One need not cheer Grokster. Much of what went on there was indeed illicit. But there are two key things to understand about peer-to-peer networks. The first is that they are hard to police. They have multiple nodes. That is why they work. It means they will have both infringing and noninfringing uses, and the noninfringing uses will be centrally connected to our deepest values of free speech and cultural decentralization. 84
The second feature of peer-to-peer networks is even more basic. They are networks and thus subject to the laws of network economics. In short, they only work well if many people use them. A person who uses a peer-to-peer system that no one else uses is in the position of the person who owns the only fax machine in the world. Peer-to-peer networks provide cheap and unregulable audiovisual or data-heavy "speech" to a mass audience. And if the past is any guide, those networks will also carry large amounts of illicit material, just as photocopying machines (and VCRs) are widely used to violate copyright. The Grokster case makes it harder, but not impossible, to have successful, widely used peer-to-peer systems that are not themselves illicit. If they are widely used, there will be infringing content. If you try to police them and filter them, you will know more about that infringing content and thus might be liable—that was the point of the Napster case. If you do not, you will be failing to take precautions. That was the point of the Grokster case. What is a poor peer-to-peer network to do? Apart from making sure that the last four letters of your service's name are not "-ster," I am hard-pressed to advise you. 85
A decision does not need to make an activity illegal in order to impede it. It only needs to make it uncertain. Already, for example, the free—and so far as I could tell, entirely well- meaning—service "bonpoo," which allowed you to send large file attachments to many people at once, has shut down all of its capabilities except photo transfer. That is simply one trivial instance of a larger harm. Lots of new communications technologies will remain undeveloped because of the uncertainties left by this ruling. 86
My colleague Jennifer Jenkins gave one useful hyperbolic illustration, drawing on earlier work by the Electronic Frontier Foundation: if one were launching the iPod today, it is not clear how it would fare under Grokster's standard. Of course, there is no danger that the iPod will be challenged. It has become respectable and the music companies ended up sanctioning it. But how does it fare if we simply apply the tests laid down in the Grokster case? There is Apple's "tainted" advertising campaign, urging users to "Rip, Mix, and Burn." Does this not suggest complicity, or even intent? There is the fact that the iPod does not restrict itself solely to proprietary formats protected by digital rights management. It also allows uncontrolled MP3 files despite the fact that this format is "notoriously" used to transfer files against the wishes of the copyright owner. This, surely, is a "failure to police." And finally, there is the fact that it would cost about $10,000 to fill an iPod with songs downloaded from iTunes. Clearly Apple must be aware that much of the music that fills iPods is illicitly copied. They are profiting from that fact to drive demand for the product, just as Grokster was profiting from the attractions of illicit traffic to drive people to use their service! 87
No one is going to sue Apple now, of course. In fact, established players in the marketplace are probably fairly safe (and have better lawyers). But what if a product as good as the iPod were being developed now by some upstart company? What if it were no more and no less likely to be used for infringing purposes? Would the business plan ever see the light of day? Or would it be quietly smothered due to legal uncertainty? I have little sympathy for Grokster the company, but the decision that doomed it is a bad piece of technology policy. 88
There is a second reason to dislike the Grokster decision. Despite some of the angst-ridden announcements made when the decision was handed down, the Supreme Court has not killed peer- to-peer systems. The concept is far too well entrenched. But the decision will mean that there are fewer of them that are widely used, easy to operate, and made by responsible and reputable people you can trust. This will probably lessen, but not end, illicit copying online. But that effect comes with a price—it makes our communications architecture a little bit more tightly controlled, reducing but not removing the availability of methods of mass distribution that are entirely outside centralized public or private control. It is another—relatively small—step toward an Internet that is more like cable TV or iTunes, a one-way flow of approved content. One might decide that such a price was well worth paying. But where is the limiting principle or end point of the logic that led to it? 89
There is no provision in U.S. statutory copyright law that imposes liability for contributory or vicarious infringement. None. The patent statute has such a provision; not the Copyright Act. The courts have simply made the scheme up themselves. Then they made up limitations—such as Sony—in order to rein it in. In Grokster, the Supreme Court went further. It made up a new type of "inducement" liability. Fine. As I have tried to indicate here, the decision is not as dreadful as it is reputed to be. But so long as there is any unregulated space in our communications network, some portion of it will have illicitly copied content on it. The more the system is free of central control, the more it is open to use by any citizen, the cheaper it gets—all very desirable characteristics—the more illicit content there will be. That is the premise of the Internet Threat—the belief that control must rise as copying costs fall. I have tried here to suggest an alternative interpretation, the Sony Axiom: without a strong internal set of limitations over copyright, cheaper copying and the logic of the Internet Threat will always drive us toward giving control over our communications architecture to the content industries. 90
There was one particularly striking moment in the Napster oral argument. The lawyer for the recording companies was arguing that Napster was illegal. The judges interrupted, as they often do, and there was a back-and-forth debate about the likely reach of any ruling that would shut down Napster. "I am not trying to say the Internet is illegal," said the lawyer. There was a pause as everyone weighed those words carefully. 91
My response would be "Really? Why not?" The logic of the Internet Threat leads to the position that a network is either controlled or illegal. The better and cheaper the network, the tighter the control needed. The Internet itself could have been designed differently. There could have been more centralized control, filtering of content, a design based on one-way transmission, closed protocols that allow users only a limited number of options. Indeed there were such systems—the national French Minitel system is an example. The Internet represents the opposite set of choices—freedom from centralized control, absence of intervention. In a famous article, Saltzer, Reed, and Clark provided the argument that an "end-to-end" network that is "dumb" and leaves processing to the "ends"—the smart terminals at either end of the wires—will be stable and robust.31 But it will also be remarkably uncontrolled and it will lower global copying costs close to zero for digital content. It is that principle that has made it successful. To put it tersely: the logic of the Internet Threat runs in exactly the opposite direction to the Internet itself. The logic of control is not the logic of the Net. 92
Here is one last thought experiment. Apply the same test I suggested for the iPod to the Internet itself.32 Imagine you knew nothing of the Net. (Those of you who are over twenty-five may actually be able to remember when you knew nothing of the Net.) Imagine that you are sitting in a room somewhere discussing—perhaps with a group of government bureaucrats or some policy analysts from the Commerce Department—whether to develop this particular network. The scientists are enthusiastic. They talk of robustness and dumb networks with smart terminals. They talk of TCP/IP and HTML and decentralized systems that run on open protocols, so that anyone can connect to this network and use it any way they want to. You, of course, know nothing about the truly astounding outburst of creativity and communication that would actually flower on such a system, that would flower precisely because it is so open and no one country or company controls it or the protocols that run it. You do not know that millions of people worldwide will assemble the greatest factual reference work the world has ever seen on this network—often providing their information for free out of some bizarre love of sharing. You do not know about Amazon.com or Hotornot.com or the newspapers of the world online, or search engines, automatic page translation, plug-ins, or browsers. You cannot imagine free or open-source software being assembled by thousands of programmers worldwide. E-mail is only a dimly understood phenomenon to you. Teenagers in your world have never heard of instant messaging—a nostalgic thought. 93
As the scientists talk, it becomes clear that they are describing a system without centralized direction or policing. Imagine that your decision is framed by the logic of control I have described in this chapter, by the fears that the content industry has had for at least the last thirty years—by the logic of the suit they brought in Sony. Imagine, in other words, that we make the up-or-down decision to develop the Internet based on the values and fears that our copyright policy now exhibits, and that the content industries have exhibited for thirty years. There is no way, no way at all, that a network like it would ever be developed. It would be strangled at birth. You would be told by the lawyers and policy wonks that it would be a haven for piracy and illegality. (And it would be, of course—though it would also be much, much more.) You would be told that the system needed to be designed to be safe for commerce or it would never attract investment, that it would need to be controlled and centralized for it to be reliable, that it would need to be monitored to stop it being a hotbed of crime. With the copyright lawyers in the room, you would end up designing something that looked like cable TV or Minitel. The Internet would never get off the ground. 94
The Internet is safe now, of course, because it developed so fast that it was a reality before people had time to be afraid of it. But it should give us pause that if we had our current guiding set of policy goals in place, our assumption that cheaper copying means we need greater regulation, we would never have allowed it to flourish. As Jessica Litman points out, we are increasingly making our decisions about technology and communications policy inside copyright law. We are doing so according to the logic of control that I have sketched out in this chapter. But the logic of control is a partial logic. It blinds us to certain possibilities, ones that have huge and proven potential—look at the Internet. 95
The law has not been entirely one-sided, however. The Sony case drew a line in the sand which promised to halt the inevitable drift toward greater and greater control over communications technology by content owners. It turned out the heavens did not fall. Indeed, the content companies thrived. Perhaps that line was drawn in the wrong place; reasonable people can disagree about that. But Grokster smudges the line without drawing a clear new one. If that new line is drawn according to the logic of control, what technologies will we never see? Could they be technologies that would transform our lives as unimaginably as the Internet has since 1995? 96
I have described the story line—the cluster of metaphors and images and concerns—that pervades our copyright policy. I labeled it "the Internet Threat." In the next chapter, I discuss an alternative story line, a different way of understanding our current policies. The subject of that story line is the best- known example of contemporary attempts to control the digital world, the Digital Millennium Copyright Act or DMCA.
