Unedited copy, scheduled to be published in September 1998 issue of Communications of the ACM, copyright ACM, posted on the Internet with permission
We have all gone to a computer store to purchase packaged software, and after paying for it, taken the box home and loaded the software on our computer. We typically throw into the trash the printed forms that describe themselves as licensing agreements, although we may have glanced at one long enough to know that these documents construe breaking open the plastic shrinkwrap or installing the software on a computer as assent to the terms of the license. These documents typically disclaim all warranties, tell you that some remote jurisdictionâs law will apply in the event of a dispute over the software, and sometimes deny you authority to make backup copies, modify or resell the software, or decompile the code for any purpose. Most of us do not consider ourselves bound by these agreements because we never really agreed to them. If we go ahead and make backup copies or modify our copy of the software, we do so because it is a reasonable thing to do.
Up until now, the law has backed up our reasonable expectations about packaged software distributed in the mass market. Most judicial opinions have refused to enforce shrinkwrap licenses because consumers have not meaningfully assented to the terms. They have regarded a contract of sale as having taken place at the time the consumer put down cash or a credit card and the store rang up the sale. The so-called 'license' in the box, they have said, may be a proposal to modify the nature of the transaction or add new terms, but the consumer must separately agree to the revised contract and typically she hasnât done so.
From where, then, does the right of consumers to make backup copies of programs, to resell the software to someone else if the original purchaser no longer needs it, or to decompile the program in order to fix a bug or make an interoperable program come? It does not derive from pure reason alone; rather, its legal source is copyright law which has a longstanding tradition of balancing the interests of copyright owners and those of the consuming public. Although Louisiana and Illinois once passed laws to validate software shrinkwrap licenses, neither statute survived closer review. In the Vault v. Quaid decision, federal judges refused to enforce the Louisiana law insofar as license terms interfered with consumer rights under federal copyright law. The Illinois software shrinkwrap license enforcement statute was subsequently repealed due to industry dissatisfaction with it.
FACTORS THREATENING THE STATUS QUO
In Internet time, the status quo on shrinkwrap licenses goes back practically to the Norman Conquest. (In England, the phrase 'from time immemorial' expressed the concept of practically forever, the referent for which was the Norman Conquest in 1066.) Three factors currently threaten to upset this status quo.
The most significant factor is proposed Article 2B of the Uniform Commercial Code (UCC) about which I have written twice previously (Legally Speaking, May and October 1997). UCC2B is a model law designed for enactment by state legislatures. It would not only validate shrinkwrap and other mass-market licenses of information; it would also set rules about electronic contracting for information products and services, establish new rules about warranties, and allow software developers to embed in the software 'technical self-help' features which can be triggered if licensees fail to pay royalties for the software. Proponents hope to get approval of the draft at the July 1998 meeting of the National Conference of Commissioners on Uniform State Laws (NCCUSL), after which they expect state legislatures will begin considering its adoption. Previous UCC model laws, such as Article 2 which regulates sales of goods and Article 2A which regulates leases of goods, have been widely enacted by state legislatures with generally only minor variations (Louisiana once again being an exception). Although the American Law Institute, which is the other body of lawyers to which UCC drafts are generally submitted for approval, has expressed serious reservations about UCC2B, this does not seem to have slowed down the NCCUSL process or the intention of proponents to seek fast legislative adoption of the model law.
While UCC2B is unquestionably the most important factor in the shifting environment about shrinkwrap licenses, two other factors deserve mention before we come back to a more in-depth discussion of UCC2B. A second factor is a trend toward new distribution channels for software. Some companies have already started to distribute their software via the Internet. To purchase software in this way, consumers have to click 'I agree' to the thousand or more words' worth of contract terms which they may view if they really care to (NOT!, as Wayne would say) before the downloading will proceed. This 'click' will look to some judges like the meaningful assent to the license that has been missing from store-bought software. But consumers may still look to copyright policy as a legal basis for continuing to make reasonable uses of their software.
This brings us to the third factor which may change the status quo on shrinkwrap licenses: the 1996 Pro-CD v. Zeidenberg decision. Written by an influential appellate court judge, it enforced a shrinkwrap license restriction permitting only ãhome useä of a CD-ROM of telephone directory information. Judge Frank Easterbrook not only ruled that the shrinkwrap restriction was enforceable as a matter of state contract law, but also rejected Zeidenberg's argument that posting of the telephone data on one's website should be permitted because copyright policy favors the free flow of published information.
