Fill and Sign the Homeowners Quote Sheet Form

Project Looking Forward Sketching the Future of Copyright in a Networked World Final Report May 1998 Prepared for the U.S. Copyright Office by Professor I. Trotter Hardy College of William and Mary School of Law Library of Congress Cataloging-in-Publication Data Hardy, I.Trotter. Project looking forward: sketching the future of copyright in a networked world, May 1998 / final report by I. Trotter Hardy. p. cm. 1. Copyright and electronic data processing—United States. 2. Internet (Computer network). I. Title. KF3030.1.H37 1998 98-27418 346.7304’82—dc21 CIP No copyright is claimed in this report by the U.S. Copyright Office or Professor I.Trotter Hardy. Any use of the material contained herein should be accompanied by a citation to its source. Contents—Summary Executive Summary _______________________ 11 What is the Internet?_____________________________ 11 Technologies relevant to copyright ________________ 13 Today’s legal issues______________________________ 15 Tomorrow’s issues ______________________________ 18 Three patterns of copyright and technology_________ 24 Copyright in a rapidly changing environment_______ 27 1. Introduction___________________________ 29 Methodology ___________________________________ 30 About names used in this Report __________________ 31 About predictions in this Report __________________ 31 2. How the Internet Works Today ____________ 33 2.1 Mechanics of switching _______________________ 35 2.2 Internet geography __________________________ 38 2.3 Internet switching technology _________________ 40 2.4 User capabilities _____________________________ 43 2.5 Who pays for the Internet? ____________________ 46 2.6 How big is the Internet? ______________________ 51 3. How the Internet Will Work Tomorrow _____ 53 3.1 Mechanics __________________________________ 54 3.2 Players and Organizations ____________________ 98 3.3 User Capabilities ___________________________ 105 4. Today’s Legal Issues ___________________ 119 4.1 Web posting _______________________________ 121 4.2 Caching ___________________________________ 124 4.3 RAM copies________________________________ 128 4.4 Intermediaries’ liability______________________ 133
 Contents—Summary 4.5 Wide-spread copying _______________________ 150 4.6 Digital registration and deposit _______________ 151 5. Tomorrow’s Issues _____________________ 153 5.1 Technology and new issues __________________ 154 5.2 “Pure” copyright issues _____________________ 157 5.3 Copyright and other laws____________________ 217 6. Analysis _____________________________ 237 6.1 Three Patterns of Copyright and New Technology238 6.2 History and Analysis of the Three Patterns _____ 244 6.3 Copyright in a rapidly changing environment __ 277 7. Conclusion ___________________________ 283 Today’s Legal Issues ____________________________ 284 Tomorrow’s Issues _____________________________ 285 Three patterns of copyright and new technology ___ 287 Copyright in a rapidly changing environment______ 289 8. Appendices ___________________________ 291 8.1 People interviewed _________________________ 293 8.2 Presentations_______________________________ 299 8.3 Glossary___________________________________ 301
 Contents—Details Executive Summary _______________________ 11 What is the Internet?_____________________________ 11 Technologies relevant to copyright ________________ 13 Today’s legal issues______________________________ 15 Tomorrow’s issues ______________________________ 18 Three patterns of copyright and technology_________ 24 Copyright in a rapidly changing environment_______ 27 1. Introduction___________________________ 29 Methodology ___________________________________ 30 About names used in this Report __________________ 31 About predictions in this Report __________________ 31 2. How the Internet Works Today ____________ 33 2.1 Mechanics of switching _______________________ 35 2.2 Internet geography __________________________ 38 2.3 Internet switching technology _________________ 40 2.4 User capabilities _____________________________ 43 Intranets _____________________________________________43 Search engines ________________________________________44 2.5 Who pays for the Internet? ____________________ 46 Backbone networks ____________________________________46 Regional networks_____________________________________48 Local networks________________________________________49 Network access points _________________________________49 2.6 How big is the Internet? ______________________ 51 3. How the Internet Will Work Tomorrow _____ 53 3.1 Mechanics __________________________________ 54 Economics and pricing _________________________________57 Encryption ___________________________________________61 Electronic copyright management systems________________69
 Contents—Details Digital containers, objects____________________________ 71 “Superdistribution” ________________________________ 75 Proprietary viewers ___________________________________ 76 Fingerprinting, watermarking___________________________ 78 Handles _____________________________________________ 84 Repositories __________________________________________ 86 Convergence of media _________________________________ 86 Dispersed works ______________________________________ 90 Image searching ______________________________________ 91 Push versus pull technology ____________________________ 92 Customized information delivery_____________________ 94 Software agents ____________________________________ 95 3.2 Players and Organizations ____________________ 98 Content providers and carriers __________________________ 98 Everyone will not be a publisher _______________________ 100 Commercial vs. academic entities_______________________ 103 3.3 User Capabilities ___________________________ 105 Internet as backdrop__________________________________ 105 “Ubiquitous computing” ______________________________ 106 Software subscriptions ________________________________ 108 Reading from books versus computer screens ____________ 112 Real time audio and video _____________________________ 115 Interactive “chat” ____________________________________ 117 4. Today’s Legal Issues ___________________ 119 4.1 Web posting _______________________________ 121 4.2 Caching ___________________________________ 124 4.3 RAM copies________________________________ 128 4.4 Intermediaries’ liability______________________ 133 Intermediaries: analysis _______________________________ 135 Direct infringement___________________________________ 137 Strictness in copyright cases ________________________ 139 The tailoring of remedies as a compromise ____________ 142 Vicarious and contributory liability _____________________ 143 Intermediaries: conclusions ____________________________ 149 4.5 Wide-spread copying _______________________ 150 4.6 Digital registration and deposit _______________ 151 The Copyright Office as central repository _______________ 151
 Contents—Details 5. Tomorrow’s Issues _____________________ 153 5.1 Technology and new issues __________________ 154 5.2 “Pure” copyright issues _____________________ 157 Non-public posting ___________________________________157 “Live” information and display forms ___________________162 Factual information ___________________________________164 Avatars _____________________________________________169 Internet broadcasting _________________________________170 In-line linking and framing ____________________________171 Authorization and extra-territoriality____________________178 Works of visual art ___________________________________183 RAM copies and personal privacy ______________________186 RAM copies and junk e-mail and other inflows ___________190 Hypertext links out ________________________________193 Changes in pricing structure ___________________________194 Computer-generated works ____________________________195 Multiple authorship __________________________________197 Libraries, archives, search sites _________________________199 Statutory provisions for archives_____________________201 Facsimile copies of printed text ______________________202 Facsimile copies of digital text _______________________205 Search sites as reproductions ________________________206 Metered use _________________________________________210 Filtering_____________________________________________214 5.3 Copyright and other laws____________________ 217 Copyright and communications law_____________________219 Copyright and bailment law ___________________________224 Copyright and contract law ____________________________229 Copyright and the Uniform Commercial Code____________231 6. Analysis _____________________________ 237 6.1 Three Patterns of Copyright and New Technology238 New subject matter ___________________________________238 New use of existing works _____________________________240 “Decentralized infringement” __________________________240 Another pattern: contract interpretation _________________242 6.2 History and Analysis of the Three Patterns _____ 244 Subject matter: issues _________________________________244 Subject matter: analysis________________________________246 New uses: issues _____________________________________248
