Forgot your password?
typodupeerror
Programming Operating Systems Government Software BSD IT Linux Politics Technology

Theo de Raadt On Relicensing BSD Code 613

Posted by kdawson
from the read-the-berne-convention dept.
iBSD writes "KernelTrap has an interesting article in which Theo de Raadt discusses the legal implications of the recent relicensing of OpenBSD's BSD-licensed Atheros driver under the GPL. De Raadt says, 'it has been like pulling teeth since (most) Linux wireless guys and the SFLC do not wish to admit fault. I think that the Linux wireless guys should really think hard about this problem, how they look, and the legal risks they place upon the future of their source code bodies.' He stressed that the theory that BSD code can simply be relicensed to the GPL without making significant changes to the code is false, adding, 'in their zeal to get the code under their own license, some of these Linux wireless developers have broken copyright law repeatedly. But to even get to the point where they broke copyright law, they had to bypass a whole series of ethical considerations too.'"
This discussion has been archived. No new comments can be posted.

Theo de Raadt On Relicensing BSD Code

Comments Filter:
  • Still confused (Score:3, Interesting)

    by Kelson (129150) * on Thursday September 13, 2007 @12:08PM (#20589421) Homepage Journal
    For the last decade+, people have been claiming that the BSD licenses are more free than the GPL, because they effectively place no restrictions on what you can do with the code. Now we're being told that there are restrictions on what you can do with the code.
  • Re:Sure, but (Score:4, Interesting)

    by Anonymous Coward on Thursday September 13, 2007 @12:14PM (#20589529)
    The GPLers didn't strip the licence. The code was available under either licence. What gets me is that Theo De Raadt is arguing that both licences apply so they couldnt remove one. He seems to not realise that if both licences applied then they couldn't use the code in FreeBSD anyway because they would be bound by the GPL parts of the licence. Funny that they kicked him off the NetBSD team for being so hard to work with. I wonder why ...
  • by huckamania (533052) on Thursday September 13, 2007 @12:25PM (#20589745) Journal
    And are you a lawyer? Cause it seems like you understanding of the law is that only lawyers can claim understanding.

    Law is understandable. It really is just common sense. That's why the jury is so important. When you think of the law, try to imagine what 10 average people would think, not what some ambulance chasing, paper pushing, pencil necked, money grubbing lawyer would think. Average people want the law to work so they can go on with their lives and not feel bad about their decision, average lawyer wants the law to work for them so they can become a partner and buy a new bmw.

    Doctors, lawyers, scientists, politicians were all held in high esteem by past generations and usually for good reason. Now, it seems like they are all just out to make a buck.
  • Re:Still confused (Score:3, Interesting)

    by fsmunoz (267297) <fsmunoz.member@fsf@org> on Thursday September 13, 2007 @12:29PM (#20589807) Homepage

    In this case, it seems that the GPL zealots have failed to comply even with that requirement in their zeal to rebrand the code as GPL.
    They didn't failed to comply with the BSD licence because they weren't bound by it but by the GPL, since it was a dual-licensed work with a "you can choose which license applies" wording. And this is really the most important issue here, Theo's interpretation of dual licensed software as a kind of "you must comply with both", and this is what warrants further discussion. Saying that "zealots failed to comply" assumes that a regular BSD licensed software was changed, when this is *not* the case.
  • by budword (680846) on Thursday September 13, 2007 @01:22PM (#20590761)
    The guy who wrote the code in question said he intended for the code to be used under either/or license, and he is just fine with there being a gpl only version. Theo is way out of line this time. He is just upset that he won't be getting any changes back but the gpl will be receiving any improvements for as long as anyone writes them.
  • Re:Sure, but (Score:3, Interesting)

    by B'Trey (111263) on Thursday September 13, 2007 @02:17PM (#20591799)
    You are wrong. The copyright holder himself relicensed the work under the GPL while stripping the BSD. He alone is free to do so.

    Huh? If that's indeed the case, then Theo definitely has no legitimate beef. I haven't seen this claim anywhere. In fact, I haven't heard the author of the driver code weigh in at all. Do you have a pointer?

