"France lends support to new open-source license CeCILL license, compatible with GNU GPL, aims to be better suited for French laws" http://www.infoworld.com/article/04/07/09/HNfrancelendssupport_1.html
What does FSFE-supporters think about this?
Creative Commons has a strong-copyleft license (Share-Alike) which they are translating and adapting to local laws.
Why doesn't FSFEurope adopt a same strategy? Why would European government or public institutions release their own software under a GPL that is only partially valid in many european countries?
Wouter Vanden Hove www.vrijschrift.org www.opencursus.org
Brazil had some legal issues regarding to GPL, mainly because it says for example that the distributor is not responsible for the software (like "provided as is"), but in brazilian law there are some severe laws to protect consumers that does not allow this. Another problem is that only foreign licenses with translations made by official law translators are valid, and FSF was not accepting other versions of the GPL than the english version. that's wy we created the CC-GPL http://creativecommons.org/license/cc-gpl?lang=en
[]s, gandhi
Wouter Vanden hove escreveu:
"France lends support to new open-source license CeCILL license, compatible with GNU GPL, aims to be better suited for French laws" http://www.infoworld.com/article/04/07/09/HNfrancelendssupport_1.html
What does FSFE-supporters think about this?
Creative Commons has a strong-copyleft license (Share-Alike) which they are translating and adapting to local laws.
Why doesn't FSFEurope adopt a same strategy? Why would European government or public institutions release their own software under a GPL that is only partially valid in many european countries?
Wouter Vanden Hove www.vrijschrift.org www.opencursus.org
Discussion mailing list Discussion@fsfeurope.org https://mail.fsfeurope.org/mailman/listinfo/discussion
On Mon, 12 Jul 2004, Ricardo Andere de Mello wrote:
Brazil had some legal issues regarding to GPL, mainly because it says for example that the distributor is not responsible for the software (like "provided as is"), but in brazilian law there are some severe laws to protect consumers that does not allow this.
In the GNU General Public License (11. and 12.), there is a specific note for that : TO THE EXTENT PERMITTED BY APPLICABLE LAW (11.) and UNLESS REQUIRED BY APPLICABLE LAW (12.).
I don't think this was a real issue.
Another problem is that only foreign licenses with translations made by official law translators are valid, and FSF was not accepting other versions of the GPL than the english version. that's wy we created the CC-GPL http://creativecommons.org/license/cc-gpl?lang=en
Are you sure that the translation is expressing correctly the GNU General Public License ? If the case is outside the Brazil (for example the major author is outside Brazil), which version will they use ?
The translation starts like a good objective but this is a major issue because of adding a layer complexity. Compatibility between the translated version and the case law that will sometimes uses one of the translation or the original version. This is a real mess.
Having a single legal text *is really good point*.
Regarding the CCs, the GNU General Public License is better for software (and functionnal) author works. For example, the NoDeriv or NC option is clearly non-free for software and computer program (doesn't respect the 4 freedoms). Attribution and Share-Alike is respected by the GNU General Public License.
Just my 0.02 EUR,
adulau
On Mon, 2004-07-12 at 09:24 +0200, Alexandre Dulaunoy wrote:
On Mon, 12 Jul 2004, Ricardo Andere de Mello wrote:
Brazil had some legal issues regarding to GPL, mainly because it says for example that the distributor is not responsible for the software (like "provided as is"), but in brazilian law there are some severe laws to protect consumers that does not allow this.
In the GNU General Public License (11. and 12.), there is a specific note for that : TO THE EXTENT PERMITTED BY APPLICABLE LAW (11.) and UNLESS REQUIRED BY APPLICABLE LAW (12.).
I don't think this was a real issue.
Besides that, it's a literary work. Does Brazil have legal issues with books not providing responsability towards damage (for instance, one could gain vision problems reading certain books)?
Rui
Rui Miguel Seabra wrote:
On Mon, 2004-07-12 at 09:24 +0200, Alexandre Dulaunoy wrote:
On Mon, 12 Jul 2004, Ricardo Andere de Mello wrote:
Brazil had some legal issues regarding to GPL, mainly because it says for example that the distributor is not responsible for the software (like "provided as is"), but in brazilian law there are some severe laws to protect consumers that does not allow this.
In the GNU General Public License (11. and 12.), there is a specific note for that : TO THE EXTENT PERMITTED BY APPLICABLE LAW (11.) and UNLESS REQUIRED BY APPLICABLE LAW (12.).
I don't think this was a real issue.
Besides that, it's a literary work. Does Brazil have legal issues with books not providing responsability towards damage (for instance, one could gain vision problems reading certain books)?
No, the consumers law protect against bad products, poor delivery, etc... You can easily have your money back or another product. In brazilian traditional law the person must proove that is innocent, in consumers law, the company has to proove that is not guilty.
[]s, gandhi
On Mon, 2004-07-12 at 12:35 -0300, gandhi_quilombo wrote:
No, the consumers law protect against bad products, poor delivery, etc... You can easily have your money back or another product. In brazilian traditional law the person must proove that is innocent, in consumers law, the company has to proove that is not guilty.
I have no doubt about that.
However... how is a book a bad product? Is a book refundable?
I find that extremely hard to believe, and since software is ruled by copyright law... I find that that can't be demanded from software anyways.
