5 Va. J.L. & Tech 11 - Facilitating Collaborative Software Development: The Enforceability of Mass-Market Public Software Licenses

1. Imagine a somewhat technologically capable individual who just downloaded the hippest new investment-decision-making software program that guarantees to beat the Dow by ten percent. During installation of the program, the following message appears on her computer screen:

LICENSE: By operating or retaining a copy of this program, the user hereinafter agrees to abide by all the terms and conditions of the following agreement (that everybody is expected to read but nobody ever does), as well as Robert’s Rules of Order and such other terms and conditions, real, imaginary, subjective, objective or offensive, as the Software Distributor shall deem appropriate or desirable, including the right to perform unreasonable search and seizures of the user’s computer, home, person and pets, whether there is sunshine when she’s gone or if New York is indeed the city that never sleeps, for better or for worse, to death do us part, would you like fries with that, don’t put the cart before the horse even if you do lead it to water and try to force it to drink, in God we trust, unless of course, you are atheist or agnostic, thank you very much and don’t forget to put all of your savings into our inflated stock.

Whether or not software mass-market licenses such as this are enforceable is currently a very unsettled issue.

2. Section II presents a background of the computer software industry and an overview of the intellectual property rights that arise in software to lay a foundation for the discussion of this issue. Divergent views of the proper model of software development exist. Some companies believe that all software development should take place within their building and protect the resulting programs with the utmost secrecy. The most familiar company that follows this closed model of software development is Microsoft.[2] Other companies believe that software developed across entity lines is superior and, thus, promote a collaborative environment of sharing software developments and advancements. This model may seem unorthodox, but it actually describes how software originally developed some decades ago. Modern companies that follow this open model include Netscape and Red Hat.[3]

3. Section III introduces general mass-market software licenses and discusses the current analysis of their enforceability. Although most recent courts have held such agreements valid, much legal scholarship and public sentiment undermines the certainty of those holdings. Many argue that mass-market licenses are invalid because they take advantage of the relative weak bargaining position of the consumer to the software company, are merely contracts of adhesion, and are otherwise procedurally unconscionable. In addition to the procedural issues, mass-market licenses are also subject to having their terms found substantively unenforceable. Not only do the terms of these agreements often seem exceedingly broad and restrictive, they may also be found unenforceable due to preemption by federal intellectual property laws. Any provision of a mass-market license that attempts to extend the rights of intellectual property owners beyond the policy bargain struck by Congress is invalid. Just as with procedural and substantive enforceability, the preemption of mass-market licenses is unsettled.

4. The dichotomy discussed in section II between the open and closed models of software development presents a classic collective action problem in that companies which tell the world their ideas are subject to free riding by companies which keep their ideas secret. However, section IV concludes that open model actors should not feel destined for failure in the marketplace, because the law of contracts may provide the perfect solution to this collective action problem. Through licensing agreements, open model companies can disclose to the whole world their software secrets while at the same time preventing closed model companies from benefiting through incorporation of those ideas in closed model programs. The two most notable public software licenses are the GNU General Public License promulgated by the Free Software Foundation and the Mozilla Public License written and used by Netscape Communications.

5. Although numerous claims of infringement and threats to seek legal resolution of software copyright issues have been made,[4] no court has yet ruled on the enforceability of public software licenses. As a result, companies desiring to follow the open model of software development must bear the cost of this legal uncertainty, which, in turn, reduces the ability of these companies to compete in markets occupied by closed model firms. Section V addresses every conceivable argument against the enforceability of public software licenses. Based on current relevant doctrine and prevailing public policy interests, public software licenses that adhere to distinct procedural requirements are enforceable.