Imagine that a bustling group of colonists has just moved into a new area, a huge, unexplored plain. (Again, assume the native inhabitants have conveniently disappeared.) Some of the colonists want to farm just as they always did in the old country. "Good fences make good neighbors" is their motto. Others, inspired by the wide-open spaces around them, declare that this new land needs new ways. They want to let their cattle roam as they will; their slogan is "Protect the open range." In practice, the eventual result is a mixture of the two regimes. Fields under cultivation can be walled off but there is a right of passage through the farmers' lands for all who want it, so long as no damage is done. This means travelers do not need to make costly and inefficient detours around each farm. In the long run, these "public roads" actually increase the value of the private property through which they pass. They also let the ranchers move their cattle around from one area of pasture to another. The ranchers become strong proponents of "public, open highways" (though some people muse darkly that they do very well out of that rule). Still, most people want open highways; the system seems to work pretty well, in fact. 2
Two new technologies are introduced. First, the automobile is developed. Now thieves can drive through the farmers' fields, stop quickly to grab some corn or a lettuce, and be back on the highway before they can be caught. Of course, the farmers' costs have also fallen dramatically; now they have tractors to work their fields and trucks to take their products to distant markets. The farmers do not dwell on the benefits of the new technology, however. Understandably, they focus more on the profits they could reap if they could get all the advantages of the technology and none of its costs. They demand new legal protections aimed at producing that result. "What's good for agriculture is good for the nation," they say. But now comes the second technological shock—the development of barbed wire. The cost of erecting impassable barriers falls dramatically. The farmers begin to see the possibility of enclosing all of their land, roads and fields alike. This will help them protect their crops from pilfering, but it will also allow them to charge people for opening the gates in their fences—even the gates on public roads. That is a nice extra revenue stream which will, the farmers say, "help encourage agriculture." After all, more fences mean more money for farmers, and more money for farmers means they can invest in new methods of farming, which will mean everyone is better off, right? 3
What is to be done? Assume that each side presents its case to the legislature. There are three obvious possibilities: 4
First, the legislature can simply tell each side to work it out amongst themselves. The law will continue to forbid trespass, but we are neither going to make it a crime to put up a barbed wire fence if it blocks legitimate public rights of way nor to make it a crime to cut a barbed wire fence, unless the fence cutter is also a trespasser. The farmers can attempt to enclose land by putting barbed wire around it. Ranchers and drivers can legally cut those fences when they are blocking public rights of way. Trespass remains trespass, nothing more. 5
Second, the legislature could heed the ranchers' fears that barbed wire will permit the farmers not only to protect their own land, but to rob the public of its existing rights of way, turning open highways into toll roads. (The ranchers, of course, are more concerned with the rights of cattle than people, but most drivers agree with them.) As a result, the state could forbid the erection of a barbed wire fence where it might block a public right of way—classing it as a kind of theft, perhaps. 6
Third, the legislature could take the farmers' side. Theorizing that this new automobile technology presents "a terrible threat to agriculture, because of rampant crop piracy," the state could go beyond the existing law of trespass and make it a crime to cut barbed wire fences wherever you find them (even if the fences are enclosing public lands as well as private, or blocking public roads). To back up its command, it could get into the technology regulation business—making the manufacture or possession of wire cutters illegal. 7
The state picks option three. Wire cutting becomes a crime, wire cutters are classed with lock picks and other "criminals' tools," and the people who make wire cutters are told their business is illegal. A storm of protest arises in the rural driving community. The wire cutter manufacturers claim that their products have lots of legitimate uses. All to no avail: the farmers press on. They have two new demands. Cars should be fitted with mandatory radio beacons and highways put under constant state surveillance in order to deter crop theft. In addition, car trunks should be redesigned so they can hold less—just in case the owner plans to load them up with purloined produce. Civil libertarians unite with car manufacturers to attack the plan. The farmers declare that the car manufacturers are only interested in making money from potential thieves and that the civil libertarians are Nervous Nellies: no one has anything to fear except the criminals. "What's good for agriculture is good for the nation," they announce again. As the barbed wire gates swing shut across the highways of the region, the legislature heads back into session. 8
BETWEEN PARANOIA AND REALITY: THE DMCA 9
I have argued that confusing intellectual property with physical property is dangerous. I stand by that argument. Yet analogies to physical property are powerful. It is inevitable that we attempt to explain new phenomena by comparing them to material with which we are more familiar. While the content companies' tales of "theft" and "piracy" are the most prevalent, they are by no means the only such analogy one can make. In this chapter I try to prove that point. 10
The Farmers' Tale is my allegorical attempt to explain the struggle over the single most controversial piece of intellectual property legislation in recent years, the Digital Millennium Copyright Act, or DMCA.1 The DMCA did many things, but for our purposes its crucial provisions are those forbidding the "circumvention of copyright protection systems," the technological measures that copyright holders can use to deny access to their works or control our behavior once we get access. These measures include encryption, controls on how many times a file can be copied, password protection, and so on. Copyright protection systems are, in other words, the digital equivalent of barbed wire, used to add an additional layer of "physical" protection to the property owner's existing legal protection. But, unlike barbed wire, they can also control what we do once we get access to the property. 11
The rules that forbid circumvention of these systems are logically, if not elegantly, referred to as the anticircumvention provisions. They are to be found in Section 1201 of the Copyright Act, an ungainly and lumpily written portion of the law that was inserted in 1998 as part of the complex set of amendments collectively referred to as the DMCA. I will explain the significance of these rules in a moment. My hope is that the analogy to the Farmers' Tale will make them a little easier to understand—at least for those of you for whom talk of digital rights management, anticircumvention provisions, and network effects is not second nature. 12
Notice the differences between this allegory and the "Internet Threat" story line I described in the last chapter. There are two sets of bad guys in the Farmers' Tale. The greedy thieves (who are still thieves in this story—not heroes) and the greedy farmers who use a genuine if indefinite "threat" posed by a new technology to mask a power grab. The Internet Threat is the story of an industry devastated by piracy, in desperate need of help from the state to protect its legitimate property interests. By contrast, the Farmers' Tale is the story of a self-interested attempt not only to protect property but to cut off recognized rights of public access in a way that will actually make the whole society worse off. The legitimate role of the state in protecting private property has been stretched into an attempt to regulate technology so as to pick winners in the marketplace, enriching the farmers at the expense of consumers and other businesses. In the long run this will not be good for business as a whole. A patchwork of private toll roads is an economic nightmare. 13
That is not the most worrying part of the story: the farmers' proposals are moving in the direction of regulating still more technology—the mandatory radio beacons and constantly monitored roads conjure up a police state—and all to protect a bunch of hysterical vegetable growers whose political clout far outweighs their actual economic importance. 14