You can, however, demand assurance from a supplier...
Rui
Rui Miguel Seabra wrote:
I have no doubt about that.
However... how is a book a bad product? Is a book refundable?
Every book has a lot of marketing on it, saying that the book has the latest information about that subject, that is the best of the area, etc... If the consumer "feel" he was deceived by the vendor, he can have a refund. Actually, if you say that you bought in an impulse, that it was not a planned cost, you can have the money back. By law, you have one week to gave up from something you bought (since it is in good state), and three months of guarantee for most products. (of course the company can give more time for guarantee, this is the minimum).
[]s, gandhi
By law, you have one week to gave up from something you bought (since it is in good state), and three months of guarantee for most products.
Interesting. Could you please check what guarantee offer the proprietary licenses in your country? It would be a good subject for comparison.
In Italy the minimum guarantee is 2 years, I think. Still, software has no guarantee. Well, the _media_ is guaranteed, if you get a bad CD you can have it replaced (still not refunded, I think), and shipping back and forth may be an expense of yours (just happened to me with a broken frame grabber, the italian distributor took months to replace it and I got to pay for both shippings).
thanks /alessandro
Alessandro Rubini wrote:
By law, you have one week to gave up from something you bought (since it is in good state), and three months of guarantee for most products.
Interesting. Could you please check what guarantee offer the proprietary licenses in your country? It would be a good subject for comparison.
In Italy the minimum guarantee is 2 years, I think. Still, software has no guarantee. Well, the _media_ is guaranteed, if you get a bad CD you can have it replaced (still not refunded, I think), and shipping back and forth may be an expense of yours (just happened to me with a broken frame grabber, the italian distributor took months to replace it and I got to pay for both shippings).
thanks /alessandro
hum... I really don't know exactly how this works with software, but I know here we have some problems because software sometimes is considered a product and sometimes not. The proprietary licenses say that crack just like the MS eula, "provided as is", and this can be contested, but the laws are so dubious that this would take some years in court. The software laws are very new and don't cover everything. Just to give you an idea, hacking into another computer, a pair of years ago was fit in a law about "energy theft". When we don't have a law about something, we try to fit in an existent, but this does not solve the problem. Another problem is that in Europe, you only make a law when you are absolutely sure you can enforce it, here we do a law because we think we should, but this not guarantees there will be a regulator to this law, or that it will be enforced. There's a lot of discussion here about software quality, mainly at the government.
On Monday 12 July 2004 23:29, gandhi_quilombo wrote:
Another problem is that in Europe, you only make a law when you are absolutely sure you can enforce it,
You must be kidding, or you must have missed some recent italian pearls ;) We have laws asking people to attach special stickers to every authorized copy of copyrighted work (this one has been mostly dropped now, after a long debate), or asking people to send and register in two national libraries a copy of every version of everything published on the internet, or asking people to make sure that every informative and regularly updated site on the Internet has a state-sanctioned (through an exam) known responsible person. And don't get me started on laws where almost all MPs said "it's is a really broken law, so right now we'll approve it and then we'll change it later". Of course it has yet to be changed but police has been instructed not to enforce it. ;)
On Mon, 2004-07-12 at 17:01 -0300, gandhi_quilombo wrote:
By law, you have one week to gave up from something you bought (since it is in good state), and three months of guarantee for most products. (of course the company can give more time for guarantee, this is the minimum).
Warranties for books do not exist, that I know of, except from some guy saying: I assure you that you will love this book.
Then you probably can go after that person, but not the publisher or the author (unless he's that person).
I'll admit, however, total ignorance of Brazil's law, so you may have warranties on literary works!
Return != warranty.
Rui
Rui Miguel Seabra wrote:
On Mon, 2004-07-12 at 17:01 -0300, gandhi_quilombo wrote:
By law, you have one week to gave up from something you bought (since it is in good state), and three months of guarantee for most products. (of course the company can give more time for guarantee, this is the minimum)..
Warranties for books do not exist, that I know of, except from some guy saying: I assure you that you will love this book.
Then you probably can go after that person, but not the publisher or the author (unless he's that person).
I'll admit, however, total ignorance of Brazil's law, so you may have warranties on literary works!
Return != warranty.
Rui
For example, suppose you buy a book to give as a gift, and you wait one month for it. Then, you give the gift and discover that it is missing 10 pages. This is a "Factory" problem, so you can have your money/book back.
On Mon, 2004-07-12 at 18:32 -0300, gandhi_quilombo wrote:
For example, suppose you buy a book to give as a gift, and you wait one month for it. Then, you give the gift and discover that it is missing 10 pages. This is a "Factory" problem, so you can have your money/book back.
But that's not warranty over the book, but over the medium it is printed on (cover, pages, uncorrupted content, etc...).
So it's equivalent to having a warranty over the medium the software resides (CD, net availablity, etc...) and not a warranty over the software itself, see?
Rui
On 2004-07-12 04:32:08 +0100 Wouter Vanden hove wouter.vanden.hove@pandora.be wrote:
"France lends support to new open-source license CeCILL license, compatible with GNU GPL, aims to be better suited for French laws" What does FSFE-supporters think about this?