II. Traditional Software Licensing

A. Software Terminology

6. Before discussing the evolution of software licensing, a brief foray into terminology helps to assure that popular misconceptions do not cloud the points made here. For starters, computers contain only two elements: hardware and software.[5] Hardware is every part of a computer that is visible, whether with a blind eye or a microscope. Software is the set of instructions that control the hardware. At its core, software is nothing more than 1s and 0s in a sequence of any length that controls the sending of electricity to the hardware, much like a light switch controls the sending of electricity to the light bulb. To put it all together, think of it as 1 turning the switch on and 0 turning the switch off. Hardware only responds to sequences of 1s and 0s.[6] A sequence of 1s and 0s is called object code, which is also referred to as the executable form of the program.[7]

7. However, programmers do not write software in executable form because not even the brightest programmer is smart enough to figure out exactly what combination of 1s and 0s will produce the desired hardware responses for complex and sophisticated applications. A very important fact relevant to the discussion below is that programmers cannot ascertain from object code the result that is produced by a computer following that code. This results in a communication problem as to how to get computers to produce desired results since one cannot possibly produce the necessary object code directly. Much like any other difference of language problems, the solution lies in translation.

8. In fact, for those desiring to produce object code, the current method involves not one, but two translations. First, a computer programmer translates her native language, say English, into source code which is written in one of several computer programming languages, such as Basic, Fortran, Pascal, or C++.[8] Second, the programmer enters the source code into a compiler program that, upon command, translates the source code into object code.[9] To recap, programmers write source code that compiler programs translate into object code, a sequence of 1s and 0s, which, when executed by the hardware of a computer, produces desired results. Once written and compiled, the next decision is whether and how to put the software to use.

B. The Birth of a Software Mass-Market

9. Before NASDAQ and millionaire making IPO’s, the computer industry occupied a reclusive corner of mainstream America. Before the 1980s, computers were large and expensive, thus restricting ownership to the government, universities and a few major corporations.[10] As a result, software was tailor-made for the specific needs and applications of each owner.[11] This phenomenon resulted in the production of software that was highly idiosyncratic and lent itself to use by only a limited number of users.

10. During this era, programmers openly shared their solutions and advancements with one another.[12] What can be labeled a quasi-donation to the public domain by software programmers proved economically more efficient than a proprietary regime of secrecy because the cost of achieving programming goals independently, without the aid of others, outweighed any commercial benefits of claiming ownership of potential intellectual property in the software.[13] The early software market involved the commercial exploitation of software through providing a service of software programming to customers rather than delivering a pre-made product to end-users.[14]

11. Software licensing was not an issue in this era.[15] The programming was typically completed through vertical integration within the entity of the computer owner.[16] In other words, the computer programmer received compensation by her employer for her development of software.[17] In return, the employer benefited not from selling the code to other entities, but by using it himself.[18] In the circumstances where the work was performed through contract, the terms of the agreement resolved any issues as to the ownership of the created software.[19] Perhaps the largest reason why software licensing was not an issue in this era was that a market for copies of the same software did not yet exist.[20] The market demanded that experts apply their training to idiosyncratic situations where previously completed solutions would be only tangentially applicable at best.[21] There simply did not exist thousands of potential customers who demanded the same program.[22]

12. Beginning in the 1980s personal computers became increasingly affordable and the number of computers capable of running the same software exploded.[23] A new market emerged comprised of customers who no longer needed experts to create custom tailored software.[24] A fork in what was once a unified software market appeared: on the one prong existed the service type market of before with heterogeneous customers, and on the other prong existed a new product type market of homogeneous customers.[25] The satellite propulsion software for NASA and missile guidance software for the B-2 bomber, for example, fall into the former category, whereas word processing and spreadsheets fall into the latter. A mass-market for computer software was born.[26]

13. Expansion of the mass-market led competitive companies to explore other methods of commercially exploiting software.[27] The suppliers of software moved away from an open attitude towards computer programming to a more secretive and closed mind-set. Firms made this move to help protect their market share and reap profits. By keeping others from knowing or using its programs, a software company could prevent other entities from free riding on the large costs of research and development associated with the creation of software.[28] This prevention of free riding is the most important benefit of following the closed model of software development.

