Hi Til: Thanks very much for this thoughtful analysis. Are you planning to post to the ftf list as well?
G. On Sep 10, 2013, at 10:24 AM, Till Jaeger <jae...@jbb.de> wrote: > Dear list, > > Bradley and Larry have asked me to share my view as a European lawyer on the > question if linking of software components (necessarily) results in a > "derivative work" as understood by the GPL. In a nutshell, my thoughts are > the following (a more comprehensive overview can be found at > http://www.ifross.org/Druckfassung/Ziffer%202.pdf, unfortunately in German > only): > > 1. > As far as I know there is no relevant case law on the question of what may > be considered a "derivative work" under European copyright law for software. > > European software copyright law has been harmonized > (http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2009:111:0016:01:EN:HTML) > since 1991. > > In my opinion "derivative work" in software law should have a different > meaning than in other fields of copyright law. > > Software is typically interacting with other software, and dependencies > (e.g. an application running on an operating system) do not necessarily mean > that two components form a derivative work. > > 2. > GPLv3 refers to copyright law ('To “modify” a work means to copy from or > adapt all or part of the work in a fashion requiring copyright permission, > other than the making of an exact copy') whereas GPLv2 might be interpreted > in a way that the understanding of "derivative work" is broader. In this > regard the GPLv2 seems to be a bit contradictory to me. On the one hand it > defines 'a "work based on the Program"'as “either the Program or any > derivative work under copyright law", on the other hand sec. 2 contains a > more detailed explanation of what the term "derivative work" is supposed to > mean within the scope of the GPLv2 ("If identifiable sections of that work > are not derived from the Program, and can be reasonably considered > independent and separate works in themselves, then this License, and its > terms, do not apply to those sections when you distribute them as separate > works."). Apparently, a computer program which is _not_ derived from GPL > code has nonetheless to be licensed under the GPLv2 when the original GPL > code and the program are not distributed "as separate works". > > If you do not want to ignore that language you have to find a meaningful > interpretation for this sentence in sec. 2 of the GPLv2. To me, it makes > sense to understand "distribute them as separate work" as a formal > criterion, i.e. distributing one binary blob makes it "one work" instead of > two or more "separate works". Of course, other interpretations are possible. > > 3. > I think it is very difficult to predict how the European Court of Justice > (ECJ) would interpret the phrase "adaptation, arrangement and any other > alteration of a computer program" as used in Article 4.1 (b) of the > Directive 2009/24/EC. > > The only hint you may find is Article 6 which says that decompilation is > allowed under certain circumstances to "achieve the interoperability of an > independently created computer program with other programs". There is a > definition of interoperability in recital 10: 'The parts of the program > which provide for such interconnection and interaction between elements of > software and hardware are generally known as "interfaces". This functional > interconnection and interaction is generally known as "interoperability"; > such interoperability can be defined as the ability to exchange information > and mutually to use the information which has been exchanged. ' > > Therefore, my understanding of the directive is that software, that is > independently created and exchanges information with other software through > an interface, is independent software and not a derivative work. > > However, it is unclear which kinds of interfaces fall within the scope of > the directive. The text is from 1991 when Java and other object oriented > programming was not known at that time (or not as common as it is today). > > 4. > If linked software should be considered a derivative work (under the > GPLv2 and GPLv3) is truly difficult to judge. With regard to the > aforementioned criteria I come to the following conclusions: > > a) > From the perspective of copyright law the way how two parts of a program > interact _technically_ with each other may provide an indication about the > derivative work question. However, the technical fact by itself that two > components are linked with each other does not necessarily lead to the > conclusion that the combination is or is not a derivative work. > > b) > If a developer modifies an existing program and puts the added code in a > library instead of the existing files the code in the library would still be > a derivative work. A modified program is a modified program, and one might > not circumvent this legal effect just by moving code into a library. > > However, the situation might be different if an independently created > application uses an existing standard library. You could argue > that the application uses the interface of the library, and linking is just > a matter of interoperability, which seems convincing to me. But you might > also consider that there is a widely accepted opinion that linking results > regularly in creating a derivative work under the GPL, and accordingly a > customary business practice has been brought into existence. > > c) > The situation might be different in the case of statically linked libraries. > If you agree with the interpretation of the GPLv2 I proposed above, the > program and the statically linked library are not distributed "as separate > works" and therefore the copyleft applies. > > 5. > The situation is far from being clear. I do not claim to have the > "right" answer to the question on hand. But I think that we need more > exchange between lawyers and software engineers for developing our view on > the issue of linking, an issue courts will have to deal with one day. > > Best regards, > > Till > > -- > Dr. Till Jaeger > Certified Copyright and Media Law Attorney > > > JBB Rechtsanwälte > Jaschinski Biere Brexl Partnerschaft > Christinenstraße 18/19 | 10119 Berlin > Tel. +49.30.443 765 0 | Fax +49.30.443 765 22 > Sitz der Gesellschaft: Berlin | Registergericht AG Charlottenburg | PR 609 B > www.jbb.de > > > > > Am 29.08.2013 22:33, schrieb Bradley M. Kuhn: >> Larry, I will be more direct since you aren't getting my subtle hints. If >> you think I've misquoted Till and am somehow damaging his professional >> reputation, then just say that, simply, to Till, and give him the source, and >> I'm sure Till will take the matter up with me if he agrees with you. >> >> Larry, I think what you're actually doing is wasting Till's valuable time. >> >> Till, just in case you do want to see it, without having to search around, >> the >> email where I referenced you is at: >> http://projects.opensource.org/pipermail/license-discuss/2013-August/001181.html >> >> The specific text that I wrote that mentions you is: >>>> BTW, if you are interested in how the European lawyers view this question, >>>> I refer you to an excellent talk by Till Jaeger at FOSDEM 2013: >>>> http://www.faif.us/cast/2013/mar/26/0x39/ >> >> As you can see, Larry, I didn't, as you claim, represent that Till supported >> any of my positions. >> >> Till, thanks again for giving that excellent talk on our track at FOSDEM >> 2013! >> I'm truly sorry that you've been dragged in to this conversation, and I had >> no idea that sharing the audio of your useful talk with others would cause >> these sorts of unsolicited emails from Larry. >> >> -- bkuhn >> . >> > _______________________________________________ > License-discuss mailing list > License-discuss@opensource.org > http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss _______________________________________________ License-discuss mailing list License-discuss@opensource.org http://projects.opensource.org/cgi-bin/mailman/listinfo/license-discuss