Both the Internet Threat and the Farmers' Tale are, of course, ways to understand what is currently going on in the intellectual property wars. In the digital realm, the part of the farmers is played by the content companies, the recording industry associations, the movie and software trade groups. Pointing to the threat of digital piracy, they demanded and received extra legal protection for their copyrighted content. Unlike earlier expansions—longer copyright terms, more stringent penalties, the shrinking of exceptions and limitations, expansions in copyrightable subject matter—this was not a protection of the work itself; it was a protection of the digital fences wrapped around it, and a regulation of the technology that might threaten those fences. 15
What is the significance of this? The digital revolution makes it easier to copy copyrighted content. It also makes it easier to protect that content, and to do so in a more granular and precisely calibrated way. Imagine being able to sell a paperback book that could only be read by the original purchaser or a song that could only be listened to by a particular person in a particular room. Digital rights management technology makes it a lot easier to do these things. Suddenly the copyright owners have considerable physical control over their songs, e-books, and software, even after they have sold them. It is as if the recording industry or the publishers had a representative in your living room. They can use that control not merely to prevent illicit copying but to control and limit usage in ways that go far beyond their exclusive rights under copyright. All of this happens without the law or the state doing anything. Like barbed wire, this is a technological protection measure. 16
Like the farmers, the content companies were not content with their barbed wire alone. They wanted legally protected barbed wire in addition to their existing legal rights under copyright. Under the Digital Millennium Copyright Act, it became illegal to circumvent a technical protection measure such as encryption—the digital barbed wire behind which content companies secrete their work—even if what you did with the content when you got past the barbed wire was a fair use; excerpting a fragment of a film for a school presentation, for example, or making a copy of an encrypted audio file for personal use in another device. In other words, by using digital barbed wire, the content companies could prevent citizens from making the "fair uses" the copyright law allowed. This undermines some of the limitations on their exclusive rights that the Copyright Act explicitly carves out in Section 107, and thus shifts the balance of power that the Copyright Act establishes. Cutting barbed wire became a civil wrong, and perhaps a crime, even if the wire blocked a public road. Under most circumstances, making wire cutters was also now against the law. 17
The ranchers—whose digital equivalents are communications companies and hardware manufacturers—chafed under these new rules. The most powerful groups managed to get special dispensations. Internet service providers, for example, got a qualified immunity from copyright infringement that occurs over their networks. But ordinary citizens, librarians, and civil libertarians also complained, and they were not as well represented in the legislature. It is true that the new rules may help to prevent illicit copying, but they also strike a blow against the exercise of fair use rights—rights that are important both to free speech and competition. Even if the content companies were absolutely right about the threats from digital piracy, this consequence should make us pause. But critics of the DMCA say that there is little evidence that the content companies are right. They quote some of the empirical studies I mentioned in the last chapter, particularly the ones that show no net negative effect from unauthorized music downloading on CD sales. They claim—and they are on strong ground here—that even if there are some losses from the new copying technologies, there are also benefits. Like the farmers, the critics would argue, the content companies take the benefits of the new technology for granted, but wish the law to step in to ameliorate the harms it also creates. And like the farmers, they are not yet satisfied. Their new proposals go even further—scarily further. Thus runs the critics' argument. 18
The critics of the DMCA conjure up a world in which it will be illegal to lend each other books or songs, where it will be impossible for us to copy even small fragments of digital work for criticism or parody, where encryption research will be severely "chilled," and where large quantities of the public domain will be enclosed together with the copyrighted content that the DMCA is supposed to protect. (The Electronic Frontier Foundation's "Unintended Consequences" studies give concrete examples.)2 They think the DMCA undoes the balance at the heart of copyright law, that it can be used to entrench existing businesses and their business methods, that it threatens speech, competition, privacy, and innovation itself. In short, they think the DMCA is the worst intellectual property law Congress has ever passed and view the adoption of similar laws around the world with a reaction little short of horror. 19
Those who supported the DMCA disagree, of course, and do so honestly. They see rampant piracy as a reality and the threat to fair use as some kind of academic hypothetical rarely encountered in reality. What's more, many of them do not think fair use is that important economically or culturally. If markets work well, users could be made to pay for the rights that fair use gives—but only if they wanted them. One could buy expensive digital books which one was allowed to share, quote, or copy for classroom use, and cheaper ones which one had to keep to oneself. Remember that for many of the people who supported the DMCA, fair use is something of a "loophole"; certainly not an affirmative right of the public or a reserved limitation on the original property grant from the state. (Remember the Sony Axiom from Chapter 4?) They find the analogy of fair use to a public road ludicrous. This film, or book, or song, is mine; anything you do with it, or to it, you do at my sufferance. (Remember the Jefferson Warning from Chapter 2?) 20
How has the DMCA worked in reality? Which group's attitudes were vindicated? Two case studies may help us to answer these questions. 21
Infectious Speech: The DMCA and Freedom of Expression 22
Jon Johansen, a 16-year-old Norwegian, was the unwitting catalyst for one of the most important cases interpreting the DMCA. He and two anonymous helpers wrote a program called DeCSS. Depending on whom you listen to, DeCSS is described either as a way of allowing people who use Linux or other open source operating systems to play DVDs on their computers, or as a tool for piracy that threatened the entire movie industry and violated the DMCA. 23
A little background is in order. When you play a commercial DVD, your actions are partly controlled by a simple encryption scheme called CSS, or the Content Scramble System. The DVD Copy Control Association licenses the keys to this encryption system to the manufacturers of DVD players. Without a key, most DVDs could not be played. The manufacturer then embeds this key in its hardware design in such a way that it is easy for your player to decode and play the movie but hard, at least for a person of average technical competence, to copy the decoded "stream." 24
Because the DVD Copy Control Association will only license keys to manufacturers whose DVD players conform exactly to their specifications, the CSS scheme can also be used to control viewers in other ways. For example, DVD players are required to have one of six "region codes," depending on where in the world they are sold. Region 1 is the United States and Canada. Region 2 is Japan, Europe, South Africa, the Middle East, and—bizarrely—Greenland. Region 3 is South Korea, Taiwan, and Hong Kong; and so on. The CSS scheme can be used to restrict a movie to a player with the appropriate region code. If you try to play a movie coded for region 6 (China) in a DVD player from region 1, it will not play. This allows filmmakers to distribute different versions of films to different regions at different times based on sequential release in cinemas, or simply to distribute DVDs with different prices to different regions without worrying about whether the cheaper DVDs will "leak" into the more lucrative markets. CSS and the hardware scheme that unlocks it can also be used to prevent you from fast-forwarding through the commercials at the beginning of the movie if the copyright owner does not want you to, or from skipping the FBI notice. The machine will not do it. In fact, it is deliberately built so that it cannot do it. 25
What we have here is a digital fence that is partly used to prevent copying. Movie studios are understandably worried about the worldwide circulation of perfect digital copies of their movies. CSS was supposed to help to prevent that, or at least make it much harder. But because almost all movies are encrypted with CSS and access to the keys comes with conditions, CSS also allows a more fine-grained control over consumers. Manufacturers are not allowed to make players which can view movies from all region codes or skip portions of the DVD that the owners do not want you to skip. The licensing body puts it this way on its Web site: "Q. Under the terms of the CSS licensing agreement, is it legal for a licensed manufacturer to produce and sell a product which allows a user to disable any CSS protections? A. No. Such products are not allowed under the terms of the CSS license. They are illegal."3 A technology introduced to protect intellectual property rights allows control in ways that those rights alone do not. 26
