I informed Ms. O'Gara, and she has now printed a correction regarding the author's credentials and has also printed a rebuttal letter from a real lawyer, explaining how the original letter writer was wrong on the facts and on the law. It's called "Practicing Law Without a License", and I recommend it if you were the least influenced or worried about the original attack on the GPL. It's here, on LinuxWorld, who should have known better than to print it in the first place, and who have compound their fault by printing a followup letter by the same person, also about the GPL. The letter from the lawyer is from Celia Santander, Esq., Adjunct Professor at Duquesne University School of Law, teaching intellectual property law, and who is former director of legal affairs for TimeSys Corporation (an embedded Linux developer), and former vice-president of technology asset management for PNC Bank.

The basic premise the "convert" wrote about was that the GPL is a contract and not a license. This was his same argument when he would post on Groklaw, and many tried to explain it to him. Now Ms. Santander answers that the GPL would be effective either way. Besides, the retired physicist got all mixed up about the law, she points out, because of wrongly thinking patent law and copyright law work the same way. They don't. Here is a brief segment of a long letter well worth reading in full:

Mr. Wallace correctly describes a number of different legal principles, including a fair description of the rights involved in derivative works. However, his key assumption - the foundation of his entire argument - has absolutely no basis in case law, statutory law or even reasonable legal analogy. His argument is based on one key principle: That in order to grant a unilateral license, the licensor may only impose conditions to the license that involve the licensor's exclusive rights, and that the licensor may not impose any conditions that involve any of the licensee's rights.

He cites a Supreme Court case (General Talking Pictures Corp v Western Electric Co Inc) for this principle. There are a number of problems with this case and Mr. Wallace's interpretation. Firstly, as a threshold issue, you cannot analogize patent law to copyright law. It is like comparing apples to zebras. . . .

Perhaps what is confusing Mr. Wallace is this: It is true that you cannot interfere with the rights of a copyright holder without his/her permission. Mr. Wallace, I believe, views the GPL as interference by one party (the licensor) with the copyrights of another party (the licensee) because the license is conditioned upon the licensee taking certain actions regarding his copyrights. He presents the following alleged conundrum: Either the second party (licensee) is agreeing to this interference, in which case the GPL cannot be a unilateral permission since now we have two parties agreeing with each other (sounds like a contract); or, the second party (licensee) is not agreeing to this interference in which case the GPL is not valid because it is a violation of copyright law. (Mr. Wallace ignores the issue of whether or not a valid contract has been formed under the first scenario.) The answer to Mr. Wallace's purported conundrum is this: The second party is consenting to the interference by fulfilling the condition of the unilateral permission, but merely fulfilling such a condition does not change the nature of the unilateral permission as unilateral.

So that is one lesson. Another lesson would be, in my opinion, that allowing your emotions to influence your thinking may be normal human behavior, but it can certainly interfere with your ability to think, no matter how smart you are.

I am reminded of an old Armenian saying. It goes like this: If one person tells you that you are a donkey, ignore him. If two people tell you you are a donkey, consider it and if you are sure they are mistaken, then forget about it. When a third person tells you that you are a donkey, go get a saddle. Mr. Wallace could have spared himself this humiliation had he just listened when so many on Groklaw tried to explain why his "legal" interpretation was not correct.

And the final lesson: not every media outlet using the name "Linux" is necessarily a friend. Whether Linux World's motivation in printing the letter was a desire for hits or a desire to do harm to the GPL or just from ignorance might make an interesting conversation. But does it really matter? They surely need to redeem themselves if they wish the community to support them.

Update: 2010 - Neither article is still available at the original links, and for fear of losing this piece of the entire saga, I am reproducing them here, first the Dan Wallace letter, and then the response:

***********************************

MAIL BOX: SCO Wins Convert to its GPL-is-Invalid Argument
January 30, 2004

Summary

"Linux cannot be distributed under any kind of 'viral license' for future development," writes Daniel Wallace. "All those enterprise users will be stuck with the version of Linux they are now running with no way to repair or upgrade because the license is fatally flawed," he adds. Read his full Letter to the Editor here.