 Contents—Details New uses of music ________________________________ 248 Radio ____________________________________________ 250 Cable ____________________________________________ 252 New uses: analysis ___________________________________ 256 Decentralized infringement: issues _____________________ 259 Private guidelines _________________________________ 260 Statutory immunization and compulsory license _______ 261 Fair use versus contributory infringement ____________ 262 Decentralized infringement: analysis____________________ 264 Do nothing _______________________________________ 264 Increase penalties _________________________________ 265 Educate the public_________________________________ 266 Make it lawful ____________________________________ 267 Technological responses____________________________ 269 Summary and conclusion on decentralized infringement __ 276 6.3 Copyright in a rapidly changing environment __ 277 7. Conclusion ___________________________ 283 Today’s Legal Issues ____________________________ 284 Tomorrow’s Issues _____________________________ 285 Three patterns of copyright and new technology ___ 287 Copyright in a rapidly changing environment______ 289 8. Appendices ___________________________ 291 8.1 People interviewed _________________________ 293 8.2 Presentations_______________________________ 299 8.3 Glossary___________________________________ 301
 List of Figures Figure 1: Map of Internet connectivity in June, 1996, from on February 23, 1997. ________________________________________38 Figure 2: Map of Internet connectivity in June, 1997, from on January 25, 1998. ____________________39 Figure 3: MCI’s 1995 Internet backbone network, from on October 28, 1996.____________________________________________________47 Figure 4: MCI’s planned 1996 Internet backbone network, from on October 28, 1996._________________________________________________48 Figure 5: Map showing Internet “Network Access Points,” where various backbone networks connect to each other, from on October 10, 1996. ____________________50 Figure 6: Screen capture of one form of digital object technology from the IBM Corporation, from in the Fall of 1996.____________________________________________________73 Figure 7: One non-watermarked, and one invisibly watermarked, images from the DigiMarc Corporation, from on December 12, 1996. _______________________________________79 Figure 8: Screen capture of a visibly watermarked image from the Corbis Corporation, on December 12, 1996.____________83 Figure 9: Screen capture showing live television broadcast of a soccer game in Spain, from on September 6, 1997.________________________________________88 Figure 10: Screen capture of TotalNews home page, from on March 9, 1997.____________173
 List of Figures Figure 11: Screen capture showing what appears after clicking on the “CNN Interactive” button in the left frame of the TotalNews home page, from on March 9, 1997.____________ 174 Figure 12: Screen capture of the TotalNews web site, from on February 23, 1998. ________ 176 Figure 13: Screen capture of the TotalNews site after clicking on the “CNN Interactive” button. Notice that a new window has opened and that the CNN site is no longer framed. From (in the background) and (in the foreground) on February 23, 1998. _________________________ 177 Figure 14: Information gathered and stored at the Iinternet Archives, from on March 14, 1997. _______________________________________________ 200 Figure 15: Screen capture from , showing the site as it normally appeared on November 11, 1996. _________________ 273 Figure 16: Screen capture of the Web page in Figure 15 saved with a “single button press” (actually using a browser’s “File Save As” menu command and reloaded into the browser as a file from the author’s disk storage). ___________________________________________ 274
 Executive Summary Executive summaries are usually one or two pages. This one is longer—about sixteen pages—but it includes the major points of this Report. A much shorter summary appears in section 7. Conclusion, at page 283. What is the Internet? The Internet is not a “thing” so much as a word that describes both a loosely connected set of computers and the technology that allows them to communicate. It provides a means of transmitting anything that can be represented in “digital” form—text, images, music, videos, cartoons, animations, voice, diagrams, photographs, etc. Millions of computers and tens of millions of people are connected with this technology. The amount of information residing on all the computers connected in this way around the world is quite large, though often overestimated: all the computers connected to the Internet as of mid-winter, 1996, contained less than the amount of information residing on a single large information service like Lexis. Circuit and packet switching. Networks before the Internet, such as the telephone network, were usually “circuit switched.” That is, when a call was placed, telephone company equipment would set up an electrical path from calling party to called party. The entire conversation would flow over that same path; the path would be reserved for the use of the caller for the duration of the call. Internet and other computer networks work differently; they are based not on “circuit switching” but “packet switching.” Transmission is effected by breaking up whatever information is   
 Executive Summary to be sent into small “packets” that are individually routed. Routing is done by computers that are connected to the Internet. A single transmission of, e.g., a twenty-page document, might be accomplished with several dozen packets, each of which might travel a different path through the network. Computers at the sending end are responsible for creating the packets; at the receiving end, computers re-assemble the packets back into a single document. Network technology is a subject of constant research, but the basic idea of packets that are individually routed or “switched” among computers is likely to remain for years to come—if for no other reason than the inertia of current huge investments and reliance on this particular technology. These networks will, however, continue to operate at faster and faster speeds. Flat-rate pricing to disappear. Most users today pay a single flat rate for basic access to the Internet. This will almost certainly change. In time, users will not buy generic “Internet access”; they will either buy a type of access, such as “e-mail access” or “telephony access” or “video access,” or they will buy different amounts of transmission capacity called “bandwidth,” or a combination of these two. Prices for the lower-end services will likely be very low—within the reach of nearly anyone. Intranets. New uses of the Internet and digital technology generally are being developed and refined rapidly. The same technology that allows messages to be sent around the world can also be used to send messages down the hall. Many organizations today are using this technology to support “in-house” e-mail, document transfer, and so on. These in-house Internets are often called “intra-nets.”   