There are recent discussions on fsfe-france and debian-legal which may interest you, starting at URLs http://lists.gnu.org/archive/html/fsfe-france/2004-07/msg00016.html and http://lists.debian.org/debian-legal/2004/07/msg00067.html (from http://mjr.towers.org.uk/blog/2004-6.html#cecill )
In general, this seems like a good thing. I have heard from researchers that some are nervous in using a licence written for foreign laws. I assume this is a bigger problem for using a US licence in France, where differences in legal style seem greater than between the US and UK.
That said, some key questions have been raised: * Doesn't specify what GPL it means: Grombat Public Licence compatible too? * Apparently commits the Holder to keep the software available forever? * Forces anyone the Holder sues to be represented in France:
Also contains mild anti-free-software "techies-only" propoganda in the preamble, which I missed on first reading.
Creative Commons has a strong-copyleft license (Share-Alike) which they are translating and adapting to local laws.
No CC licences are free software licences, or even free-software-compatible, so I think we're off-topic there, much as I wish it otherwise.
Why doesn't FSFEurope adopt a same strategy?
Maybe you should cc team or office on just that narrow question, for an official reply. I'm not sure who reads this list and how fully (fx: prepares to be snowed on).
Why would European government or public institutions release their own software under a GPL that is only partially valid in many european countries?
Mindshare, basically, or they might think the valid bits are sufficient. Also, they may have a US entity that they can use to publish it there first, which I think means the GPL would be as fine as for FSF.
On Mon, Jul 12, 2004 at 10:15:34AM +0100, MJ Ray wrote:
On 2004-07-12 04:32:08 +0100 Wouter Vanden hove wouter.vanden.hove@pandora.be wrote:
Why doesn't FSFEurope adopt a same strategy?
Maybe you should cc team or office on just that narrow question, for an official reply. I'm not sure who reads this list and how fully (fx: prepares to be snowed on).
That is good advise to raise the chances of an answer. I usually read the list but sometimes with huge delay.
On Mon, Jul 12, 2004 at 05:32:08AM +0200, Wouter Vanden hove wrote:
"France lends support to new open-source license CeCILL license, compatible with GNU GPL, aims to be better suited for French laws" http://www.infoworld.com/article/04/07/09/HNfrancelendssupport_1.html
What does FSFE-supporters think about this?
It is not a good idea to write new licenses in general. http://www.dwheeler.com/essays/gpl-compatible.html Please, where possible, use an existing widely-used license for your software that is known to be compatible with the General Public License (GPL),
I advise people to choose between GNU GPL, GNU LGPL and X11 style.
Creative Commons has a strong-copyleft license (Share-Alike) which they are translating and adapting to local laws.
Why doesn't FSFEurope adopt a same strategy?
Personally I think that creating new licenses does not solve many problems. The three licenses mentioned above are reasonable valid in all countries I have heard of and they are well understood.
Why would European government or public institutions release their own software under a GPL that is only partially valid in many european countries?
Because the non-working portion is neglectable in face of the drawbacks of writing and establishing a new license.
Bernhard
On Mon, Jul 12, 2004 at 07:54:49PM +0200, Bernhard Reiter wrote:
It is not a good idea to write new licenses in general. http://www.dwheeler.com/essays/gpl-compatible.html Please, where possible, use an existing widely-used license for your software that is known to be compatible with the General Public License (GPL),
You can read everywhere about this 'gpl-compatibility' thing. But what is compatibility in fact? IMHO this has to mean one license is compatible to another and back. Which is not the case with the GPL. Of course it is always possible to create a new project which uses some different licensed code and release the whole thing under the GPL. Yes. But it is *not* possible to create a new project, use some GPLed code and release the whole thing under the BSD license, because of the unique infecting behavior of the GPL. So, if you talk about gpl-compatibility, you should clearly state that you talk from the GPL point of view, not in general.
kind regards, Tom
On Mon, 2004-07-12 at 23:24 +0200, Thomas Linden wrote:
You can read everywhere about this 'gpl-compatibility' thing. But what is compatibility in fact? IMHO this has to mean one license is compatible to another and back. Which is not the case with the GPL.
"Compatible" means "works together". The BSD licence is as much GPL- compatible as the GPL is BSD-compatible.
But it is *not* possible to create a new project, use some GPLed code and release the whole thing under the BSD license, because of the unique infecting behavior of the GPL.
You can never relicence other people's code. The GPL doesn't allow you to do that either - so, I don't think you're right to call the GPL "infecting". If you combine two pieces of software that are licenced differently, you must fulfil the obligations of *both* simultaneously: you must respect the BSD licence as much as the GPL.
Cheers,
Alex.
On Mon, Jul 12, 2004 at 10:52:49PM +0100, Alex Hudson wrote:
You can never relicence other people's code. The GPL doesn't allow you to do that either - so, I don't think you're right to call the GPL "infecting".
The GPL forces me to relicense code - my code. If I write some new code, which I want to license some way (but not under GPL) and use a couple of .c files of an existing gpl'ed project, the GPL *tells* me that the resulting code must also be released under the GPL. The BSD license on the other hand does not tell me how I do have to license my code. I can use bsd-licensed code however I like, I don't have to release the resulting source under some bsd license; this allows me to create a new GPL project which includes e.g. FreeBSD code. So, there is definitively a difference which makes them incompatible. This is the cause why most (I'm not sure if this is the case for all) BSD base systems do not include GPL code. They always put such stuff into the ports, not the base, to make sure the base will always be licensable under the BSD license.
