C o n t r o l l i n g t h e W i r e d

( a n d H e n c e t h e C o n t e n t L a y e r )

In the last chapter, I argued that there is a tension between control at the physical layer and freedom at the code layer, and that this tension affects the incentives for innovation. The original freedom built a commons; more control can undermine that commons; the tragedy is our forgetting the value of the free in our race to perfect control. The “wired” who are affected by the changes I am describing here are not exactly the same “wired” who built the open source and free software movements that I spoke about in chapter 4.
But in a critical sense, they are the same. Both innovate by building on the content that has gone before. Both therefore reveal how much creativity depends upon the creativity that has gone before. Both show, that is, innovation as adding something to the work of others. the Net emerged into the popular press, there was an anxiety among many about what the Net would make possible. People could do things there that we had discouraged or made illegal here. Pornography was the most dramatic example of this anxiety. The freedom of the Net meant, the world quickly learned, the freedom of anyone— regardless of age—to read the obscene. The news was filled with instances of kids getting access to material deemed “harmful to minors.” The demand of many was that Congress do something to respond.
In 1996, Congress did respond, by passing the Communications Decency Act (CDA).1 Its aim was to protect children from “indecent content”
in cyberspace. The act was stupidly drafted, practically impaling itself upon the First Amendment, but its aim was nothing new. Laws have long been used to protect children from material deemed “harmful to minors.” Congress was attempting to extend that protection here.
Congress failed. It failed because the CDA was overbroad, regulating speech that could not be regulated constitutionally. And it failed because it had not properly considered the burden this regulation would impose upon activity in cyberspace. The statute required adult IDs before adult content could be made available. But to require sites to keep and run ID machines was to burden Internet speech too severely. Congress would have to guarantee that the burden it was imposing on the Internet generally was no greater than necessary to advance its legitimate state interest—protecting children.
In 1998, Congress tried again. This time it focused on clearly regulable speech—speech that was “harmful to minors.” And it was much more forgiving about the technology that would permissibly block kids from “harmful to minors” speech. Still, federal courts struck down the law on the ground that the burden it would impose on the Internet generally was just too great.2
These cases evince a distinctive attitude. Though the state’s interest in protecting children is compelling, courts have insisted that this compelling state interest be pursued with care. In effect, a demonstration that the regulation won’t harm the Net too broadly is required before this state interest can be promoted. Facts, and patient review, are the rule in this area of the law of cyberspace.
The state was free to advance its compelling state interest; but it was required, in so doing, not to kill the rest of the Net. the same time that parents were panicking about porn on the Net, copyright holders were panicking about copyright on the Net. Just as parents worried that there was no way to keep control over their kids, copyright holders worried that there was no way to keep control over copyrighted content. The same features of the Internet that made it hard to keep kids from porn also made it hard to keep copyrights under control.
Both forms of panicking were premature. While it is true that the Net as it was originally built made it hard to control content (by either keeping it from kids or keeping it from being copied by kids), the Net as it was originally built is not the Net as it must be. Code made the Net as it was; that code could change. And the real issue for policy makers should be whether we can expect code to be developed that would solve this problem of control.

In Code I argued that in the context of copyright, we should certainly expect such code to be developed.3 And if it were developed as its architects described, then the real danger, I argued, is not that copyrighted material would be uncontrolled: the real danger is that copyrighted material would be too perfectly controlled. That the technologies that were possible and
that were being deployed would give content owners more control over copyrighted material than the law of copyright ever intended.
This is precisely what we have seen in the past two years, but with a twist that I never expected. Content providers have been eager to deploy code to protect content; that much I and others expected. But now, not only Congress but also the courts have been doubly eager to back up their protections with law.
This part I didn’t predict. And indeed, in light of Reno v. ACLU, one would be justified in not predicting it. If parents must go slowly before demanding that the law protect their kids, why would we expect Hollywood to get expedited service? The answer to that question is best left until after we have surveyed the field. So consider the work of the courts, legislatures, and code writers in their crusade to expand the protections for a kind of “property” called IP.

