Lee v. Deck the Walls, Inc., 925 F. Supp. 576 (N.D. Ill. 1996), aff'd sub nom. Lee v. A.R.T. Co., 125
F.3d 580 (7th Cir. 1997).

[FN67]. See H.R. REP. NO. 100-609 (1988).

[FN68]. See 17 U.S.C. s 1202(c)(2).

[FN69]. Cf. id s 1202(c)(5) (requiring that the information identify the screenwriter, performers, and director of an
audiovisual work; audiovisual works can be works made for hire under s 101).

[FN70]. See id. s 512(l).

[FN71]. Id. s 512(a), (b), (c).

[FN72]. 907 F. Supp. 1361 (N.D. Cal. 1995).

[FN73]. By contrast, service providers are insulated from liability for defamation, see Communications Decency Act,
47 U.S.C. s 230(c)(1) (1998):

(1) Treatment of publisher or speaker

No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information
provided by another information content provider.

[FN74]. See discussion supra notes 13-22 and accompanying text.

[FN75]. 17 U.S.C. s 512(c)(1)(A), (B).

[FN76]. Id. s 512(c)(1)(A)(iii).

[FN77]. Id. s 512(c)(1)(C).

[FN78]. Id. s 512(d).

[FN79]. See id. s 512(g)(3).

[FN80]. Id. s 512(g)(2)(C).

[FN81]. Id.

[FN82]. See id. s 512(g)(1). Moreover, any person who knowingly misrepresents that material or activity is infringing
will be subject to damages incurred by any person injured as a result of the service provider's removal of the material,
see s 512(f).

[FN83]. Id. s 512(g)(3)(D).

[FN84]. Cf. id. s 512(h) (authorizing issuance of subpoena to identify subscriber, request must be made to U.S.
District Court), discussed infra note 88 and accompanying text.

[FN85]. If the subscriber does not send a counter notification, it will not have consented to jurisdiction under s
512(g)(3)(D). For application of the minimum contacts standard to website operators, see, for example, Jane C.
Ginsburg & Myriam Gauthier, The Celestial Jukebox and Earthbound Courts: Judicial Competence in the European
Union and the U.S. over Copyright Infringements in Cyberspace, 173 REVUE INTERNATIONALE DU DROIT
D'AUTEUR [R.I.D.A.] 61 (1997).

[FN86]. See 17 U.S.C. s 512(j).

[FN87]. Id. s 512(g)(3)(D).

[FN88]. Id. s 512(h).

[FN89]. This is not to suggest that the U.K. is, or will be, a server haven. The draft Proposal for a European Directive
on Certain Legal Aspects of Electronic Commerce in the Internal Market, COM(98)586 final, 1999 O.J. (C 30) 4, art.
14, also provides for a notice and take down regime.

[FN90]. One might fear that even if they are subject to the jurisdiction of U.S. courts, some foreign host service
providers may be judgment-proof in the U.S., and that no meaningful relief will be available. Anticipating the potential
difficulties of enforcement against off-shore service providers, s 512(j)(1)(B)(ii) authorizes injunctive relief against the
local U.S. access provider to block access to the foreign host. See discussion infra note 91 and accompanying text.

[FN91]. See 17 U.S.C. s 512(j)(1)(B)(ii).

[FN92]. Id. s 512(e)(1).

[FN93]. Id. s 512(e)(1)(A)-(C).

[FN94]. See Digital Performance Right in Sound Recordings Act, 17 U.S.C. ss 106(6), 114(d)-(j), 115(c)(3) & (d).

[FN95]. 17 U.S.C. s 112(a)(2) (emphasis added).

[FN96]. Id. s 114(d)(1)(A) (emphasis added).

[FN97]. Id. s 114(j)(3). The Communications Act defines a "broadcast station" as a "radio station" that disseminates
"radio communications intended to be received by the public ...." 47 U.S.C. s 153(5), (6) (1998).

[FN98]. See 17 U.S.C. s 114(d)(2)(A).

[FN99]. The 1998 webcaster amendments and the 1995 Digital Performance Right in Sound Recordings Act appear,
as a practical matter, to limit the scope of the audio private copying privilege under s 1008. That provision precludes
copyright infringement claims "based on the noncommercial use by a consumer of a device or medium for making
digital musical recordings." Under this provision, home taping off the radio, for example, would not be infringement.
It is not clear that a similar conclusion would apply to home downloading, since a general-purpose computer is not a
"digital recording device" as defined in s 1001(3). It is therefore doubtful that a computer should be considered a
"device for making digital recordings" under s 1008. In any event, the 1995 and 1998 amendments would limit the
opportunities for home users to employ their computers to make digital recordings from webcasts.

[FN100]. Id. ss 114(f)(2) [prior to 1998 amendments], 801(b)(1).

[FN101]. Id. s 114(f)(1)(C)(iii)(2)(B).