    You are free to continue using code derived from the original BSD licensed works under the terms of the BSD license, but the newly released work, from the original author, both has no dependence on the BSD and grants none of the rights available under the BSD.

    Double huh? You can use newly derived code under whatever license it's released under. The license of the parent code may or may not restrict the choice of license available for the release of the derived code. If the derived code is not released under the BSD license, you're certainly not free to use it under a BSD license.
  • by redelm (54142) on Thursday September 13, 2007 @02:20PM (#20591879) Homepage
    I think I know what Theo is saying (always a question with him): the changes to the Altheros wireless drivers from OpenBSD to Linux are not substantial enough to meet the criterion of creating a derivative work.


    OK, fine. That's one opinion, particularly valid from the PoV of a code-writer. But the PoV of a code-reader/user is very different: either the code crashes or it does not. Any change, even trivial in terms of creative effort, has enormous impact on the value of the work.


    Or put another way, deRaad doesn't think debugging has enough value to be granted the rights of creating a derivative work.


    I might agree adding a comma or changing a word on page XX in a novel does NOT create a derivative work. However, code is very fussy and utilitairian. Unless the code will run _exactly_ as-is, any mod is a derivative work.

  • by Anonymous Coward on Thursday September 13, 2007 @02:41PM (#20592241)

    The guy who wrote the code in question said he intended for the code to be used under either/or license, and he is just fine with there being a gpl only version. Theo is way out of line this time.


    Which guy and which code are we talking about here? Because from what I can tell there are multiple developers whose copyrights had been stripped, and on some files the diffs I've seen only had the BSD license (specifically some Sam Leffler's header files).
  • BTW, I am an avid Linux user, and I think that the Linux kernel developers who made the error made an honest mistake. Let us treat it as such.

    Unfortunately, the more I do my own research, the more worried I am about Theo's main complaint-- that the SFLC may be giving out advice that seems questionable to me.

    While IANAL, I say so based on my own understanding that it is nearly impossible to sue lawyers for malpractice and so we *all* need to develop a basic understanding of the law in areas which are relevant. Here are specific points I would make:

    1) While the BSDL and related licenses clearly do not have the intent to force sharing of code, they clearly *do* have the intent to provide the downstream recipients of the original elements of that code with the rights listed in the license. So Theo is right that you cannot simply wrap the BSDL in the GPL.

    This is particularly relevant to the GPL3 because it introduces potential license incompatibilities between BSDL-code and GPL3 code (see section 7 on removing additional permissions *without* asserting copyright).

    2) Copyright law seems even in the US holds that nonexclusive licenses are clearly indivisible and do not automatically grant sublicense rights (a sublicense being a new license issued by a licensee). Some BSD-like licenses (like the MIT License) explicitly allow sublicensing the code and in this case, wrapping it in the GPL would be allowed. Otherwise, it seems difficult to make this case. Whether exclusive licenses are divisible is not yet a settle matter of law as far as I can tell (you have the Gardner v. Nike case which suggests that they exclusive licenses are indivisible, but that is the only case I can find).

    BTW, Mr Moglen dismisses the above issue without providing any substantive argument against it.

    3) Some BSD-like licenses seem to be addressed to all downstream users and do not include the right to sublicense. The ICU licnese, for example, and the X.Org licenses start out "Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files..." and does not specifically state a sublicensing right.

    Thus I am not sure that the advice that these can be automatically sublicensed under the GPL is advice that is sound.

    For these reasons, I have been suggesting that open source project leaders should seek unbiased legal advice from people outside the community.
  • I think you are confusing BSD-Licensed code with public domain code. In public domain code, nobody has a copyright so there are no terms or conditions to abide by in the license.

    In the BSDL, you have to reproduce the license on the code, include the copyright notice, the disclaimer of warranty, and optionally a couple of other clauses (non-endorsement and/or advertising).

    You can still use the code. You can still release the code.

