Version 1.1 of the definition has been released. Please help updating it, contribute translations, and help us with the design of logos and buttons to identify free cultural works and licenses!


From Definition of Free Cultural Works
< Talk:Definition
Revision as of 00:30, 20 June 2006 by (Talk) (Transparent copies)

Jump to: navigation, search

Rewording Edit

I submitted a relatively large change. While the diff itself may be big, especially in the preamble, but I don't think I've made significant changes to the tone of the content. As a result, I think it is probably alright coming this late.

Here's a summary of the larger changes or set of changes that I made:

I removed a number of "therefores", and a number of comma-separated clauses acting as parenthetical asides that I thought it could survive without. I tried to break up a few long sentences. This was all purely stylistic.

I dropped the "in addition to a requirement of author attribution" from the preamble because it's clear without it (IMHO) and because it seems to qualify a statement about essential freedoms which we speak out against below. I think it's best to state clearly that there are essential freedoms and then there are some extra restrictions that do not in fact have an impact on these essential freedom. Attribution is one of these but need not be specially cases. This way it doesn't sound like an exception but an explanation that it is by definition free.

In two paragraphs of the first three paragraphs there are these two lists:

artistic works, scientific and educational materials, commentary, reports, and documents
Works built by communities collaborating as volunteers, art created for the purpose of shared enjoyment, essential learning materials, scientific research funded through taxpayer money, and many other works do not benefit from artificial scarcity.

They seem to be redundant so I've tried to reduce it to just one list where the the first list was.

I've removed "regardless of their profession, their beliefs, their country of origin, or any other criteria" and just say "anyone, anywhere, for any purpose." I think it's just as clear and more general.

I switched the strange switch into the second person in the paragraph with the god like creator. It read nice but was a jarring switch.

I added a line or two mentioning that some licenses are also used purely to take away people's freedoms. ASCAP is a licensing organization. The last draft talks about licenses as if they are only used to give away freedom but they were first used to sell freedom in restrained ways.

I was confused by the line: "Indeed, depending on the nature of the work, it may even be unethical to deny any of the enumerated freedoms above." It seems like our argument in this document is that it's always unethical, on some level, to restrict essential freedom. I think that this sort of confuses the issue and, in any case, don't add much where it was.

I removed this phrase:

This definition only covers freedom in terms of copyright law; usage of a work may be restricted by other laws.

I guess I have two issues here. First, it's not clear what these other laws and this sounds very legalistic for a statement of principles. More importantly, it seems like we would want to oppose other types of freedoms articulated in other types of law as well. The FSD is useful in opposing patents just as well as copyright (which the overly forumlate OSD could not). I don't think it's necessary.

I've removed this phrase:

"Explicitly, it must not limit commercial use of the work."

Because I basically moved it into the list of essential freedoms above. There was already an example there and I think that this is important enough it's worth dealing with a little higher up in the document. I think it's now more clear and doesn't need to be mentioned below.

-- Benjamin Mako Hill 02:01, 30 April 2006 (CEST)

I'll have another edit later. The main problem I have is with the changes in the first few paragraphs about "information goods". First, I don't like this phrase because it essentially adopts the language of information as property, a commodity. I find it somewhat amusing that you would choose this language given your objections to the word "content". ;-) I find the argument against attempts to equate information with physical property very persuasive, and am inclined to remove this phrase entirely from the document.
The second problem is with the change to the examples enumerated in the beginning. The ones chosen - "Works built by communities collaborating as volunteers, art created for the purpose of shared enjoyment, essential learning materials, scientific research funded through taxpayer money, .." - were chosen explicitly because the argument that these works should be free is strongest and most persuasive. Reworded, it essentially sounds like a declaration against copyright on "anything that can be represented as a sequence of bits". This can definitely not stay this way.--Erik Möller 02:54, 30 April 2006 (CEST)
Sounds good. "Information goods" was clumsy and we're better off with out it.
I'm still worried by the list of things. We're suggestion, but not defining, the scope of the document here. Something to think about and work out in drafting period. --Benjamin Mako Hill 05:36, 1 May 2006 (CEST)


Is it worth describing the threats to traditional freedom or to the possibilities for new freedom that it is necessary to define and defend freedom against? Stallman described software freedom and adopted the economically extreme tactic of copyleft in the face of a generation of software being made proprietary.

Culture is faced with the expansion of copyright, trademark and patent laws. DRM and end-user licenses. And even appropriation artists and sampling musicians are seeking to assert their "rights" against others who follow them.

So is it worth explicitly making this a call for freedom in the face of increasing unfreedom?