Before the DMCA, the movie companies could have done exactly this. They could have wrapped their movies in a digital fence. The consumer electronics companies that wanted to could license a key and be allowed to use a trademark that indicated that they were approved by the DVD Copy Control Association. But what if a manufacturer of DVD players felt that American consumers wanted to be able to play their Japanese anime movies without buying another DVD player to do so? Or what if they thought people were antsy and did not want to watch the FBI notice before every film? The manufacturer could have tried to "reverse engineer" the CSS system, to figure out how it worked. If they succeeded, they could make a player that was free of the restrictions that the CSS licensing authority imposed. 27
Of course there were some legal limitations even before the DMCA. Our hypothetical manufacturer could not break into the safe where the CSS code was being held or bribe an employee to provide it. (That would be a trespass or a violation of trade secret law.) It could not violate copyright laws over the various types of software that controlled DVD players. It could not use the trademarks of any of the entities involved, including any seal of approval granted by the DVD Copy Control Association. But it could—at least in the United States—try to reverse engineer the product so as to make a competing product with features that the customers liked more. It would be no more illegal than a company making a cheaper generic razor cartridge that fits my expensive Gillette Mach 3 razor, a generic printer cartridge to replace the expensive one in my Lexmark printer, or, for that matter, a generic remote control for my garage door opener. In each case, of course, the original manufacturer would prefer that I use their products rather than the unlicensed ones. They can design their product to make it hard to use a generic replacement or even tell me that my warranty will be void if I use one. But they cannot say that the unlicensed product is illegal. We are back in option one of the Farmers' Tale, before the legislature acted. The farmers can put up their wire, and even use it to block passage that would be otherwise legal, but it is not a crime to figure out a way through the fence unless the fence cutter is also a trespasser. The DMCA, however, might have changed all of that. 28
Let us return to Mr. Johansen, the 16-year-old Norwegian. He and his two anonymous collaborators claimed that they were affected by another limitation imposed by the CSS licensing body. At that time, there was no way to play DVDs on a computer running Linux, or any other free or open source operating system. (I will talk more about free and open source software later.) Let's say you buy a laptop. A Sony Vaio running Windows, for example. It has a slot in the side for DVDs to slide in and software that comes along with it which allows the DVD reader to decode and play the disk. The people who wrote the software have been licensed by the DVD Copy Control Association and provided with a CSS key. But at the time Mr. Johansen set out to create DeCSS, the licensing body had not licensed keys to any free or open source software developers. Say Mr. Johansen buys the Sony Vaio, but with the Linux operating system on it instead of Windows. The computer is the same. The little slot is still there. Writing an open source program to control the DVD player is trivial. But without the CSS key, there is no way for the player to decode and play the movie. (The licensing authority later did license an open source player, perhaps because they realized its unavailability gave Mr. Johansen a strong defense, perhaps because they feared an antitrust suit, or perhaps because they just got around to it.) 29
Mr. Johansen and his supporters claimed strenuously that DeCSS was not in fact an aid to illicit copying. In fact, they argued that CSS was not really designed to protect DVDs against illicit copying. Commercial DVD "pirates" do not need to crack the CSS encryption. Quite the contrary: they produce exact copies of the DVD, CSS encryption and all, and the buyer's player dutifully decodes it and plays it. Mr. Johansen claimed that his goals were very different from those of the pirates. 30
The motivation was being able to play DVDs the way we want to. I don't like being forced to use a specific operating system or a specific player to watch movies (or listen to music). Nor do I like being forced to watch commercials. When your DVD player tells you "This operation is not allowed" when you try to skip commercials, it becomes pretty clear that DRM really stands for Digital Restrictions Management.4 31
In Mr. Johansen's view, CSS was simply an attempt to control consumers, an attempt which should be a valid target for legal reverse engineering. He has a point. There were indeed other ways to copy DVDs which did not require DeCSS and which gave you files of more manageable size. CSS was indeed more than a simple anticopying device. The entire scheme—the keys, the licenses, the hardware requirements—was designed to give movie studios greater control over their movies in a number of ways, some of them unrelated to copying. On the other hand, he overstated the point. One function of CSS was indeed to make it harder for the average person playing a DVD on a computer to copy the file from the DVD to her hard disk and give it to a friend. It is very easy for the average 14-year-old to take a commercial music CD, change the songs into smaller files in the MP3 format, and share them with a friend. It is not as easy to do the same thing to a DVD—not impossible, just harder—and CSS is one of the reasons why. 32
Mr. Johansen's program, DeCSS, was quickly made available worldwide. Mirror sites provided copies of the program and lists of such locations were easy to find using standard search tools. One such list was provided by the online site run by a magazine called 2600: The Hacker Quarterly. The magazine features everything from pictures of pay phones from around the world to tips on how to hack into computer or telephone systems. Its publisher is one Eric Corley, who goes by the name Emmanuel Goldstein—the resistance leader in George Orwell's 1984. 33
In 1999, Universal City Studios brought suit against a number of individuals for distributing DeCSS. The case was called Universal City Studios v. Reimerdes et al. Corley was among the defendants. The suit prominently included a claim that the defendants were violating the DMCA. It was in this case that the DMCA received its first major legal challenge. 34
Depending on the characterization of the facts, the case seems to be about very different things. It could seem a classic First Amendment fight. ("Plucky magazine publisher told copyright law forbids him from linking to other sites on the Internet!") Or it could seem the very essence of illegal activity. ("Shadowy site which unashamedly caters to computer 'hackers' tries to spread access to the burglar's tools of cyberspace!") 35
Of course, most lawsuits involve conflicts over facts. Much of what lawyers do is put the same facts into different conceptual boxes. But here, merely describing what Corley does, what hackers are, or what 2600 magazine is all about involves one in a profound culture clash. The best way to capture the clash may be to quote from an early entry about Corley in Wikipedia, the remarkable online encyclopedia. 36
The encyclopedia first quotes the description of 2600 magazine from Judge Lewis A. Kaplan, the federal district court judge who decided the Reimerdes case. 37
"2600: The Hacker Quarterly has included articles on such topics as how to steal an Internet domain name, how to write more secure ASP code, access other people's e-mail, secure your Linux box, intercept cellular phone calls, how to put Linux on an Xbox, how to remove spyware, and break into the computer systems at Costco stores and Federal Express. One issue contains a guide to the federal criminal justice system for readers charged with computer hacking. In addition, 2600 operates a web site located at 2600.com (http://www.2600.com), which is managed primarily by Mr. Corley and has been in existence since 1995." 38
The Wikipedia article then continues as follows: 39
While the judge's tone is clearly disapproving, others would point out that bookstores, movies and television channels are filled with material on how to commit murder . . . and that without the efforts of the hacker community, however illintentioned, computer insecurity would be even more of a problem than it already is.5 40