 Executive Summary Technologies relevant to copyright A lot of new technologies have copyright significance. Some important ones are described here. Encryption. Packet switching technology means that messages travel through many different computers, each of which can be owned and controlled by almost anyone located almost anywhere. Consequently, security of Internet communications, including some assurance of privacy, is an altogether different matter than it has been for telephone or paper mail communication. Encryption is therefore an essential element of the Internet of tomorrow. Electronic Copyright Management Systems. Some of today’s technology research is directed toward making digital works harder to copy or easier to license. These schemes are often lumped under the single heading “Electronic Copyright Management Systems” or “ECMS.” The most basic systems rely on “secure transmission” of digital materials between sender and recipient. A stronger degree of protection can be created with a centralized source of access to copyrighted material. For example, a user might gain access to an entire lengthy work like a novel or movie, but with access only for purposes of making use of the work while “connected” to the source site. Digital objects. A “digital object” is a unit of information such as a story, a movie, an image, a game, a computer program, or any other informational work, that is encrypted and then “wrapped” inside a software “envelope.” Anyone receiving a copy of a digital object would be able to read the “wrapper.” Access to the encrypted contents would, however, be conditioned on acceptance of terms specified in the wrapper, such as payment of a royalty fee. Proprietary viewers. With some technologies, a digital object that is unencrypted (after, say, payment of the appropriate fee)   
 Executive Summary becomes a digital work that is “in the clear”: no longer encrypted. The work can be further copied or distributed without authorization. Another technology, called “proprietary viewer” technology, changes that outcome. A proprietary viewer is a computer program that keeps a digital object always under its control. For example, if a buyer of an encrypted novel satisfied the conditions for decrypting the novel, the decryption would remain under the control of the proprietary viewer program. The proprietary viewer program would not allow the user to make unauthorized uses of the work. For instance, if the user had paid for reading only, the proprietary viewer would prevent the user from printing out the work, or from making additional copies, and so on. Watermarks. A digital watermark is a small, almost unnoticeable alteration to a digital work like an image, a photograph, or a sequence of sounds.1 The watermark cannot be perceived with the human eye, but can be detected with a computer program designed for the purpose. Watermarks can be used to embed identifying information into the digital work. Moreover, software for working with images can automatically detect the hidden markings and act accordingly: not permit copying, for example. If the watermark contains a serial number, any given copy of a watermarked work can be logged and recorded somewhere, allowing the author to track down the source of unauthorized copies. Dispersed works. Web pages are often more than a static repository of text and graphics. A given “page” may consist of some material in a single computer file along with links to information stored in many other files. These other files may reside on the same computer as the primary “page” or they may A work of text cannot as readily be watermarked because any alteration of the bits would alter the letters or punctuation and show up as an error. A slight alteration to an image is far less noticeable. 1   
 Executive Summary reside anywhere on the Internet. In addition, much of the page may not “reside” anywhere, but be generated “on the fly” under the control of computer programs (often written in the programming language called “Java”). These well-known facts have a less well-known consequence: much of the material on the Internet cannot be copied by a “single button press.” There is not a single “thing” to be copied. Real time audio and video. A variety of technologies exists to provide audio in digital format. Many radio stations are now broadcasting over the Internet simultaneously with the over-theair broadcasts. Other Internet sites have arisen that offer to play CD’s of the user’s choice, on demand. Similarly, video signals are also being transmitted over the Internet on a regular basis. Today’s legal issues The rate of change in Internet and digital technology is too rapid for any easy classification of issues as “today’s” or “tomorrow’s.” The use of these terms in this Report is as a short-hand for issues that are either widely recognized (“today’s”), or less well known (“tomorrow’s”). Web posting as “publication.” Putting material on a Web site and thereby making it available to a wide audience is not the same thing as distributing thousands of books or magazines to retail stores or consumers. At one time, these differences might have caused some to think that unauthorized Web posting of copyrighted material was consequently not an infringement. However, today it seems clear that such a posting violates the copyright owner’s rights. Less clear is the question whether such a posting should be considered a public “distribution,” or “performance,” or a “reproduction.” Less clear as well is the issue whether such a posting constitutes “publication” for copyright purposes.   