--Rob Myers 20:17, 14 June 2006 (CEST)

Fair Use

Angela made the point that we should position our definition vs. the existing rights of fair use. This echoes some of the sentiments of Larry's response. So we should think about that before we take it live.--Erik Möller 17:09, 30 April 2006 (CEST)

Made some further edits as per the above. Avoided explicitly mentioning fair use for now to avoid confusion, might add that later.--Erik Möller 02:28, 1 May 2006 (CEST)

I'm not entirely clear defining our position in regards to fair use should mean here. I'm concerned because (a) fair use is only in the US a few other countries. Some other juristictions have similar "fair dealing" law but there are important differences. It's also a balancing act left up to judges and can sometimes be very unclear. I'm also concerned because (b) fair use is a set of compromises that is supposed to balance what would be overly restricted copyright. Our argument here is that copyright is always overly restricted and that we need to not settle for fair use rights but for something more meaningful. Or am I completely off-base here? --Benjamin Mako Hill 05:44, 1 May 2006 (CEST)
No, I think you're right. From an ethical perspective, I think that it is important that fair use rights are also protected and broadened, but I'm not sure if this belongs into this definition. For now, I have added the phrase "Only very limited freedoms are granted to others" to clarify that copyright is not absolute; perhaps that is sufficient.--Erik Möller 05:47, 1 May 2006 (CEST)
I think an argument can be made that it is important that any definition expressly acknowledge fair use (US) or fair dealing (rest of the world) and/or any other exception or limitation to copyright law. This is because of recent judicial decisions that tend to find, for example, that one can contractually override exceptions to copyright law (such as reverse engineering). Thus, I think it could be important to clarify that licenses that satisfy the free content/expression definition are "fair use/fair dealing plus" and grant a layer of permissions in addition to and on top existing exceptions to copyright law. Happy to leave this up to community discussion, however.--Mia Garlick 10:12, 1 May 2006 (PDT)

Fair use is not unified at all among countries, so referring to it in the definition would introduce uncertainty and complication (unless the definition also gives its own definition of fair use). Moreover, fair use is completely covered by the rights specified in the definition AFAIU. --Antoine 20:43, 1 May 2006 (CEST)

Fair use can give more rights than a copyleft license. Suppose that you copy a page from a copylefted textbook and give it to a student. Without fair use, you'd have to distribute source for the page (or maybe you could get away with giving the source for the whole book). With fair use, you just make the photocopy and are done.--Sanjoy Mahajan 00:17, 16 May 2006 (EDT)

I would submit that the issue of whether a specific acknowledgement that free licenses apply in addition to and on top of existing exceptions and limitations that exist at law should not be decided based on whether the term "fair use" is used or not. One can use the terminology that I have just used "exceptions and limitations at law." This encompasses jurisdictional differences. And it should not be the case that "fair use" or any other exceptions or limitations are covered by the rights specified in the definition. In fact, I would have thought that the entire point of free licenses is to carve out a space that exists beyond what is currently provided for by law. --Mia Garlick 2:23, 2 May 2006 (PDT)

Hi Mia. As far as I've understood you correctly, I don't agree. Your proposal doesn't encompass jurisdictional differences, it just hides them which is vastly different. If I read "exceptions and limitations at law", I don't know what the text is talking about because those exceptions and limitations are different depending on the jurisdiction. Consequently, the text has no normative value because people from different parts of the world will understand different things.
Let me stress it, because it is very important: the Definition should be totally independent of any local specifics. There must be no international insecurity as to what legal system the definition is referring to when it is using a given expression. The obvious way to achieve this is to avoid any reference to regionally varying legal concepts (or only as a side note as is the case for moral rights; then fair use or legal exceptions to copyright might also be mentioned as a side note, not a constitutive part of the definition).
Please take example from the Free Software Definition: it is worded in terms which are clear for everyone in the world (assuming they speak English ;-)), it does not refer to any non-essential legal concept. Actually even the notion of copyright is not essential in the Free Software Definition: it is mentioned only as a common way of implementing the FSD, not as a constitutive part of the definition.--Antoine 01:05, 3 May 2006 (CEST)
Hi Antoine, I think your argument is inconsistent. On May 1 you argued that fair use should not be referred to because it is not a feature of the laws of all countries. In response, I proposed wording that addresses the basic concept - namely, that there are exceptions & limitations to rights granted at law & that "free licenses" should be explicitly acknowledged not to limit those but to provide a level of permissions & freedom that apply in addition to & on top of those limitations and exceptions. Then on May 3 you state that my proposal does not encompass jurisdictional differences (except that is does) & the Definition must be free of local specifics. According to your logic, by adopting language that is not jurisdictionally-specific, jurisdictional differences become hidden. Um, yes, isn't that the point? However, then your argument continues that the "obvious way" to avoid "international insecurity" as to which legal system is being referred to is "to avoid any reference to regionally varying legal concepts." Yes, again the point. As far as I can tell, the only way to let everyone know exactly which exceptions and limitations are intended to be referred to by the definition _and_ to avoid being jurisdictionally-specific - other than the flexible & encompassing language that I suggested - is for the Definition to provide a catalogue of every single exception & limitation that exists in each jurisdiction and confirm that it intends its freedoms to apply in each jurisdiction in addition to the specified list of exceptions and limitations that exist in that jurisdiction. This would seem to be a very cumbersome if not impossible result. If we are agreed with the principle - that Free Licenses must grant freedoms that do not just reinforce existing exceptions & limitations to copyright law - then let's find a good way, that is not jurisdictionally specific, to express that. No doubt a person will not necessarily know which exceptions & limitations exist in other jurisdictions, but they should know those that apply in their own & thus, this statement will have meaning as they apply to their own activities. If they do not know the exceptions & limitations that exist at law in their own jurisdiction, then they can educate themselves as to them - after all, isn't part of the purpose of creating this Definition to educate? --Mia Garlick 13:39, 7 May 2006 (PDT)
Hi Mia. I think it will be clearer if we break the down the discussion in several points.
- Should the definition refer specifically to fair use?: fair use is quite local. I think we both agree that the definition needn't refer to "fair use".
- Should the definition make a comprehensive list of exceptions that exist in any juridiction: of course not.
- Should the definition refer to the general notion of legal exceptions to copyright: this is where we perhaps disagree. I think no, not in the body of the definition at least. Perhaps there can be a separate page about how the Definition interacts with the current legal systems. Moral right could go there too; and also the various questions about trademarks, right to self image, right to privacy, limitations to public expression (defamation, etc.).
It is obvious we agree that free licenses should provide broad freedoms to the user, but I'm not sure what you imply precisely by that statement: "Free Licenses must grant freedoms that do not just reinforce existing exceptions & limitations to copyright law". Do you mean that free licenses must take into account, in their design and their wording, existing exceptions and limitations? (I think most don't. I also don't see how this could be a problem... is there something I'm missing?)
Regards. --Antoine 00:29, 8 May 2006 (CEST)