By Maureen O'Gara

To the Editor:

Why is the Free Software Foundation given a pass on the issue of contract enforcement under state law on binding legal agreements like the GPL? The consequences are dramatic indeed for the commercial enterprise environment.

When the Free Software Foundation speaks of unilateral permissions or bare license law enforcing the GPL, they are referring to a long line of case law concerning patents that was summarized by the Supreme Court in General Talking Pictures Corp v Western Electric Co, Inc., 305 US 124,125.

The principle involving a "bare license" or "unilateral permission" is that a patentee may condition his own reward of exclusive rights as he chooses. After all, the rights are his alone and he may condition them as he sees fit. The conditions he places must involve only his exclusive rights and not the exclusive "of parties involved and there is no privity requirement. The legality of this principle has never been questioned (i.d. at 125).

A derivative copyright work by the definition of sec. 103 (b) contains exclusive rights for two distinct parties, the authorizing "preexisting material" author retains his rights and the "contributed material" author gains separate, disjoint and exclusive rights to the new material he has contributed. There is no analogous "derivative patent" creation under patent law.

The "pre-existing material" author has the exclusive right to authorize and prepare derivative works of his "pre-existing" work but has no copyrights whatsoever in the "contributing" author's material.

Since the derivative work contains both the "pre-existing material" and the "contributed material" it is obvious that if both sets of "material" have exclusive disjoint rights that two separate, exclusive permissions are required to distribute the derivative work as a whole - the "pre-existing" author's permission and the "modifying" author's permission. Both authors must agree to permit distribution of the derivative work as a whole.

This is where the unilateral permission model fails. A "pre-existing" work licensor cannot place conditions on "contributing" authors' exclusive rights. It's utterly outside the scope of definition of a unilateral permission.

The "pre-existing" author has exclusive rights and no conditions can be placed on his exclusive rights without his agreement to do so. The "contributing" author has exclusive rights and no conditions can be placed on his exclusive rights without his agreement to do so. They are after all, exclusive rights. Only some form of "binding agreement" between both parties can supply both permissions required to distribute a derivative work.

The IBM legal team in their Amended Counterclaims is absolutely aware of this problem. In their Amended Counterclaims they describe the GPL as a "public agreement" cast in "a binding legal form." Ask yourself this question: If IBM describes the GPL as "a binding legal form" - what is being "bound?" The answer is "copyright permissions." That's what the GPL is about. Exclusive copyright permissions.

Unilateral license permissions do not need state enforcement. Unfortunately "binding legal forms" are enforced under state action as there is no recognized federal authority in this area.

How then, do you permit a derivative work to be distributed? This is usually done at the time the "pre-existing" author authorizes the derivative work by way of a "binding legal agreement" of some form with the "contributing" author.

When the GPL asserts:

" 2. You may modify your copy or copies of the Program or any portion of it, thus forming a work based on the Program, and copy and distribute such modifications or work under the terms of Section 1 above, provided that you also meet all of these conditions:..."

It is the fact that copyright law recognizes two disjoint, mutually exclusive copyrights in the same derivative work that frustrates license law as promulgated in General Talking Pictures Corp v Western Electric Co mentioned above. This type of "derivative" creation was never anticipated in the evolution of patent licensing law.

There is an exclusive right for an original author to "prepare a derivative work," but there is no exclusive right to distribute a derivative work. The Copyright Act is absolutely silent concerning the "distribution of derivative works." It's left to contract law to control the distribution of derivative works.

Now things get worse. Unilateral permissions do not require privity. "Binding legal forms" do require privity. It is obvious IBM is attempting to solve the privity problem by describing the GPL as a "public covenant." "Public trusts" do not require privity.