In fact, Judge Kaplan was not entirely disapproving. He mentions articles in 2600 that cover laudable or innocuous tasks, as well as others about tasks that most readers would find objectionable and rightly think to be illegal. But the anonymous volunteer who wrote this version of Corley's Wikipedia entry clearly saw the issue differently. Wikipedia does not portray the hacker community as universally benevolent ("however ill-intentioned"), but that community is also seen as providing a useful service rather than merely a set of how-to guides for would-be digital burglars. 41
To most people, pointing out vulnerabilities in computer security systems seemed, at least in 1999, like telling the world that your neighbor has forgotten to lock his door and all his possessions are there for the taking. But to the online community, it is by no means so clear. From the perspective of those who are knowledgeable in the field, there is a moral continuum. There is clearly legitimate computer security and cryptography research, which includes attempts to break into computer systems to test their defenses—that is how one finds out they are secure, after all. Then there are "hackers." This term could be used to describe those who merely like to program. Richard Stallman, for example, the originator of the free software movement, describes himself thus. But the term could also be used for those who are interested in security or interoperability—making two systems work together. That was Mr. Johansen's declared goal, after all. But some self-described hackers go further. They believe that exploring and disclosing the weaknesses of supposedly secure systems is intellectually fulfilling, practically important, and protected by the First Amendment. They disclaim both moral and legal responsibility for the consequences of their disclosures. (Or at least the negative consequences; they frequently take credit for the positive consequences, such as improved security.) Finally, there are "crackers," whose interest in gaining entry to computer systems is malicious or for financial gain. At what point on this continuum does the activity become legally, or morally, unacceptable? As the Reimerdes trial went on, it became clear that the answer the DMCA gave might not be the same as the one given even by undeniably legitimate computer scientists. 42
A large number of legal arguments were involved in the Reimerdes case, but for our purposes here the most important ones dealt with the relationship between copyright and the First Amendment. What is that relationship? 43
In one obvious sense copyright actually aids free speech. By providing an incentive to create works, copyright "add[s] the fuel of interest to the fire of genius,"6 and thus helps to create the system of decentralized creative production and distribution I described in Chapter 1. But copyright also restrains speech. At its base, it allows an individual to call upon the state to prevent someone from speaking or expressing themselves in a particular way. This may involve a simple refusal to let the speaker use some text, picture, verse, or story in their message, or it may involve a refusal to let them transform it in some way. 44
Neither copyright law nor the American Constitution is blind to these dangers. Copyright has a number of built-in safeguards. The most important of these is that copyright only covers "original expression"—both the ideas and facts in this book can be used by anyone without my permission. Thus, goes the theory, the speaker's freedom of expression is never truly restrained. The only thing I am barred from is using your words, your exact plot, your photograph, your music—not your facts, your ideas, your genre, the events you describe. 45
That is not always enough, of course. Sometimes the problem is that the speaker cannot paraphrase around the restraints posed by copyright. He needs to use the particular text or image in question to convey his message. The ideas, the facts, or a mere paraphrase of the expression would not be enough. In cases like that copyright's answer is "fair use." A politician could not prevent journalists who disagree with him from quoting his autobiography in discussing his life. If an African-American author wishes to tell the story of Gone With the Wind from the slaves' perspective, she may do so in the face of the copyright holders' attempts to stop her. Even fair use, though, may not cover every concern about free expression. Before World War II, Alan Cranston—later a U.S. Senator—wanted to convince American readers that the version of Hitler's Mein Kampf published in the United States was distorted. He believed it to be slanted toward American sensibilities, downplaying both anti-Semitism and German expansionism. His solution? To publish his own English translation, taken direct and uncut from the German edition. He wanted to prove, with Hitler's own words, that the United States had a dangerously distorted version of the German leader. But this is the kind of thing copyright law forbids and it is not clear that fair use allows. (In the end he did it anyway.)7 46
For the moment though, it is enough to realize that copyright law is not immune from the First Amendment or from free speech concerns more generally. If we do not notice that most of the time, it is because the internal limitations of copyright—fair use, the idea-expression distinction, and so on—generally take care of the First Amendment issue, not because the issue was never there. 47
So what First Amendment issues did the DMCA present? Most obviously, the DMCA gave a new right to copyright owners. By using a few simple technological measures, they could distribute a work in a particular format and yet, because of their new intellectual property right, they could make illegal an otherwise lawful process of gaining access for the purposes of making fair use. Of course, the First Amendment allows me to make fair use factually impossible. I can do that without raising any constitutional issues by hiding my manuscript and never letting you see it or just by using unbreakable encryption on my digital products. It allows me to use existing conventional property rights to make fair use illegal. If I own the only copy of the book and it is inside my house, it would be trespass for you to enter. No First Amendment problem there. But in passing the DMCA, Congress had created a new intellectual property right inside copyright law itself, a law aimed directly at expression, that made it illegal to get access for the purpose of making fair use even when you legally bought the physical book, or the physical DVD, and now wish to quote it or parody it. Even that is not the problem. It is that Congress cannot grant the exclusive rights of copyright without simultaneously accompanying them by the limitations of fair use.8 Regardless of what physical constraints and tangible property rights might do to limit my ability to make fair uses, Congress had now, by law, allowed a copyright owner to distribute a particular work with the exclusive rights but without some of those limitations. 48
Imagine that Congress had passed the following law instead of the DMCA: "Any copyright owner can make it illegal to make a fair use of a copyrighted work by putting a red dot on their books, records, and films before selling them. It shall be a crime to circumvent the red dot even if, but for the dot, the use would have been fair." That would be clearly unconstitutional. It gives copyright owners a new intellectual property right to "turn off fair use" in copyrighted works distributed to the mass market. Is the DMCA not the same thing? 49
This was the issue in Reimerdes. True, if I cut through the digital fence on a DVD in order to excerpt a small portion in a critical documentary, I would not be violating your copyright, but I would be violating the anticircumvention provisions. And DeCSS seemed to be a tool for doing what the DMCA forbids. By providing links to it, Mr. Corley and 2600 were "trafficking" in a technology that allows others to circumvent a technological protection measure. DeCSS could, of course, be used for purposes that did not violate copyright—to make the DVD play on a computer running Linux, for example. It enabled various noninfringing fair uses. It could also be used to aid illicit copying. But the alleged violation of the DMCA had nothing to do with that. The alleged violation of the DMCA was making the digital wire cutters available in the first place. So one First Amendment problem with the DMCA can be stated quite simply. It appeared to make it illegal to exercise at least some of the limitations and exceptions copyright law needs in order to pass First Amendment scrutiny. Or did it just make it very, very difficult to exercise those rights legally? I could, after all, make a videotape of the DVD playing on my television, and use that grainy, blurry image in my documentary criticizing the filmmaker. The DMCA would not be violated, though my movie might be painful to watch. 50
The other possible First Amendment problem with the DMCA was that in regulating programs such as DeCSS, the DMCA was actually regulating "speech." The first challenge to the DMCA was that, by making tools like DeCSS illegal, the DMCA took away a constitutionally necessary escape hatch to copyright, thus making copyright law as a whole violate the First Amendment's guarantee of freedom of speech. The second challenge was different. The problem was that the program itself was speech and the DMCA was regulating it illicitly. 51