Judge Easterbrook offered three reasons why copyright policy should not override the shrinkwrap license restriction in this case. First, he observed that contract rights are different in nature and legal effect from property rights such as copyright. Contracts generally bind only the two parties who enter into the agreement, unlike property rights that generally are good against the world. Second, he asserted that the extra element of the user's agreement to the contract made the contract claim Pro-CD brought against Zeidenberg different from a copyright infringement claim. (The equivalence of a copyright and contract claim would have led to the state law claim being preempted by federal copyright law and policy.) And third, Judge Easterbrook observed that Pro-CD would be unable to recoup its considerable investment in compiling, double-checking, formatting, and updating the telephone directory data if users could just upload the data onto their personal websites.
Naturally, the proponents of Article 2B of the UCC are enthusiastic about the Pro-CD v. Zeidenberg decision both for endorsing the idea that shrinkwrap restrictions can be enforced and for rejecting the copyright policy override argument. They see in that decision what they have all along -- rather wishfully, it seems to me -- asserted to be true.
But the Pro-CD ruling has not been without critics, especially as regards its copyright policy analysis. Given the ubiquity of shrinkwrap licenses in the mass-market for software and given the intent of licensors to bind the entire market, the commercial effect of enforcing these licenses would make them resemble property rights (that is, rights good against the world) more than contract rights (good only against the two parties to the contract). It is, moreover, a legal fiction to say that opening a package or installing software constitutes an agreement to the terms of a shrinkwrap license. Insofar as the act of unauthorized copying would necessarily underlies any contract or copyright claim in a case like Pro-CD, the two claims seem equivalent. This would seem to require preemption of the state contract claim by federal copyright law.
Some intellectual property scholars believe that the market failure argument (based on concerns about Pro-CD's ability to recoup its investment if people like Zeidenberg can upload its data to the web) is the only persuasive rationale for a ruling in Pro-CD's favor. This rationale supports a more limited interpretation of the Pro-CD decision. Under it, a court might enforce the 'home use' restriction in Pro-CD because, as applied, it was a reasonable restriction and averted market failure. Yet, the same court could decide not to enforce restrictions on backup copying or the like because Congress decided that backup copying was reasonable and doesn't threaten market failure.
Whatever one's view on the Pro-CD decision, it has unquestionably changed the intellectual landscape about shrinkwrap licenses and copyright policy. It gave a tremendous boost to the forces that are currently pushing for adoption of Article 2B of the UCC. So let's take a closer look at UCC2B and what it has to say on these and related issues.
DANGERS LURKING IN UCC2B LICENSES
The paradigmatic transaction of UCC2B is a license, that is, a limited transfer of rights to use information on stated terms and conditions. Contrast this with the dominant paradigm of the manufacturing age, namely, the sale of copies. Sales involved a complete transfer of ownership rights in particular copies from the vendor to the purchaser, following which the purchaser could largely do with her copies whatever she wished. If you own a copy of a copyrighted work, you can sell or give it away to friends. However, you can generally redistribute a licensed copy only if you have specially contracted for the right to do this. If UCC2B gains the universal adoption to which its proponents aspire, this model law may contribute to a fundamental transformation in the way information is disseminated in our culture, indeed, throughout the world. UCC2B may herald the shrinkwrapping of information of all kinds: books, magazines, CDs, movies, you name it.
The problem is that UCC2B would validate mass market as well as negotiated licenses for information as long as a consumer has manifested token assent to the license by such acts as clicking 'I agree' or loading the information onto her computer after an opportunity to review the often lengthy and sometimes incomprehensible terms of the license. Only if the terms of a license are 'unconscionable' (shockingly oppressive) will UCC2B refuse to enforce them. Proponents of Article 2B seem to think that it should be permissible to waive all of the default rules of copyright law, such as the right to make backup copies, in mass market as well as in negotiated licenses. Article 2B is structured to ensure that this occurs.
Suppose, for example, that Microsoft decided that it didn't want you to make fair uses of its software when you acquired one of its products. Microsoft should be free, so the theory goes, to offer you software on terms that call upon you to waive this right. Microsoft argues that this promotes freedom of contract and that competition can be counted on to ensure that information providers will be responsive to consumer demands. If consumers find such terms unacceptable, then they'll decide not to license information from certain vendors. If enough consumers feel the same way, the market will, in theory, correct the imbalance. (Of course, the theory doesn't work if the market is dominated by one firm or a small number of firms that adopt the same onerous license terms.)