Even if privity were granted, things get worse yet. The GPL purports to restrict the exclusive rights of the "pre-existing" owner(s) in a derivative work from authorizing a new derivative work by "bargaining for permissions" from new "contributing" authors in a continuous sequence of new derivative works (ad infinitum). Remember the phrase "You must cause any work that you distribute...to be licensed" in the GPL sec 2(b)?

The interpretation of "binding legal forms" is left to state action for enforcement. So this "binding legal form" is clearly an attempt to create a "new right against the world" concerning copyrights. The FSF calls this new right "copyleft." The attempt to create a "new right against the world" enforced under state action because of the "binding legal form" triggers preemption by sec. 301. This legal principle of preemption is clearly described in ProCD Inc v Zeidenberg, 86 F.3d 1447 (7th Cir.).

When SCO's attorney Mark Heise said copyright law "preempted" the GPL and only allowed "one copy" he was being somewhat vague. He wasn't referring to the "number of copies," he was referring to the "number of successors." SCO is indeed correct that the GPL is invalid.

Promissory estoppal will keep everyone from suing each other short-term but Linux cannot be distributed under any kind of "viral license" for future development. All those enterprise users will be stuck with the version of Linux they are now running with no way to repair or upgrade because the license is fatally flawed. "Copyleft is not possible."

A quick check by a competent attorney with the citations above will quickly confirm the accuracy of these conclusions.

Sincerely,
Daniel Wallace
[retired physicist, associate member of the FSF]

*************************************
*************************************

Mail Bag: Practicing Law Without a License
February 6, 2004

By Maureen O'Gara

The letter from Daniel Wallace, a retired physicist we misidentified as working for Insight Communications, that we ran last week created a little firestorm judging from the calls and e-mail we got. The following letter, which came in at press time, is indicative of the reaction.

To the Editor:

I am an Adjunct Professor at Duquesne University School of Law teaching upper-level intellectual property law, former director of legal affairs for TimeSys Corporation (an embedded Linux developer), and former vice-president of technology asset management for PNC Bank. In addition to teaching, I currently advise a number of corporate clients regarding information technology matters with related intellectual property issues, and, in particular, risk management issues related to Linux and other open source programs.

I was disturbed to read your recent posting of a letter to the editor authored by Daniel Wallace. I was equally disturbed by the lack of legal expertise reflected in most of the posted comments. I feel strongly that a proper legal response is required, and needs and deserves equal exposure in your paper - not a buried posted response.

So - I am requesting that you publish the following response to Mr. Wallace's letter in the next available issue.

Most sincerely,

Celia Santander Esq.

Mr. Wallace correctly describes a number of different legal principles, including a fair description of the rights involved in derivative works. However, his key assumption - the foundation of his entire argument - has absolutely no basis in case law, statutory law or even reasonable legal analogy. His argument is based on one key principle: That in order to grant a unilateral license, the licensor may only impose conditions to the license that involve the licensor's exclusive rights, and that the licensor may not impose any conditions that involve any of the licensee's rights.

He cites a Supreme Court case (General Talking Pictures Corp v Western Electric Co Inc) for this principle. There are a number of problems with this case and Mr. Wallace's interpretation.

Firstly, as a threshold issue, you cannot analogize patent law to copyright law. It is like comparing apples to zebras. As just a couple of the many examples of why this is true, consider some fundamental differences between the two bodies of law: patents allow each co-inventor to exercise her exclusive rights without any accounting to the other(s) while copyright law requires co-authors to report to and compensate as appropriate their fellow co-authors for any exercise of rights that results in revenue generation; patents cover ideas while copyrights cover expressions of ideas; copyrights arise naturally - they exist without the author taking any formal action at all, while patents involve formal processes of application and prosecution via the US Patent and Trademark Office to have a patent granted. These are just a few of the many, many differences that make patent case law wholly inapplicable to copyright law.

We could stop right here on solid ground that Mr. Wallace's arguments are without merit, having been based entirely on patent case law. However, just for the sake of argument, let's analogize anyway, and see if we can follow Mr. Wallace's train of thought.