I don't want to flame against the GPL (I use it in my own projects), but it is not the holy grail, and I don't see, why people "should" use a gpl-compatible license.
- Tom
On Tue, 2004-07-13 at 00:04 +0200, Thomas Linden wrote:
On Mon, Jul 12, 2004 at 10:52:49PM +0100, Alex Hudson wrote:
You can never relicence other people's code. The GPL doesn't allow you to do that either - so, I don't think you're right to call the GPL "infecting".
The GPL forces me to relicense code - my code.
No, it doesn't. You can take someone else's BSD code and combine it with someone else's GPL code and release the lot. You don't need to relicense any code, which is why they are "compatible". You just have to respect *both* licences. It doesn't require you to relicense code - either yours, or anyone else's (which you cannot relicense, even if you wanted to).
the GPL *tells* me that the resulting code must also be released under the GPL.
It says that you have to respect the GPL licence (i.e., that you release it complying with the terms of the GPL). In practice, this means that the licence terms for the whole cannot be more restrictive than the GPL, but if you disagree with that then it's not the GPL you have a problem with but Copyleft in general.
The BSD license on the other hand does not tell me how I do have to license my code. I can use bsd-licensed code however I like, I don't have to release the resulting source under some bsd license; this allows me to create a new GPL project which includes e.g. FreeBSD code.
That's not right. If you take someone else's BSD code, you cannot relicense it. You have no rights to do that - you must respect the BSD licence. Hence the "problems" with the original BSD licence that had an advertising clause. If you could relicense BSD code, that problem would have vanished as people just relicensed things to a different licence.
So, there is definitively a difference which makes them incompatible. This is the cause why most (I'm not sure if this is the case for all) BSD base systems do not include GPL code.
BSD systems tend not to include GPL code because they are a cohesive whole (an entire operating system), and they wish to have uniform licensing. The BSD licence is the one they have chosen, that's why they do not accept BSD code. All big projects will only accept code licenced under certain terms; Linux only accepts GPL code, for example. BSD just happens to include many separate utilities.
I don't want to flame against the GPL (I use it in my own projects), but it is not the holy grail, and I don't see, why people "should" use a gpl-compatible license.
GPL compatibility is important for a number of reasons. Firstly, there is a huge amount of free software licenced under the terms of the GPL - so, for purely practical reasons, something which is licenced compatibly with the GPL is likely to be usable in combination with as much other software as possible. This is ignoring any potential ideological reason you might have for choosing one over another.
Also, if it's GPL compatible then it's using a well-known licence that people understand - if it was under some new licence that was not GPL compatible, people would have to analyse the licence to make sure that it really was a free software licence.
Cheers,
Alex.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 13 Jul 2004 at 0:04, Thomas Linden wrote:
On Mon, Jul 12, 2004 at 10:52:49PM +0100, Alex Hudson wrote:
You can never relicence other people's code. The GPL doesn't allow you to do that either - so, I don't think you're right to call the GPL "infecting".
The GPL forces me to relicense code - my code. If I write some new code, which I want to license some way (but not under GPL) and use a couple of .c files of an existing gpl'ed project, the GPL *tells* me that the resulting code must also be released under the GPL.
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
Technically speaking, I could take a GPLed C file, alter two lines and so long as I have clearly demarcated those two lines as being copyright to me and therefore under my magic special license, that's fine. Of course, my magic special license must supply source with any binary release plus meet all the other GPL requirements.
This is the cause why most (I'm not sure if this is the case for all) BSD base systems do not include GPL code. They always put such stuff into the ports, not the base, to make sure the base will always be licensable under the BSD license.
I have the sneaking suspicion that bits & pieces of LGPL code may have found their way into the 5.x series FreeBSD kernels. It hardly matters though the GPL itself would as depending on how strictly you interpreted what constitutes a derived work, you could argue that all BSD applications must therefore be GPL. Of course on Linux the convention is that there is a line drawn at the edge of kernel space.
I don't want to flame against the GPL (I use it in my own projects), but it is not the holy grail, and I don't see, why people "should" use a gpl-compatible license.
Neither do I. Indeed, the GPL is a very poor license to choose for certain kinds of project though to even suggest that it isn't perfect is flamebait to certain kinds of mentality. I only use the GPL for augmentations to existing GPLed works - I use LGPL or better for all new code.
Cheers, Niall
On Tue, 2004-07-13 at 00:46 +0100, Niall Douglas wrote:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
Technically speaking, I could take a GPLed C file, alter two lines and so long as I have clearly demarcated those two lines as being copyright to me and therefore under my magic special license, that's fine. Of course, my magic special license must supply source with any binary release plus meet all the other GPL requirements.
Where on earth did you take that absurd idea from?
2. You may modify your copy or copies of the Program or any portion of it, thus forming a work based on the Program, and copy and distribute such modifications or work under the terms of Section ^^^^^^^^^^^^^^^^^^^^^^^^^^ 1 above, provided that you also meet all of these conditions: ^^^^^^^
And what is section 1?
