User:JWSchmidt/Blog/1 December 2007
The Wikimedia Foundation Board of Trustees is exploring, "longstanding community concerns about issues of compatibility between the GNU Free Documentation License and the Creative Commons CC-BY-SA license" (Board resolution).
What does, "longstanding community concerns about issues of compatibility between the GNU Free Documentation License and the Creative Commons CC-BY-SA license," really mean? What are the longstanding complaints about use of the GFDL for Wikimedia wiki project content?
Is the GFDL suited for wikis?
Some people claim to be concerned that the GFDL is not suited for media files and is not even really suited for text at wiki websites. The GFDL says it applies to, "work, in any medium" but it also says that it was designed for "manuals for free software". In my view, the license has two parts. One part of the GFDL license is the intent and the other part are some implementation details written specifically with software manuals in mind. Like any other document, the GFDL license should be interpreted with common sense and adequate mental flexibility.....we need enough flexibility to apply the intent of the license to "works" and conditions that were not imagined when the license was created.
The intent of the license is to make sure that "works" licensed under the GFDL can be freely copied, with or without modification, either commercially or noncommercially, with author attribution and any derivative works must also be copyleft.
Include a copy of the GFDL
The GFDL license says that people are free to copy works licensed GFDL as long as the copies all reproduce the license. This makes sense for large software manuals where including a copy of the GFDL would not significantly add to the size of the document. However, we are now in the internet age. People are now familiar with the idea that information is available on the internet. We no longer have to make copies of texts, we can just point to web pages that hold texts. If I license an image or a sound file under the GFDL, it makes no sense to include a copy of the GFDL every time I make a copy of the media file. It would be a trivial matter for a new version of the GFDL to acknowledge this.
Derivative works must be GFDL
Some people complain that "works" licensed under the GFDL must have any derivative works also made available "under precisely this license". However, the intent of the GFDL is that derivative works of be free in the sense of being copyleft. There is nothing that prevents us from saying that a work is available under multiple copyleft licenses. It has been common practice to dual license media files uploaded to wiki websites under both the GFDL and the Creative Commons attribution share-alike license. It would be a trivial matter for a new version of the GFDL to acknowledge that these two licenses have the same intent and formally acknowledge that people should feel free to use either or both licenses.
The flip side: complaints about Creative Commons
Just as there are complaints about the detailed language of the GFDL, there are also legalistic complaints about the Creative Commons attribution share-alike license (see this thread).
"Laurence Lessig has posted multiple times claiming that it is acceptable to take illustrations licensed under CC-By-SA and produce combined works which are not freely licensed." <-- The Creative Commons attribution share-alike license does not use the term "combined works" but it does use the term "Collective Work". I think it is the case that the rules for both the CC-By-SA "Collective Work" and the GFDL "AGGREGATION" agree that "combined works" do not have to be freely licensed. I could use a restrictive license on part of a "combined work" (for example, place it under a "no commercial re-use" license) that includes copyleft components. If this is a complaint against CC-By-SA, it is also a complaint against the GFDL.
What counts as a "modification" under the FDL? <-- I'd be interested to see this position argued in a court of law. Yes, there is a gray area between "aggregate" and "derivative", but the way I read the GFDL it is not as "viral" as other people seem to think.
"When someone is really unwilling or unable to freely license their derivative" <-- A derivative is different than a "combined work" and both the GFDL and the Creative Commons attribution share-alike license make clear that derivative works must be copyleft.
"expect to be able to exactly stipulate how attribution is provided" <-- I think wiki participants tend to be lax about attribution because many wiki pages have many authors. Attribution often means providing a way to trace back to a complex edit history for a webpage....and sometimes multiple webpages that have been split and merged over an extended period of time. Other copyleft works might be the copyrighted work of one person and it is not unreasonable for such a copyright holder to expect clear attribution. Again, I do not see this as being significantly different for the GFDL and the Creative Commons attribution share-alike license.
"these by-attribution licenses don't actually provide attribution if a service provider specifies so in their terms of service" <-- I think this is not a reasonable reading of the Creative Commons attribution share-alike license. The license says that the terms of service for a website might be one mechanism used to provide some information about attribution of some of the website's content. It does not say that attribution need not be given if the terms of service say the website does not give attribution.
