Licensing unrelated to any intellectual property right has been a common practice for generations - since at least the time of the civil war.  But there are some who think of licensing as only appropriate if there are rights to license and that, if the rights do not exist, the contractual license terms must be abrogated.  This is wrong, but a recent case indicates that drafting agreements in this context must be done with care.

A “license” is a contract.  It sets conditions on use of informational assets.  When intellectual property rights are involved, part of the action of the license consists of a covenant not to sue for conduct that would be infringement if not within the license.  But that is not an essential feature of an agreement that constitutes a license.  The contractual agreement is the essential factor.

            The lower court in C.B.C. Distribution and Marketing, Inc. v. Major League Baseball Advanced Media, L.P., 2007 WL 2990366 (8^th Cir. 2007) got the analysis terribly wrong, but luckily was abrogated by the Court of Appeals.  The lower court argued that contract restraints on use of non-secret data without underlying IP rights violated fundamental public policy.  What policy?  Why would that policy over-ride contract law?  The opinion gave little insight into this. 

The case involved a license from Major League Baseball to C.B.C. of the right to use player names and statistics in interactive media fantasy baseball systems.  The contract broke down, but CBC continued to use the data in violation of the license which specifically prohibited such use.  The District Court held that enforcement of this clause was precluded because a “strong federal policy favoring the full and free use of ideas in the public domain … prevails over [those] contractual provisions.”  The District Court was apparently referring to a general First Amendment right, but that right can be and frequently is contractually waived.  It also referred to the ancient Supreme Court decision in Lear v. Adkins, but that case dealt with patent law and has not been expansively applied by courts since originally announced during an era of judicial hostility to intellectual property rights.

The appellate court in a 2-1 decision avoided the lower court’s error by reading the license to create a warranty of title in a way that will give drafters of such licenses pause.  The “warranty” was breached, thus precluding enforcement of the non-use clause because the subject matter of the license was not covered by intellectual property rights.  The warranty language? “[The] Players Association “is the sole and exclusive holder of all right, title and interest” in and to the names and playing statistics of … major league baseball players.” But the interpretation of all right and title is, at best uncertain in a setting where both parties must have understood that the data (as compared to the names) involved was not covered by intellectual property rights.   

            The dissent came closer to getting it right.  It concluded that there was no warranty.  Instead, the contract said “that the parties “agree” that MLPBA is the sole and exclusive holder of all right, title and interest in and to the Rights. CBC surely can “agree,” as a matter of good business judgment, to bargain away any uncertain First Amendment rights that it may have in exchange for the certainty of what it considers to be an advantageous contractual arrangement.”  As to the public policy issue, the dissent observed that there was no federal statute that addresses state-law contract obligations on this issue and, thus, “no indication that Congress sought to abrogate contracts in this area that are otherwise enforceable under state law.”

            So, the lower court’s wrong analysis is gone as a matter of precedent, but we have a precedent that reads a common contract clause in a way that makes severe trouble for any license drafting regarding data. 

            The far better result would have been to have both courts recognize that nothing is strange about contractually licensing information that is not within the scope of intellectual property law.  That is and has long been the law.  See Ross, Brovins & Oehmke, P.C. v. Lexis Nexis Group, a Div. of Reed Elsevier Group, PLC, 463 F.3d 478, 80 U.S.P.Q.2d 1518 (6th Cir. 2006).

I hope not, but that is one risk created by the Supreme Court’s decision in the Ebay case and by the actions of some courts who have denied permanent injunctions in successful infringement cases.  But the fact that a permanent injunction does not issue after a judgment of infringement does not mean that the infringer (by losing the case) obtains a right to use the patent owner’s property in the future.  It simply means that the court declined to add the coercive force of an injunction to the statutory right to exclude as to future infringing conduct.

            Although most patent cases that result in a finding of infringement yield a permanent injunction, in some circumstances, injunctive relief will be more difficult to obtain.  This is especially true if the defendant is not in competition with the patent owner or is a patent holding company (sometimes referred to as a patent troll) that engages in no business other than licensing patents it acquired from third persons.

But what is the appropriate remedy if the court holds that a permanent injunction is not warranted, but that infringement has occurred and may occur in the future? 

One response is simply to assess damages as to past infringement, leaving any future use of the patent for a voluntary agreement of the parties (a license) or a subsequent infringement suit for the subsequent infringements.  That is clearly the preferable option, although it does raise limited issues of judicial economy.

A second alternative is to permit subsequent use by the defendant subject to the payment of a reasonable royalty imposed by the court.  This is a form of compulsory licensing that rewards the wrongdoer, unless the remedy has been requested by the patent owner.  Nevertheless, a panel of the Federal Circuit indicated that such a remedy may be appropriate.  One wonders why.

            In Paice LLC v. Toyota Motor Corp., 2007 WL 3024994 (Fed. Cir. 2007), after a finding of infringement and damages of $25 per use of infringing drive trains for hybrid motors, the District Court denied the patent holder’s request for a permanent injunction, and instead sua sponte issued an order permitting Toyota to continue to produce the infringing items subject to payment of a $25 per motor royalty.  The refusal to grant an injunction was not appealed, but the Federal Circuit noted that the court’s decision was based on at least the following factors:

·         In context, the absence of an injunction would not adversely affect the patent holder’s ability to license the patent to others. 

