James V. Delong on GPL3

by on March 29, 2007 · 26 comments

I thought this was interesting and with permission I quote in its entirety from ipcentral:

Having examined the latest draft of the Free Software Foundation’s General Public License version 3 (GPLv3) several times, and having looked over the Rationale document, I have come to a diagnosis.

If GPLv3 were a human being, one would say that it has delusions of grandeur. It thinks it is a law rather than a license.

Legally speaking, GPLv3 is a license, which is a form of contract. It specifies the terms on which the holder of copyrights or patents on software will permit others to make use of it. It is a bit of a special case because it is open to the world at large; anyone may use it, without payment, as long as they abide by its terms, which is unusual in contract law. However, there are doctrines of promissory estoppel and third party beneficiaries that take account of such things, and GPLv3 is firmly within the legal genre of contract.

But the GPLv3 was apparently drafted on the assumption that it is something quite different — that it is a regulation controlling a range of general behavior by software users, and that it is being promulgated by a governmental body with law-creating power.

The difference between a contract and a regulation is extremely important.

The most basic principle of drafting a contract is to strive for clarity. All parties involved should know clearly their rights and duties. Failure to achieve clarity raises transaction costs of all kinds as interested parties squabble and struggle to decipher the inscrutable and cope with unforeseen risks, and it often ends in bitterness and litigation.

Tricky lawyers have known for centuries that ambiguity in a contract can be useful. Sneak some murky language in and then insist that it really means something to your advantage that the other party did not intend to agree to.

And judges have known for centuries that there are tricky lawyers. So a tenet of contract law is that ambiguities are construed against the party that drafted the document, especially in the case of form contracts, such as software licenses, which the buyer (or licensee) must take-or-leave without negotiation.

Laws or regulations are subject to different interpretive rules. Legal doctrines of administrative law give government officials great latitude in interpreting ambiguous laws and regulations, never mind that they may be the chief architects of the confusion.

In consequence, government drafters are often motivated to seek ambiguity rather than clarity.

A simple reason is that various parties with a veto over the product may not agree on what it should say. To get something out the door, the drafters elide the issues, kicking the can to later parts of the process.

A more Machiavellian process may also be at work. Courts give considerable deference to government interpretations or rules, so substantial rewards of power await the official who induces Congress to draft a confusing law, or his fellow agency staffers a muddy a regulation. He/she can then issue “guidance” or “interpretive rules” or even a legal complaint in an enforcement action in which some theretofore hidden standard is unveiled. (See, e.g., Out of Bounds, Out of Control: Regulatory Enforcement at the EPA (Cato Institute 2002).)

GPLv3 bears the marks of such a process of deliberately crafted ambiguity. designed to enable future surprises, and not pleasant ones for the regulatees. Virtually every important provision is subject to multiple interpretations.

For example, with respect to the important question whether Digital Rights Management systems could be used in conjunction with programs released under GPLv3, the language of the license is murky and numerous ancillary statements by the drafters express strong hostility to DRM. On the other hand, one of the key figures in the debate recently published a note saying that everyone was misunderstanding the situation and that DRM is indeed compatible with GPLv3.

So why has the issue not been made clear? The debate has been going on for two years. Any sensible corporate GC will say that he has quite enough to do in dealing with real government regs without adding the FSF to his roster of inscrutable oracles.

Why the FSF should have a Machiavellian motivation is easy to explain. As many have noted, the Free Software Foundation is hostile to all forms of property in software and other creations of the intellect, and eager to use the GPLv3 as an instrument of this antipathy.

Confusion substantially raises the transaction costs of users who want to combine proprietary and open programs, and the FSF can selectively enforce of the terms of the license to promote or hobble perceived friends and foes. IBM, which pours money into the open source movement, can be given indulgence; Microsoft better keep its lawyers working overtime.

This “FSF is a government equivalent” mentality was at work with respect to GPLv2, which is bedeviled by the definitional problems about what constitutes a “derivative work.” In explaining it, FSF has made references to FAQs and other extraneous materials, similar to a regulatory agency invoking legislative history. The approach as been radically expanded in the GPLv3, which was originally billed as an effort at clarification.

This strategy of seeking ambiguity has a problem, however. Whatever grandiose fantasies the FSF has about itself as a government, possessed of authority to draft regulations, it is not. FSF is a private body that has drafted a form contract. It does not get to into court and enforce new meanings. Instead, on murky points, a judge will decide what it means, and the FSF should lose.

And once the ambiguities are stripped out, there is not going to be much left of the GPLv3.

=====================
Additional credentials: Long before his IP career, James DeLong was the Research Director of the Administrative Conference of the United States, which was dedicated to studying issues of administrative law. Before that, he helped draft and issue regulations for the Federal Trade Commission. In fact, he once co-wrote a regulation that caught the eye of a Washington Star column called “Gobbledygook,” which featured particularly striking examples of this bureaucratic art form. Unfortunately, we were not, that time, trying for obscurity

  • http://bennett.com/blog Richard Bennett

    If GPLv3 were a human being, one would say that it has delusions of grandeur. It thinks it is a law rather than a license.

    Yup.