One of Microsoft's licensing officials, Robert Gomulkiewicz, has put the new paradigm of UCC2B succinctly: "The product is the license." (Think how weird it sounds to say, "let's go to Fry's and shop for some licenses.") In the new world order of UCC2B, you will get what you pay for. But all you will get is a license with highly restrictive terms, any breach of which will terminate any rights you have under the license. One minute you will be a licensee; the next minute you'll be an outlaw. If information ever wanted to be free, it must have changed its mind because under UCC2B, information seems intent on being licensed.
(Incidentally, in case you have any doubts about what people at Microsoft think about fair use, let me relate this story: While working at Microsoft during the summer of 1997, one of my students went to the annual summer intern hot-dog party at which Bill Gates invited questions from the interns. My student asked Bill Gates what he thought about fair use. Bill's first response was to laugh. And then he said, "you don't need fair use; we'll give you fair use rights when you need them." That's what's got some of us worried, Bill.)
THE CLASH BETWEEN UCC2B AND INTELLECTUAL PROPERTY POLICY
Intellectual property professionals who have studied proposed Article 2B of the UCC have challenged the notion that software developers (or any other information provider) can unilaterally bypass public policy limitations imposed by copyright law by adopting highly restrictive mass-market licenses. At a conference hosted by the Berkeley Center for Law and Technology at the University of California at Berkeley in April 1998, copyright scholar David Nimmer --who is, by the way, no relation to Ray Nimmer, the Reporter for UCC2B -- pointed out that computer software firms are far from the first information providers to attempt to use restrictive licensing agreements to control mass- market distributions of their products.
Book publishers and sound recording companies once tried to restrict what purchasers of their products could do with them by "licenses", but fortunately the courts didn't let them get away with it. (Take a look at an old Victrola recording jacket and youâll see that it purports to license use of the recording to one Victrola machine and to deny authority to retransfer one's copy of the recording.) One important case was Bobbs-Merrill v. Straus. Bobbs-Merrill sued Straus because he sold copies of Bobbs-Merrill books in violation of a license restriction that conditioned the right to retransfer copies of the books on an agreement to charge at least $1 per copy. The U.S. Supreme Court treated the license restriction as ineffective as a matter of copyright policy. The Bobbs-Merrill decision contributed to the emergence of the "first sale" or "exhaustion of rights" doctrine in copyright law, under which publishers lose authority to control redistributions of copies of their works when, in commercial reality, the transaction is a sale. In the aftermath of this and similar cases, publishers and sound recording companies abandoned these overly restrictive practices.
While explanatory notes to UCC2B say that 2B will not affect the sale of books and newspapers, they do not cite to Bobbs-Merrill. Nor do they indicate that publishers or sound recording companies should not expect success if they attempt to displace copyright's first sale doctrine by means of mass-market shrinkwrap license restrictions. Because UCC2B presumes licenses of all sorts are enforceable, it is reasonable to interpret the draft as intending to provide print publishers with the same freedom of contract as it would provide to software developers. Yet, many copyright professionals would strongly disagree with that result and believe that this should not be the end of the matter.
Proposed UCC2B recognizes the potential for conflicts between its provisions and federal intellectual property law. Section 2B-105 says that "[a] provision of this article which is preempted by federal law is unenforceable to the extent of such preemption." The current draft of UCC2B claims it is "neutral" on the issue of preemption, seeming to leave to litigation the ultimate determination of the contours of preemption.
In a paper prepared for the Berkeley conference on the intersection of intellectual property policy and Article 2B, Ray Nimmer, Reporter for the UCC2B project, demonstrates that he's not neutral on the subject of preemption. His paper states that there are no circumstances in which Article 2B or any license governed by it might be preempted. Even when federal copyright law has explicitly made a certain right of authors nonwaivable by contract, the Reporter thinks it can be avoided by a 2B license of perpetual duration. On this issue, the Reporter is unquestionably wrong. On the more general issue of copyright preemption of terms of contracts, it is useful to compare the two Nimmer articles [1,2]. David, the copyright specialist, offers many examples of contract terms in copyright licenses that are likely to be preempted. It remains to be seen whose prediction about preemption will be borne out, but my bet is on David.