·        The patent holder did not actually manufacture any goods, and there was no threat that it would lose name recognition or market share without an injunction.

·        Monetary damages were adequate as suggested by the fact that the patent owner had offered to license Toyota during the post-trial period.

·        The balance of hardships favored Toyota since issuing an injunction would disrupt not only its business, but that of suppliers, dealers and others.

This gives you some idea of the type of issues likely to adversely influence a patent owner’s ability to obtain injunctive relief.  They depict a plaintiff company whose only interest in control of the patent is to obtain licenses or judgments of infringement, rather than to build or protect an existing or potential market.

            Should a denial of permanent injunctive relief automatically require judicially forced licensing?  The answer is no.

The Federal Circuit suggested that courts have long had the ability to order ongoing royalties as a remedy in appropriate cases, but carefully said that the District Court may do so, not that this result is ever required.  Even then, the cases it cited – an antitrust remedies case and an infringement case in which the ongoing royalty was not contested – do not support the conclusion that broad judicial power exists to make an award that effectively takes away the winning party’s property without being requested by the prevailing party.  Indeed, the Patent Act talks about granting injunctions to prevent violation of the property rights, while this remedy permits use of the invention.

But even if a court has power to make such an award, what is an appropriate case?  The one “appropriate” case the Federal Circuit mentioned was where, after the judgment of infringement, the parties attempt but fail to agree on a license for future use.  The court commented that that this may be a case for granting ongoing royalties as a remedy, not that this was required.  The court commented:

But, awarding an ongoing royalty where “necessary” to effectuate a remedy … does not justify … such relief … whenever a permanent injunction is not imposed. In most cases, where … a permanent injunction is not warranted, the district court may wish to allow the parties to negotiate a license amongst themselves … before imposing an ongoing royalty. Should the parties fail to come to an agreement, the district court could step in to assess a reasonable royalty in light of the ongoing infringement.

But why is a failed bargain as to future conduct grounds for the court to step in and impose a bargain of its own forcing the injured party to allow continued use of its property, even though the owner did not request that remedy?  The court did not explain this strange reversal or comment on whether a rights owner who does not want the court to control its future income would be better off not negotiating an attempted license at all.

            In Paice, the court remanded because the lower court’s order contained nothing to show why the selected royalty was appropriate.  But there is a more fundamental issue.  Judge Radar, writing a concurrence, directly approached that issue.  To him, giving the parties an opportunity to negotiate a license should be a precondition to the court having any power to impose an ongoing royalty.  He commented:

[This] court should require the district court to remand this issue to the parties, or to obtain the permission of both parties before setting the ongoing royalty rate itself. District courts have considerable discretion in crafting equitable remedies, and in a limited number of cases, as here, imposition of an ongoing royalty may be appropriate. Nonetheless, calling a compulsory license an “ongoing royalty” does not make it any less a compulsory license.

The preconditions should be both an opportunity to negotiate a license and, failing a bargain, a request by both parties for the court to impose a royalty as part of the remedy for infringement.  A patent creates a right to exclude and, where the patent owner prefers to exercise that right, it should not be forced into a licensing arrangement resulting from a case in which it prevailed on the infringement claim.  There may be some cases in which vital public policy interests justify this result, but those cannot be grounded simply in the fact that the court denied a permanent injunction or the parties have not agreed to license terms.  A remedy should not penalize the person to whom the remedy is awarded.

No.  At least not according to three judges of the Second Circuit.  And as a result, one company’s valuable market estimates expressed in numerical form were subject to comprehensive misappropriation by a competitor.  But the court allowing this result forgot that numbers can be as expressive as words and sometimes are much more communicative.

The broad issue is whether a publication or database that consists primarily of numbers can be protected not only as to the arrangement or selection of the subject matter as a whole, but as to the individual numbers themselves.  This issue is important for any database of estimated or summarized market values of products or commodities. 

For example, if I examine four hundred technology stocks and publish my estimates in numerical form of the price the stocks will be one month from today, can you copy those estimates and publish them in competition with me?  Contracts solve some issues.  But we are talking about copyright. In copyright, the issue turns on whether the numbers are protected expression and, if so, whether the manner of the defendant’s taking of them constitutes infringement or (for example, fair use).  They are not protected expression if they constitute facts or if the number is merged with the idea it expresses.

So, what is a fact and what is expression? 

This question has tantalized generations of copyright law scholars.  But in our context, the answer seems simple.  Information in “print” cannot be a “fact” unless it describes or purports to describe something that exists in the real world.  “Facts” and “fiction” (or estimates) are opposites.

The Second Circuit has held that predictions of market value for automobiles were expression when based on the estimators’ judgment and not summarizing actual sale prices.

This is the right approach.  A numerical database is like to a textual work containing various expressive words and phrases.  This leaves ordinary doctrines such as “fair use” or “substantial similarity” the task of determining whether a specific appropriation of the numerical text is infringement.  It thus protects against substantial copying for competitive purposes while not precluding references to any aspect of the numerical text.