[FN102]. See 63 Fed. Reg. 25,394 (1998); the rate is 6.5% of "gross revenues resulting from residential services." 37
C.F.R. s 260.2(a) (1998).

[FN103]. See Council Directive 93/98 Harmonizing the Term of Protection of Copyright and Certain Related Rights,
1993 O.J. (L 290) 9 [hereinafter Duration Directive].

[FN104]. Or from 100 years from creation to 120 years from creation (if counting the term from creation yields a
shorter period than counting from publication). See 17 U.S.C. s 302(c).

[FN105]. The prospect of an additional twenty years of protection was particularly unappealing to libraries; Congress
ultimately modified s 108 of the Copyright Act, which permits several kinds of library photocopying, to enhance the
opportunities, during the last twenty years of copyright, for preservation, scholarship, and research reproductions in
facsimile or digital form. See s 108(h).

[FN106]. See 17 U.S.C. s 304(c) (note, however, that authors did not retrieve their rights with respect to previously
licensed and executed derivative works); see generally Mills Music, Inc. v. Snyder, 469 U.S. 153 (1985). Creators of
works made for hire are not entitled to terminate grants, whether under the nineteen-year or the twenty-year extended
renewal periods, see s 304(d).

[FN107]. See 17 U.S.C. s 304(d).

[FN108]. U.S. Const. art. I, s 8, cl. 8.

[FN109]. Id. (emphasis added).

[FN110]. See Duration Directive, art. 7.

[FN111]. See 17 U.S.C. s 302(c).

[FN112]. See Duration Directive, art. 2.

[FN113]. Eldred v. Reno, complaint available at <http:// cyber.law.harvard.edu/eldredvreno/complaint.html> (visited
May 17, 1999) [hereinafter "Complaint"].

[FN114]. See Complaint, supra note 113, ¶ 33.

[FN115]. See Berne Convention, art. 7.

[FN116]. See U.S. Const. art. II, s 2, cl. 2.

[FN117]. See, e.g., Richard A. Epstein, Congress's Copyright Giveaway, WALL ST. J., Dec. 21, 1998, at A19. See
generally Richard A. Epstein, The Public Trust Doctrine, 7 CATO J. 411 (1987).

[FN118]. See Berne Convention arts. 2, 6bis - 12.

[FN119]. This is not to suggest that there are no constitutional limits on Congress' power to expand or reinforce
copyright. For example, if Congress abolished the fair use defense, the First Amendment would still constrain
copyright owners' prerogatives.

[FN120]. U.S. Const. art. 1, s 8, cl. 8.

[FN121]. See Fortnightly Corp. v. United Artists Television, Inc., 392 U.S. 390 (1968); Teleprompter Corp. v.
Columbia Broad. Sys., 415 U.S. 394 (1974).

[FN122]. See 17 U.S.C. s 304(c)(5).

[FN123]. See letters made part of the floor debates on S. 505, 144 CONG. REC. H9952-54 (daily ed. Oct. 7, 1998),
discussed infra.

[FN124]. 17 U.S.C. s 110(5).

[FN125]. See H.R. REP. NO. 94-1476, at 87 (1976). See also Letter of Marybeth Peters, Register of Copyrights, to
William Hughes, Chair, House Subcommittee on Intellectual Property, Sept. 28, 1994, included in floor debate on S.
505, 144 CONG. REC. at H9953-54.

[FN126]. See Sailor Music v. Gap Stores, Inc., 668 F.2d 84 (2d Cir. 1981). But see Edison Bros. Stores, Inc. v.
Broadcast Music, Inc., 954 F.2d 1419 (8th Cir. 1992); Broadcast Music, Inc. v. Claire's Boutiques, Inc., 949 F.2d 1482
(7th Cir. 1991) (both holding that s 110(5) applied to a retail chain's distribution of radio sets to each of the stores in
the chain).

[FN127]. 17 U.S.C. s 110(5)(B).

[FN128]. See supra note 123.

[FN129]. Agreement on Trade-Related Aspects of Intellectual Property Rights, April 15, 1994, 33 I.L.M. 81 (1994).


[FN131]. See, e.g., Jessica Litman, Copyright and Information Policy, 55 L. & CONTEMP. PROBS. 185 (1992);
Jessica D. Litman, Copyright Legislation and Technological Change, 68 OR. L. REV. 275 (1989); Jessica D. Litman,
Copyright, Compromise, and Legislative History, 72 CORNELL L. REV. 857 (1987).

[FN132]. See discussion supra text following note 41 (negative inference to draw from plethora of specific limitations
on circumvention of access controls is that the new access right is otherwise impregnable).

[FN133]. Cf. Jessica Litman, Debate on Digital Copyright Issues, Some Background to the Debate: Choosing
Metaphors, in AALS Workshop on Intellectual Property (March 11-13, 1999) at 87 (where one comes out on the scope
of digital copyright depends to some extent on what metaphors one employs) (copy on file with author). The
characterization of copyright owners as denizens of the nether reaches, however, is (facetiously) mine.