1. You may copy and distribute verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the notices that refer to this License and to the absence of any warranty; and give any other recipients of the Program a copy of this License along with the Program.
Hence, derivate must be GPL'ed.
Rui
Rui Miguel Seabra wrote:
- You may modify your copy or copies of the Program or any portion of it, thus forming a work based on the Program, and copy and distribute such modifications or work under the terms of Section ^^^^^^^^^^^^^^^^^^^^^^^^^^ 1 above, provided that you also meet all of these conditions: ^^^^^^^
And what is section 1?
- You may copy and distribute verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice and disclaimer of warranty; keep intact all the notices that refer to this License and to the absence of any warranty; and give any other recipients of the Program a copy of this License along with the Program.
Hence, derivate must be GPL'ed.
ONLY if it is to be distributed. I would welcome more clarification on what constitutes a distribution. If a small company modifies and builds and runs GPL software on a single computer they are not required to release the source as the derivative work is not being distributed. Arguably a larger company (having more than one PC) may use the modified GPL software across the company without releasing source - even if they software is used to provide a public or commercial service - because the derivative work is not being distributed.
Certainly if the GPL derivative code were made available to another company this would count as a distribution. What about to a different department? What about a different group company? What about to club members? What if the club is a company? What if the club is a company AND a club of companies?
Finally, to clarify Thomas Lindens case (I'm not sure if he misunderstood the GPL here or just badly demonstrated the different between GPL and BSD licenses):
Thomas Linden wrote:
The GPL forces me to relicense code - my code. If I write some new code, which I want to license some way (but not under GPL) and use a couple of .c files of an existing gpl'ed project, the GPL *tells* me that the resulting code must also be released under the GPL.
If from that you can conclude that the GPL forces you to relicense code I can only conlcude that you thought the GPL forced you to use the "couple of .c files" as well.
The truth is, you have absolutely no right to distribute derivative works based on those .c files (due to copyright) without permission and that GPL'd is one way you are granted permission with certain conditions which aren't as free as the BSD grant.
Users who don't like the conditions may just avoid use the "couple of .c files", the GPL isn't going to make anyone use those .c files, but those who do should respect the copyright of the authors and respect the law.
The GPL can't save the world from everything but it is a political holy grail for those who want software to remain free (as in liberty), and I suspect that because of this it is capable of inspiring the hot zeal that exists more among linux advocates that BSD advocates. Because the GPL grants and maintains rights it is something to actually get excited about.
Sam
On Tue, 2004-07-13 at 09:29 +0100, Samuel Liddicott wrote:
Rui Miguel Seabra wrote:
(...)
Hence, derivate must be GPL'ed.
ONLY if it is to be distributed.
Yes. But isn't that the case we're talking about?
I would welcome more clarification on what constitutes a distribution. If a small company modifies and builds and runs GPL software on a single computer they are not required to release the source as the derivative work is not being distributed. Arguably a larger company (having more than one PC) may use the modified GPL software across the company without releasing source - even if they software is used to provide a public or commercial service - because the derivative work is not being distributed.
Certainly if the GPL derivative code were made available to another company this would count as a distribution. What about to a different department? What about a different group company? What about to club members? What if the club is a company? What if the club is a company AND a club of companies?
It may belong to a group, but they are distinct companies. Other than that, most of what you said are considered private circles in most about anything.
Rui
Rui Miguel Seabra wrote:
On Tue, 2004-07-13 at 09:29 +0100, Samuel Liddicott wrote:
I would welcome more clarification on what constitutes a distribution. If a small company modifies and builds and runs GPL software on a single computer they are not required to release the source as the derivative work is not being distributed. Arguably a larger company (having more than one PC) may use the modified GPL software across the company without releasing source - even if they software is used to provide a public or commercial service - because the derivative work is not being distributed.
Certainly if the GPL derivative code were made available to another company this would count as a distribution. What about to a different department? What about a different group company? What about to club members? What if the club is a company? What if the club is a company AND a club of companies?
It may belong to a group, but they are distinct companies. Other than that, most of what you said are considered private circles in most about anything.
What if the employee doing the code was empoyed by the group and not one of the group companies? Does collaboration (say cvs) between developers at different companies of the same group count as distribution? This is significant as cvs distributions are based on patches and diff's? If ed diff's are used it may be possible to collaborate without transmitting any of the original code; so this collaboration arbuably would not constitute a distribution; but its just a special form of the question: can I distribute a non-GPL patch script to a GPL product?
If company A buys company B (on paper) and then after merging some computer systems immediatly sells company B would this count as a distribution? Yeah, I am taking things to desperate limits, but desperate times may call for desperate measures; one day something like this may happen?
I would like to know what "distribution" means.
Sam
On Tue, 2004-07-13 at 10:22 +0100, Samuel Liddicott wrote:
What if the employee doing the code was empoyed by the group and not one of the group companies?
It may depend on local country law, but I don't think this is possible.
Does collaboration (say cvs) between developers at different companies of the same group count as distribution?
Probably so, but it would only grant the rights to who got the code. Not to everybody.
This is significant as cvs distributions are based on patches and diff's? If ed diff's are used it may be possible to collaborate without transmitting any of the original code; so this collaboration arbuably would not constitute a distribution; but its just a special form of the question: can I distribute a non-GPL patch script to a GPL product?