But, in New York Mercantile Exchange v. Intercontinental Exchange, 497 F.3d 109 (2nd Cir. 2007), the Second Circuit muddied the waters in what was admittedly unnecessary dicta.   The case involved estimated “closing” prices for commodities future contracts on the Exchange.  The estimates did not rely on the price of the last recorded trade.  Indeed, for many futures contracts on many days, there were no trades.  Yet, the court suggested that it was a close question of whether the estimates were “facts.”  This was supposedly true for estimates on days where many trades had occurred and, thus, the estimated price was grounded in substantial data.  But of course that is nonsense.  Regardless of how much data were used, an estimated price is just that – an estimate, not a fact.   The court also suggested that on heavily traded days for particular contracts, the might be only one “correct” estimate.  But, again, that is nonsense.

At worse, New York Mercantile undermines prior law.  At best, its dicta suggest a line difficult if not impossible to draw in practice. 

The better rule is that expression of estimates in numerical form are protected expression and that whether infringement occurred when some or all estimates are copied hinges on issues of fair use, which in this case were unlikely to protect the defendant.

After rambling about facts and expression, the court decided the case on a different basis, holding that the numerical expression merged with the idea being expressed; thus, it could not be protected under copyright law.  If expression is so “merged” with the idea that protecting the expression effectively precludes others from expressing the same idea, then the copyright statute precludes protection of the expression.

But what was the “idea” expressed?  The court never defined it.  Let me try.

·        Perhaps the idea is that: “an estimated closing price can be stated in numerical terms.”   

·        Alternatively, the idea is: “providing guidance for persons who will engage in commercial transactions regarding futures contracts.”

If either of these or many other “ideas” apply, they can be expressed in thousands of ways in numerical form.  The expression does not merge with the idea.

But the court revealed its true purpose – to deny protection of the numerical estimates.  Why?  Because the court did not think that protection was appropriate since the database producer would continue to produce the data regardless of copyright protection.

Bad law for a bad purpose.  Copyright law seeks to create incentives for creative authors, but it does so in general terms.  The analysis is any particular case is not whether the particular author will continue to produce its works without copyright, but whether denying copyright to an entire set of works will reduce the incentive and the production generally for works of the type.

The Supreme Court has accepted cert in the LG Electronics case.  The issue involves patent exhaustion doctrine.  This is an opportunity for the Court to affirm the right of a patent (or copyright) property owner to control the terms on which its invention or work is provided to the public – by license or sale.  But unless the rights-owner community pitches in, the Court may get it wrong.

The issue is simple – if I own property, should I be able to apportion rights in transactions relating to that property or am I limited to choosing between either selling outright or never providing it to the public or third parties?  The obvious answer is that a property rights owner should have the right to choose the terms on which it makes that property available for others to use.

            The LG appeal challenges that concept as it applies to patents. 

LG licensed Intel to use a portfolio of patents and to sell products covered by several of the LG patents.  The license provided that Intel sales did not cover buyers’ use of the product in combinations covered by other LG patents.  Intel was required to notify buyers from it that the sale to them did not give them a license to use the products in combinations that controlled by LG patents.  Those notices were part of the Intel-buyer contracts.

So it seems simple.

            Property owner gave a limited license to Intel.  Intel produced products under that limited license and conveyed them to third parties pursuant to the license and under a contract limiting the scope of the sale transaction.  Transferees went outside that scope.  While transferees are shielded from some infringement claims by exhaustion doctrine, that doctrine applies to authorized unconditional sales; these transactions were not that because there were contractually effective restrictions on the rights transferred.

            The Supreme Court has accepted cert on the following issue:

Whether the Federal Circuit erred by holding … that respondent's patent rights were not exhausted by its license agreement with Intel Corporation, and Intel's subsequent sale of product under the license to petitioners. 

The license here did not permit sales by Intel that would enable the purchaser to use the product in combinations that would infringe the licensor’s other patents.

            Of course such transactions did not “exhaust” the patent.  They were a contractual method of commercializing patents under license agreements that commercially apportioned rights or permissions.   If I rent my condo in Taos for a specified, limited use, have I exhausted my rights of ownership as to the lessee for any uses it desires to make of the premises?  Of course not.  The Court needs to make this clear.

            So what was the argument by the losing party below?  The following is quoted from their brief:

Under the patent exhaustion doctrine that this Court has applied for more than 90 years …. an authorized first sale of a patented article exhausts the patent owner's rights in that article, and nullifies any “conditions” that the patent owner has tried to attach to its use or resale. Beginning with its decision in Mallinkrodt v. Medipart, Inc., 976 F.3d 700 (Fed. Cir. 1992), however, the Federal Circuit has steadily eroded the exhaustion doctrine. In this decision the Federal Circuit held that exhaustion is entirely optional, and easily nullified by a “notice” announcing that the patent owner would prefer that it not apply. That is an unprecedented and extremely dangerous expansion of the patent monopoly, in direct conflict with numerous decisions of this Court.

Frankly, this misrepresents the Federal Circuit’s actual ruling, prior law, and what the case involves.

The basic rule is and has long been that ideas of exhaustion and first sale only apply where there has been an authorized unconditional sale of a product or a copy or an unconditional and unrestricted license.  As the Federal Circuit commented:

The theory behind this rule is that in [an unconditional sale], the patentee has bargained for, and received, an amount equal to the full value of the goods.  This exhaustion doctrine, however, does not apply to an expressly conditional sale or license. In such a transaction, it is more reasonable to infer that the parties negotiated a price that reflects only the value of the 'use' rights conferred by the patentee."