The reasoning went like this. A computer program is a form of expression and communication. The source code can even be read by human beings. True, it is an abstract form of communication—like musical notation and mathematical algorithms. But those are clearly protected by the First Amendment. Congress could not make Schoenberg's twelve-tone scale illegal or punish mathematicians for physics equations that seemed to support a theory of the universe's origin other than the creationism that is currently so popular. True, the source code is a description of a method of doing something, and the code can, if run on a computer, produce a result—but one could argue that those attributes do not affect the First Amendment's protection. Neither a recipe for hash brownies nor a player piano roll for the Nazi "Horst Wessel" song could constitutionally be prohibited, even though actually to make the hash brownies would be illegal, and even though the piano roll is functional (it "makes" the player piano play the tune). True, most people cannot read computer code, but speech does not need to be common or accessible to be protected. In fact, the courts have even held that the choice to communicate in a particular language is constitutionally protected in some settings. 52
On the other hand, software code is undeniably functional. Lots of functional articles can be said to have some expressive content—a gun, an airbag, a crash helmet, a set of burglar's tools, a computer virus. And many actions have expressive content: a terrorist bombing, for example. Surely these could be regulated by Congress? To the defendants, DeCSS looked like a physics equation, a musical score, or a recipe. To the movie studios, DeCSS had all the First Amendment significance of a crowbar, lock pick, or, for that matter, a car bombing. The same argument was repeated over the hyperlinks that Corley and others provided to sites which carried the DeCSS program. Speech or function? To the defendants, forbidding 2600 to link to these sites was like preventing the Washington Post from describing the availability of drugs on certain blocks of 16th Street. To the movie companies, the hyperlinks were the equivalent of loading potential buyers into a van, taking them down there, and giving them enough money to make the purchase. 53
Which of the two First Amendment arguments is more convincing? That the DMCA is a congressionally created off-switch for fair use? Or that software code is speech and the DMCA restricts it? Like a lot of scholars, before Reimerdes went to trial, I thought that the first argument was by far the more powerful. I still do. I thought the odds of the court buying the "code is speech" argument were low. About that I was wrong, though it turned out not to matter. 54
A number of the reports noted that after some initial skepticism, Judge Kaplan had been impressed by the defendants' expert witnesses, particularly those who had testified that code was speech. When the ruling came out, this impression was confirmed. Judge Kaplan agreed that code was a form of speech or expression. But celebration was premature. Having done so, he disagreed with the defendants' claim that it could not be regulated. 55
Computer code is expressive. To that extent, it is a matter of First Amendment concern. But computer code is not purely expressive any more than the assassination of a political figure is purely a political statement. Code causes computers to perform desired functions. Its expressive element no more immunizes its functional aspects from regulation than the expressive motives of an assassin immunize the assassin's action. In an era in which the transmission of computer viruses— which, like DeCSS, are simply computer code and thus to some degree expressive— can disable systems upon which the nation depends and in which other computer code also is capable of inflicting other harm, society must be able to regulate the use and dissemination of code in appropriate circumstances. The Constitution, after all, is a framework for building a just and democratic society. It is not a suicide pact.9 56
Judge Kaplan is right in saying that there cannot be a bright- line rule immunizing computer code from regulation merely because it has expressive elements. The First Amendment does not protect computer viruses. But the defendants were not arguing that computer code was constitutionally inviolable, only that any law that regulated it had to be subject to First Amendment scrutiny. After all, the government makes the description of how to make a nuclear weapon classified information. That is clearly "speech," but its regulation is also constitutional. The First Amendment is not, and never was, an absolute guarantee of freedom of speech. Instead, the question is whether the law is within the realm of "the freedom of speech" guarantee, which in turn depends on what kind of a law it is. Where does it fit in the "levels of scrutiny" that courts have constructed to discriminate between types of legislation affecting speech? Is the DMCA a "content-based" regulation, such as a law forbidding labor picketing but allowing other kinds of demonstrations? Content-based regulations are given the highest and most demanding level of scrutiny. Alternatively, is it a "content- neutral" regulation, such as a law that forbids talking—about any subject—in a library? To Judge Kaplan, the answer was clear, and grounds for sarcasm. 57
The reason that Congress enacted the anti-trafficking provision of the DMCA had nothing to do with suppressing particular ideas of computer programmers and everything to do with functionality—with preventing people from circumventing technological access control measures—just as laws prohibiting the possession of burglar tools have nothing to do with preventing people from expressing themselves by accumulating what to them may be attractive assortments of implements and everything to do with preventing burglaries. 58
I agree, though it is worth noting that the burglar tool analogy is a disputed one. Johansen claimed DeCSS was more like a screwdriver—something with both licit and illicit uses. 59
So the DMCA was content-neutral regulation. That means it still has to pass a fairly daunting legal threshold. It will only be upheld if "it furthers an important or substantial governmental interest; if the governmental interest is unrelated to the suppression of free expression; and if the incidental restriction on alleged First Amendment freedoms is no greater than is essential to the furtherance of that interest."10 Judge Kaplan felt that the DMCA satisfied that standard. I am not so sure. Yes, the governmental interest in protecting copyright holders' rights is important. And yes, I must disagree with some of my friends in the civil liberties world and say that the government's interest is unrelated to the suppression of free expression. But is "the incidental restriction of First Amendment freedoms no greater than is essential to the furtherance of that interest"? In other words, could the DMCA have achieved its goals without imposing as great a limitation on the expression of people like Mr. Johansen and Mr. Corley? 60
Congress could have passed many laws less restrictive than the DMCA. It could have only penalized the use of programs such as DeCSS for an illicit purpose. If it wished to reach those who create the tools as well as use them, it could have required proof that the creator intended them to be used for illegal purposes. Just as we look at the government's intention in creating the law, we could make the intent of the software writer critical for the purposes of assessing whether or not his actions are illegal. If I write a novel detailing a clever way to kill someone and you use it to carry out a real murder, the First Amendment does not allow the state to punish me. If I write a manual on how to be a hit man and sell it to you, it may. First Amendment law is generally skeptical of statutes that impose "strict liability" without a requirement of intent. But Judge Kaplan believed that the DMCA made the motives of Mr. Johansen irrelevant, except insofar as they were relevant to the narrowly tailored exceptions of the DMCA, such as encryption research. In other words, even if Mr. Johansen made DeCSS so that he and his friends could watch DVDs they purchased legally on computers running Linux, they could still be liable for breaking the DMCA. 61
The DMCA's breadth goes further than its treatment of intent. The statute could have only made it illegal to provide a program yourself. But Judge Kaplan interpreted it to prohibit even linking to a site where the program is to be found. No requirement of intent. No requirement that you actually supply the infringing program. That is a pretty broad interpretation and one which he admits restricts expression. How could he conclude that restrictions this broad were "no greater than essential"? From his rhetoric, the answer is clear. Judge Kaplan believes the story of the Internet Threat I discussed in Chapter 4. He sees DeCSS as a poison. In fact, he thinks it is worse than a poison because it may spread to infect others. It is a disease, a virus. The DMCA is the stern and harsh quarantine required to control it—a digital public health measure. His reasoning is worth quoting at length. 62