  • http://bennett.com/blog Richard Bennett

    If GPLv3 were a human being, one would say that it has delusions of grandeur. It thinks it is a law rather than a license.

    Yup.

  • http://enigmafoundry.wordpress.com enigma_foundry

    Hmmm…The right wing is all for voluntary contracts, until someone comes up with a voluntary contract that they don’t like.

    Well, Solveig and James, I suggest you not release any of the software you care to write under the GPL 3.0.

    So the very big difference between GPL and a government is the GPL is a VOLUNTARY contract.

    If you don’t like it, don’t use it, and don’t use any GPL code.

    It really is THAT SIMPLE.

  • http://enigmafoundry.wordpress.com eee_eff

    Hmmm…The right wing is all for voluntary contracts, until someone comes up with a voluntary contract that they don’t like.

    Well, Solveig and James, I suggest you not release any of the software you care to write under the GPL 3.0.

    So the very big difference between GPL and a government is the GPL is a VOLUNTARY contract.

    If you don’t like it, don’t use it, and don’t use any GPL code.

    It really is THAT SIMPLE.

  • http://bennett.com/blog Richard Bennett

    You seem a little testy, enigma. Does it bother you that Linus Torvalds says he’s totally not into GPL 3?

  • http://mcgath.blogspot.com Gary McGath

    The implied words in enigma_foundry’s statement are, “If you don’t like it, don’t use it, and don’t use any GPL code — AND KEEP YOUR MOUTH SHUT.” But the fact that something is a voluntary contract, and possibly a bad one, is not a reason to exclude it from public discussion and criticism.

  • http://bennett.com/blog Richard Bennett

    You seem a little testy, enigma. Does it bother you that Linus Torvalds says he’s totally not into GPL 3?

  • http://mcgath.blogspot.com Gary McGath

    The implied words in enigma_foundry’s statement are, “If you don’t like it, don’t use it, and don’t use any GPL code — AND KEEP YOUR MOUTH SHUT.” But the fact that something is a voluntary contract, and possibly a bad one, is not a reason to exclude it from public discussion and criticism.

  • Doug Lay

    According to CNet, Torvalds is “pleased” with the latest revisions to GPL 3.0, although still noncommittal about migrating the Linux kernel to the new license. Apparently the anti-DRM language has been greatly toned down in the latest revision, which seems prudent. Stallman and the FSF will compromise when they feel it’s in their long-term interest (LGPL, anyone)?

    I don’t think Linus has spoken out against the anti-software-patent language in GPL 3, but I could be wrong. It may be that the kernel developers are willing to accept language that targets “side” IP covenants like the MS-Novell deal.

  • http://www.codemonkeyramblings.com MikeT

    enigma_factory,

    That’s why some of us generally avoid GPL software when there is something better. Have you not noticed that most good OSS is in fact dual-licensed or licensed under a BSD or MIT license? Apache’s projects, OpenOffice, Mozilla’s projects, need I go on?

  • Jeff Macdonald

    Richard, you haven’t seen the latest statement from Linus regarding the lastest draft:

    “Linus Torvalds is ‘pretty pleased’ with the current GPL v3 draft”

    http://linux.slashdot.org/article.pl?sid=07/03/29/1331237

    And Gary, those implied words are yours, and only yours unless enigma_foundy says otherwise.

  • Jeff Macdonald

    MikeT, none of those products use BSD or MIT license. They all have their own license. OpenOffice actually uses the LGPL. I know of 2 products that use the GPL but offer dual licenses, MySQL and SleepyCat DB (Berkely DB). They offer dual licenses for a price. GPL is free, another license costs $$.

    I’m curious, what is your defination of ‘something better’? Is it something along the lines of ‘I can claim the code as my own’?

  • Doug Lay

    According to CNet, Torvalds is “pleased” with the latest revisions to GPL 3.0, although still noncommittal about migrating the Linux kernel to the new license. Apparently the anti-DRM language has been greatly toned down in the latest revision, which seems prudent. Stallman and the FSF will compromise when they feel it’s in their long-term interest (LGPL, anyone)?

    I don’t think Linus has spoken out against the anti-software-patent language in GPL 3, but I could be wrong. It may be that the kernel developers are willing to accept language that targets “side” IP covenants like the MS-Novell deal.

  • http://www.codemonkeyramblings.com MikeT

    enigma_factory,

    That’s why some of us generally avoid GPL software when there is something better. Have you not noticed that most good OSS is in fact dual-licensed or licensed under a BSD or MIT license? Apache’s projects, OpenOffice, Mozilla’s projects, need I go on?

  • Jeff Macdonald

    Richard, you haven’t seen the latest statement from Linus regarding the lastest draft:

    “Linus Torvalds is ‘pretty pleased’ with the current GPL v3 draft”

    http://linux.slashdot.org/article.pl?sid=07/03/

    And Gary, those implied words are yours, and only yours unless enigma_foundy says otherwise.

  • Jeff Macdonald

    MikeT, none of those products use BSD or MIT license. They all have their own license. OpenOffice actually uses the LGPL. I know of 2 products that use the GPL but offer dual licenses, MySQL and SleepyCat DB (Berkely DB). They offer dual licenses for a price. GPL is free, another license costs $$.