The LG‑Intel license expressly excluded granting a license to Intel’s purchasers to combine Intel's parts with other components in ways that would infringe LG’s other patents.  This conditional agreement required Intel to notify customers of the limited scope of what they were purchasing, which it did.  Although Intel was free to sell the products, those sales were required to be conditional, and Intel's customers were expressly prohibited from infringing LG's combination patents.

            This is the proper rule. If rights owner chooses to commercialize its invention by apportioning rights or permissions, rather than simply selling product, that choice should be respected.  It is not an effort to circumvent ideas of exhaustion or first sale, but simply a decision to not engage in transactions to which those theories apply.

            In Nagrampa v. Mailcoups, Inc., 469 F.3d 1257 (9^th Cir. 2006) the Ninth Circuit held unconscionable as to a California franchisee,  a mandatory forum and arbitration clause picking Massachusetts as the venue for disputes in a franchise agreement . The alleged procedural problems were the unbalanced bargaining positions and a lack of actual bargaining. But contracts are seldom bargained as to matters other than price and quantity or scope. 

             The alleged substantive problems? 

(1) The clause lacked mutuality because the franchisor could use judicial process in some cases (e.g., to protect its trademark rights), while the franchisee could not. But contract law has never required that each clause have mutuality, only that the entire contract reflect a mutuality of obligation, which this one did.

(2) “[The] contract would require a one-woman franchisee … to fly across the country to arbitrate a contract signed and performed in California…. She may not be able to maintain her claim … The forum selection provision has "no justification other than as a means of maximizing an advantage over [franchisees]."” But the Supreme Court has held that mere cost is not enough to invalidate such clauses. Also, there was a justification for the clause – without it, the franchisor would be forced to litigate in all of the states in which it granted franchises and, thus, to charge all franchisees a premium to enable it to do so.

            Nagrampa is bad enough, but the Ninth Circuit later continued the exegesis of its view of law in Douglas v. U.S. District Court, 2007 WL 2069542 (9^th Cir. 2007), a mandamus action. There, after holding that an amendment of an online agreement lacked assent, the Ninth Circuit gratuitously commented that even had there been assent, the terms for class action waiver and mandatory arbitration were unconscionable. The procedural flaw: the clause was a non-bargained standard form. The court said: “The district court held that the arbitration clause [was] not procedurally unconscionable … because Douglas had meaningful alternative choices for … service. Under New York law [this] forecloses any procedural unconscionability claim. However, [in Nagrampa] we noted that California “has rejected the notion that the availability ... of substitute ... services alone can defeat a claim of procedural unconscionability.” [A] contract can be procedurally unconscionable if a service provider has overwhelming bargaining power and presents a “take-it-or-leave-it” contract to a customer - even if the customer has a meaningful choice as to service providers.”

            But– why is that true? If the other party has a valid option that would allow it to choose another provider and it does not exercise that option, why is the result procedurally unconscionable?  The court did not provide any answer to this.

            Then in Davis v. O'Melveny & Myers, 485 F.3d 1066 (9^th Cir. 2007), the Ninth Circuit held that an arbitration clause proposed by a law firm and assented to by an employee was unconscionable. It was procedurally unconscionable not because of any lack of actual notice, use of misleading language or terms, or lack of options for the employee. It was procedurally unconscionable because the law firm did not offer the employee the option of rejecting the clause and continuing as an employee with an unmodified contract. The law firm “merely” offered the employee a three month period of employment in which to decide and accept the term or find a new job.

            This collection of cases is bad law that transfers making and amendment of contracts to courts, rather than parties acting in a marketplace. Procedural unconscionability is properly associated with lack of notice, deception, advantage-taking, and quasi fraud. As the doctrine itself states: “The principle is one of prevention of oppression and unfair surprise and not of disturbance of allocation of risks because of superior bargaining power.”

             Under the Ninth Circuit view, however, procedural unconscionability morphs into mere use of non-negotiated standard forms and sets the stage for a court hostile to the use of arbitration and class action waivers (or other terms) in consumer and employee cases to invalidate otherwise valid agreements. 

The appropriate role should be facilitative of competition, rather than to resolve technological competition before the market chooses. Where two (or more) technologies or products have viable market presence, the proper place to resolve that competition lies in the marketplace, not on the technologist’s computer screen or the bureaucrat’s regulations.

This calls for deference to the market with respect to standards not associated with safety, health or other over-riding public policy issues. Standards imposed by an SSO in a competitive market should occur only based on true consensus, including among all competing firms, or one’s framed in a manner that does not alter the competitive balance. This can be accomplished in numerous ways, including by recognizing that competing approaches exist and adopting only standards for individual competitors or by adopting standards that reflect both (all) competing systems, leaving the separate standards compete in the same fashion as the basic products and technologies compete.

The closer a standard moves toward resolving an existing, competitive issue in a mandatory or quasi-mandatory way, the more predictable it is that the standards-setting will in fact be driven by competition issues. Resolving such issues is not an appropriate role for standards setting organizations.

The intrusive impact of standards into competitive contexts is directly related to the type of standard involved. As the standard moves more toward a permissive framework, its competitive impact lessens and the likelihood of competitive influences diminishes.