 Executive Summary Caching. Nearly everyone wishes the Internet were faster. Technologies to speed things up are therefore popular. If a message (or text file, or image, etc.) can originate at a point closer to the ultimate consumer rather than farther, or from a computer that is faster or less congested than another, the consumer will be able to obtain it more quickly. Mechanisms to do that—to store information temporarily “closer” to the consumer or on a more powerful or less congested computer, in order to speed up access—are referred to as “caching.” When information is temporarily stored midway between sender and recipient, it is also copied. Copying invokes copyright, and the issue of whether temporary storage is a fair use, or impliedly licensed. RAM copies. “RAM” stands for “random access memory.” It refers to a type of computer memory, the solid state or “internal” memory of computers. With present technology, a digital computer cannot run any of its programs without effecting some sort of copying of information and data into the computer’s internal RAM memory. A number of cases have held that loading instructions from a disk into RAM memory constitutes the making of a “copy” of the program for copyright purposes. That is, the process results in the creation of a copy that if not expressly or impliedly authorized or within some exception such as fair use, is a potentially infringing violation of the copyright owner’s rights. Any access to or use of digital information means access to or use of information that is under the control of a computer. Access to digital information therefore entails the running of one or more computer programs. If a computer program is run, a copy is created in the computer’s internal memory. There is a link, in short, between access to digital information and copyright law: the former implicates the latter. Absent statutory amendments to the contrary, judicial decisions will likely result in copyright law’s becoming the means for governing that access and use.   
 Executive Summary Intermediaries’ Liability. The Internet has given rise to many organizations whose purpose is to provide individuals with online information services and access to the Internet. These “Online Service Providers” or “OSPs”—or for that matter, any computer on the Internet that forwards information from one point to another—are thus situated between their usersubscribers, and the Web sites to which these subscribers connect and browse. Copyright law has a tradition of liability for innocent infringers. No case so far has held that such an intermediary is liable in circumstances where the intermediary’s actions were wholly innocent. Nevertheless, a current issue is whether and to what extent OSPs as intermediaries ought to be liable for copyright infringement when their users engage in copyright infringing activities. Wide-spread copying. The most obvious concern about copyright and the Internet today is that a lot of copying takes place over the Internet. Web pages, graphics, news articles, e-mail messages, and other digital works are frequently taken from one source and used in another or circulated to large numbers of people. The phenomenon of falling reproduction costs resulting in widening dispersal of reproduction activities is labeled “decentralized infringement” in this Report. Decentralized infringement has a long history: photography, photocopying, analog audio tapes, digital audio tapes, video tapes, computer software, and so on. Responses to the phenomenon have included the application of fair use principles; the development of fair use guidelines; statutory compulsory licenses; application of the doctrine of contributory infringement; and new technological developments to raise the cost of unauthorized copying. Digital registration. Project Looking Forward did not address the technical issues of digital deposit or registration because the Copyright Office is actively pursuing those concerns through other more focused projects like CORDS. But interviews conducted as part of the project did reveal some misconceptions in the technical community about the nature and function of the   
 Executive Summary Copyright Office as a central repository. In particular, some think that the Copyright Office’s function is to authenticate authors: to verify that anyone claiming to have authored a work is in fact the author. This view can lead to the conclusion that if digital techniques allow a work to be “self-authenticating,” a central registry is no longer needed. Much less understood or appreciated is the fact that the Office’s files provide a means for the public to track changes of ownership after initial authorship. Digital selfauthentication alone cannot accomplish this type of tracking. Tomorrow’s issues Non-public posting. It is generally conceded that the unauthorized posting of information on a world-wide-accessible web site constitutes infringement. Whether denominated “distribution” or “display” or “performance,” the act of such posting violates the copyright owner’s rights. Crucial to this conclusion, though, is a finding that something “public” has happened: either a public distribution, a public display, or a public performance. Any of those things done privately would not constitute infringement. Increasingly, Web technology is being used for less than world-wide access. Corporate “intranets” exemplify this trend. The narrowing range of accessibility raises the “slippery slope” issue of when a posting is “public” and when private. The slippery slope problem is common to all law, and certainly to copyright law, but the matter is made more difficult here because the copyright definition of “public” is tied in part to physical places. “Live” information and display forms. The Internet will commonly feature links to changing databases of information. Users accessing this information through a World Wide Web “browser” will be able to pull up constantly updated information that is reformatted on the fly for Web display. The techniques for doing this live, constant updating of Web page information are fairly straightforward today, and will only get easier tomorrow.   
 Executive Summary Copyright law, and particularly the Copyright Office registration process, will be faced with an enormous amount of frequently changing information displays, raising the issues of both copyright registration of rapidly-changing works, and the copyright of derivative works. Factual information. Facts are not copyrightable. Much of what makes the Internet useful is the ability to pull down timely information on a moment’s notice. Much timely information is factual: weather, sports scores, locations, directions, forecasts, departure times, maps, etc. The general prohibition against copyrighting facts, and the Feist case’s prohibition against “sweat” copyrights, will prevent copyright protection for these works. In the absence of other forms of protection, such as for factual databases, Web site owners who would like to charge for essentially factual material will find it difficult to do so. Avatars. Computers can create three-dimensional figures and animate them. Experiments are already underway on the Internet in allowing users to create animated figures to represent themselves for purposes of on-line interactive discussions. Some obvious questions will arise from this technology: Can one copyright one’s avatar as a fictional creation? Does this bring into play other legal issues like trademark and right of publicity? Internet broadcasting. Today, a number of radio stations are broadcasting their performances directly over the Internet, simultaneously with their over-the-air broadcasts. Some video broadcasting is also being undertaken. The quality of both is rapidly improving. Because the cost of becoming an Internet radio or television station is substantially less than becoming a broadcast station, we will see a large increase in the number of such stations. For those that are already broadcast entities, and have been paying ASCAP and BMI royalties, an issue may arise as to whether that existing royalty payment means that royalties have “already been paid” and further royalties are not owed. For new Internet stations, the issue will be whether such stations   