"Creative Commons branding, which many people feel is exploitative" <-- I'm not sure I can reconstruct this complaint. Apparently some people feel that saying their work is under a Creative Commons license is "advertising" for the Creative Commons "brand". I guess I can understand that some people who do not understand copyleft could have this sort of reaction to licensing their work. I do not see why the Creative Commons attribution share-alike license is any different in this regard than the GFDL. If I wanted, I could simply say that I want my copyrighted work to be available according to the generic rules of copyleft. I imagine that there are advantages to me if I use a known copyleft license. If people do not feel that way, then I guess it is time to educate them about copyleft licenses.
I think it should be clear from my comments, above, that I am not really moved by complaints about either the GFDL or the Creative Commons attribution share-alike license. I think they have the same intent. I do like the sound of a wiki-specific copyleft license. I suspect none of the hand-wringing over specific copyleft licenses will really matter until there is a history of lawsuits that define how copyleft works in "real world" situations of conflict over rights to content. Will someone ever really go to court over a failure to adhere to some picky rule in a copyleft license? We are now several decades into exploration of the copyleft concept. I suspect the true power of copyleft is growing out of the internet and the power of online collaboration. There will always be theft of intellectual products, including people who will make derivatives of copyleft works that are sold and not placed under a copyleft license. I think we should be exploring ways to create institutions that will go to court when such theft is discovered. I'm a supporter of the general copyleft idea, not any one license.
Update - December 2
Yes, that is a rather pessimistic section heading....I doubt if the FUD will ebb any time soon over the idea of shifting Wikimedia licensing (apparently by way of a new version of the GFDL) towards a more Creative Commons-type of license.
"The FSF interprets the GFDL so that e.g. a photograph embedded into an article would require the article to be 'copyleft' under the GFDL" (source) <-- I'm not exactly sure how to construct such an interpretation of the GFDL. It must be something like, "a page of text that includes a GFDL-licensed image must be copyleft because the page is a derivative of the image" or "the whole page is a 'work' and you cannot license the two parts independently". I can imagine some situations in which the text and image would be best viewed as a single combined work under one license, but such situations would be relatively rare....not a "requirement" of all combinations of text and image. For most situations there will be sensible interpretations in which a page of text with an image are best viewed as an aggregate of two works (the text is one, the image is another). The text could be copyleft, but it does not have to be. Putting a GFDL-licensed image into a non-copyleft document does not remove the need to properly attribute the image as being copyleft. If some people have imagined that works under the GFDL cannot be aggregated with non-copyleft works I think they are fundamentally mistaken. Do some people really imagine that GFDL-licensed works are "viral" to the extent that the GFDL automatically spreads to anything that GFDL touches? I just to not see how people can believe such a state of affairs. I wonder what such people think the GFDL section on aggregate works means? I admit that this language is terrible: "A compilation....is called an 'aggregate' if the copyright resulting from the compilation is not used to limit the legal rights of the compilation's users beyond what the individual works permit". I interpret that to mean that nobody can make an aggregate and in so doing try to claim that their copyright on the aggregate cancels the copyleft licensing of works in the aggregate.
"I believe what is really needed is a license that gives authors the choice of "strong copyleft" for embedded media: the work into which the media are embedded (whether either work is text, sound, film, a rich media mix, or whatever) should be licensed under a copyleft license." (source) <-- I admit that I do not make a habit of trying to regularly follow arguments about licenses, but it is rather surprising that I've never heard any of this line of reasoning previously. I guess I can understand the idea that some copyright holders might not want their work to be aggregated with non-copyleft works. Further, I suppose there is a gray area between what constitutes an aggregate of works and what constitutes a derivative work. Some people might want their work to only be used under a very strict definition that makes what I would call an aggregate be legally viewed as a derivative. It will be interesting to see how far people can push their desired interpretation of the GFDL as "strong copyleft" on other people. It will be interesting to see if people can actually create and enforce a copyleft license that forbids aggregation of the licensed works with non-copyleft works. It is hard for me to see this as a priority for the Wikimedia Foundation. I doubt if there are many people who object to their copyleft works being aggregated with non-copyleft works.