14. However, there must be an economic trade-off to firms that follow the closed model of software development and choose to restrict public use of their ideas. The cost to companies comes in the form of increased research and development costs because the companies must discover and create solutions through means other than resorting to the computer programming community-at-large. As a result, a quasi-prisoner’s dilemma results: firms that continue to make their developments available to others are subject to free riding, while other firms that refuse to make their developments available to others are forced to bear the costs of performing all of their own research.

15. Deciding whether the open-academic or the closed-secret model of software development is most beneficial to a specific firm or society in general is not a simple question.[29] Under certain circumstances, either model may be more efficient.[30] Further, even given the same market circumstances, both models may prove equally efficient and thus competitive with one another.[31] The proper role for law in this area is to advance the ability of companies, regardless of which model is used to enforce the underlying mechanisms in order to achieve their own development at minimal transaction costs.

16. Before moving on to further introductory discussion, it may be helpful to put real life names together with these concepts. Microsoft Corporation is a closed model software company that distributes the WindowsNT operating system for networks. Red Hat is an open model software company that distributes the Linux operating system for networks. Both are successful companies in the network computer operating system market.[32] However, Linux poses no competitive threat to Microsoft in the personal computer operating system market.[33] Further, due to the characteristics of the market, open model software firms are economically prohibited from competing in the personal computer operating system market.[34]

17. Likewise, the Internet operations program market similarly follows one model of software distribution, but with open model firms supplying the great majority of this market.[35] Every time Internet surfers enter a domain name address into their browser, a software program translates the combination of alphanumeric letters to the actual digital Internet web location.[36] DNS (Domain Name System), the software program that completes this operation, and Sendmail, the program which routes nearly 80% of all e-mails, are software developed under the open model.[37]

18. As noted above, a more in depth discussion of the public policy implications, resulting from the potential competitive aspects of the two models, appears below. At bottom, economically feasible reasons exist for firms to follow either model depending on market characteristics. Having presented concrete examples in an attempt to improve the sense of real world applicability of the topics discussed here, a return to more introductory material is timely.

C. The Creation of Intellectual Property Rights in Software

19. Rights in intellectual property in general are based on the idea that individuals who add value to society through invention or creation of new work deserve a reward. Put in another way, in order to expand the economy, the law should provide incentives for individuals to advance the interests of society, either by making productive use of the intellectual property or by increasing the size of the knowledge in the public domain.

20. This idea is not merely altruistic, it is constitutional.[38] Intellectual property law provides such an incentive to inventors and artists by granting exclusionary property rights of limited duration in new creations to the inventor or author.[39] In setting these limitations, a public policy balance must be struck between providing sufficient incentives to create new works and providing the public benefit from those creations.[40] This balance plays a role in the enforceability analysis applied by courts in judging issues of intellectual property statutory intent and in determining the proper scope and terms of licensing agreements.[41] The American justice system repeatedly invalidates attempts to expand intellectual property rights beyond congressionally proscribed limits.[42]

21. Once ideas are created or expressed, inventors must make decisions as to what forms of intellectual property protection to pursue and to what extent they wish to disclose their ideas to others. Patents protect inventions that are useful, novel, and nonobvious.[43] Due to its intangible nature, courts once thought computer software was ineligible for patent protection.[44] However, a widely accepted recent case held that anything made by man can be the subject of a patent, including software.[45] As a result, patent applications for software and related developments – including methods of doing business with computers and Internet related processes – began flooding the United States Patent and Trademark Office.[46] Another consequence of this new type of intellectual property protection for software is that licensing agreements must now deal with the potential underlying patents to the particular software.

22. If the inventor desires a patent, she cannot follow a closed model of software development. This is true because the patent laws require an applicant to publish a detailed specification of the program that enables one of ordinary skill in the art to make and use the program and that describes the best mode known to the programmer for effectuating the program.[47] The programmer cannot, at the same time, seek patent protection for software and strictly follow the closed model of development because the patent laws require that the world be told what the program is, how it is made, and how to put it to most productive use.[48] For this reason, actors following a closed model would presumably choose not to pursue patent protection, but rather resort to copyright and trade secret protection of their software.[49]

23. Federal law grants copyrights to authors who fix expressions of an idea in a tangible medium.[50] Unlike the issue of patent eligibility, the Copyright Act explicitly states that software is copyrightable.[51] A programmer who writes out the design or protocol of a computer program has fixed an expression in a tangible medium and, as such, has a copyright to that expression. Source code written by the programmer based on the design or protocol is also copyrighted because it likewise is a fixed expression of an idea in a tangible medium.[52] Once compiled into object code, the same programmer also has a copyright in the object code. Therefore, the same program can have three copyrights, at least.

24. If this seems odd or unfair as being too large of a grant, think about the writing of a paper involving revisions and translations. Each revision and translation of the paper is itself copyrighted. Similarly, each translation of the computer program from human language to source code and from source code to object code is copyrighted. Thus, software transactions that intend to transfer the complete package of copyrights in the software must grant rights in each copyrighted version of the program. If the license does not cover all of the copyrights, the licensee is limited to using only that particular manifestation of the software covered by the license.

25. To tie all this in, think back to the description of the two models of software development presented above, the open, sharing amongst entities, model and the closed, protecting against free-riders, model. In comparing the two models and their relative merits from a public policy standpoint, put aside for the moment any argument based on the impact each model may have on the incentive-based bargain underlying intellectual property rights. This paper argues that, contrary to much academic scholarship and public perception, neither model of software development has a perverse impact on the public policy balance struck by Congress under any of the intellectual property statutes. Rather, the opportunity for software programmers to choose between the two models, or even form hybrids, allows for efficiencies that neither model can provide by itself. Next, a discussion of the decisions commonly made by actors following each model concerning the distribution of their intellectual property rights proceeds.

D. The Distribution of Intellectual Property Rights in Software

26. Once a patent is issued on software, there is a difference in the choices made by actors under either of the two models with respect to use of the patent. By definition, a patent tells the public how to achieve the result protected by the patent. Assuming economic decision-making, actors under both the open and closed models will put the issued patents to their most profitable use through licensing or blocking. The decision with respect to which option is most profitable for a particular firm will differ according to the underlying model of distribution adhered to by that firm.

27. An economically rational open model actor seeks to grant non-exclusive licenses to the full extent of market demand, up to and including the world.[53] A closed model actor finds it more economical to use its patent to block competitors by filing infringement suits.[54] Since the patent provides protection of the program for twenty years from date of application, and most software programs are obsolete well before twenty years, the patent acts as a virtual lifetime protector of its underlying intellectual property.[55] Even so, many closed model actors do not seek patent protection because of the high cost of enforcing patent rights against infringers, the substantial likelihood that the patent will be declared invalid during litigation, and the fact that, commonly, only a reasonable royalty is awarded to the patentee as damages even if the patent is held valid and infringed.

28. Unlike patent law, there is no requirement of publication in order to obtain or enforce a copyright.[56] Therefore, under both the closed and open models, programmers seek and receive copyrights in their software. The choice each makes with respect to the level of secrecy or disclosure of the copyrighted work differentiates actors under the two models. A closed model actor keeps secret as many of the copyrights as possible, while an open model actor makes available all of its copyrights to the public.

29. Under the closed model, only a minimally necessary amount of intellectual property is disclosed. Software is most commonly distributed through the sale of computer disks containing only the executable form of the program. Consumers install and operate the program by running the executable form of the program contained on the CD-ROM or 3.5” floppy disks. A license to the object code that is contained on the disks accompanies the program.[57] Licenses are never granted to the source code or any other translation of the program, allowing closed model actors to rest assured that their intellectual property is protected in two ways. First, all of the closed model actor’s intellectual property is protected under the utmost secrecy leaving only the bare minimum amount of copyrights necessary in the program being released to the public. Second, from a technology standpoint, competitors find it highly difficult to reverse-engineer the program from only its commercially available object code.[58]

30. Under the open model, all of the software copyrights are available to the market. A company can do this in two ways: abandon the intellectual property rights to the public domain or license each of the copyrights. Although abandoning the copyright can bring indirect economic gain through notoriety or prestige within the community, most economically prudent open model actors prefer to license their copyrights instead.[59] This is so because abandoned copyrights can become the basis of a derivative work of another work, which itself becomes copyrighted.[60] In order to reap economic benefit and protect their copyrights from being hijacked by closed model actors, the open model actor prefers to license all of their copyrights to others. To prevent the exportation of the programs to the closed model, a term of open model license agreements must disallow derivative works from being licensed under the closed model.[61]

31. Think of two actors, each acting under one of the opposing software distribution models: Red Hat follows the open model and Microsoft follows the closed one. If Red Hat abandons all copyrights in program X to the public domain, Microsoft can write program X+ and receive full copyright protection in all of forms of the software, whether underlying design, source code, or object code. Microsoft can then license its copyrights under the closed model as described above. Thus, within literally hours of releasing a program to the public, an uneducated open model actor can find her intent being thwarted. True, the original program X will always and forever be part of the public domain, but this means very little in a technological age where drastic changes and improvements are made to software on a virtually instantaneous basis.

32. A better way for Red Hat to advance its open model desires would be not to abandon its copyright, but rather to license it with restrictions on what the licensee can do with the copyright.[62] Red Hat can assure an open model of distribution for its program by requiring all works that incorporates any part or derivative work of the copyright be distributed in an open model manner. Therefore, to truly advance an open model of software development, the copyright to the software should not be abandoned to the public domain, but rather licensed with promises by the licensees that they will perpetuate the copyright and developments resulting from it only under an open model.

E. Why Software is Licensed and Not Sold

33. Why is software licensed as opposed to sold? When I go into a store, I buy the software and consider myself to own it, so you can’t really be telling me that the software developer actually owns the software and is just letting me borrow it, can you?

34. Under the agreements that accompany the software, software companies almost universally retain ownership to the intellectual property incorporated by the software and extend a non-exclusive license to the consumer to use the intellectual property - patents, copyrights, trade secrets and trademarks - in accordance with the terms of the agreement.[63] Through license agreements, software distributors can place greater restrictions and limitations on the use of the intellectual property contained in the software than if the transaction consummates a sale. Although this distinction between license and sale may seem formalistic, the consequences are potentially severe. Under the UCC, as well as the federal patent and copyright statutes, sale of a product drastically limits the control individuals can retain over that product.[64] However, such drastic limitations do not arise if the individual retains ownership and merely licenses the product to end-users.[65]

35. Placing tight restrictions on the use of software “purchased” by consumers may seem intuitively unfair, but the public benefits from allowing this to occur. By being able to limit the rights of customers, firms can offer the software at lower prices. Presumably, a firm will charge less for a program that can only be operated on Sundays than the same program that can be run any day of the week. Similarly, a program that cannot be redistributed by the consumer will cost considerably less than the same program that does not prohibit redistribution by the consumer. Tailoring the grant of rights to fit the demands of various markets through the agreements attached to the purchased software allows software companies to charge lower prices to those markets which desire less rights. Economically, this ensures a more efficient distribution of the program to all consumers. To summarize, in software, it is the rights granted by the license, not the program itself, that is the consumed product.