 Executive Summary should be treated just like broadcasters for copyright purposes, or in some other way. In-line linking and framing. “Framing” is a technique whereby one Web site’s pages are made to appear in a smaller window inside another Web site’s page. The copyright issues in framing have already arisen in litigation, though have not been resolved: does the provision of such links involve the creation of a “copy?” A “public display?” A “derivative work?” Authorization and extra-territoriality. Copyright’s traditional commitment to “territoriality”—that United States copyright law does not apply outside the territory of the United States—has led to judicial decisions holding that an authorization made within the United States of the use of copyrighted works outside the United States does not violate an owner’s right to “authorize” uses of a work under section 106. With the sharp rise in easy transmission of information around the globe, we will likely see a sharp increase in the number and variety of such circumstances. The increase will accordingly put new stress on the interpretation of the “authorization” right in section 106. Works of visual art. Parts of the current Copyright Act continue to be tied to the notion that works of authorship exist in printed or paper form. The moral rights that inhere in “works of visual art,” for example, depend on definitions of “works of visual art” that are directed to traditional media like paper and canvass, rather than to digital works fixed in electronic form. The application of these rights to the electronic medium will therefore raise new issues of interpretation tomorrow. RAM copies and other issues. When a user browses the WWW with browser software, the various sites browsed may be able to gather personal information about that user from the user’s computer. Copyright does not apply to most personal information because such information is factual and cannot be considered an original work of authorship. Yet, information extracted   
 Executive Summary automatically from a user’s own computer will typically cause some form of computer processing on that user’s computer. That computer processing may entail the creation of copies of the programs in the RAM memory of the user’s computer. Can the user argue that the copying of programs in that way is a copyright violation? Similar arguments can be made about the receipt of unwanted e-mail—receipt takes place under the control of computer programs on the user’s own computer and hence may cause some form of copying into the user’s RAM memory. Can the user argue that the receipt of unwanted e-mail therefore causes an unauthorized copy to be made of computer programs? Similar arguments can also be made about unwanted hypertext links to one’s own computer. When those links are followed by others, they cause the running of certain computer programs, and perhaps the copying of those programs. If unauthorized, is that a copyright infringement? Changes in pricing structure. The Internet seems to represent a “convergence” of different media into one digital transmission stream. Oddly, the Internet will likely also begin to exhibit “divergence,” with different service quality levels becoming available at different prices. This shift may be perceived in some quarters as reducing access to high quality information by poorer schools, individuals, and other organizations. Quite possibly what is perceived as a reduction in access will encourage some to press for changes in copyright law. The pressure will take the form of urging alterations in copyright’s protection to ensure greater access to information, as by expanding the exceptions to copyright’s subject matter. Computer-generated works. Humans can program computers to create, and the resulting computer programs can clearly be original and copyrightable; but what about the output of the computer program? The traditional answer has been that a computer cannot be a copyright “author” because computers are not human. Yet, computers are getting better all the time. If computer poetry today is not very good, then it will be better   
 Executive Summary tomorrow. Computers can draw maps, for example. Maps can be copyrighted. Though not conceptually a problem, the copyrightability of computer-created works is nonetheless likely to cause factual disputes over which portions of such works are of human origin and hence copyrightable. Multiple authorship. Collaboration among multiple authors is a phenomenon of long standing. In this sense, multiple authors of digital works raise no new issues. But lower communications costs may expand the opportunities for multiple authorship well beyond what we have seen in the past. How should works by many authors, perhaps very many authors, be treated for copyrighted purposes? Libraries, archives, search sites. Huge amounts of information are circulated daily on the Internet. Organizations have begun keeping archives of much of this material. These archives exist and function by making “copies” of the materials they store and serve, invoking copyright issues. Provisions of the current Act deal specifically with “Libraries and Archives,” but these provisions are directed to reproduction of works that were originally fixed on paper and are of uncertain application to works originally created in electronic format. Metered use. The use of digital information can be monitored by a computer. That is, given that the information is “computerized” in the first place, it requires a computer to display or perform or copy or otherwise make use of it. This allows a very fine level of metering and billing for information access. In the past, we have been used to being charged for larger units of information: hundred-page books, two-hour movies, fifteen-song CD’s, etc. Arguments can be made that current copyright provisions like fair use were created in a world of that sort of “coarse-grained” metering, and that those provisions should be changed to recognize a different world if policies underlying the provisions are to be preserved.   
 Executive Summary Filtering. Filtering technologies, or “filters,” consist of computer software that screens all material received over computer communications lines. Filters, designed originally to screen out pornography, can also be used to provide automatic inclusion or exclusion of elements and pieces from various information sources. They might be used to combine two or more WWW pages together, for example, or re-arrange their elements. They might also be used to delete advertising from a WWW page. When an Internet user makes use of filtering software to alter a copyrightable work such as a Web page, does that user infringe the right of the copyright owner to authorize the preparation of derivative works? Copyright and other laws. Copyright provides a form of “property” rights in some forms of information and permits the exchange of information “products.” Other laws like those surrounding broadcasting and telephony were conceived primarily as regulations of activities rather than the distribution of products. As the Internet brings about an increasing overlap between information sold as a “thing” and information distributed as part of a service, conflicts between fundamentally different legal regimes will arise. Even the notion of information as a “product” will cause conflicts. Many of our laws evolved over the centuries to deal with tangible objects in a world in which tangible objects and information products seemed obviously different. As the number and variety of transactions in information products grows, more of these transactions will have the earmarks of transactions in tangible products. That will lead to more frequent and complex interactions between two sets of laws: those that evolved to handle tangible goods, and those— principally copyright law—that evolved to handle information goods. The intersection of copyright law and Bailment law is one example. Contract law, and especially proposals to add a new section on software licensing to The Uniform Commercial Code, may also give rise to issues of conflict between copyright and other legal regimes.   