As long as the resulting product is GPL'ed... which would the non-GPL patch rather pointless, I guess.
If company A buys company B (on paper) and then after merging some computer systems immediatly sells company B would this count as a distribution?
I don't think it would make sense.
Yeah, I am taking things to desperate limits, but desperate times may call for desperate measures; one day something like this may happen?
I saw nothing of desperate in there, unlike the tactics to promote the DMCA and EUCD.
I would like to know what "distribution" means.
I think you're raising a storm on a glass of water.
Rui
I raised this a while ago and it has been brought up again in relation to Sveasoft who are making "pre-release" distributions with extra conditions. [The point of this message is not to discuss whether what Sveasoft are actually doing is legal, but to discuss what the GPL permits and requires]
Creative commons have a very nice wizard http://creativecommons.org/license/ that helps you choose a license.
The GPL is only one license but it has options, perhaps the FSF should produce a web page that lets you select some options and then shows you your obligations and then gives borderline examples of what would and would not fulfil those obligations.
Read on to see the confusion that is ensueing with people who are trying to understand the GPL,
Samuel Liddicott wrote:
I would welcome more clarification on what constitutes a distribution. If a small company modifies and builds and runs GPL software on a single computer they are not required to release the source as the derivative work is not being distributed. Arguably a larger company (having more than one PC) may use the modified GPL software across the company without releasing source - even if they software is used to provide a public or commercial service - because the derivative work is not being distributed. Certainly if the GPL derivative code were made available to another company this would count as a distribution. What about to a different department? What about a different group company? What about to club members? What if the club is a company? What if the club is a company AND a club of companies?
There is clearly still a lot of confusion over what constitues a "distribution". If you want to get a better idea of the confusion that exists read the full slashdot story and related links at http://slashdot.org/article.pl?sid=04/07/21/2255239&tid=193&tid=1
This issue is extremely important as many rights and obligations of the GPL center at the act of "distribution"
I wish someone from the FSF legal team would offer some clarification here before some wider precedents are made in relation to the recent strategy of: "If you excercise your rights under the GPL you are excommunicated from your supplier and are alone with your source."
This has been tried to some degree by RedHat with their enterprise linux, I understand that if you made a distribution of the code you received with your support contract that your support contract was then terminated. I hope to be corrected in this if I am wrong, but it would be a worrying trend; - but even worse would be the closed-club sharing that is not quite a distribution at all.
A good summary of the actual situation is at: http://slashdot.org/comments.pl?sid=115310&cid=9766597
which quotes from Sveasoft: "/We release two versions of firmware at Sveasoft, public and pre-release.
Subscribers can redistribute public versions of firmware to anyone they choose to without any change in their subscription rights. When you redistribute public firmware you must offer both source code and binaries or you violate the GPL license. Other than this caveat you can redistribute whenever and to whomever you choose.
The policy for pre-release firmware is different. You can also choose to redistribute pre-release firmware under the GPL. You must also offer both source code and binaries as with the public releases. Should you choose to redistribute pre-release versions however, your subscription rights terminate and you will not have access to the Sveasoft forums or future firmware pre-releases afterwards."
/Sveasoft also seem to charge $49 for a CD-ROM of the source to the "pre-release"
The slashdot user "BJH" goes on to say: "2) Sveasoft make a distinction between pre-release and public release versions of the firmware. This is probably based on the idea that you do not have to provide source code if you're distributing software purely within your own organization (i.e. not publically distributing it). However, that right is not clearly defined in the GPL, and indeed section 6 states that: /6. Each time you redistribute the Program (or any work based on the Program), the recipient automatically receives a license from the original licensor to copy, distribute or modify the Program subject to these terms and conditions. *You may not impose any further restrictions on the recipients' exercise of the rights granted herein.* You are not responsible for enforcing compliance by third parties to this License./ In my opinion, Sveasoft's artificial distinction between pre-release and public release firmware comes into conflict with this section."
user "mdsft13" writes: " "Now under the terms of the license for access to the server they can cancel your server account if you redistribute."
It's still a restriction on *distribution* which they can't do and be complying with the GPL. "
user "black mariah" sees it thus: " They're not putting restrictions on the redistribution. They're putting restrictions on their own subscriptions. In order to be a subscriber, you have to abide by their terms. One of those terms is not redistributing the source. If you choose to distribute the source, it simply terminates your subscription. This may not be in the spirit of the GPL (not that I care much), but at least from my interpretation is is well within the letter of the GPL. "
"hazem" points out in: http://slashdot.org/comments.pl?sid=115310&cid=9766834 " "T/he GPL only kicks in when you DISTRIBUTE your modifications."/
My thoughts exactly. Now suppose someone like Sveasoft wants to have beta testers of their product, and to become one, you pay $49. You're now part of the club and maybe could be considered part of the organization.
So, they send you binaries to test. Are you considered internal to the organization now? Or is this an external distribution? If it's the former, they may not have to give you the source at all. And if they do give you the source, they could "kick you out of the club" if you chose to distribute that source that is in beta form. If it's the latter, then what actually constitutes "in-house" vs a public distribution?"
and the reply to that is: "Irrelevant. Any such arrangement (such as a contract of employment) is an additional restriction and violates the GPL. There's no such thing as 'in-house' distribution under the GPL."
http://slashdot.org/comments.pl?sid=115310&cid=9767294 Is also worth reading.