Leery of leaving the preemption issue entirely to litigation -- partly because litigation is such a time-consuming and expensive process and partly because state courts do not have much experience ruling on copyright policy matters in view of the exclusive jurisdiction Congress has conferred on federal courts in copyright cases -- some have sought a more direct way to resolve the copyright preemption issue presented by UCC2B mass-market licenses. At the American Law Institute meeting in May 1997, Professor Charles McManis made a motion to amend the mass-market license provision of UCC2B to clarify that these licenses canât override fair use and other user rights under specified provisions of federal copyright law. This motion passed at the ALI meeting. But at a subsequent NCCUSL meeting, after heavy lobbying by information industry groups, a counter motion passed which asked the ALI to reconsider the McManis motion.
Meanwhile, in Congress, Representatives Rick Boucher (D-Va.) and Tom Campbell (R-Ca.) have included a provision very similar to the McManis motion in H.R.3048 (the good copyright bill before Congress this year). But despite having forty-some sponsors, the Campbell-Boucher bill has not achieved the momentum of a rival bill, H.R. 2281, which major copyright industry groups and the Clinton Administration favor. H.R. 2281 contains no provision on preemption of mass-market license terms. So it looks like litigation will be the way that the great preemption debate posed by Nimmer vs. Nimmer will be resolved.
There are a host of reasons why UCC2B should not be adopted in its present form. Many of these, as well as many pro-2B statements, can be found on the website maintained by Carol Kunze at http://www.SoftwareIndustry.org/issues/guide/.; The ALI has insisted that substantial changes be made to the draft before it will consider endorsing the model law. A few changes have already been made to respond to the ALI concerns, but much work remains to be done.
At the Berkeley conference on UCC2B, Michael Froomkin, a law professor from the University of Miami, likened 2B -- a formidable document which is currently 200 pages long (single-spaced, no less) and has many novel provisions, including some that would enforce contracts made by "reasonably configured" electronic agents -- to a beta version of a huge new operating system program. Is it really a good idea to trust the health and vitality of our emerging information economy to such an untested beast?
While one can admire the grand ambition underlying the UCC2B effort to anticipate and engrave in legal stone a complex set of rules for the global information economy, one can also question whether it is really possible to get it right at this point in time. Wouldn't it be better to start more modestly with a few key provisions that may be needed to get electronic commerce off the ground, and then revisit the rules periodically to conform them to the electronic and information commerce that achieves market success? Commercial law rules, said contracts scholar Grant Gilmore, should be "accurate, not original." UCC2B violates this maxim in a number of respects.
Unfortunately, state legislatures are far too likely to assume that the years that have gone into the development of Article 2B must mean it has produced a very good product. They may also fall for slick arguments by lobbyists for information industry groups that support 2B (not all do, by the way) who may assert that the stateâs early adoption of UCC2B will cause information businesses from all over to flock to that state. This is the sucker tactic that apparently persuaded the Louisiana legislature in the early 80's to pass the shrinkwrap license enforcement statute that the Vault v. Quaid case refused to enforce because of its conflict with federal copyright policy. If these tactics worked once, they may work again. Why else would the legislatures of Wisconsin and Iowa supposedly be champing at the bit to enact UCC2B, as proponents of the model law claimed at the Berkeley conference?
The debate about copyright preemption of shrinkwrap license terms boils down to whether copyright owners can have their cake and eat it too. Copyright law has always (and wisely) insisted on there being a balance between the rights conferred on authors and the rights reserved to the public. The balance has been key to ensuring that copyright law fulfills its constitutional mandate of "promot[ing] the progress of science and [the] useful arts." Lately that balance has been under assault, as high protectionists have sought overbroad copyright legislation, new forms of legal protection against extractions of information from databases, and now mass-market licenses capable of overriding the public policy balances of copyright law. We shouldn't let them get away with it. If publishers want the rights that copyright confers, they must take the responsibilities along with the rights.
Pamela Samuelson is a Professor of Information Management and of Law at the University of California at Berkeley and is Co-Director of the Berkeley Center for Law & Technology
Thanks to Mark Lemley, Laurel Jamtgaard, and Peter Huang for their helpful suggestions.
 Nimmer, D., Brown, E., & Frischling, G.N. The Metamorphosis
of Contract Into Expand, Calif. L. Rev. (forthcoming 1998).
 Nimmer, R. Breaking Barriers: The Relation of Contract and Intellectual Property Law, Berkeley Tech. L.J. (forthcoming 1998).