The SSO process potentially displaces the market in choosing preferred technology. In many circumstances, however, rather than being neutral, adopting a standard may advantage one competitor over another without allowing the market to judge the competing approaches. Standards-setting has become (perhaps always has been) a competition environment that impacts commercial markets, but that functions outside their influence.

Why is it wrong, one might ask, to have technocrats deciding for a market based on technical criteria, rather than the market deciding. The answer in part lies in faith in the market and in innovation that is not prematurely channeled into a particular path by a “standard.” The simple fact is that the market is the best guide to consumer value. A standards-based choice risks choosing the wrong one and foreclosing alternatives. Once implemented, the switching costs to an alternative are high and perhaps insurmountable.

But the question is not simply why it is wrong to have technocrats decide. A pure technological decision in ongoing market competition is impossible to achieve. Any standards-setting will draw competition-based reactions. And so it should. SSO procedures should attempt to exclude competitive themes in a “neutral” SSO process, but that is likely to be impossible to enforce when significant competitive issues are involved.

The true question is – how should SSOs deal with the fact that for important standards there will competitive advocacy based on obtaining market advantage? The SSO setting is not a market, but a company whose competitive position may be affected by a standard will act to protect that position or enhance it.

SSOs should take the following steps:

• First, SSOs must recognize the inevitability of competition issues whenever a proposed standard would resolve current competition in the market place.
• Second, consensus standards-setting should in fact reflect a true consensus and, particularly, agreement among all major competing firms in the affected market before a mandatory or quasi-mandatory standard is promulgated. This gives veto power to all competitors and runs counter to the goal of those who seek expansive scope for standards drafted on the basis of pure technology, but it recognizes the inevitable competition impact that such mandatory standards may have.
• Third, where standards are proposed in environments where two or more approaches are in active competition, any standard adopted should be permissive by focusing on only characteristics shared among the competitors, are unlikely to adversely affect one or another of the competitors, or are formulated solely for one of the approaches and not mandated for the others.

That software and other intangibles are not within the idea of goods is also recognized by many courts. The Eleventh Circuit in Systems Unlimited, Inc. v. Cisco Systems, 2007 WL 1047064 (11th Cir. 2007) explicitly recognized the distinction between tangibles (a refrigerator or a diskette) and intangibles (software). It held that a sale of software assets was not a sale of goods. Because it was not a sale of goods, it did not contain an express or implied obligation to deliver copies of the software to the transferee.

There are two messages here. One, as the nature of software becomes increasingly clear, courts (and legislatures) are unlikely to make a false analogy between the software and the physical media on which a copy might exist. Two, lawyers and parties who do not recognize this may leave huge gaps in their transactions.

The contract was a settlement agreement between Systems and Cisco. Cisco: “granted, bargained, sold, transferred and delivered … the following: Any and all of [Cisco]'s right, title and interest in any … intellectual property of any kind associated with any software, code or data, including [mentioning specific software] and including any and all available copies thereof and any and all books and records related thereto by [Cisco]....”

Cisco never delivered copies of the software to Systems.

The court held that there was no express obligation to deliver copies.

Furthermore, there was no implied obligation under Article 2 of the UCC because, well simply because Article 2 did not apply. The court correctly explained: “Systems also argues that the UCC imposes a duty on Cisco to deliver the software. We will assume without deciding that Systems' reading of the UCC is correct. Even so, the provisions of the UCC only apply to contracts that deal predominately with “transactions in goods.” The sale of intellectual property, which is what is involved here, is not a transaction in goods. Thus, the UCC does not apply.”

The U.S. Supreme Court dealt with the same type of issue under a different law, but also acknowledged the distinction between copies and software in Microsoft Corp. v. AT & T Corp., 2007 WL 1237838 (2007). The statute here provides that patent infringement liability if a person provides components of a patented invention to persons outside the U.S. without authority. (271(f)) Microsoft provided master disks to international licensees containing an allegedly infringing invention. The court concluded that this did not satisfy the requirement of providing “components”.

The Court emphasized that software code in the abstract is merely an idea and not a “component” of a machine. Thus, when Microsoft supplied a master disk to recipients in other countries, it was not “supplying” a component, but merely an abstraction, even though the recipient was expected to and did in fact make copies from the master disk into the machines or copies it resold.

So we have an increasing recognition that the software code and functionality it creates is not within the idea of being a tangible “good”. Providing a transferee with the right to copy software (in general or in its own computer) is not a transaction in goods. Making exercise of that right possible by providing a copy of that software does not change the nature of the transaction – it is a transaction in information, not in goods.

An intriguing part of GPL 3.0 is how it singles out particular practices or parts of an industry for special treatment. The drafters do not like anti-circumvention laws, so they try to use their license to circumvent those laws. § 3. They do like some companies and licenses that do not wish to follow pure GPL restrictions, so they validate many “Additional Terms.” § 7 The license attempts to commit patent owners who distribute “free” software to granting licenses to both current and any future patents.

The drafters do not like commercial arrangements like that between Microsoft and Novell, so they include terms trying to block that practice in the future, while not precluding Novell from adopting version 3.0 even as to its existing agreement. § 11. They believe that if 3.0 is used by Novel for future transactions, then Microsoft will be captured into broad patent licenses not contemplated in their original agreement. I think they are wrong.