 Executive Summary Three patterns of copyright and technology Copyright law has had to accommodate new technologies repeatedly over the two centuries of its existence. This accommodation gives rise to broadly similar issues time and time again. Those issues can be summarized as issues of new subject matter, new uses of existing copyrighted works, and decentralized infringement. New subject matter. Some technologies create a new type or a new medium of expression. They give rise to the “subject matter” question: should the new type of expression, or the expression that is recorded in a new medium, fall within copyright’s subject matter—i.e., be appropriate for copyright’s protection? Subject matter issues are less in evidence today than in the past. Congress worded the 1976 Act’s concept of subject matter much more broadly and generally that it had been before. As a result, most works fixed in new types of media no longer give rise to subject matter questions: it is the “work” that is protected, not the medium. Questions can still arise over new types of works, though. The menu command structure in the Lotus v. Borland case can be analyzed as one such type of new subject matter. New use of existing works. Second, some technologies create a new way of using existing copyrighted works. These technologies give rise to the “new use” question: does the new use of an existing copyrighted work infringe the author’s rights? This has been a troublesome issue for copyright law over its history, arising in connection with phonograph recording of music; radio air play of music; cable-casting of broadcast programs; and others. The issue will likely recur with Internet transmission of a variety of digital works. It is more troublesome than the subject matter issue because the Copyright Act defines “rights” (and hence   
 Executive Summary “infringement”) in ways that are far more technology- and medium-dependent than is its definition of subject matter. Almost invariably, the arguments over new uses take the same form each time. Copyright owners argue that if others are making money from their works, they deserve a share. Those making the new use argue that their use only advertises or extends the copyrighted works to a wider audience and so should not be burdened with a royalty obligation. Just as invariably, the arguments overlook the central—and largely unanswerable— question: will the new use eventually grow to supplant the old use? If it will, there are strong arguments that royalty payments will be needed to preserve incentives. If it will not, those arguments are much weaker. An appreciation for possible future market effects of a new use technology is essential to good copyright decision making in the present, but seldom in evidence. “Decentralized infringement.” Finally, some technologies neither create new forms of expression nor allow new ways of using existing expression, but rather make methods of infringement far cheaper than before and also harder for copyright owners to discover. This sort of technological development raises what can be called the issue of “decentralized infringement.” Decentralized infringement occurs, for example, when the development of photocopy machines makes reproduction of printed materials much cheaper and easier than before, or the development of home tape recording makes the reproduction of music much easier than before, or the development of the personal computer makes the reproduction of computer programs easier, or the development of the Internet makes copying a wide variety of digital materials easier than before, etc. Decentralized infringement usually raises issues of fair use, proposals for statutory amendment, or compulsory licenses. It typically motivates copyright owners to seek counter-measures: technological developments the purpose of which is to raise the   
 Executive Summary cost and inconvenience of making unauthorized uses of their works. Often, the issues are never definitively resolved. This may be an appropriate outcome, given that developments in the technologies that raise or lower costs will proceed unevenly and unpredictably. Digital vs. analog: really different? Arguments that digital works are fundamentally different from analog works depend on an assumption that digital works are easily and cheaply copied. They are therefore arguments about decentralized infringement. For copyright purposes, however, the differences between digital and analog works is not one of technology, but of the cost of unauthorized uses. If technological developments raise the cost or inconvenience of making unauthorized uses of digital works, the assumed differences between analog and digital works will shrink proportionally. Many developments sketched in the Report have just that effect, including watermarks, encryption, dispersed works, proprietary viewers, digital objects, and others. From a copyright perspective, these technologies may end up making the world of tomorrow more like the world of yesterday than like the world of today.   
 Executive Summary Copyright in a rapidly changing environment Will copyright decrease in importance in tomorrow’s world of digital communications? A conclusion that copyright is less necessary in tomorrow’s digital world rests on a crucial assumption: that the digital world will continue to evolve in the way that some parts of that world seem to be evolving now. That evolution features new business models such as giving away certain digital works for free and earning revenue from other services: technical support, updating, advertising. We should not be optimistic, however, about our ability to foresee the future evolution of either technology or new business models, especially those that relate to the Internet: we were not too good about predicting the rise of the Internet and the World Wide Web in the first place. Individuals and businesses may choose to produce things for which copyright is important, or they may choose to produce other things. The existence of copyright law gives them that option. A copyright possessed by an owner can be either asserted or waived as the public demand dictates. A copyright not possessed can only be “waived,” as it were—it cannot be unilaterally created even if the public’s good makes the assertion of copyright desirable. Copyright functions in a changing world not to enable a particular technology, business model, or market; but rather to preserve choices among different technologies, business models, and markets. Whether it should function this way or not, however, is a policy determination beyond the scope of this Report.   