Hi. I would like to raise two questions re: fair use.

-(Fair Use Plus question) Should the definition require a free license to grant some permission that the copyright exceptions & limitations in the applicable law do not without taking away any such right? Let me assume that the answer is yes.

-(Interpretation of the implicit question) Should the definition require a free license to make clear that the license grants does not mean to take away anything that is already given to the licensee via fair use, fair dealing, and/or exceptions and limitations of applicable copyright law? I think it is not feasible to require explicit reference at this point. GFDL does not mention fair use explicitly for one thing. But if there is no explicit reference, can we reasonably agree if a given license text indeed gives more than the applicable law already does without taking away anything? If we cannot, then we will have hard time discussing if a license is free or not.

Let me bring up an example. Japanese copyright law has a provision that an art work that is permanently placed in outdoor space that is open to public can be photographically reproduced without permission, so long as the main purpose of the reproduction is not to sell it as a reproduction, and the reproduction is not sold. In short, you are not allowed to sell a postcard which prominently features the art work, but you can use the photo as a part of your book that you sell.

When there is a seemingly free, share-alike type license, does it prohibit this kind of use under Japanese law, unless the copies are also under the license? Tomos 17:56, 8 May 2006 (CEST)

Hi Tomos. It is important to remark that you don't have to accept a free license. Not accepting the license still grants you the rights that the law grants you (in the Japanese case, rather broad ones although not as broad as in the FCD).
It also can be argued that a license can only apply where copyright applies (i.e. not in the exceptions such as stated above). The license states in what conditions the author gives up his/her monopoly of rights. Where there is no monopoly of rights (i.e. in the exceptions), there is nothing to give up thus the license simply does not interfere. However, I don't know if this is the right interpretation (IANAL).
As for the exact nature of exceptions and the way they can interact with a license or contract, not everyone agrees. In France, some people argue that the "right to private copy" (the right to make a copy for personal and family use) is really a right, others say it's an exception. It looks like the distinction has important consequences upon whether the user can demand for the right to be respected (but again the subject is too technical for me). --Antoine 04:09, 10 May 2006 (CEST)
Indeed, one often has an option of not accepting a license and go with fair use (though not always - in case of, for example, a shrink-wrap license that a user has to accept before even gain an access to the full content). An argument one can make from there is that because of that, a license can be called free when it puts restriction on licensee's fair use/fair dealing/other exceptions& limitations. ::Second point you raise - that a license cannot restrict copyright exceptions/limitations in a legally effective way - can also lead one to argue "therefore, even when a license includes some restrictions on fair use, etc., it could still be a free license."
I am not a lawyer, but in Japanese legal context, there is an ongoing discussion at the government regarding which part of the copyright law could be overridden by a contract. Can an author transfer his moral rights to a company if the both so agrees? Which of the exceptions and limitations of copyright law could possibly be canceled by a contract between an author and a user of a copyrighted work? It seems even experts do not know/agree on some of the answers to these questions.
I am inclined to feel these two arguments are a bit weak, but I am also curious to learn what others think.
But the main reason why I hesitate to suggest that a free license must be "fair use plus something" is that I expect it very hard to tell if a given license text places any restriction on any fair use/fair dealing/copyright restrictions& limitations. It would be easy when there is an explicit statement in the license that the license does not restrict any of those. CCPLs do have the statement, but GFDL does not, if I remember correct. It might be possible if a license is tied to a specific national law. iCommons licenses are tied to specific national law, but the generic ones, and GFDL are not. Tomos 13:38, 12 May 2006 (CEST)
Well, I agree with Erik's opinion elsewhere that the Definition should say what a free content is, not what a free license is. Of course we are giving a list of free licenses somewhere, so we have to provide some kind of answer to that question too, but it needn't be as definitive and authoritative as the answer to "what is a free content".
Also, as you mention, most free licenses make no statement about copyright exceptions/limitations. I think it's not reasonable to demand that all those licenses adopt the same kind of complicated wording and national adaptation as CC licenses do (which gives them their own problems). --Antoine 16:27, 12 May 2006 (CEST)
Okay, so the question is if free content should be usable within a range of freedom that is inclusive of and broader than that given by the copyright law's fair use/fair dealing/copyright limitations and exceptions.
How do you think about the idea that "fair use plus" should be in the "recommendations" section, because it is not feasible to "require" it, but is still important? Tomos 14:44, 14 May 2006 (CEST)