So let's look at the patent case that Mr.. Wallace relies on - General Talking Pictures Corp v Western Electric Co, Inc., 305 US 124. This case does not discuss in any detail whatsoever the scope or nature of permissible conditions to unilateral license. The case merely cites another case (United States v General Electric Co) to affirm that a patent license may be require, as a condition of the license, certain performance of the licensee provided that such performance reasonably relates to the benefits that the patent holder can expect to receive from his patent. The case cited (United States v General Electric Co) has since been called into question by the Supreme Court and its holdings are no longer good law.

Moving right along -

The other lines of patent cases that Mr. Wallace refers to (aside from being patent cases and therefore not applicable to copyright law) merely reinforce that licenses can be conditioned on the performance of the licensee; still other lines of cases discuss the principle that once a tangible commodity is produced from a patent, the patent holder's ability to control what happens with that tangible commodity is greatly diminished and eventually disappears. The only thread of reason from all of these cases is that a condition imposed on a patent license by the licensor must have some relationship to the rights of the patent holder. For example, if I license my patent to ABC Manufacturing, I can make the license conditional upon ABC using only top-grade materials, conforming to ISO standards, selling only to Australia, etc, but I cannot make the license conditional upon the president of ABC standing on his head for one hour every Tuesday morning at 9:00 am. Why? That condition has no relationship to the benefits I can expect to gain from my status as a patent holder. I suspect this is the principle that Mr. Wallace grossly misconstrued as "the conditions he places must involve only his exclusive rights and not the exclusive [rights] of parties involved..."

If we take this principle and apply it to the GPL, the original licensor is conditioning his license to the world at large on one simple condition - "If you want to copy and make derivative works of my code, you have to license those derivatives under a license like this one...." If Mr. Wallace insists on analogizing to patent law, there is nothing wrong with this condition. It relates directly to the benefits the copyright holder can expect to obtain. A copyright holder benefits by having sole control of his copyrighted work, to do with as he sees fit, including, if he so desires, to donate the copyright to the public domain, a common practice of universities and government entities. The copyright holder in this case is merely conditioning his license to create derivative works on the condition that such derivatives be licensed under similar terms, to prevent the licensee (for example) from copying his code and incorporating it into a proprietary product for which she can then charge money, competing with the original copyright holder, and denigrating the benefits the copyright holder derives from his copyright. This is a direct correlation to the rights of the copyright holder - a perfect example of exactly the type of condition that is appropriate for a licensor to impose on a licensee. This is no different from a patent licensor conditioning manufacture of his surgical device being restricted for sale in the US and/or manufactured only from surgical stainless steel.

Now let's look at the common sense definitional view. Mr. Wallace states that for a licensor to place conditions on a licensee that involve the licensee's "exclusive rights" falls "outside the definition of a unilateral permission." However, nowhere does Mr. Wallace attempt to define a "unilateral permission" nor does he cite any support whatsoever for this statement. So let's take a look. A unilateral permission is a permission granted by one person that a second person can choose to accept or not accept. I realize that "unilateral" signifies one person only, but simply because I give you my unilateral permission to swim in my swimming pool does not mean that you have to or that you will. You can choose to do so or not, and your actions do not make my permission any less "unilateral."

Unlike a contract, a unilateral permission does not require the second person (accepting the permission) to do anything to signify acceptance. In a contract, the second person would have to sign his name on a document, click a web site icon, begin performance, send a check, or any other number of acts that signify acceptance. It is this performance of an act signifying acceptance, and the first person (offeror's) awareness of this act (acceptance) that concludes formation of the contract. What makes a unilateral permission different is that the second person does not need to do anything to accept the permission - he does not need to perform any act signifying acceptance, and the first person (offeror) does not need to ever know if the second person accepted the permission or not. That is the difference.

When a unilateral permission is conditioned upon something, however, the condition must be fulfilled in order for the license to be valid. This means the licensee must do something to fulfill the condition. However, the licensor does not ever need to know about it, and doesn't care. Either the condition is fulfilled and the permission (license) is valid, or the condition is not fulfilled and the permission is not valid.