 1. Introduction This is the final report from Project Looking Forward. That project represents one avenue by which the Copyright Office has sought to predict the future evolution of the Internet and related digital communications technologies, and to identify the copyright issues that might arise as a result. This Report is not a report from any individuals or organizations or focus groups or anyone other than myself. Though I have been informed by a great many people, all of whom I thank and express my utmost appreciation to,2 this Report reflects solely my own views. During the course of this study, I gave a number of presentations to various audiences on “Internet and Copyright” issues. At many In addition to those formally listed in Appendix 8.1 People interviewed, I express special appreciation to members of the U.S. Copyright Office who were helpful in countless ways, in particular to Register Marybeth Peters, Sandy Barnes, David Fernandez, Mary Gray, Shira Perlmutter, and Jerry Tuben; I also thank William and Mary law student Carrie Schneider for proof reading; and the College of William and Mary, then-law school Dean Thomas Krattenmaker, and most especially Shirley Aceto, for finding a way for me to take the leave necessary for this project. 2  
 Introduction of these presentations, I was able to get some very helpful questions and suggestions about project Looking Forward. I have listed these presentations, though not the attendees individually, in the Appendices, at Section 1.1. Note, however, that I always spoke at presentations as a William and Mary faculty member, not as a Copyright Office member; views expressed at these presentations were my own, not those of the Office. The Office requested an identification of issues, but not any proposal for their resolution. The resolution of most copyright issues touches on matters of important public policy, matters that can only be addressed with careful deliberation and consultation with affected parties and with Congress. I thank the Copyright Office, particularly the Register, Marybeth Peters, and the Associate Register for Policy & International Affairs, Shira Perlmutter, for initiating the project, for giving me the opportunity to participate in it by conducting this study, and for providing on-going support and encouragement. Methodology I relied in the preparation of this Report on several things. Most of all, I have spent a good deal of time talking with people who have an interest in and knowledge of the Internet. I have also continued to conduct research of my own, both into the legal issues and into the Internet as a technology. I use the Internet on a daily basis, and have found the net itself to be helpful in learning about what is happening with the technology and what is likely to be upcoming in the near future. In addition, in cooperation with the Copyright Office, I planned and led three small “focus group” sessions, two of them “live” and in-person; the third conducted by e-mail. The first live session was held on the campus of Stanford University, in conjunction with the Stanford law school and in particular with Prof. Carey Heckman, the Director of the Law and Technology Policy Center  
 About names used in this Report there. The second live session was held at the Copyright Office itself in Washington, D.C. The e-mail conference was, of course, in “cyberspace.” About names used in this Report It is helpful in explaining things to give real-world examples. Often when an actual example would be useful, I refer to particular companies or products or services by name, frequently with a World Wide Web address included. I do not endorse or support or have any affiliation with any of these—or any other— commercial organizations. This Report also includes a number of images that are “screen captures,” that is, images taken directly from a computer screen. Most of these images are of pages from different sites on the World-Wide Web, pages that may themselves be copyrighted. Having concluded that for purposes of this Report the reproduction of these page images is a fair use, I have not obtained permission for their use. Naturally, the copyright status of this Report overall does not affect the copyright status of these images. About predictions in this Report The Copyright Office asked me to look at the future evolution of the Internet and related digital technologies, and then to try to predict what new copyright issues that future will raise. This should help the Office to be informed and hence to be prepared to address tomorrow’s issues and controversies. To accomplish that goal, one need not accurately predict a single “future.” One can predict many futures, or many different aspects of “the future” and then suggest some, perhaps even differing, copyright consequences. To that end, I have not tried to sketch out a single vision of digital communications of tomorrow, but rather  
 Introduction looked at various pieces of digital communications technology: different mechanisms, functions, user capabilities, and so on. Nor have I tried to ensure that all these various elements are consistent with each other and with a single evolutionary path; perhaps they are, perhaps they are not. My assumption is that it will help the Copyright Office to know what it can about the possible copyright issues of tomorrow, not what the world of tomorrow will look like in general.  
 2. How the Internet Works Today An appreciation for the “mechanics” of the Internet’s functions is very helpful to anyone who wants to understand the legal issues. The salient point about the net is its distributed nature—that is, information flowing through the Internet travels over many different paths from the point of origin to the destination, even during a single “session” of Internet use. This mechanism contrasts sharply with the existing voice telephone network, over which phone calls travel over a single path for the duration of a call. Two important consequences are that Internet information passes through a great many computers, each of which may be controlled by different individuals or organizations, either within or without the United States; and information travelling over the Internet is “copied” dozens of times as it progresses from origin to destination. This section of the Report explains the mechanics without drawing conclusions as to copyright issues. Later sections  
 How the Internet Works Today discussing copyright issues explanations that appear here.  
 will draw on the technical Mechanics of switching 2.1 Mechanics of switching The Internet is a network for carrying computerized (“digitized”) data from place to place. We have had computer networks for many years, long before the rise of the Internet. What makes the latter type of network so different? Earlier computer networks were centrally “switched,” so that all messages between any two points on the network were sent through the central switching computer. These networks are today called “star” networks, because there is a central point in the network—the central switching computer—that has wires “radiating” out to all other computers. Though all computer networks of the time were “star” shaped in their architecture, different switching technologies were often employed by each of them. In the late 1960’s, the Department of Defense, like other computer network users, relied on star-shaped networks for military communication.3 The Department understood, however, that those networks had at least two problems. First was that they were highly vulnerable. Anything that rendered the central switching computer inoperative—whether a bomb, sabotage, or just “down time”—would simultaneously render the entire network inoperative. Second, because different star networks used different technologies for switching messages internally, they could not communicate with each other. Messages were confined to the network from which they originated. Problems with star networks 3 More on the history of the Internet is available from the Internet Society’s Web site. See Barry M. Leiner, Vinton G. Cerf, David D. Clark, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Larry G. Roberts, and Stephen Wolff, A Brief History of the Internet, , available as of February 23, 1997.  