Effective Technical Measures

In regards to the little back-and-forth about "effective technical measures." I now realize that you are using the term as a little legal term-of-art to distinguish between TPMs and things like formats or compression. However, I'm afraid it still blocks things like GPG/PGP encrypted email. I'm going to suggest a phrasing like, "technical measure designed to restrict the freedoms above from being exercised by the person to whom the work is distributed."

This phrasing doesn't block things like encrypted email, compression, formats, etc. and doesn't require using a relatively obscure legal term. --Benjamin Mako Hill 05:52, 1 May 2006 (CEST)

Physical works (non-digital)

User:Rgladwell said: How does the definition handle digital works (such as images, documents, etc) versus non-digitual works (such as hard-copy books, paintings, sculptures, etc)?

A requisite would be to replace "modified versions" with "modified copies". If one is allowed to make a "modified version" of a physical, autographic work of art, then the original is destroyed in the process. Not something we would like to encourage. --Antoine 22:02, 1 May 2006 (CEST)
I'm not sure we can really consider physical works to be covered by the definition: concepts like distribution, copying and modifications do not seem to really apply. --Ricardo Gladwell 23:50, 1 May 2006 (CEST)
You can read the Free Art License. It has been worded very carefully by artists and lawyers so that it does apply to physical works. The key is that there is a distinction between the original and copies. Only copies can be distributed and modified (of course, this applies seamlessly in the digital world where any transmission of data is implicitly a copy). By recursivity, a copy becomes itself an original for the person who receives it, and can make copies of the copy (modified or not).
All licenses need not be worded as carefully, but the Definition itself would be sub-optimal if its wording left this aspect onto the table.--Antoine 00:07, 2 May 2006 (CEST)
In which case I would note that when they say a copy they seem to be talking about a converting an original to a digital work for which the freedoms can then apply. The freedoms mentioned in the above license do not seem to apply to originals at all, in which case I would argue that Art Libre are really talking about digital works in their license and avoiding the whole question of physical works. --Ricardo Gladwell 11:47, 2 May 2006 (CEST)
You understood wrongly. Copies in the Free Art License can be analog copies (e.g. you can make an analog copy of a painting, which gives you a different painting). --Antoine 12:50, 2 May 2006 (CEST)
I'm not sure it does: the license also talks about the freedom to modify copies which implies that the copies themselves must have the property of modifiability which, in my mind at least, implies they should be digital (certainly, painting copies would not seem to fall in this category). I'm not sure you can talk about free software/content freedoms in respect to non-software works without really talking about digital works. --Ricardo Gladwell 14:09, 2 May 2006 (CEST)
Modified copies means the modification can take place in the course of making the copy itself. Like if you make a parodic copy of Mona Lisa: you don't need to first make a verbatim copy and then modify it. However, you can modify a non-digital work (a painting, a piece of sculpture...). --Antoine 14:21, 2 May 2006 (CEST)
I think the license authors should probably clarify that, the license seems to talk about the freedom to "modify the copies", but says nothing about the freedom to modify the reproduction process. Maybe the two things are the same thing but it would seem open to intepretation.
As for modifying painting and sculptures, you are right, of course, you can modify physical works although it's not a freedom I think a lot of people would use for sculptures and it's a freedom generally implicit in physical property law: once you buy a painting or sculpture you are free to do what you like with it. OTOH, digital works need this freedom thanks to advent of things like DRM and non-modifiable file formats such as object code and PDFs. --Ricardo Gladwell 14:34, 2 May 2006 (CEST)
Mmmh, you should not mix up physical property law and "intellectual property" law. Even if you bought the physical medium of an artwork, it doesn't give you the right to modify the work, because the "intellectual property" still belongs to the author (unless the rights were also granted or transferred by contract).
This is especially true in countries where the moral right of the author is protected by the law, by the way. In France, if you buy a painting and then decide to destroy or modify it, the author can sue you (and he will win). --Antoine
Have no fear, Antoine, I'm not mixing the two up at all and explicitly differentiated between them in my remarks above. My point was that when you purchase physical property you are implicitly free to what you like with it (i.e. modify it). That the Art Libre license makes reference to modification of copies seems to imply they are talking about digital works. But I think you didn't address all the points I made above.
Re physical works not being covered by the definition...and "concepts like distribution, copying and modifications"...I understand the problem, but perhaps if we included the concept of a derivative...So, there should be some way that, if I create a physical piece of art, that I can indicate to the world that I am OK with making a derivative (or simple copy where possible) in either the physical or digital realms. Since I like copyleft licenses myself, this should be possible to indicate as well. Getting to a physical object should not be a legal dead end. (zotz - drew Roberts)