The Microsoft-Novell agreement provides that each company shall covenant not to sue the other company’s customers for patent infringement on account of the customer’s use of certain products. It provides for payment from each party to the other based on how many such customers receive the commitment and for a means of recording this (e.g., using coupons). This is an unusual cross-licensing agreement since it protects the customers, but does not protect either company from a lawsuit by the other.

So why does the Free Software Foundation (FSF) dislike an arrangement that protects users of open source works from patent claims by Microsoft? It is not clear to me that the reason is anything other than a visceral dislike of Microsoft as epitomizing the so-called proprietary software realm and the fact that this arrangement allows Microsoft to participate more in the open source without adopting all premises set out by FSF. Beyond that, the only complaint is that the protections are not comprehensive – neither party gives the other’s customers a covenant not to sue as to several limited types of uses.

GPL 2.0, the most widely used open source license, does not preclude arrangements like the Microsoft-Novell agreement. This is because GPL 2.0 focuses solely on the rights and limitations received by the transferee of the software from the licensor. Microsoft is not a licensor of Novell software and Novell is not a licensor of Microsoft software under this arrangement. Rather, both have merely agreed with the other to waive rights as to third parties. In legal terminology, the customers are third-party beneficiaries of the MS-Novell agreement. GPL 2.0 does not deal with such relationships.

GPL 3.0, however, has two provisions that try to prevent this type of an arrangement (or at least change it to fit FSF goals). The problem for FSF is that Microsoft is not bound by either provision when it is not itself distributing software under or covered by GPL 3.0 and that Novell is affected only if it chooses to use GPL 3.0.

The first is paragraph 6 of § 11. It says basically that if you (as licensor) “convey or propagate by procuring conveyance of” a covered work, and grant a patent license to some parties receiving a GPL 3.0 work, “the patent license is automatically extended to all recipients of the covered work and works based on it.” The FSF apparently believe that if Novell adopts GPL 3.0, Microsoft will have procured conveyance of the GPL software by giving coupons to its customers and a patent waiver to Novel customers, thus falling within the provisions of paragraph 6. But this language is in GPL 3.0 – arguably a contractual document, but certainly not a body of law good against the world. If I am not bound by that contract, I am not bound by that provision.

Look at this from Microsoft’s perspective. It never agreed to GPL 3.0 in a transaction for a product covered by the Novell-Microsoft agreement. There is a debate about whether GPL creates a contract at all, but if GPL does, the contract is solely between the parties who agreed to it. That is one of the most basic tenets of contract law. Judge Easterbrook many years ago got it exactly right: contracts are not rights good against the entire world, but only create obligations between the contracting parties.

The GPL method for obtaining agreement is that the party agrees to the license by modifying or “propagating” the program. The word “propagate” is new to Version 3.0 and to law generally – it is another of many concepts that may spawn litigation in the future. But GPL 3.0 defines it to mean doing anything that would entail infringement (direct or indirect) under copyright law except executing in on a computer or modifying a private copy.

Microsoft neither modifies nor propagates in the type of transaction we are discussing. The customers who receive the Microsoft covenant not to sue are customers of Novell. Microsoft may give Novell covenants not to sue to its own customers or inform its own customers that it will not sue if they use Novell open source software within the Microsoft-Novell agreement, but it does this in transactions using its own licenses and software, not GPL. Microsoft (or any other party in a similar arrangement) is not bound by the excessively broad requirement in paragraph 6, even if it’s contracting partner decides to use GPL 3.0 for its own customers.

If GPL does not create a contract, the situation is even more clear. Non-contractual purported restrictions on the use of a transferred copy may not even bind the transferee. They certainly do not bind a third party.

The situation for Novell (or similarly situated party) is less clear as to its own patents. Novell may choose to both agree to GPL 3.0 and to grant a patent license for its own patents to its transferees. If it does so, then the broad paragraph 6 undertaking may apply. But Novell does not and cannot grant patent licenses to Microsoft patents. Indeed, the Microsoft-Novell agreement seems to assume that Microsoft commits to a patent enforcement waiver only for a defined group of Novell customers.

The second provision is paragraph 7 of § 11. This paragraph states that a party who has software covered by GPL 3.0 cannot convey the software to another person if that party is part of an arrangement defined specifically to track the Microsoft-Novell arrangement where the patent license agreed to is “discriminatory” (e.g., does not give or conditions the rights available under GPL). GPL 3.0 provides that this non-conveyance rule only applies to arrangements made after March, 2007. The Microsoft-Novell agreement is thus excluded to “tactical” reasons – that is, FSF wants Novell to shift to GPL 3.0 and believes that Microsoft is caught up within paragraph 6 for any GPL 3.0 product.

But as we have seen, they are not. Thus, this paragraph 7 is a provision that narrows the settlement and commercial agreement options for all subsequent developers.

The lower court had concluded that this fact and the potential that there would be a market for such images in other contexts indicated that the defendant’s use had a market impact. The court of appeals, however, suggested that this hypothetical market was insufficient given that there was no evidence that users of the search engine downloaded the thumbnail images for cell phone or other use.