I hope you're not too confused, I know I am.
Sam
Am Die, den 13.07.2004 schrieb Niall Douglas um 1:46:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The derived code *must* be GPL.
Technically speaking, I could take a GPLed C file, alter two lines and so long as I have clearly demarcated those two lines as being copyright to me and therefore under my magic special license, that's fine. Of course, my magic special license must supply source with any binary release plus meet all the other GPL requirements.
No. There is no separate license for your two lines of code.
Neither do I. Indeed, the GPL is a very poor license to choose for certain kinds of project though to even suggest that it isn't perfect is flamebait to certain kinds of mentality. I only use the GPL for augmentations to existing GPLed works - I use LGPL or better for all new code.
The GPL is a perfect license for most of the project. For a very small number of project, the LGPL is is better.
Thanks,
Reinhard Mueller wrote:
Am Die, den 13.07.2004 schrieb Niall Douglas um 1:46:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The derived code *must* be GPL.
To be clear: ONLY if you distribute or publish the derivative work. AFAIK you can still provide public and commercial services using the derivative without distributing the derivative.
What does "to all 3rd parties" mean? It doesn't sound like it means "people you distributed the derivative work to" which is what I previously understood. This ressurects cases of embedded linux routers without source; who must supply the source to the end customer? The re-saler? The importer? The manufacturer? Or the developer of the embedded code who supplied it ONLY to the manufacturer?
Sam
Just a warning about the GPL and distributions:
You may think it cool that you are being paid to work developing GPL code as you get to keep access to your changes, but there is nothing cool about this if the new work is not distributed publicly.
You may not have the right to distribute the work you have made as copyright may belong to your employer; but unless your employer either publicly distributes the work under GPL or to someone else who will make it public under the GPL the changes you make will not necessarily be available to you or anyone else.
My tip is to arrange to commit code changes back to a public repository as often as possible.
Sam
On Tue, 2004-07-13 at 09:43 +0100, Samuel Liddicott wrote:
My tip is to arrange to commit code changes back to a public repository as often as possible.
But in certain countries or under certain job contracts, you may need to have explicit permission from the company to do so.
Rui
Am Die, den 13.07.2004 schrieb Samuel Liddicott um 10:32:
The derived code *must* be GPL.
To be clear: ONLY if you distribute or publish the derivative work.
Yes, that is correct.
What does "to all 3rd parties" mean? It doesn't sound like it means "people you distributed the derivative work to" which is what I previously understood.
Here it means "all people you distribute the derivate work to". You are not forced to give a derivate work to anybody you don't want to.
This ressurects cases of embedded linux routers without source; who must supply the source to the end customer? The re-saler? The importer? The manufacturer? Or the developer of the embedded code who supplied it ONLY to the manufacturer?
Most logical would be "the one who commercially distributes the binaries". So actually this would be the reseller, but I am not 100% sure about that.
Thanks,
Reinhard Mueller wrote:
Am Die, den 13.07.2004 schrieb Samuel Liddicott um 10:32:
This ressurects cases of embedded linux routers without source; who must supply the source to the end customer? The re-saler? The importer? The manufacturer? Or the developer of the embedded code who supplied it ONLY to the manufacturer?
Most logical would be "the one who commercially distributes the binaries". So actually this would be the reseller, but I am not 100% sure about that.
The problem being when the reseller doesn't have the source and you have no legal relationship with the person who has so their obligation is not to you; and the reseller isn't inclined to pursue the case with any degree of vim.
Sam
What does "to all 3rd parties" mean? It doesn't sound like it means "people you distributed the derivative work to" which is what I previously understood.
Let me quote 2b again:
b) You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
This means that the _license_ is valid for everyone else. It doesn't mean that you must give (or make available) a copy to anyone else.
But if you distribute your program or derived work to company A and later find company B has a copy, you can't say "no, I didn't give permission to company B.
This ressurects cases of embedded linux routers without source; who must supply the source to the end customer?
The one who gave the binary to the customer (this is the distributing company, not the shop clerk, usually). It's their problem if they didn't get the source with their binary, not the user's.
/alessandro
--------------010502020007000600000609 Content-Type: text/html; charset=us-ascii
PS: please avoid sending two copies of your mail message. The html is definitely unneeded and often exceedingly long. You should configure your mailer to only send the text version of your messages.
At Tue, 13 Jul 2004 09:09:45 +0200, Reinhard Mueller wrote:
Am Die, den 13.07.2004 schrieb Niall Douglas um 1:46:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The derived code *must* be GPL.
Technically speaking, I could take a GPLed C file, alter two lines and so long as I have clearly demarcated those two lines as being copyright to me and therefore under my magic special license, that's fine. Of course, my magic special license must supply source with any binary release plus meet all the other GPL requirements.
No. There is no separate license for your two lines of code.
Software doesn't have to be under one license only, it can have multiple licences. And that's what happens here.
Of course, the whole program will be under the GPL. That's what the GPL requires. But that doesn't prevent someone from saying that all the lines of codes they added to the program can be used under the BSD license. So, in fact, those lines of code are than licensed under both the GPL and the BSD license.