Now here is the key point and a basic fact of law - Anytime a person does something that they are not legally obligated to do, or voluntarily refrains from doing something that they have a legal right to do, they are acting in a manner that affects their personal and exclusive "rights" - they are giving up one of their "rights." Therefore, anytime a person voluntarily fulfills a condition in order to obtain a unilateral permission (license), they necessarily are performing acts that involve their own personal exclusive rights. After all, they don't have to fulfill the condition - there is no legal obligation to do so - but they choose to give up a right that they have in order to receive the benefit of the unilateral permission that was so conditioned. It matters not one whit that the exclusive rights we are talking about here are copyrights.

So - Mr. Wallace's basic premise - that the GPL is unsound because it requires the licensee as a condition of the unilateral permission of the licensor to do something affecting the licensee's "exclusive rights" is an impossible assertion. He concedes that unilateral permissions can be conditional, but ignores the fact that you cannot fulfill a condition without affecting your rights (copyright or otherwise)(your right to free speech, your right to seek employment, you right to go or not go to the supermarket, your right move to Alaska, etc). We all have innumerable rights and have to exercise at least one of them to fulfill any condition.

Perhaps what is confusing Mr. Wallace is this: It is true that you cannot interfere with the rights of a copyright holder without his/her permission. Mr. Wallace, I believe, views the GPL as interference by one party (the licensor) with the copyrights of another party (the licensee) because the license is conditioned upon the licensee taking certain actions regarding his copyrights. He presents the following alleged conundrum: Either the second party (licensee) is agreeing to this interference, in which case the GPL cannot be a unilateral permission since now we have two parties agreeing with each other (sounds like a contract); or, the second party (licensee) is not agreeing to this interference in which case the GPL is not valid because it is a violation of copyright law. (Mr. Wallace ignores the issue of whether or not a valid contract has been formed under the first scenario.) The answer to Mr. Wallace's purported conundrum is this: The second party is consenting to the interference by fulfilling the condition of the unilateral permission, but merely fulfilling such a condition does not change the nature of the unilateral permission as unilateral.

Re: Mr. Wallace's quotations from IBM's amended counterclaims, IBM's reference to the GPL as a "public agreement" cast in a "binding legal form," has nothing to do with the characterization of the GPL as a unilateral permission. It has been a common and unchallenged legal practice for countless years for copyright holders who wish to donate their copyrights to the public to do so by unilateral permission, generally in the form of a notice contained in the copyrighted work to that effect. Members of the public can "agree" or not "agree" to accept the permission by copying or not copying the proffered work. This is what IBM is talking about when it refers to a "public agreement" cast in a "binding legal form" and, yes, they are talking about copyright permissions. That is the subject, is it not?

Finally, Mr. Wallace's discussion of federal preemption is not applicable. Certainly, federal law preempts state law if there is a conflict. However, there is no conflict here, and the FSF's use of a euphemism "copyleft" does not somehow create a conflict in laws where one does not otherwise exist. The GPL is founded and can only exist on the basis of copyright law and the rights of authors to do what they want with their own work and to condition licenses to others as they see fit. There is no conflict, merely the fear of the uniformed when encountering a new exercise of freedom and rights granted by the same laws that have governed us for almost a century.

As an aside to Mr. Wallace's comments, I believe that the GPL functions as either a unilateral permission or a contract, depending upon how it is being implemented by any given party. So, for example, if a company offers you an online download of Linux and asks you to click an "I agree" button acknowledging that the download of Linux you are about to receive is subject to the GPL (and you have been given an opportunity to read the GPL before clicking), and you click, seems to me you've just entered a binding agreement called the GPL no different from a Microsoft EULA effected in the same way. By contrast, simply placing the GPL in a tarball with the code in my mind falls squarely into the category of "unilateral permission." I see no conflict here - merely a difference in administration, both equally effective.