I N CREASING CONTROL

Copyright Bots

The content in these dorm rooms is being used without direct compensation to the original creator. No doubt, no permission was granted for the taping of the LPs. Posters displayed to the public are not displayed with the permission of the poster producers. Books may have been purchased, but there was no contract forbidding passing them to other friends. Photocopying goes on without anyone knowing what gets copied. The lyrics from songs copied down from a recording are not copied with the permission of the original author.Cartooncharacters, the exclusive right of their authors, are not copied and posted, on walls or on computer desktops, with the permission of anyone. All these uses occur without the express permission of the copyright holder. They are unlicensed and uncompensated ways in which copyrighted works get used. Not all of these uses are impermissible uses. Many are protected by exceptions built into the Copyright Act. When you buy a book, you are free to loan it to someone else. You are free to copy a small section of the book and give it to a friend. Under the Audio Home Recording Act, you are free to copy music from one medium to another. Taped recordings of records are therefore quite legal.
But some of these uses of copyrighted works may well be illegal. To post the poster may be a public display of the poster not authorized by the purchase.4 To use icons on your computer of Simpsons cartoons is said by Fox to violate its rights. And if too much of an assigned text has simply been copied by the student, then that copying may well exceed the scope of “fair use.”
The reality of dorm rooms, however—and, for that matter, most private space in real space—is that these violations, if they are violations, don’t matter much. Whether or not the law technically gives a student the right to have a Simpsons cartoon on his desktop, there is no practical way for Fox Broadcasting Company to enforce its rights against overeager fans. The friction of real space sets the law of real space. And that friction means that for most of these “violations,” there is no meaningful violation at all.
Now imagine all this activity moved to cyberspace. Rather than a dorm room, imagine that a student builds a home page. Rather than taped LPs, imagine he produces MP3 translations of the original records. The Simpsons cartoon is no longer just on his desktop; imagine it is also on his Web server. And likewise with the poster: the rock star, we can imagine, is now scanned into an image file and introduces this student’s Web page.
How have things changed?
Well, in one sense, one might say the change is quite dramatic. Now, rather than simply posting this content to a few friends who might pass through the dorm room, this student is making this content available to millions around the world. After all, pages on the World Wide Web are available anywhere in the world. Millions use the World Wide Web. Millions can now, for free, download the content that this student posted.
But there’s a gap in this logic. There are millions who use the World Wide Web. But there are billions of Web pages. The chances that anyone will stumble across this student’s page are quite slight. Search engines balance this point, though that depends upon what’s on a particular page. Most Web pages are not even seen by the author’s mother. The World Wide Web has amazing potential for publishing; but a potential is not a million-hit site. Thus, in reality, this page is effectively the same as the student’s dorm room. Probably more people view the poster on the dorm room window than will wade through the student’s Web page. In terms of exposure, then, moving to cyberspace doesn’t change much.

CPHack

There ’s lots of junk on the World Wide Web. And there’s lots that’s worse than junk. Some of the stuff, for some people, is offensive or worse. The worse includes material deemed obscene or, and this is a very different category, “harmful to minors”—aka pornography. As I’ve described, there’s a long and tedious history of Congress’s efforts to regulate porn in cyberspace.11 I’m not interested in that story here. I’m interested here in the efforts of companies to regulate porn in cyberspace by producing code that filters content. The code I mean is referred to affectionately as “censorware.” Censorware is a class of technology intended to block access to Internet content by forbidding a Web browser to link to the blocked sites. Censorware companies make it their job to skim the Web looking for content that is objectionable, and they then add the link to that content to their list. Their list of banned books is then sold to parents who want to protect their kids. There is obviously nothing wrong with parents exercising judgment over what their kids get to see. And obviously, if the choice is no Internet or a filtered Internet, it is better that kids have access to the Internet. But this does not mean that censorware is untroubling. For often the sites blocked by censorware systems are themselves completely unobjectionable. Worse, sites often are blocked merely because they oppose the technology of censorware. In December 2000, free speech activists at the civil rights group Peacefire reported that a number of censorware systems had begun to block Web sites affiliated with Amnesty International.12 This is just the latest in an endless series of similar cases. They all point to a technology that is fundamentally at odds with the openness and free access of the original Net.


DeCSS

The lawyers for Mattel relied directly upon copyright law. But there was another tack they might have taken—one that will prove much more important as time goes on. In 1998, Congress passed the Digital Millennium Copyright Act (DMCA).18 That act strengthened copyright in a number of ways, but one way was particularly troubling. This was its “anticircumvention” provision.
The anticircumvention provision regulates code that cracks code that is intended to protect copyrighted material. There are two parts to the provision—one that restricts the cracking of code that protects copyrighted material, and one that forbids the creation of code that cracks code that protects copyrighted material. In both cases, the aim of the law is to lend legal support to the tools that copyright holders deploy to protect their copyrighted material.In the ordinary case—with ordinary property—there can be little in this to complain about. It is a crime to steal my car. But obviously, that isn’t enough to stop car theft. So many people install a burglar alarm in their car to further inhibit car theft. But obviously again, that too isn’t enough. So if a legislature, wanting to reduce the risk of theft even more, passes a law that makes it a crime to disable burglar alarms, or to sell tools whose sole purpose is to disable burglar alarms, there can’t be any complaint about these rules, either. If it is wrong to steal a car, and permissible for people to protect their property, it is wrong to crack technology designed to protect the property.
But this story about real property doesn’t map directly onto intellectual property. For as I have described, intellectual property is a balanced form of property protection. I don’t have the right to fair use of your car; I do have the right to fair use of your book. Your right to your car is perpetual; your right to a copyright is for a limited term. The law protecting my copyright protects it in a more limited way than the law protecting my car. This limitation is not just laziness on the part of Congress. The limits on the law’s power to protect copyright are inherent in the clause granting Congress power to regulate copyright, and in the First Amendment’s restrictions on Congress’s power. Copyright law, for example, cannot protect ideas; it can protect only expression. The law’s protection can extend only for limited times. And fair use of copyrighted works is understood to be constitutionally required.