Things which are not works of the mind

The word "content" is good for non-material commons, but what about "material commons" like grains, electromagnetic spectrum, genetic information, "commons" in general?. They need also a "free/freedom" definition. --Vjrj 02:13, 3 May 2006 (CEST)

The Definition is about works of the human mind (and craft). It is not only a legal category, but also a philosophical one: creation of works - art works, software works, whatever - is a well-defined philosophical concept. The additions you are proposing do not belong to this category. Trying to find a "one-size-fits-all" ethical message only destroys the meaning of the message and transforms it into a meaningless slogan. But staying inside the boundaries of a clearly defined category of things helps us remain meaningful, and powerful.
Of course, this does not preclude someone else from giving a definition for "freedom of genetic information", "freedom of water resources", "freedom of electromagnetic spectrum", etc. Only, the issues are very different and it would be sterile to try to explain them in the same terms as free contents.
(btw., this kind of thing probably belongs in a FAQ ;-)) --Antoine 02:27, 3 May 2006 (CEST)
Ok, I get your pragmatic point of view. But only one comment. Think about seeds like code: used, improved, copied, studied in the past by every generation. Now we are changing from a free model to a private model (Monsanto's Nightmare and similar). I prefer to use the term Commons that includes Free Content (non-material commons) and material commons. --Vjrj 04:22, 4 May 2006 (CEST)
I understand your opinion, and I could agree if we were trying to write a Manifesto (which can be vague, broad, and very encompassing). But we are trying to write a Definition, which must be precise and based on firm (conceptual) ground ;-)) --Antoine 14:23, 4 May 2006 (CEST)

Source code

User:Erik_Möller said: Is it possible for something to be free content without the "source code" (or something equivalent) being available? Under the current definition, it is. Perhaps we need to find a wording that requires source availabiliy where such sources are essential to modifying the work.