But there is a troubling circularity to the market effects analysis here. The market for the thumbnails could include licensing their use in search engines, except that the court’s fair use analysis eliminates this potential market and does so based on the analysis that the market does not exist. This circular analysis deprives the copyright owner of the opportunity to full exploit its work by developing markets over time, since it suggests that another party that enters a market defined by modified versions of the original work (e.g., derivative works of it) before the copyright owner may in fact preempt that field of use. That cannot be what is meant by market impact in the fair use standard.

What is wrong in the court’s analysis is that it failed to consider the likelihood that a copyright owner in general or that this copyright owner in specific would opt into a market defined by online use of thumbnail versions of its works. In fact, in this case and in this era, that likelihood is strong.

Licensing Law Materials (provides brief summaries of selected licensing contract law cases and articles)

Current copyright literature (lists recent copyright law articles with reference to pdf of first page)

Findlaw (broad service providing access to cases, model contract, address, and other resources)

Groklaw (Running commentary on various legal issues; links to numerous sites)

Info Law.org (Registration and other materials on the 20th Annual Intellectual Assets and Technology Law Institute)
I
nternet Law and Regulation (Articles, cases, and other extensive materials relating to Internet law)

The Authors' Guild (News and positions regarding various copyright law and related issues from an author's rights perspective)

I wanted to begin by being general on how things have changed over the years about the idea of “electronic commercial law.” Simply stated, courts and commercial parties have recognized that electronic commercial transactions and licensing are now a central part of commerce, but legislatures and agencies seem to want a regulatory world for this form commerce. We need to fight that tendency.

Electronic commerce refers to the use of digital systems to create or perform transactions and enforce rights in digital information. The term covers a wide variety of commercial practices that were once new in ordinary commerce: they are now staple elements of it.

What does that mean for law?

For contract law, the meaning should be simple. Contract law is a practical discipline that supports the marketplace and should respond supportively to changes in commercial methods and subject matter. We have a different transactional framework (licensing), a new subject matter (digital information), and a new way of doing business (electronic interactions). Contract law needs to continue to adjust to support this new environment.

The relationships have changed, not just the mechanics. Ethan Katsh once commented: “[A] new system of communication … leads to the creation of new relationships and, most importantly, changes our attitudes, expectations, and ways of thinking about law.” M. Ethan Katsh, The Electronic Media and the Transformation of Law 22 (1989). He was right. New relationships. New attitudes. New expectations. These are the grist for the mill of contract law and practice. The new technologies have fundamentally changed the character of relationships. These new relationships need to be governed by new supportive contract law themes, not by themes developed from an old, now gone era.

Where does all that leave contract law?

Contract law creates (or denies) legal obligations associated with voluntary interactions. When the interactions and their subject matter change, contract law serves the same function but its tools and outcomes must differ from the past.

In our world significant change seldom flows smoothly. While many embrace change, others resist it. Some of the resistance is due to what Lewellyn explained years ago: “You wake up then to the fact that the throne your subject matter once occupied is overshadowed”; that is a fearful situation for many. The costs imposed on commerce by reaction to that fear are extravagant and harmful.

In my view, rather than protecting the status quo, the role of law generally should be to establish a responsive body of rules that support change and that limit regulation to cases where actual clear abuse otherwise exists. This has been the tradition of U.S. commercial law. But it has not consistently been the way in which law related to electronic commercial transactions has evolved. Instead, we have seen an explosion of new law, often regulatory in nature, dealing with data security breaches, “spyware”, privacy policies, “spam”, etc. Too often, political arguments and interest group politics weigh in toward the view that the proper role of law is to regulate commerce, rather than to support it. Much of this lies simply in a grab for position enforced through law, rather than in the marketplace. Also, this stems in part from the regulators’ view that preexisting regulations were already inadequate to safeguard consumer interests and that the transition to a new commerce provides an opportunity to impose more extensive regulation. In addition, some believe that there are enhanced risks of abuse in e-commerce that require proactive regulation even before abuse actually occurs,

But when a regulatory approach is taken in a period of rapid social change, the result is an enormous expansion of new law and we pay a huge price for this. Its short-term effect lies in the creation of an often-bewildering array of new rules and regulations with which commercial entities must deal, and which seldom reflect sound or considered legal or social policy.

Digital media have enabled a wide range of new ways of doing business and new commercial subject matter. There are numerous instances in which the balance between technological capability to engage in new transactions, and legal infrastructure that permits or impedes implementing aspects of the new technologies is being drawn. We need to exercise care in such contexts.

The messages are simple ones

· Electronic commercial transactions have engaged a fundamental change in commercial and consumer relationships, expectations and subject matter. The relationship between a licensor and licensee, an access provider and its clients, and other modern information-based transactional relationships does not flow from models of relationships between the buyer and seller of a toaster or television set or from a desire to regulate and impede change.

· Electronic commerce might once have been reasonably treated as a small sub-part of contract law and commercial practice. That is no longer true. This is, indeed, modern commercial law and needs to be respected on its own terms with its own unique demands.

· The business of law in our market economy is not to comprehensively regulate and control, but to support technological and commercial evolution, limiting regulation to cases of proven and palpable abuse.

More on these themes later.