    The argument is when one may add additional restrictions to the code that is released. In general:

    1) Must one have a valid copyright to enforce on the code? If Theo is right (and I think he is) then BSDL-code is incompatible with the GPL3 at least as other components in the Corresponding Source are concerned (because you can only put the files in the Corresponding Source if they can be *relicensed* under the GPL3 *without* asserting additional copyrights).

    2) What consitutes a valid copyright? In the US one must have substantial original, expressive elements in the work. Reformatting, correcting typos, etc. doesn't count. Adding a large block of code definitely does. In between, it may be a gray area.

    Let me flesh out my reasons for agreeing with Theo on point 1. IANAL.

    1) A sublicense is a new license agreement between the one licensee and a downstream licensee. If you sublicense my code under the GPL, the parties to the agreement are you and the downstream users. You might or might not have statutory relief (IANAL), but you would probably have tort relief. To issue a sublicense, one must either be a copyright owner or have permission from the copyright owner to create such a contract.

    2) Prior to the 1976 Copyright Act, sublicenses were not seen as implied in copyright licenses of any sort. The copyright was seen to be indivisible, and so was any agreement short of a total transfer of copyrights(cf Harris). After the 1976 Copyright Act, non-exclusive licenses were still clearly indivisible (i.e. not sublicensable or transferrable by default), but there is some debate about exclusive licenses (see Gardner v. Nike where the 9th Circuit ruled that exclusive licenses were indivisable and not transferrable without the consent of the original copyright owner). Hence if a license does not have a sublicensing right attached to it, no sublicensing is possible at all from a nonexclusive license, like the BSDL.

    3) Many of the BSD-like licenses are clearly addressed to all downstream recipients of the code (of course this only applies to projects that release the code). This means that you cannot remove rights simply by saying so, which makes it problematic from a GPL3 perspective (reread secton 7 again carefully). Note that this is only problematic for dependencies covered in the corresponding source definition. Copying in BSDL code would still be OK as long as the file as a whole was still a distinct work from the copied code.

    So I don't think that most BSD License variants (exceptions include the MIT License) allow this sort of relicensing.

    In short, with a BSD License, you are not forced to release your code, but if you do, you must pass along the same rights to the original code as were in the original BSD code. Derivative works can be under other licenses, however.

    Once again, IANAL. If this matters to you, I suggest you seek competent and unbiased legal advice from a copyright attourney without an agenda relating to Free Software.
  • by Anonymous Coward on Thursday September 13, 2007 @05:50PM (#20595587)

    No, it isn't. A Linux developer proposed a diff that, had it been accepted, might have been an act of copyright infringement in some ways. The OpenBSD team went overboard in response, acting as if it had been accepted into the kernel, acting as if there were more than there actually were (including the OpenBSD team deliberately ignoring dual licensing clauses and claiming they meant something other than what the authors intended and have publicly stated they intended), and have been rude and abusive.

    It's not the end of the story because Theo and the more hysterical of the OpenBSD group have actually turned it into a whole new story.

    Today on undeadly.com:

    - Celebrity guns down newspaper office after being called "fat cow" by gossip columnist: Newspaper committed libel, end of story!
    - Some guy is murdered by a police death squad after failing to stop at a red light: The guy ran a red light, end of story!
    - Millions dead after Iran nuked after the Iranian premier described Bush as "A silly man": Iran insulted Bush, end of story!

  • by IgnoramusMaximus (692000) on Friday September 14, 2007 @09:41AM (#20602391)

    IANAL but I think that taking parts of the code and relicensing it under the GPL is not colliding with any of these rules.

    Someone just pointed out to me that the BSD license requires its verbatim inclusion in the source distribution, which would then render all source containing the BSD-licanese ... BSD-licensed. So, if the BSD code is included in any other code which must be distributed as source under any other license, say GPL, the licenses would immediately conflict. This problem does not occur with licenses which allow no-source distributions (even DRM encumbered for-profit ones). So in that context the BSD apparently does appear to say "its OK to steal it and sell it for profit, but not to borrow and share it".

    This presents an interesting perspective of the BSD phillosophy.

It seems that more and more mathematicians are using a new, high level language named "research student".

Working...