The common position among the Free Art License people (not that I always agree with them ;-)) is that providing the source code is a subcase of allowing to study the work. They argue that the source code is necessary to study software, while there is no such necessity for works of art.
One could counter-argue that even if source code is not necessary to study a piece of non-software content, it is nonetheless very practical for doing modifications (the GNU GPL defines source code as "the preferred form of the work for making modifications to it").
Of course another problem is that some kinds a work do not allow any clear notion of source code. --Antoine 22:11, 1 May 2006 (CEST)
From my own perspective, as a representative of the Free RPG Community there are many examples where a source code is a real requirement. A pen-and-paper roleplaying game can be both an artistic work as well as a functional work. Roleplaying games can be converted into software computer games so source code can be required for this. Many players frequently make house rules and ammendments so access to source code is requisite for this - both to the rules and to the . There are so many reasons to have a source code even from just our narrow field.
This is quite aside from the fact that access to the source code of a free content work might be necessary not only to study the work, but to study how the work was put together. Trivially, this could mean as little as being able to examine a word processor document to see what fonts are being used, etc. A more important example could be being able to examine a layered image file to see how a image effect was put together.
I'm not sure which works you are referring to that do not have a clear notion of a source code: the only such works I am aware of are non-digital works in which case the other freedoms may not apply anyway (see my remarks above). Please see my remarks here for more infomation: Talk:Definition --Ricardo Gladwell 00:00, 2 May 2006 (CEST)
Ricardo, I fullheartedly agree that the distinction between artistic and functional works is a fallacy (I have already said this to various people, including RMS). That is why, in my opinion, the same broad freedoms should apply to free contents as well as free software (the wording of the definition can differ of course, and that's why we are here ;-)).
As for freedom for physical works, you can read my answer to your remark ;-))
But even for digital works, can simple "transparent" binary data be considered the source code for everything? Let's say I have the WAV recording of a concert. What do the bytes tell me about how the guitarist played the strings of his instrument, or even what precise notes he played, with what effects etc.?
Software writing is a symbolic activity (writing code is writing text according to certain conventions). As such, its representation as text (i.e. ASCII or Unicode bytes) is a perfect mirror of the way it functions. It is not necessarily the case with other types of works.
But I do agree that access to some kind of "source code" is important in many cases. Complicated isn't it? --Antoine 00:23, 2 May 2006 (CEST)
It is indeed complicated and I'm glad you also disagree with RMS false distinction between "artistic", "functional" and "political" works.
IMHO I always found the GPL definition of source code (BTW, why isn't the GPL listed as one of the free content licenses?) to be the simplest and most flexible. For example, when start talking about things such as music recordings, photographs, etc, in other words works that originate with physical objects, it is often better to apply the freedoms to the digital copies only. In your example, it is prohibitive to provide the recording studio, instruments, etc that produced the original work, but it is simple enough to provide the "preferred form of the work for making modifications to it", in other words, the uncompressed WAV files or the music in some other modifiable format of the digital copy. --Ricardo Gladwell 12:01, 2 May 2006 (CEST)
The problem with source code in works of art is that it's always hard to determine what is source code. Look at flash movies made by they use photos and drawings to create animated flash movies with words and music which usually are themselves derivative works. So what is source code? Original song with it's score? Original photos? Code they wrote using this or that program for creating flash animations? All of above? I'm not sure such a requirement will have a positive impact on culture. It is important to have an access to work without DRM restrictions. But asking an author for all the pieces he used in the process of creaton is not necessery and practical. There we have a problem of tools: most advanced programs for creating digital works are proprietary. Asking for a source code of a work is of little practical value if we do not have access for tools itself. Following that logic we could ask authors to use only free software tools what itself is a good thing, but this seems simply going too far. My opinion is, that source code is not necessity, much more important is using open formats whenever possible (and please note, that even now it's not easy if you create multimedia works). JaroslawLipszyc 17:24, 2 May 2006 (CEST)
I couldn't disagree more strongly: from my own experience the freedom to modify or study a work is strongly predicated on having access to the source. Without the access to the source, these freedoms are meaningless license verbiage. Again, drawing from my own experiences in the pen-and-paper RPG world, the recent "open gaming movement" used the Open Gaming License was a (partial) copyleft license with no source requirement used to license the Dungeons and Dragons roleplaying system (more popularly called the d20 System). The original d20 system documents whre provided in a modifiable format. However, without the source requirement downstream publishers simply published derivative works in PDFs and hardcopies only, effectively creating a dead-end for the works, taking from the commons but not giving back to it.
I also dispute that it isn't clear what the source code is: the GPL clearly defines this as "preferred form of the work for making modifications" which, in the cases you mention above, obviously resolves the issue for the sources of digital works. In the case of the flash animation, the source is clearly the original SWF file, the image file is its own source, and I don't see an amiguity over that (whether requirements to provide all the elements used to compose a work should also be needed is a seperate discussion).
The argument regarding properietary programs also seems spurious: in the software world not all compilers are free software, cheap or easy to access but that doesn't have any bearing on whether you license software written for these compilers as free software. Whether requiring free software tools for free content is a completely seperate discussion. --Ricardo Gladwell 11:34, 3 May 2006 (CEST)
Ricardo, what you say is "in case of works of art source code means open format". Thus basically what we have to say is very similar, we define terminology differently. JaroslawLipszyc 12:10, 3 May 2006 (CEST)
That's interesting, please expand on your definition of open format: do you think the issue can be resolved by simply requiring publication in open formats, such as ODT? --Ricardo Gladwell 12:30, 3 May 2006 (CEST)