Ray

Coexistence implies two or more persons or products being together in the same place or system peacefully despite their differences. For software, it entails being able to have software that interacts with, or uses, FSOS software despite differences in licensing philosophy and approach. But coexisting with free and open source software entails coping with FSOS license terms in each license used in a system.

Many open source licenses make this easy because they are permissive, like the BSD license. Permissive licenses approximate openly dedicating the code to the public domain. They should be philosophically preferred.

Other licenses are pass-through. In a pass-through license the terms must be passed through to only as to the original licensed material. License terms for new material are at the option of the party who adds the material.

Then there are the viral licenses. These are the most restrictive. They provide that, in defined circumstances, a licensee's use is conditional on applying the license to the original work and to new code. The GPL claims: "To protect your rights, we need to make restrictions ... These restrictions translate to certain responsibilities for you ...." Shades of the novel: 1984.

Let's look at paths of coexistence with viral licenses.

Path 1: End user modification
Since current licenses do not impose terms on end user modifications, end users are safe today. Distributors can distribute works that are not derivative of an FSOS work, but that enable the end user to create a derivative for its use.

Path 2: Distributing non-expressive elements
Viral provisions often are limited to works that are copies or derivative works of the original. A path to coexistence is to design software distributed in a manner that does not contain expression of the original work.
This may seem trivial, but the most valuable aspects of software are often not protected expression. Ideas and approaches are not copyrighted expression. Thus, code extracted from another's work is not expression if it fits one of three characteristics:

• What is used is so abstract or so changed from the original that it simply uses the unprotected idea, rather than the expression.
• What is used is so narrow and related to machine functions that it is merged with the function and an unprotected process.
• What is used is found in so many unprotected works that it does not constitute expression.

Path 3: Distribution outside the scope
Viral licenses impose their terms only on code that connects to the free software in a particular way. Avoiding that particular connection avoids the viral effect.
For example, viral terms often cover "derivative works." This is a copyright law term. A work is not a derivative work unless it contains expression from the work from which it allegedly derives. It is not a derivative work simply because it interacts with the other work or uses ideas from it, regardless of how the links or other connection occurs technically.

Path 4: Relying on exceptions
One can also rely on exceptions from the owner of the copyright or obtain a separate agreement devoid of FOSS restraints. The task is to identify sources of code and to obtain clearance from that source. Free software licenses are private documents. The person who controls the copyright can grant broader rights or exceptions. In fact, many FOSS programs are dual licensed - available with and without viral terms.
The most significant exception deals with the Linux kernel. Torvalds made clear early that the viral terms of GPL did not apply to applications operating in "user space" even though those applications made calls to or otherwise linked with the operating system kernel.

Path 5: Distancing strategies
A distancing strategy places valuable code in a safe harbor, even if other code is used to link to it and, perhaps, constitute a derivative work. The connecting code is "sacrificed code" that may fall under viral terms. Whether, as distributed, the code constitutes a derivative work does not matter, so long as the interaction between it and the "Protected Program" falls into a safe harbor of permitted coexistence.

Summary
So, there are ways around the threat to coexistence. But I wonder why they are needed. Why does a movement that refers to itself as free and open, try to impose restrictions on others to act freely or openly?

The Court's decision in Independent Ink permanently (I hope) set aside the fiction that rights under IP law equate to market power. Virtually all neutral observers have urged burial of that concept, as have virtually all lower courts. The only people supporting the fiction were plaintiffs seeking antitrust damages and a coterie of people and organizations that believe commercialization of IP should be tightly constrained. That's where the fiction began - in a desire years ago to constrain those rights, a hostility to them, and a fear of misuse. To me, those attitudes seem today arcane and extravagantly misguided.

It took the Supreme Court years to eliminate this one artifact of an old world view not because the issue was in doubt, but because the Court can only rule on cases presented to it. I suspect that Congress thought that it had eliminated the fiction in 1988 when it legislated that no person could be held to have misused a patent by requiring another license or purchase as a condition for the license, unless the patent gave it market power in the antitrust sense. But that statute only referred to misuse, not antitrust, and did not refer to copyright or trademark cases. Under any fair reading of the Independent Ink decision, however, there is no basis in misuse or antitrust law for continuation of a market power assumption.

Other artifacts of that old world-view linger in modern law because the Supreme Court has not revisited them. The signal from the modern Court is quite clear. Courts should also discard these other remnants. Among the top on my list to be discarded are:

· The idea that patent or copyright misuse can occur without circumstances that support an antitrust claim.
· The idea that agreeing to pay royalties after a patent expires is misuse.
· The idea that an agreement to not contest the validity of a licensed right is unenforceable.
· The idea that licensing copyrighted works or patents only on the condition that a whole package of numerous works or patents be taken is suspect or perhaps per se invalid.
· The idea that requiring a licensee to compute royalties on a base that is broader than the licensed patent is an expansion of the patent, rather than simply a means of pricing the contract.

This is not to say that any of these practices might not be objectionable in some contexts, but that they all should be evaluated under competition law standards and a rule of reason, rather than the remnant of old hostility to IP and IP owners. As Judge Posner commented: "If misuse [and other] claims are not tested by conventional antitrust principles, by what principles shall they be tested? Our law is not rich in alternative concepts of monopolistic abuse; and it is rather late in the day to try to develop one without in the process subjecting the rights of patent holders to debilitating uncertainty."