36. While many lawyers have agreed with the line of reasoning behind the questions posed above, some judges have gone so far as to disregard the “license” label and find that the transaction was actually a sale, thus exempting the consumer-owner from the restrictions contained in what was claimed to be a license.[66] Holdings that negate the limitations placed on software purchasers have devastating effects on the ability of firms under either model to tailor software products for various markets. If software cannot be effectively licensed in a discriminating manner, software firms are left with the decision of which price point to offer the “sale” of their product. Entities that can afford that price can purchase the software and have all the rights accorded them under the contract and intellectual property laws. However, many small businesses and individuals will not be able to afford to “buy” the software.

37. For instance, under the first sale doctrine, an owner of a piece of software can transfer her program to whomever and for whatever she desires. The use of a license prevents this doctrine from applying, which allows computer programming firms to price-discriminate between customer characteristics. If Microsoft wants to give Windows software to public schools at a cost blow the production cost and the transaction consummates a sale, the first sale doctrine would apply, and the school could resell the programs at a higher price to a corporation, retaining the difference. This would cause Microsoft to charge all customers one price, either by lowering its price, forcing it to run at a loss, or raising its price, thus making the program unavailable to schools and other meagerly funded organizations. This result is economically inefficient and would most assuredly be politically unpopular.

38. Through the formalistic distinction recognized between a license and a sale, contract law provides a mechanism to effectuate the desired price discrimination discussed above. Firms under either the open or the closed model of software distribution provide licenses to the market while retaining ownership in the underlying intellectual property. This allows software distributors to effectuate their business goals. Although this discussion may seem simplistic and well settled, many courts have not adhered to these market norms in resolving software license validity disputes. In fact, until four years ago, every court decision on the issue and the vast majority of legal scholarship staunchly opposed enforceability of license agreements in the mass distribution of software. The next section presents an in-depth discussion of the legal development of software mass-market license enforceability.

III. Mass-Market Software Licenses: Shrinkwrap and Clickwrap

39. Software developers desiring to license the intellectual property in their software set forth terms of the proposed agreement in writing. For programs distributed to a limited number of consumers, the parties carefully draft and negotiate customized license agreements.[67] However, as the number of licensees rises, the cost of forming idiosyncratic agreements with each customer becomes expensive.[68] Much like commercial contracting in a multitude of other markets, use of standard form contracts in the software industry produces a reduction in transaction costs.[69] The term “mass-market licenses” refers to the licensing agreements used with widely distributed software.[70]

40. If a consumer purchases a physical copy of the program, the license is typically included in print form either on or within the packaging. The license gives the consumer the opportunity to review the license and return the software for a full refund if she does not wish to abide by its terms. Notice of the license and the requirement that the consumer consent to its terms typically appears on the exterior of the software packaging.[71] This method of mass-market licensing is referred to as “shrink-wrap” licensing.[72]

41. If the consumer downloads a program via an Internet web site or some other digital means, the screen usually displays the license agreement. Alternatively, the consumer may be prompted to “click” to another screen to review the license agreement. Upon prompting, the user must point and click on a button labeled “I accept” or “Agree” in order to proceed with use of the program. This method of mass-market licensing is referred to as “click-wrap” licensing.[73]

42. Although the technological processes involved in effectuating mass-market licenses are well established, there still exists a large risk attached to their use.[74] The state of the legal enforceability of software mass-market licenses has yet to stabilize. Just as one authority holds a particular view, opposition soon responds with an increasingly intensive argument for the contrary position. Since mass-market public software licenses are similar in procedure and substance to other types of mass-market licenses, legal analogies can be drawn. Therefore, an analysis of the highly unpredictable law with respect to mass-market licenses is necessary before discussing the legal enforceability of public software licenses.