AFAIK, the common definition of open format is a format which is openly documented and (preferably) unencumbered by patents. It is insufficient w.r.t. your criteria, because by this definition PDF and Postcript are open formats (anyone can read and implement the PDF specification). --Antoine 13:15, 3 May 2006 (CEST)
Please note, that sometimes also formats under "reasonable and non-discriminatory license" (RAND) are treated as open, which is a big lie. I'm not sure what status PDF and Postcript have, but it seems that they are simply free. Why it is hard to extract data from PDF? I'm trying to understand in what situation open format is not enough, maybe i'm missing something. JaroslawLipszyc 18:11, 3 May 2006 (CEST)
As Antoine clarified, an open format requirement is fine, but more important is a modifiable, source code format. The problem with PDF and Postscript is that it is possible extract source-like information from the file, but, like software object code, I believe it's tricky and lossy. Can someone who is more familiar with PDF clarify/verify this? --Ricardo Gladwell 18:58, 3 May 2006 (CEST)
The "problem" (actually an understandable design decision) with PDF is that it does not have any notion of textual content (let alone semantic structure). It is a purely graphical rendering language. Say you have text "FOO BAR" in a PDF, the PDF code actually says "render letter F at centimeters (15.34,24.12), render letter O at centimeters (16.78,24.12)", etc. So even a "simple" task like extracting raw text from a PDF needs heuristics to distinguish words, paragraphs, etc. (I had the pleasure of trying to do this in pure PHP and actually get 90% good results for many documents ;-)) --Antoine 22:12, 3 May 2006 (CEST)
So there is no problem. It's lot like a comic strip: PDF is a text in a hostile (graphic) envirement. But if this is authors will, it's OK. We can't outlaw comic strips, right?JaroslawLipszyc 22:33, 3 May 2006 (CEST)
Who said anything about outlawing comic strips or PDF? I'm talking about allowing people the freedom to distribute PDFs, provided that they also supply the original, structured source document which rendered the PDF. The issue here is that its a non trivial task to modify or reverse engineer PDFs. Distributing in only non-modifiable formats (such as PDF or Flash) renders the freedom to modify the work impossible to use, and therefore meaningless. --Ricardo Gladwell 23:53, 3 May 2006 (CEST)
My fault, bad wording. I should talk about declaring non-free, not outlawing (tried to use a metaphore and considerably failed :-) I think you are wrong about this issue. In culture there is always some kind of source code one level deeper. If you have digitally animated movie you have compressed file good for internet sharing, than not-compressed file, analog copies for theater screenings, than you have all the code, background rendering and heroes of your story in polygons or whatever. Without all that it's hard to make some kinds of derivative works, but other kinds are perfectly possible even with low-quality internet clip. In a long run this doesn't really matter. It's the law what is important and lack of any kinds of DRM, not a format characteristic by itself. Printed book is extremely non-modifiable kind of format, but i will not declare all printed works not-free because of that. To copy a book - yes, it is tricky, you need special machines for that, scanners and OCR. But you can do it, legally and practicaly. The same applies to PDF (print it, scan it, OCR it, done), and in 5, 10 or 50 years it will be the same with video or multimedia content.JaroslawLipszyc 17:49, 4 May 2006 (CEST)
Again, who said anything about declaring hard copy books as non-free? Provided the source files are freely distributed for the book the book can be free! Also your point about copying a book is invalid: The situation is very similar to using a decompiler to reverse engineer object code: both processes are highly lossy and difficult which is why the FSF made source code a predicate for free software, and which is why we should do the same for free content.
As I stated above, publishing in a non-modifiable format such as hard copy print or PDF only renders the freedom to modify and study a work meaningless. You either consider the freedom to modify to be as important as the other freedoms (read/display, copy, etc) and we should require a source copy where appropriate or you would prefer they not be exercised in certain situations, in which case this you should explain why you think it is subordinate to the other freedoms and "author's will". --Ricardo Gladwell 11:52, 5 May 2006 (CEST)
I think you make a mistake when you think about freedom only in a digital envirement. When i say about "authors will" i think for example about art-books - individually crafted, unique hard copy books. As long as it is posible to someone who obtained such a book to make a scan and publish it and make derivatie works i will consider the book free. You not, because it's author have chosen deliberately a format (art form) not really suitable for any kind of digitalisation and manipulation. In this example to comply with a free license as you see it author of such a book has to create a completely different art object. That simply doesn't make sense. I do not agree that format characteristic rather than legal status of work and its medium makes a work non-free. This discussion gets lenghty, i suggest to wait and see what others have to say on this topic.JaroslawLipszyc 00:07, 6 May 2006 (CEST)
My point is that, since in most cases the original source documents already exist, it does not seem particularly onorous to require these also be distrubuted. I guess there could be an exemption where no such documents exist. But, clearly the freedom to modify is a legal right enabled by the technical format: formats that restrict modification neuter the freedom to modify. But, I note you do not answer my question: do you not think the freedom to modify is subordinate to the other freedoms? How does the freedom to modify fit in with your view of free content? --Ricardo Gladwell 21:23, 6 May 2006 (CEST)
Extracting anything from CVS is tricky, and any given video codec is lossy. We cant demand user friendliness, as well as we can't demand minimal resolution of a movie. My problem is different: what to do with flash, which obviously is not an open format. Are we to declare all flash works non-free by default? JaroslawLipszyc 21:29, 3 May 2006 (CEST)
Again, the discussion about requiring open/transparent formats is seperate to the discussion about requiring the source be distributed with free content. If you would like to discuss this second issue, please create a seperate heading for it. --Ricardo Gladwell 23:53, 3 May 2006 (CEST)
I think this is strictly connected, see aboveJaroslawLipszyc 17:49, 4 May 2006 (CEST)
I don't think I read anything about open/transparent sources in your above post. You can require source code without also requiring that source be open/transparent as well, so the two issues can be clearly seperated. Of more importance is whether we require source code and I would prefer to focus on that for the time being. --Ricardo Gladwell 11:52, 5 May 2006 (CEST)