There was a time when copyright infringement could be dealt with quite adequately by focusing on the infringing act. . . . The copyright holder . . . usually was able to trace the copies up the chain of distribution, find and prosecute the infringer, and shut off the infringement at the source. In principle, the digital world is very different. Once a decryption program like DeCSS is written, it quickly can be sent all over the world. Every recipient is capable not only of decrypting and perfectly copying plaintiffs' copyrighted DVDs, but also of retransmitting perfect copies of DeCSS and thus enabling every recipient to do the same. . . . The process potentially is exponential rather than linear. Indeed, the difference is illustrated by comparison of two epidemiological models describing the spread of different kinds of disease. In a common source epidemic, as where members of a population contract a non-contagious disease from a poisoned well, the disease spreads only by exposure to the common source. If one eliminates the source, or closes the contaminated well, the epidemic is stopped. In a propagated outbreak epidemic, on the other hand, the disease spreads from person to person. Hence, finding the initial source of infection accomplishes little, as the disease continues to spread even if the initial source is eliminated.11 63
This is a very good point, and one that the critics of the DMCA sometimes gloss over too quickly. The structure of digital replication is indeed different from the old centralized model of copying and distribution. Instead of tracing all illicit copies back to a single infringing printing press, we face the fear that the machinery of piracy can be copied just as fast as the copies it allows us to make. 64
It is here that the defendants lose the battle of the metaphors. Yes, code is speech, it conveys information. But viruses are codes and they convey information too. Judge Kaplan explicitly invokes this comparison several times. Biological viruses are tools for the replication of genetic information. They subvert their hosts' cellular programming to make copies of themselves, just as a computer virus hijacks an infected computer and causes it to send out more copies of the virus. True, DeCSS requires human intervention to download the program and use it. Yet from Judge Kaplan's language it is evident that he sees the program not as an act of expression but as a virus spreading like wildfire. Seen this way, the individual "choices" to download or redistribute are simply the program's method of spreading itself, like the irritation produced by the cold virus that encourages sneezes and coughs, thereby transmitting the illness to others. Just as in an epidemic, the harshest measures are called for. There is no poisoned well here, no pirate with a printing press we can shut down. Anyone is potentially an infringer. Individuals cannot be presumed to be healthy. We cannot give their immune systems, or their motives, the benefit of the doubt. Instead we must see them as potential carriers. The healthy must be quarantined as well as the sick. Facing such a danger, Judge Kaplan agrees that Congress needed to be draconian. We cannot wait for illegal copying. We must strike preemptively at the technology that might enable it. There is no place for inquiries into "intent" here; no way that we can restrict liability to those who actually provide the program. Thus, though "code is speech" and the DMCA does incidentally restrict expression, Judge Kaplan concludes that its restraints are no greater than is necessary. 65
There are three questions here. The first is whether Congress was right. The second is whether, in the context of the movie industry, we can see evidence of the evil it needed to combat. The third question is very different: whether the DMCA is constitutional. In my opinion, the answer to questions one and two is no, for the reasons outlined in Chapter 4's analysis of the Internet Threat. Yes, cheaper copying can increase the rate of illicit copying, but it also lowers advertising costs and offers new business models—Netflix, downloads on demand, viral distribution of trailers, and so on. The technology helps as well as hurts. It does not help the movie industry as much as it might help the music industry, which can more easily distribute its products over the Internet. But the Internet also does not pose as much danger to movies as it does to music. The movie industry's doomsaying aside, there is no exact movie equivalent of Napster and there is unlikely to be one in the near future.12 66
This is not just because movies are longer and harder to download than songs. It is because most people only watch a film once. Most people do not want a library of two thousand films to play again and again. Music is a repeated experience good in a way that movies simply are not, and that social fact profoundly affects the likelihood of downloading as opposed to rental. The transient song on a radio or an Internet stream is not an adequate substitute for possessing the song permanently—something which costs a lot more. Apart from kids' movies, which can be used to induce catatonia in one's progeny time and again, and a few classic favorites, most people do not want to own movies. Watching the film on television or renting it for a night is perfectly satisfactory. Both of these involve little hassle or cost. The content industries are fond of saying "you cannot compete with free." But this is simply not true. Cheap and easily acquired goods of certified quality compete very well with free goods of uncertain quality whose acquisition involves some difficulty. This is one of the main reasons the movie companies were wrong in the Sony case. 67
Thus while Judge Kaplan's discussion of the looming digital Black Death is nicely apocalyptic, it does not seem very accurate. How many of your friends download movies illicitly over the Internet, let alone movies that were ripped from DVDs? Yes, it can be done. But the actual descriptions of the process in the Reimerdes case smack more of bathos than terror. 68
Although the process is computationally intensive, plaintiffs' expert decrypted a store-bought copy of Sleepless in Seattle in 20 to 45 minutes. . . . The decryption of a CSS- protected DVD is only the beginning of the tale, as the decrypted file is very large. . . . One solution to this problem, however, is DivX, a compression utility available on the Internet that is promoted as a means of compressing decrypted motion picture files to manageable size. . . . While the compressed sound and graphic files then must be synchronized, a tedious process that took plaintiffs' expert between 10 and 20 hours, the task is entirely feasible. . . . At trial, defendants repeated, as if it were a mantra, the refrain that plaintiffs, as they stipulated, have no direct evidence of a specific occasion on which any person decrypted a copyrighted motion picture with DeCSS and transmitted it over the Internet. But that is unpersuasive. Plaintiffs' expert expended very little effort to find someone in an IRC chat room who exchanged a compressed, decrypted copy of The Matrix, one of plaintiffs' copyrighted motion pictures, for a copy of Sleepless in Seattle. While the simultaneous electronic exchange of the two movies took approximately six hours, the computers required little operator attention during the interim. 69
So the epidemic threat that hangs over the movie industry consists of the danger that someone will spend fifteen minutes decrypting and ten to twenty hours tediously synchronizing a movie that is then available for a speedy six-hour download? 70
Admittedly, someone only needs to do the synchronizing once. There are newer tools that make the task easier. And we could improve the download time. But even so, would you bother? Faced with the colossal expense and hassle of renting the same movie at Blockbuster for $3, some consumers might prefer this process, I suppose. But I would not sell my shares in movie studios quite yet. In fact, the real threat to movie studios is the large- scale criminal distribution of illicitly copied DVDs—copied bit for bit from the original. The distributors of those do not need to use programs like DeCSS. A more distant threat comes from legal recordings from television made on TiVo's and ReplayTVs—where consumers' actions are legal and CSS is not an issue. So far as we can tell, there is no measurable effect of illicit digital downloads on sales or rentals of DVDs. We could go through the process Judge Kaplan describes, I suppose, just as when the VCR was invented we could have taped movies from television and swapped them with our friends. But as the movie studios discovered after the Sony case, most of us would rather just rent the movie. Because something is possible does not mean it will happen. 71