    I’m curious, what is your defination of ‘something better’? Is it something along the lines of ‘I can claim the code as my own’?

  • http://enigmafoundry.wordpress.com/ enigma_foundry

    Well, the reason I commented on this is that the original post said:

    “But the GPLv3 was apparently drafted on the assumption that it is something quite different —that it is a regulation controlling a range of general behavior by software users, and that it is being promulgated by a governmental body with law-creating power.”

    And my post reminded all that the GPL 3.0 is in fact a voluntary contract, because James seems to want very much to paint the FSF in an “anti-freedom” light, which is basically a smear.

    Freedom is what the GPL is all about.

    “You seem a little testy, enigma. Does it bother you that Linus Torvalds says he’s totally not into GPL 3?”

    Actually, as I had predicted here and elsewhere, the Novell deal and the inclusive revision process will bring more and more developers into GPL 3.0 camp. Some of the messiness that we see is just a community building consensus.

    As for being testy, it seems to me that those who would detract from the GPL are in fact quite testy. They really abandon their supposedly freedom-based orientation as soon as they are confronted with something unexpected.

    But the possibillity of unexpected emergent phenomena is exactly what makes freedom a good, and efforts to bound freedom, in the absence of a compelling rationale, wrong.

  • http://enigmafoundry.wordpress.com eee_eff

    Well, the reason I commented on this is that the original post said:

    “But the GPLv3 was apparently drafted on the assumption that it is something quite different —that it is a regulation controlling a range of general behavior by software users, and that it is being promulgated by a governmental body with law-creating power.”

    And my post reminded all that the GPL 3.0 is in fact a voluntary contract, because James seems to want very much to paint the FSF in an “anti-freedom” light, which is basically a smear.

    Freedom is what the GPL is all about.

    “You seem a little testy, enigma. Does it bother you that Linus Torvalds says he’s totally not into GPL 3?”

    Actually, as I had predicted here and elsewhere, the Novell deal and the inclusive revision process will bring more and more developers into GPL 3.0 camp. Some of the messiness that we see is just a community building consensus.

    As for being testy, it seems to me that those who would detract from the GPL are in fact quite testy. They really abandon their supposedly freedom-based orientation as soon as they are confronted with something unexpected.

    But the possibillity of unexpected emergent phenomena is exactly what makes freedom a good, and efforts to bound freedom, in the absence of a compelling rationale, wrong.

  • http://bennett.com/blog Richard Bennett

    Linus Torvalds doesn’t support GPL3.

    I see the GPLers are trying to twist Torvalds’ statements on GPL out of context. He said he’s “pleased” that the latest draft of GPL3 is “better than the first draft”, but he’s made it clear that it’s still not good enough for him to submit to.

    The money guys who actually fund Linux development – IBM and HP – aren’t on board either, so GPL3 is going nowhere.

  • http://bennett.com/blog Richard Bennett

    Linus Torvalds doesn’t support GPL3.

    I see the GPLers are trying to twist Torvalds’ statements on GPL out of context. He said he’s “pleased” that the latest draft of GPL3 is “better than the first draft”, but he’s made it clear that it’s still not good enough for him to submit to.

    The money guys who actually fund Linux development – IBM and HP – aren’t on board either, so GPL3 is going nowhere.

  • http://weblog.ipcentral.info/ Noel Le

    If GPL3 goes nowhere, what will happy to the great proletarian revolution?

  • http://weblog.ipcentral.info/ Noel Le

    If GPL3 goes nowhere, what will happy to the great proletarian revolution?

  • http://www.codemonkeyramblings.com MikeT

    Sorry, I meant to imply that they are licensed under licenses similar to the MIT or BSD licenses, no the GPL. By better I mean just that, better. My views on OSS are more anarchistic than Stallman’s. The plugins and scripts I have written for WordPress, for example, are released with IP anarchy in mind. I honestly don’t care if someone were to take them, modify them and sell them. I find it highly pretentious to think that using a few lines of code from an other OSS project should infect your entire code base. But, as I have said, I define liberty in terms of a vacuum of power and authority, not in policy, the way that many libertarians do. Liberty to me is best defined as a lack of being told what to do.

  • http://www.codemonkeyramblings.com MikeT

    Sorry, I meant to imply that they are licensed under licenses similar to the MIT or BSD licenses, no the GPL. By better I mean just that, better. My views on OSS are more anarchistic than Stallman’s. The plugins and scripts I have written for WordPress, for example, are released with IP anarchy in mind. I honestly don’t care if someone were to take them, modify them and sell them. I find it highly pretentious to think that using a few lines of code from an other OSS project should infect your entire code base. But, as I have said, I define liberty in terms of a vacuum of power and authority, not in policy, the way that many libertarians do. Liberty to me is best defined as a lack of being told what to do.

  • Stephen Pollei

    MikeT, I think you mean that you have Ochlocracy rather than Anarchist views. An anarchist still has rules.

  • Stephen Pollei

    MikeT, I think you mean that you have Ochlocracy rather than Anarchist views. An anarchist still has rules.

Previous post:

Next post: