Advertisement
Interested in hardware, diags, validation, Linux, C, ARM, Microcode and low level programming and blazing networks?
| |||||||||||||
|
| |||||||||||||
Another counterexample?Another counterexample?Posted Jun 23, 2008 22:32 UTC (Mon) by man_ls (subscriber, #15091)In reply to: A position statement on closed-source kernel modules by lrosen Parent article: A position statement on closed-source kernel modules So all I need to reuse your GPL software is to make it into a library (still under the GPL) and link to it. I can use as much of its functionality as I want, engrain it as deeply as I like, but as long as I just link to it I should be safe. For example I may take your painfully created GPL office software, use it as a huge GPL library, stick my proprietary user interface on it -- and the result is not a derivative work of the original package. I can have my cake and eat it. Is that so? It follows that the business models of QT and MySQL, to give two known examples, are utterly flawed. They distribute their software under the GPL, so anyone linking to it (even in utterly proprietary software) ought to be safe. Shouldn't they be worried?
(Log in to post comments)
Another counterexample? Posted Jun 24, 2008 0:23 UTC (Tue) by lrosen (guest, #31972) [Link] If their business model is based upon misinformation about the reach of the GPL, then they ought to worry. But I have it on good authority that neither QT or MySQL are dependent upon that view of GPL.
Another counterexample? Posted Jun 24, 2008 4:05 UTC (Tue) by rahulsundaram (subscriber, #21946) [Link] I would be interested to hear about their view of GPL then.
Another counterexample? Posted Jun 24, 2008 5:34 UTC (Tue) by frazier (guest, #3060) [Link] Both MySQL (now part of Sun) and QT (unless Nokia has changed this) can be proprietary licensed from their respective companies: http://trolltech.com/products/qt/orderform http://www.mysql.com/about/legal/licensing/ How this is possible is that, unlike the Linux kernel which has a lots of copyright holders, or one of many free software projects that has the copyright assigned to the FSF, the copyright is owned and controlled by a single commercial entity. This allows them to offer two (or more) sets of terms that run parallel. In the case of QT and MySQL, this allows them to offer non-GPL compliant licenses for a fee.
Another counterexample? Posted Jun 24, 2008 7:35 UTC (Tue) by michaeljt (subscriber, #39183) [Link] Now you have got me interested - are you really saying (as an informal, non-legal statement of course) that I could legally distribute a closed-source binary that links against the GPL version of Qt?
Another counterexample? Posted Jun 24, 2008 15:43 UTC (Tue) by tetromino (subscriber, #33846) [Link] Programmer A writes a Gtk+ emulation layer on top of Qt, and releases his result under the GPL. Programmer B writes a closed-source Gtk+ program, and links it to A's wrapper. A and B don't talk to each other. I am not a lawyer, but I am guessing that if you put A's library and B's program on one CD and sell it, you are not violating the GPL (since aggregation is not a derivative work, see article 2 of GPL-2).
Another counterexample? Posted Jun 25, 2008 10:23 UTC (Wed) by dwmw2 (subscriber, #2063) [Link] I am not a lawyer, but I am guessing that if you put A's library and B's program on one CD and sell it, you are not violating the GPL (since aggregation is not a derivative work, see article 2 of GPL-2).Putting two unrelated programs on a CD probably would count as "mere aggregation on a volume of a storage or distribution medium". But your reasoning in parentheses is a little suspect... ยง2 of the the GPL (v2) explicitly states that it applies to collective works — and collective works are aggregation, by definition.
The GPL goes on to make an exception for "mere aggregation on a volume of a storage or distribution medium". That's not a particularly specific definition, and people could argue about precisely what it means for ever — although that argument would be fairly pointless because there is no "right" answer until/unless a court has ruled on it. In your particular jurisdiction. However, what we can manage for ourselves, as laypeople, is to eliminate the interpretations which are meaningless and contradictory. For example, if that "mere aggregation on a volume of a storage or distribution medium" paragraph actually excuses all forms of aggregation, which includes all collective works, then the two preceding paragraphs of the GPL — the ones which talk about the GPL applying to sections of a collective work which, if distributed separately, would be considered independent and separate works in themselves — would be a particularly verbose no-op. I don't particularly want to get into a long discussion about precisely what would count as "mere aggregation on a volume of a storage or distribution medium", but I think it's a fairly safe bet to assume that a court isn't going to rule that the GPL is self-contradictory and it actually means to except all forms of aggregation from its terms.
Another counterexample? Posted Jun 24, 2008 21:11 UTC (Tue) by man_ls (subscriber, #15091) [Link] Care to elaborate? Just an authority based on your authority (a second-hand authority if you will) is not a good argument on itself.Of course linking is not the concern of copyright, just as compiling isn't; as tialaramex explains below, it is writing code and distributing it (or a derived work such as the compiled binary) that can be doubtful. But is it the joint distribution of the linked code that makes software a derived work, or is it using internal interfaces, or what specifically? The MySQL and Qt examples would point to joint distribution, since those writing proprietary packages based on them usually rely on published interfaces (such as SQL); in which case the "received wisdom" on the Linux kernel about published vs internal interfaces and GPL-only symbols does not hold. But of course this is just hearsay; a qualified opinion would be most interesting.
Another counterexample? Posted Jun 27, 2008 16:12 UTC (Fri) by giraffedata (subscriber, #1954) [Link] I had the same thought that Larry missed the point when he implied people are saying linking is a copyright infringement. It is, instead, the preparation of and distribution of a derivative work that is thought by some to be a copyright infringement. But the derivative work argument I have seen does not talk about distributing a linked combination. The derivative work is the Nvidia driver all by itself. That would mean distributing the driver alone, or even creating it in the first place is controlled by copyright law (so requires the permission of the authors of Linux). Though this seems strange, the analog that is apparently well accepted in classic copyright is writing of a sequel to someone else's book. If you use the same characters, settings, etc., you need the permission of the author of the original. Extending that to loadable kernel modules, the argument I've seen considers it significant that the module's entire purpose is to extend that one work -- just like a book sequel.
|
|
||||||||||||