Re: Re: Open Source Licensing

Bryan Taylor wrote:

>Why do you say the GPL is unclear? It seems pretty straightforward to me: if
>you make a derivitive (and distribute it), then the combined work must be
>licenced by you under the GPL. If there is uncertainty about what constitutes a
>derivite work, that is ambiguity within the Copyright law, not the GPL.
>

Notice in your text "combined work" - that is not a US Copyright concept. In US copyright, you have these concepts (relevant to what I think you mean by a combination): compilation works, collective works, joint authorship, and derivative works. Each of those are defined terms in the lexicon of US copyright. Further to those definitions, there are hundreds of cases that refine, explain (and sometimes confound) the meaning of those defined terms. In contract law (a license is a contract not to sue under stated conditions), if reference is made to a word that is defined by legal principles, such use often incorporates the meaning of that term that is found in those legal sources (this is most often done with words like "misappropriation" in trade secret law because there is a Uniform law on that, and many cases).

When the GPL or any license makes a reference to a defined lexicon and then tries to further define it by use of words like "combination" - it is impossible to me to tell what is meant, and I believe that a court would then permit parol evidence of what the parties intended. Of course, typically the parties never even discuss the GPL, much less get into the intricacy we do here. So, what is a court left to do? It then does its best job to figure out what the parties meant. When it does that, it is going to look at the words used in their context. And when it does that, it probably is going to find that the words "containing the Program" are broader than "derivative work" and so any software company that licensed software that "contained the Program [GPL'd code]" probably is subject to the terms.

Its ambiguous because of the use of terminology that does not have a meaning in the US Copyright law (BTW, GPL is also ambiguous because it just says "copyright law" and does not say its US copyright law - and the copyright laws of other countries are different - so another ambiguity is what law is being referenced). I did a memo for a client (I practice almost exclusively software licensing) a year or so ago on how bad the LGPL was, I think I counted something like 12 fundamental contractual errors or ambiguities in it. I do not claim to write perfect contracts or licenses, but I catch the big ones, and this is a big one.

>Should you choose to make such a derivitive work, you (of course) still own the
>copyright on your original contributions, and could sell separate licences for
>it (standing alone) if you so chose.
>

That is not true (under US law) if the original copyright holder did not authorize you to make the derivative. The GPL of course, authorizes the derivative, but places conditions on the authorization.

My question has always been: are the conditions placed on the author of the derivative work a misuse of the copyright? Here in the 4th Circuit, the copyright concept of misuse flourishes (one of the very few cases actually holding a copyright misuse for software is in the 4th Circuit). Misuse, however, can be purged (by deleting the offending provisions). I think there is still a decent question whether the extra-copyright based restrictions on the derivative work of a GPL's code are enforceable. I say this because by publishing all of the code, there are no trade secrets - the traditional basis on which extra-copyright based restrictions can be upheld. The restriction in the GPL I always had serious reservations about was the clause that required licensing of patents.

>As for SCO, they seem to be on crack. Their GPL defense is just comical. I hope
>the judge sanctions them if they bring it in court.
>

I wish there was more information on the SCO case. I think its fascinating, and of course, has serious ramifications for linux and the open source community if SCO prevails. So you know, judges rarely issues "sanctions" these days in hotly contested IP cases (Rule 11 has been severely watered down from the days when I did litigation). Indeed, even the discretionary right to award attorneys' fees is either not used, or if such discretion is used, often the award is overturned on appeal. You would almost have to prove SCO knew (as in actual subjective knowledge) that it had no case - very unlikely.

The one place that they might have a problem is this: they claim the linux kernel contains SCO code. If it does, that code cannot possibly be a trade secret, because its published. Even if the code was misappropriated, the genie is out of the bottle and it would be impossible for a court to reconstitute the trade secret. So, SCO should be disclosing the precise code that is claimed to have been copied, so users and resellers can determine if they are infringing. For SCO to refuse to publicly disclose that code makes them look stupid, and also, like they are asserting rights they do not have, or that are overstated. If that ultimately turns out to be the case, then I see where sanctions are possible. Who knows.

>I think it likely that SCO will yield greater damages to the IBM counterclaims
>than the value of their company. IBM niclely hit every single SCO product with
>a patent infringement claim. Ouch.
>

IBM is no spring chicken. There is doctrine in antitrust law known as in pari delicto - where a court can basically say: look, both of you have been bad (in legalese, have unclean hands), so I am not enforcing either of your rights against each other. This defense is rarely applied (heck, real antitrust cases are rarely brought), but from only the name calling and public press releases I have seen, it might be that a court leaves these parties where they were when they started . . .

>They seem to be employing the "Chewbacca Defense":
>http://www.connect-dots.com/Poofs/chewbacca.html
>

That is a funny cartoon. Sadly, I have seen something close to that in real life.

best regards, mike oliver
Bowie & Jensen, LLC Received on Thu Aug 21 2003 - 22:00:29 GMT