A. Current Enforceability of Mass-Market Licenses

43. Before June 1996, no court had held mass-market licenses enforceable. In actuality, only four cases even touched on the issue of mass-market license enforceability, and of the four, two simply relied on the predecessors.[75] The first courts to rule held that such agreements were adhesion contracts.[76] In so holding, the courts focused on the relative bargaining strength of the actors and determined that consumers had no ability to negotiate the terms of the agreement.[77] Other courts have scrutinized the agreements under UCC and common law principles of contract. These courts found the licenses unenforceable as a violation of UCC sections 2-207 and 2-209.[78] Under these cases, the terms of the license must be brought to the consumer’s attention prior to purchase in order to be enforceable.[79]

44. In ProCD v. Zeidenberg, the Seventh Circuit broke with precedent and held that standard form shrinkwrap license agreements are enforceable.[80] Speaking through Judge Easterbrook, the Seventh Circuit examined applicable copyright law, Article 2 of the U.C.C., and the commercial reality of the mass-market software transactions.[81] The court discussed the commercial justifications for the contract and rejected arguments that the contract was procedurally unconscionable.[82] The court recognized that the method of contract formation implemented by the software distributor, “may be a method of doing business valuable to buyers and sellers alike.”[83]

45. The court noted explicitly that it would not perform substantive contract analysis of the license because neither party claimed that the terms themselves were unenforceable on contract grounds.[84] The only issue raised on appeal was what terms constituted the license and whether federal law preempted such terms. In addressing the issue of federal law possibly preempting the contract, the court examined § 301(a) of the Copyright Act.[85] Following precedent from three other circuits, the court concluded that the rights created by the license were not “equivalent to any of the exclusive rights within the general scope of copyright,” because contract rights do not restrict the options of persons who are strangers to the author.[86] Therefore, under ProCD, no contractual agreement can be preempted.[87]

46. ProCD continues to be a highly controversial decision.[88] Some commentators claim that the Seventh Circuit’s preemption analysis seriously erred by combining the two prongs of a correct preemption analysis into one.[89] A more detailed discussion of this debate follows below in Subsection C. The decision has also been criticized for erroneously concluding that the transaction was a sale of a license to use the software rather than a sale of the software itself.[90] This argument presupposes that proper application of the UCC declares these transactions to be “sales.”[91] As a result, the first sale doctrine would apply, and, therefore, limitations placed on the use or further distribution of the software would be invalid.[92] Further, others argue that the Seventh Circuit erred in its commercial analysis because intellectual property licenses are “notoriously fallible” indicators of the parties’ intent.[93]

47. In M.A. Mortenson Co. v. Timberline Software Corp., the Court of Appeals of Washington followed the reasoning in ProCD to uphold a shrinkwrap license.[94] The defendant claimed the license was unenforceable for several reasons, all based on principles of contract.[95] First, the court concluded the license was not procedurally unconscionable for three reasons: (1) reasonable minds know that use of software is governed by licenses containing multiple terms; (2) the license was printed in full twice, once on the sealed envelope which contained the disk and a second time on the inside cover of the manual; and (3) notice of the license was displayed on the user’s screen each time the program was used and the consumer was given reasonable opportunity to review and reject the license.[96] The court held that the defendant manifested assent to the terms of the license by installing and using the program.[97] Second, the court addressed whether the term, here a limit on consequential damages, was substantively unconscionable.[98]

48. One case that did not follow ProCD is Novell, Inc. v. Network Trade Center, Inc.[99] The United States District Court of Utah held that the transaction, although labeled a license, was actually a sale of goods as defined by the UCC.[100] Therefore, the first sale doctrine applied, and the defendant had full ownership rights under the copyright statute.[101] However, Novell is of questionable legal value because the court has since vacated its judgment as requested by the parties.[102]

49. Each court that has ruled on the enforceability of the clickwrap form of mass-market licensing has held the license enforceable. In Caspi v. Microsoft Network, the Superior Court of New Jersey, Appellate Division, focused on three factors in determining whether a clickwrap license was enforceable.[103] These factors were whether the consumer received adequate notice of the license, whether sufficient competition in the market existed to equate the relative bargaining strengths of the transacting parties, and whether the consumer received adequate time to