 How the Internet Works Today The Defense Department undertook a research project through its research arm, the Advanced Research Projects Agency4 (ARPA) to remedy these two major drawbacks to existing network technology. ARPA research ARPA’s research led to an alternative networking technology that successfully avoided both the problems identified by Defense. ARPA created a standard format for electronic messages that could be used between networks to connect them in spite of internal differences; and it devised an interconnection method that was based on many decentralized switching computers. Any given message would not travel over a fixed path to a central computer. Rather, it would be “switched” among many different computers until it reached its destination. The network designers set a limit on the size of a single message. If longer than that limit, a message would be broken up into smaller pieces called “packets” that would each be routed individually. This new type of network switching was therefore called “packet switching.” By creating a system that relied on many decentralized computers to handle message routing, rather than one central computer as was the method for star-shaped networks, ARPA produced a network that could still operate even if many of its individual computers malfunctioned or were damaged. ARPA implemented a prototype network called “ARPANet” to test out and continue development of this new technology. Unrelated to ARPA’s work on this packet switching technology, at about the same time (the early 1980’s) the National Science Foundation (NSF) funded the creation of several supercomputer sites around the country. There were far fewer supercomputers than scientists and researchers interested in using them. NSF understood that it would be important to find ways for At times, ARPA has been known as “DARPA,” which stands for “Defense Advanced Research Projects Agency.” ARPA and DARPA are the same agency, however, so this Report will use “ARPA” throughout. 4  
 Mechanics of switching researchers to use these computers “remotely,” that is, without having to travel physically to the supercomputer site. NSF was aware of the work going on with the ARPANet, and determined that that network might provide the sort of access methods needed to link researchers to the supercomputers. NSF then funded the construction of its own network for these purposes, a network known as the “NSFNet” and built with the ARPA packet-switching technology. NSFNet connected a variety of local university networks and hence enabled nation-wide access to the new supercomputer centers. The idea of calling this sort of network an “Internet” reflects the fact that its first use was conceived primarily to allow an interconnection among existing incompatible networks; in its early incarnations, the Internet was viewed less as a “network” for its own sake, in other words, and more as a means to connect other networks together. NSFNet Thus the first practical, non-prototyped version of the Internet was designed and built to enable researchers to use distant computers. Before long, however, the users of the NSFNet began to realize that they were not limited to just sending data back and forth to computers. They could also send messages to each other. At first, these may have been messages that related to the use of the supercomputers. But it soon became obvious that a message from one researcher to another need not have anything to do with supercomputers: the NSFNet was in fact a very general purpose communications medium. Electronic mail, file transfers, and the like thus arose literally as afterthoughts to the Internet’s creation.  
 How the Internet Works Today 2.2 Internet geography This multi-purpose usefulness of the Internet has fueled its enormous growth. It is possible to connect to the Internet from almost anywhere, though of course the actual number of people and institutions that are connected is less than the number that are close enough to obtain a connection because connections carry a cost. Here is a map showing the places where Internet connectivity was possible as of June 15, 1996. Notice that all but a handful of nations fall within the “connection-is-possible” areas. Figure 1: Map of Internet connectivity in June, 1996, from on February 23, 1997.  
 Internet geography Here is the same map from a year later, in June, 1997. Note that considerable expansion of connectivity has taken place in the interim. Figure 2: Map of Internet connectivity in June, 1997, from on January 25, 1998.  
 How the Internet Works Today 2.3 Internet switching technology The way information travels across the Internet is mysterious to most people, beyond the simple knowledge that information is broken into “packets” and that these packets get relayed from computer to computer until they reach their destination. Detailed understanding about the architecture of the Internet is not central to the copyright issues, but some understanding will be useful. Voice telephones and circuit switching The Internet and its contribution to network technology is easier to understand if one first understands how the telephone network was designed to work. Voice telephony is based on “circuit switching.” A “circuit” is the particular path of physical wires over which the conversation will be carried. Typically there are many physical wires going among various locations. These wires will meet at various switching points across the country. When a call is made, it goes to a nearby switching station; there the switch must determine which outgoing wire of many the call should be routed on. That outgoing wire may in turn go to another switching station, which again must find an available outgoing wire; etc. This may happen at many different switching points. Once the various switching points have chosen all the necessary wires, they form an end-to-end path or “circuit”; that path remains constant for the duration of a phone call, whether anyone is actually talking or not. The time delay necessary for the switches to find an available line and making the necessary connection, constitutes the delay one experiences after dialing— the time spent “waiting for a connection.” The disadvantage of this mechanism, compared to the Internet’s “packet switching” mechanism, is that it “wastes” the resources of the circuit when there are pauses in the conversation. The advantage is that it ensures that the call will continue without interruption no matter how much the network may become congested. Indeed, congestion in the telephone network only  
 Internet switching technology affects those who are initially trying to place a call. If the network is highly congested, they will be unable to make a call at all and will instead receive a “fast busy” signal.5 In a packet-switched network like the Internet, however, the mechanism is radically different. All Internet information is digital—represented by a series of high and low voltage values for an electric current that in turn represent the one’s and zero’s that computers work with. This differs from the telephone system, where phone calls are represented by continuously varying voltages—a so-called “analog” communication. But that is not the crucial aspect of the Internet’s form of transmission. Rather it is the fact that digital information is broken up into small “packets” before transmission. Each packet is numbered in sequence and sent out over the network individually. When received at the ultimate destination computer, that computer must re-assemble the packets in the correct order before delivering them to the recipient. Packets may in fact arrive out of order, and quite commonly do because the packets have traveled over quite different physical paths. Packet switching mechanism in contrast to circuit switching The nature of this packet switching mechanism makes it extremely well suited to some forms of digital communication, and less well suited to others. The ideal use of packet switching is for electronic mail. E-mail communication is “asynchronous,” that is, it is not a live or “real time” exchange back and forth between two people. An e-mail message that is divided into packets arriving essentially randomly at the destination computer can therefore be reassembled without a delay that is noticeable to the message’s recipient. Typical “delays” are on the order of fractions of a second in any event, perhaps reaching seconds or minutes during periods of heavy Internet traffic. Most people are unaware 5 The usual busy signal that one gets when calling a telephone that is busy is a slow alteration between a tone and silence. When all circuits are busy—not just the telephone one is trying to reach—the same alternating tone and silence is heard, but at about twice the rate.  
 How the Internet Works Today and unconcerned whether an e-mail they receive was sent five seconds or five minutes ago. Some digital communications are not “asynchronous,” however, and hence are better suited to live or real time operation. A telephone call over the Internet is such an example. A phone call is “synchronous”—it is live and in “real time.” When an Internet user places a phone call over the Internet (as an increasing amount of new software allows one to do), the call is digitized at the sending end (converted into one’s and zero’s). It is then—like an e-mail message or any other Internet communication—broken up into packets. Like the packets of an e-mail message, the packets of a “phone call message” may well travel over different routes and may arrive out of or