These limitations distinguish copyright as property from ordinary property. And that distinction suggests the trouble with direct analogy from laws protecting burglar alarms to laws protecting code protecting copyrighted work. If copyright law must protect fair use—meaning the law cannot protect copyrighted material without leaving space for fair use—then laws protecting code protecting copyrighted material should also leave room for fair use. You can’t do indirectly (protect fair-use-denying-code protecting copyright) what you can’t do directly (protect copyright without protecting fair use).

i CraveTV

iCraveTV was a site that streamed television content over the Internet.23 The site was located in Canada, where Canadian broadcasting law made such streaming legal. Under Canadian law, anyone has the right to rebroadcast television content, as long as he doesn’t change the content in any way. iCraveTV wanted to take advantage of that right to give computer users access to TV.
Soon after iCraveTV went on-line, copyright holders in the United States brought suit to shut it down. The theory? By setting up an Internet service to broadcast TV, iCraveTV was broadcasting TV into the United States. It was therefore violating U.S. copyright law (by “publicly performing” what iCraveTV streamed to American viewers). Until it could “guarantee,” as the Hollywood lawyers put it, that no United States citizen would get access to this free Canadian TV, the Canadian site had to be shut down. There was a significant dispute about how hard iCraveTV was working to keep non-Canadians out of its site. The Hollywood lawyers hired Harvard Law School Berkman Center’s boy genius Ben Edelman to demonstrate just how easy it was to hack the iCraveTV site. But whether easy or not, the significant issue about the case is this: How much should someone in one country have to be burdened by the laws of another country? For example: Imagine the Chinese government telling the American site China Online24 that it must shut down until it is able to block out all Chinese citizens, since the content on China Online is illegal in China. Or imagine a German court telling Amazon.com that it must stop its selling of Mein Kampf until it can guarantee that no German citizen will be able to get access to that book—since that book is illegal in Germany. Or imagine a French court telling Yahoo! that it has to block French citizens from purchasing Nazi paraphernalia, since that is illegal in France. (Oops, no need to imagine. A French court did just this.25)
In all these cases, we are likely to think that the action of these foreign governments is somehow illicit. That the free exchange of the Net tilts us in favor of open and regular access. That steps to shut down foreign sites because of local laws are the very essence of what the Internet was designed to avoid.
But when it comes to copyright law, we become like the Chinese, or Germans, or French. With respect to law, we too want to insist upon local control—especially because local law here is so strong. So with respect to copyright law, we push local control. And the result is the birth of technologies that will facilitate better local control.
iCraveTV, for example, promised the court that it would develop technology to make it possible to block out everyone except Canadians. Jack Goldsmith and Alan Sykes have described the growing collection of technologies that will achieve the same end.26 These suggest that the future will
be very much like the past: life on the future Internet will be regulated locally, just as life before the Internet was regulated locally.

MP 3
Ten days after launching the service, MP3.com received a letter from RIAA attorneys.28 Its service was a “blatant” violation of copyright laws, said the letter, and MP3.com should take the service down immediately. MP3.com refused, and the lawyers did what lawyers do when someone refuses: they filed suit in U.S. district court, asking for over $100 million in damages.
The RIAA lawyers had a point, if you looked at the statute quite literally. MP3.com may have purchased a bunch of CDs, but it had clearly “copied” these CDs when it created its single, massive database. There was, on its face, then, an unauthorized copy of each of these CDs, and the question became whether or not this copy was nonetheless fair use. Applying the ordinary standard for fair use, the RIAA argued that it was clearly not. This was for a commercial purpose. Thus, fair use was not a defense, and the blatant and willful copying was then a prosecutable offense. When lawyers have such a clean, slam-dunk case, they get very, very sure of themselves. And the papers in the My.MP3 case are filled with outrage and certainty.