So in my view, Congress generally overestimated the threat posed by the digital world and underestimated the benefits. In addition, the movie industry is a weak place to make the case for the necessity of the DMCA. Fine, but that is not the legal issue here. The constitutionality of the DMCA does not turn on whether the DMCA was a good idea. That is not the court's decision to make. The question is not even whether the particular industry involved is, in reality, facing much of a threat from digital downloading. The law, after all, exists for all digital works, not just the ones at issue here. The question is whether the restriction on speech imposed by the DMCA was "no greater than is essential." And that is a harder question. 72
I still disagree with Judge Kaplan. A more narrowly tailored statute could have accomplished the DMCA's legitimate goals without impinging as greatly on expression. I think that the rhetoric of the Internet Threat blinded Judge Kaplan to some important issues and led him to overestimate the danger and thus the severity of the measures necessary to combat it. Thus, even under the "code is speech" part of the analysis, I think the DMCA fails First Amendment scrutiny. But if we are confining ourselves to the expression inherent in the software itself, I acknowledge that it is a close call. 73
Sadly, Judge Kaplan spent much less time on the other First Amendment argument against the DMCA—that it is unconstitutional because it gives copyright holders a new intellectual property entitlement, created by Congress under the Copyright Clause, a legal power to deprive users of a constitutionally required limitation on copyright's exclusive rights. In my view, he also framed the argument wrongly when he did discuss it. To be fair, these problems can partly be traced to the fact that the defendants spent most of their energy on the argument that code was expression, paying less attention to everything else. As Judge Kaplan explained it, the claim was that the DMCA might have the effect of restricting an alleged fair use right of access to copyrighted material. Predictably enough, he responded that there was no such right of access. Copyright holders could always lock up the book or restrict entrance to the gallery. In any event, while fair use of DVDs might be curtailed, he argued that most movies are also available on videotape. Even if the film were only available on DVD, the prospective fair user could write down the words and quote them, or record the sound from the screen. Finally, Judge Kaplan pointed out that even if the DMCA might allow a significant erosion of fair use to develop over time, such a problem was not present here. Those making First Amendment claims are sometimes allowed by courts to show that, even if the law as it applied to them were constitutional, it would restrict the First Amendment rights of others. Judge Kaplan declined to apply that doctrine here. In effect, he said "come back when there is a problem." 74
On appeal, the case was decided by a panel led by Judge Jon Newman. Here the fair use argument received more attention but the result was the same: "Come back when there is a problem." Significantly, both courts pointed out another concern. The DMCA could effectively make copyright perpetual because even though the copyright term would expire, the legally protected encryption would continue, and tools such as DeCSS, which would have allowed access to the public domain work, would be illegal.13 This is a major issue because it appears to violate both the First Amendment and the Copyright Clause's requirement of a limited time. The defendants did not spend adequate time on this argument, however, and the courts again left it for later consideration. 75
The court of appeals saw the defendants' argument in just the same way as Judge Kaplan had seen it: a claim that there was a fair use right of actual access to the finest version of every work in every medium, on which the DMCA put a practical limitation. Such a claim was easy to dismiss. There was no such right of guaranteed practical access. Copyright owners could restrict the practical ability to exercise fair use in many ways without the Constitution being involved. In addition, in a world where copyrighted content is frequently available in both analog and digital form, the actual effects of the DMCA might be trivial and were, in any event, constitutionally acceptable. Judge Newman repeated Judge Kaplan's point that one could always make fair use of the work in a way the DMCA did not reach, such as by videotaping a picture of the screen. 76
The fact that the resulting copy will not be as perfect or as manipulable as a digital copy obtained by having direct access to the DVD movie in its digital form, provides no basis for a claim of unconstitutional limitation of fair use. A film critic making fair use of a movie by quoting selected lines of dialogue has no constitutionally valid claim that the review (in print or on television) would be technologically superior if the reviewer had not been prevented from using a movie camera in the theater, nor has an art student a valid constitutional claim to fair use of a painting by photographing it in a museum. Fair use has never been held to be a guarantee of access to copyrighted material in order to copy it by the fair user's preferred technique or in the format of the original. 77
Once the issue is framed this way, the case has been lost. I would argue that there are three baseline errors here: a focus on "affirmative rights of access" as opposed to limits on Congress's power in handing out exclusive rights over expression without their constitutionally necessary limitations, a focus on practical effects of the provisions rather than on formal constitutional limitations on the copyright system over all classes of works, and a confusion between intellectual property rights and physical property rights that goes to the heart of the Jefferson Warning discussed in Chapter 2. The question is not whether users have a constitutionally protected right of practical access to a preferred version of a work. The question is whether it violates the First Amendment for Congress to give to copyright holders an intellectual property right to exempt their copyrighted works in some formats from fair use and other provisions that are necessary for copyright law in general to be constitutional. 78
Remember my earlier example. What if Congress amended Section 1201 to say "Any copyright owner can make it illegal to make a fair use of a copyrighted work by putting a red dot on their books, records, and films before selling them. It shall be a crime to circumvent the red dot even if, but for the dot, the use would have been fair"? This statute, I think, is clearly unconstitutional. It would be no answer to say that some owners will not use the red dot, and even for those that do, there will be older, dotless versions still available. It is irrelevant that I might be able to copy down the crucial lines of the book over your shoulder while you read it and thus claim that I, personally, had not circumvented the dot. The unconstitutionality of the statute does not turn on whether the dots might fall off because of bad adhesive, or whether there are many secondhand bookstores in the area, in which undotted volumes can be found. Even if the red dot rule were only to be applied to hardback books, or graphic novels, or cassette tapes, it would still be unconstitutional. Nor do we have to wait until the entire marketplace is dominated by red-dotted products before considering the issue. It is no answer to say that even before the red dot rule, copyright holders could always have hidden their works, or locked them in safes, or even negotiated individual contracts with the purchasers that have the effect of limiting fair use. That way of framing it just misunderstands the issue on a fundamental level. The claim is not about the happenstance of practical access or the way that a copyright holder can use physical control of an object or existing tangible property rights to undercut fair use. 79
The point is that Congress violates the First Amendment when, with respect to any work, it gives me an intellectual property right to prohibit copying and distribution of an expressive work sold in the marketplace and an additional legal power to opt out of the limitations contained in Section 107 over that work. The bundle of rights conveyed by the DMCA does exactly that. It is not the DMCA alone that we must analyze. The question is whether Congress can give the exclusive rights contained in Section 106 of the Copyright Act over a particular class of works (say digital works), if it also gives a new right to prohibit citizens from gaining access to those works for the purposes of making a fair use. If Judge Kaplan and Judge Newman are correct, then the DMCA gives an entirely new intellectual property right (technically, a legal power) to the copyright holders to do exactly that. To put it the other way around, the DMCA subtracts from the citizen's bundle of entitlements under federal copyright law, the right (technically, lawyers would call it a privilege) to gain access to a work legally in his possession for the purpose of making a fair use. It is that rule change that is unconstitutional, I would argue, and the way Judge Kaplan and Judge Newman frame that issue makes them miss the point. 80