Update - December 4
I'm late getting through the blizzard of FUD in this licensing thread....this was actually posted on the 3rd:
"If you take my cc-by-sa image and modify it, I insist that your modified version be made available under cc-by-sa. If you merely use my image *near* something else, then I do not insist that your entire work be made available under cc-by-sa, because I do not believe that your newspaper article is a "derivative work" of my photograph." (source)
This (above) is how it has always seemed to me. I still do not understand how other people claim that the GFDL or the goals of the Wikimedia Foundation imply that we should expect anything else from a copyleft license.
I'm trying to make sense of this position:
"A copyleft that doesn't substantially encourage people to make more freely licensed works is just another copyright related imposition, it's harmful and without public or even private benefit and should be avoided." (source)
As far as I can tell, Gregory Maxwell intends, "substantially encourage people to make more freely licensed works," to mean forcing people to use copyleft works only in the context of other copyleft works. In my view, Maxwell's "substantial encouragement" would itself be a copyright-related imposition that should be avoided. If I license my work under the GFDL, I want a teacher to use my work, even if all of that teacher's work is owned by their educational institution and not freely available. I'm not in the business of saying that teachers and schools cannot copy my work unless all their work is also copy left. The GFDL and the Wikimedia Foundation have never been about trying to force everyone else in the world to switch to copyleft. If people can make their work freely available, fine. If they cannot, then they are trapped in another system and it it not my mission to deny them the use of what I have freely licensed.
In my view, the attribution and "share-alike" aspects of copyleft have benefits even if people are allowed to aggregate non-copyleft and copyleft works. For internet-based projects, attribution in the form of URLs back to the original source leads people into the world of copyleft content creation. If they like what they see, they can join in. I'm baffled as to why anyone claiming to be interested in the Wikimedia Foundation would take the position that such benefits (as I see them) are not really benefits and that our practice of asking for attribution by a link back to the source "should be avoided".
Update Dec 15
At What's wrong with CC-BY-SA?, SB Johnny wrote "any work including work under the GFDL must be GFDL". We had a chance to discuss this and discovered a new interpretation of section #7 of the GFDL" AGGREGATION WITH INDEPENDENT WORKS. I've always assumed that "independent" means independently created and licensed works. You make an aggregate of independently created works. After making that assumption, I've always gone on to marvel over the strangely worded first sentence of section 7.
SBJ suggested an alternative interpretation of section 7 in which "independent" means that there has to be a semantic firewall between the independent works. For example, take two articles (one about elephants and one about Mt. Everest) and bind them together in a printed book. The two articles have no meaningful impact on each other and just happen to be bound together physically. I find it hard to take this interpretation seriously. The reason people aggregate independently created works is because they are exploring or exploiting some semantic connection between them. Its strange to suggest that the GFDL has a section on aggregating works that only applies to meaningless jumbles of works that have no synergistic meanings.
I can accept that some people might want a "super strong" copyleft license that prevents aggregation of copyleft works with non-copyleft works, but I do not think the GFDL is such a license. This is a good example of how different people can make different interpretations of a document. Hopefully the GFDL will be cleaned up and some of these issues clarified in the next version of the license.
Update November 3, 2008
paraphrasing: "the Wikimedia Foundation passed a resolution asking us to update the FDL so as to allow Wikipedia to also use the Creative Commons Attribution-ShareAlike (CC-BY-SA) 3.0 license. This new version of the license is meant to fulfill the Wikimedia Foundation's request."
"Later this month, we will post a re-licensing proposal for all Wikimedia wikis which are currently licensed under the GFDL. This re-licensing proposal will include a simplified dual-licensing proposition...." source
- GFDL suggestions - discussion of possible changes to the GFDL at the Wikimedia meta-wiki
- Wikipedia Signpost coverage.
- Let's switch to CC-BY-SA by Axel Boldt in September 2007
- The FSF On FDL Derivatives from May 2007. The FSF argument that placing a GFDL image into a text document converts the text to GFDL.
- video - Jimmy Wales talking about the shift from GFDL to CC-BY-SA
- related: w:Jamendo