But when you stand back from the outrage and ask, “What’s really going on here?,” this case looks a lot different. First, as should be clear, My.MP3 was not facilitating the theft of any music. You had to insert a real CD into your computer before you could get access to the copy on MP3.com’s server. Of course, you could borrow someone else’s CD and hence trick the system into thinking you were the rightful owner of the CD. But you could borrow someone else’s CD and copy it anyway. The existing system permits theft; My.MP3 didn’t add to that.
Second, it should be fairly clear that this service would increase the value of any given CD. Using this technology, a consumer could listen to his or her CD in many different places. Once the system recognized your rights to the music on the CD, the system gave you those rights whenever you were at a browser. That means that the same piece of plastic is now more valuable. That increase in value should only increase the number of CDs that are purchased. And that increase would benefit the sellers of CDs.
Third, it is also fairly clear that exactly the sort of thing that MP3.com was doing could easily have been done by the consumers themselves. Any number of companies have created free disk space on the Internet. Anyone could “rip” his or her CDs and then post them to this site. This ripped content could then be downloaded from any computer. And this download could be “streamed” to be just like the service MP3.com was providing.
The difference is simply that users don’t have to upload their CDs. On a slow connection, that could take hours; on a fast connection, it still can be quite tedious. And a second difference is that the duplication that would be necessary for everyone to have his or her CDs on-line would be much less. Ironically, by shutting down MP3.com, the RIAA was inducing the production of many more copies of the very same music.

Napster

Against the background of MP3.com, Napster does look a bit dicey. After all, the service at issue in MP3.com was a service to give individuals access to content that they presumptively had purchased. On Napster, the presumption is the opposite. There seems little reason for me to download music I already own. But even that is not quite correct. I’ve been a Napster user, though I am not an imaginative user, and I am generally quite lazy. I know exactly what I want to hear, and I know that because I own the music already. But it is easier simply to download and play the music I own on Napster than it is for me to go through the CDs I own (most of which are at home, anyway) and insert the one I want in a player. Thus, while I won’t say that none of the music I have listened to on Napster is music I don’t own, probably only 5 percent is.

Eldred

This latest change outraged many, and especially Eric Eldred. Eldred threatened civil disobedience—promising to publish a series of Robert Frost poems that would have fallen into the public domain. After some of us convinced him that that was a very dangerous strategy, Eldred chose instead to challenge the statute in court. In January 1999, in a federal court in Washington, D.C., Eldred filed his complaint.
Eldred’s claims were simple. If the Constitution permits Congress to grant authors an exclusive right “for limited times,” then the Framers of that power clearly intended that that exclusive right must come to an end. Permitting Congress the power to perpetually extend copyrights would defeat the purpose of the express limitation.
This was Eldred’s claim based on the language of the copyright clause of the Constitution. He also raised an argument based on the First Amendment. The First Amendment says that Congress “shall make no law . . . abridging the freedom of speech, or of the press.” Copyright is a law that certainly limits Eric Eldred’s HTML press. So how are these two provisions of the Constitution—one granting Congress the power to issue copyrights, and the other limiting Congress’s power to “abridge” the freedom of the press—to be reconciled?