Neither do I. Indeed, the GPL is a very poor license to choose for certain kinds of project though to even suggest that it isn't perfect is flamebait to certain kinds of mentality. I only use the GPL for augmentations to existing GPLed works - I use LGPL or better for all new code.
The GPL is a perfect license for most of the project. For a very small number of project, the LGPL is is better.
And for even smaller number of projects, BSD-style is better (ogg vorbis for example). :-)
Jeroen Dekkers
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 13 Jul 2004 at 9:09, Reinhard Mueller wrote:
Am Die, den 13.07.2004 schrieb Niall Douglas um 1:46:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
I'll firstly point out that I argued precisely what you've just argued over a period of three weeks. It turned out that where we were taking contrary meanings from the same text was because I had not noticed a very small phrase.
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The key phrase here is "... under the terms of this License". This is NOT the same as "... under this License".
It might seem like splitting hairs, but apparently this phrase legally makes the difference between requiring that all derived works must also be GPLed and merely requiring that the license for all derived works must be *compatible* with the GPL.
Now I am not a lawyer, but this person who I argued with for three weeks was heavily plugged into the whole FSF/GNU thing and while I'm not saying that there are members on this list who aren't, I did advance precisely your arguments and I discovered I was wrong. I just checked my email archives, I had that argument with Alex Hudson who I think is/was on this group? I've CC-ied him anyway.
Cheers, Niall
On Tue, 2004-07-13 at 19:10 +0100, Niall Douglas wrote:
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The key phrase here is "... under the terms of this License". This is NOT the same as "... under this License".
No. It is very clear, at least to me.
Theory: "Under the terms of this License" != "a compatible license"
Proof by example of falsehood:
If "Under the terms of this License" == "a compatible license" then all compatible licenses have the same terms.
X11 is compatible with the GPL and does not have the same terms (which are required).
Hence "Under the terms of this License" != "a compatible license"
q.e.d.
Rui
On Wed, 2004-07-14 at 09:19, Rui Miguel Seabra wrote:
Hence "Under the terms of this License" != "a compatible license"
Of course, but having a compatible license is a prerequisite to fall under the terms of the GPL.
Simo.
On Wed, 2004-07-14 at 08:19 +0100, Rui Miguel Seabra wrote:
On Tue, 2004-07-13 at 19:10 +0100, Niall Douglas wrote:
GPL, 2.b You must cause any work that you distribute or publish, that in whole or in part contains or is derived from the Program or any part thereof, to be licensed as a whole at no charge to all third parties under the terms of this License.
The key phrase here is "... under the terms of this License". This is NOT the same as "... under this License".
No. It is very clear, at least to me.
Theory: "Under the terms of this License" != "a compatible license"
Proof by example of falsehood:
If "Under the terms of this License" == "a compatible license" then all compatible licenses have the same terms.
X11 is compatible with the GPL and does not have the same terms (which are required).
Hence "Under the terms of this License" != "a compatible license"
q.e.d.
Argh, mornings... the proof only makes full sense inverted:
Theory: "Under the terms of this License" == "a compatible license"
Proof by example of falsehood:
Assumption: If "Under the terms of this License" == "a compatible license" then all compatible licenses have the same terms.
Verification: X11 is compatible with the GPL and does not have the same terms (which are required).
Conclusion: Hence "Under the terms of this License" != "a compatible license"
q.e.d.
On Tue, 2004-07-13 at 19:10 +0100, Niall Douglas wrote:
It might seem like splitting hairs, but apparently this phrase legally makes the difference between requiring that all derived works must also be GPLed and merely requiring that the license for all derived works must be *compatible* with the GPL.
I think the situation is a bit more subtle than that.
If you want to create composite works, they must be licenced compatibly - that is, the terms of the licences must not conflict with each other.
The composite work must be distributed according to the terms of both licences - both licences still stand, because you are not allowed to re- licence a piece of work unless you are the copyright holder.
In the case of the GPL, which specifies that a distributor may not impose more restrictive terms, you cannot create a derivative that has stricter distribution terms than those of the GPL. Also, you must respect the terms of the GPL when you distribute the derivative. So, basically, the derivative is being distributed under the terms of the GPL - but that's not to say that it is wholly licensed under the GPL, which is the subtle difference.
The argument we had previously was whether or not the GPL was 'infectious', whether it had a relicensing effect, which it does not.
Cheers,
Alex.
On Tue, 2004-07-13 at 01:46, Niall Douglas wrote:
I'll just short circuit the answer to this because I thought the same as you until quite recently. The GPL's wording only mandates that derived code must supply its source, not that the derived code must also be GPLed.
Sorry Nial, afaik you're wrong.
If you take previously existing code you only need to assure it has a GPL-compatible license to make a new work base on both a GPL piece and this piece. Instead if you make _derivative_ code, that must fall under the GPL, otherwise you have no permission to use the GPL portion of the code you're deriving from.
Adding a pre-existing portion of BSD code to a GPL project is not deriving code, as it existed well before. Instead creating code as new that depends on the existing GPL code is a derivative work.
Derivative works and other works have really a completely different legal status in most legal systems.
Simo.