COMMONS ON THE WIRES

The internet is a network of networks. In the main, these networks connect over wires. All of these wires, and the machines linked by them, are controlled by someone. The vast majority are owned by private parties— owned, that is, by individuals and corporations that have chosen to link to the Net. Some are owned by the government.
Yet this vast network of privately owned technology has built one of the most important innovation commons that we have ever known. Built on a platform that is controlled, the protocols of the Internet have erected a free space of innovation. These private networks have created an open resource that any can draw upon and that many have. Understanding how, and in what sense, is the aim of this chapter.
The internet is not the telephone network. It is a network of networks that sometimes run on the telephone lines. These networks and the wires that link them are privately owned, like the wires of the old AT&T. Yet at the core of this network is a different principle from the principle that guided AT&T. Like the principle Baran confronted, this principle affects what is allowed and what is not. And like the principle that Baran confronted, this principle has an effect on innovation.
First described by network architects Jerome Saltzer, David Clark, and David P. Reed in 1981, this principle—called the “end-to-end argument” (e2e)—guides network designers in developing protocols and applications for the network. End-to-end says to keep intelligence in a network at the
ends, or in the applications, leaving the network itself to be relatively simple.
There are many principles in the Internet’s design. This one is key. But it will take some explaining to show why.
Network designers commonly distinguish computers at the “end” or “edge” of a network from computers within that network. The computers at the end of a network are the machines you use to access the network. (The machine you use to dial into the Internet, or your cell phone connecting to a wireless Web, is a computer at the edge of the network.) The computers “within” the network are the machines that establish the links to other computers—and thereby form the network itself. (The machines run by your Internet service provider, for example, could be computers within the network.)
The end-to-end argument says that rather than locating intelligence within the network, intelligence should be placed at the ends: computers within the network should perform only very simple functions that are needed by lots of different applications, while functions that are needed by only some applications should be performed at the edge. Thus, complexity and intelligence in the network are pushed away from the network itself. Simple networks, smart apphicadions.
That was my purpose in Code and Other Laws of Cyberspace. There I argued that it was the architecture of cyberspace that constituted its freedom, and that, as this architecture was changed, that freedom was erased. Code, in other words, is a law of cyberspace and, as the title suggests, in my
view, its most significant law.
But in this book, my focus is different. The question I want to press here is the relationship between architecture and innovation—both commercial innovation and cultural innovation. My claim is that here, too, code matters. That to understand the source of the flourishing of innovation on the. Internet, one must understand something about its original design. And then, even more important, to understand as well that changes to this original architecture are likely to affect the reach of innovation here. which code matters? Which parts of the architecture?
The Internet is not a novel or a symphony. No one authored a beginning, middle, and end. At any particular point in its history, it certainly has a structure, or architecture, that is implemented through a set of protocols and conventions. But this architecture was never fully planned; no one designed it from the bottom up. It is more like the architecture of an old European city, with a central section that is clear and well worn, but with additions that are many and sometimes confused.




This design has important consequences for innovation—indeed, we can count three:
• First, because applications run on computers at the edge of the network, innovators with new applications need only connect their computers to the network to let their applications run. No change to the computers within the network is required. If you are a developer, for example, who wants to use the Internet to make telephone calls, you need only develop that application and get users to adopt it for the Internet to be capable of making “telephone” calls. You can write the application and send it to the person on the other end of the network. Both of you install it and start talking. That’s it.
• Second, because the design is not optimized for any particular existing
application, the network is open to innovation not originally imagined. All the Internet protocol (IP) does is figure a way to package and route data; it doesn’t route or process certain kinds of data better than others. That creates a problem for some applications (as we’ll see below), but it creates an opportunity for a wide range of other applications too. It means that the network is open to adopting applications not originally foreseen by the designers.
• Third, because the design effects a neutral platform—neutral in the
sense that the network owner can’t discriminate against some packets while favoring others—the network can’t discriminate against a new innovator’s design. If a new application threatens a dominant application, there’s nothing the network can do about that. The network will remain neutral regardless of the application.
The significance of each of these consequences to innovation generally will become apparent as we work through the particulars that follow. For now, all that’s important is that you see this design as a choice.
The physical platform on which the Internet took off came prewired. It was the telephone wires that linked homes to homes. But the legal right to use the telephone wires to link to the Internet did not come preordained. That right had to be earned, and it was regulation that earned it. Nothing guaranteed that modems would be permitted on telephone lines. Even today, countries in Asia regulate the use of modems on telephone lines.62 What was needed before the revolution could begin was permission to connect the Net to this net.
For certain applications, “best efforts” is not enough. Internet telephony, for example, doesn’t do well when packets carrying voice get delayed. Any delay greater than 250 milliseconds essentially makes the system unusable.67 And as content on the Net moves to real-time, bandwidth demanding technology, this inability to guarantee quality of service becomes increasingly costly.
Put differently, a pricing system for allocating bandwidth solves certain problems, but if it is implemented contrary to end-to-end, it may well do more harm than good.
That is not to argue that it will do more harm than good. We don’t know enough yet to know that. But it raises a fundamental issue that the scarcity mentality is likely to overlook: The best response to scarcity may not be a system of control. The best response may simply be to remove the scarcity. This is the promise that conservative commentator George Gilder reports. The future, Gilder argues, is a world with “infinite” bandwidth. Our picture of the Net now—of slow connections and fast machines—will soon flip. As copper is replaced with glass (as in fiber optics) and, more important, as electronic switches are replaced by optical switches, the speed of the network will approach the speed of light. The constraints that we know from the wires we now use will end, Gilder argues. And the end of scarcity, he argues, will transform all that we do.
There is skepticism about Gilder’s claims about technology. So, too, about his economics. The economist in all of us can’t quite believe that any resource would fail to be constrained; the realist in all of us refuses to believe in Eden. But I’m willing to believe in the potential of essentially infinite bandwidth. And I am happy to imagine the scarcity-centric economist proven wrong.