Wording proposal

Maybe we can try to find a wording that makes source code access mandatory where it has a meaning. Here is a rough proposal (English mistakes notwithstanding!):

  • Source code access: if the work or part of it is generated by computation of a modifiable symbolic form (e.g. textual), the user must be given access to this modifiable symbolic form (often called source code).

The important question is whether the expression "generated by computation of a modifiable symbolic form" 1) covers all (or most) cases where source code access is desirable 2) doesn't mistakingly cover cases where it is meaningless. If it does have these properties, then source code access can become a fundamental freedom. --Antoine 22:48, 14 May 2006 (CEST)

Hmmm, reading my proposal again I do not find it very good. It is too complicated and may rather belong to a separate page explaining source code.

Here is another generic attempt:

  • Freedom to modify: if materially possible, the work must be available under a form which does not make modification unreasonably difficult (see the page about source code criteria).

--Antoine 05:05, 15 May 2006 (CEST)

It's a very good start, but I feel the definition is a bit complicated and specifies digital works when, maybe, it doesn't have to. I would also note that it is not enough to require people publish in a form that "does not make modification unreasonably difficult": people should still be allowed to publish in "closed" forms provided they also supply an "open" (modifiable, structured, etc) form. To expand, we need several things from the source code:
  1. modifiable - the ability to modify all parts of a work.
  2. structured - the ability to study how the work is put together
  3. original form - the ability to rebuild the work from scratch, to allow modifications and redistribution without loss of information (for example, photocopying is a poor cousin of true redistribution from original sources). Also, the ability to not only study the work, but how the work is built.
(NB This is probably not an exhaustive list: suggestions welcome)
It's a simpler thing to copy (with some interpolation) the FSF defininition and GPL definition of source code and modify the freedoms to study modify as follows:
the freedom to study the work and to apply knowledge acquired from it. access to the work in the preferred form for making modifications to it, where available, is a prerequisite for this
the freedom to make improvements or other changes, and to release modified copies. access to the work in the preferred form for making modifications to it, where available, is a prerequisite for this
The key being the part that says "where available" meaning this definition should only be used for those works where it makes sense to have a "preferred form of the work for making modifications to it" or where such a form is available, exempting those works (such as paintings) that have no such form. --Ricardo Gladwell 11:09, 15 May 2006 (CEST)
You are right, structured is much more adequate than symbolic.
As for wording in the definition itself, I don't have any strong opinion. I hope Eric and Benjamin will soon update the unstable version according to all the suggested modifications; then we will be able to have a clearer view ;) --Antoine 19:23, 20 May 2006 (CEST)
What is wrong with the GPL definition of source code: "the preferred form of the work for making modifications to it."? It takes care of the case of an art book made from 250 35mm photos: the source code would probably be the book itself. It takes care of the common cases, like a PDF file: unless it was hand-crafted, you also need the TeX/OpenOffice/HTML/etc. document that produced it. The Open Software License adds "...and all available documentation describing how to modify the Original Work," an addition that I like. --Sanjoy Mahajan 01:59, 24 May 2006 (+0100)
I suppose the problem as stated by others is that the preferred form may not exist for all kinds of works, for example sculptures, which is why I ammended the GPL source code definition in my comments above to include the proviso "where available".
I'm not at all sure about the OSL requirement to provide "all available documentation describing how to modify the Original Work" - does this mean I have to provide all the Writer help documentation if I distribute a single ODF file? --Ricardo Gladwell 14:24, 24 May 2006 (CEST)
I suppose it must be interpreted to only cover parts of that documentation specific to the distributed work. In other words, you shouldn't have to distribute generic operating system or library documentation. However, you have to provide makefiles, specific tool options and instructions on how to generate the work in object form, etc. If the work does not require any specific instructions which don't apply generically to other works of the same type, then I suppose you don't have to ship any documentation at all.
Apparently, the exact same license is also available under the name "Academic Free License". --Antoine 23:24, 24 May 2006 (CEST)

Free expression?

User:Rgladwell said: Free expression is not the same as free content: you can have the right to free expresssion without having free content, and free content does not guarantee free expression.

Also in Spanish "Libertad de expresión" is the translation of "Free Speech", so that would be officially confusing. -- A Spanish Speaker.

In French too, "Liberté d'expression" (freedom of expression) is more widespread than "Liberté de parole" (freedom of speech). By the way, according to Wikipedia there is a similar possibility in English: "The synonymous term freedom of expression is sometimes preferred [to the term freedom of speech]". --Antoine 19:17, 9 May 2006 (CEST)

Right to use and perform

Apparently the current definition takes as implicit the first freedom of free software (the right to use/execute). I suggest to make it explicit, because it allows to include "related rights" such as the right to perform (e.g. a play, a song...):

  • The freedom to use and perform the work: The licensee must be allowed to make any use, private or public, of the work. This includes all derived uses ("related rights") such as performing or interpreting the work.

--Antoine 01:15, 2 May 2006 (CEST)

I agree that the freedom to perform is important, especially with regard to creative works such as music and film/video, and should be made more explicit. Here is my suggestion:

  • The freedom to perform the work, or otherwise display it for any purpose: The work may be performed, broadcast, webcast, or displayed in any manner in public or private.

I feel like "any use" as referred to above by Antoine is a bit too broad here, but perhaps it would make sense to add the derived uses portion?

--Elizabeth stark 20:33, 4 May 2006 (CEST)

I would advocate keeping the "freedom to use" in the freedom title as well ("use" meaning execute a program, watch a movie, etc.). It may be granted in most of today's legal systems, but the Definition should be independent from legal systems, and most readers won't know much about the law. So I think it's a healthy reminder.
Unless you think "use" is too broad a term. What problems could this entail? --Antoine 12:41, 5 May 2006 (CEST)
Regarding this point, I suggested earlier on CC-license mailinglist that GFDL does not seem to grant freedom to perform. It allows us to reproduce, create a derivative work, create a compilation, translate, distribute any of these works, but, well, is not clear about performance of various kinds. My wild guess is that it was not expected that the license would be used for something far remote from software manuals as to make it meaningful to perform a licensed document.
I agree, though, that the freedom to perform is very important. How about making it a recommendation?Tomos 14:53, 14 May 2006 (CEST)
I think freedom to perform should be part of the basic freedoms required for a "performable work" to be called "free". Then we can make more subtle decisions and commentary in our license list (like "the GFDL is not adapted for situations where the work can reasonably be performed"). It is quite a similar problem to source code: we must find a way to state that in some situations source code access is mandatory, while recognizing that there are situations where it has no meaning.
I'm not sure how to understand the GFDL. It says "Such a notice grants a world-wide, royalty-free license, unlimited in duration, to use that work under the conditions stated herein." Does "use that work" comprise public performance? The answer might depend on the jurisdiction. --Antoine 22:36, 14 May 2006 (CEST)

Right to study and apply the "information"

I find the phrasing "study and apply the information" a bit confusing. The word "information" seems to imply that only the symbolic (e.g. textual) content of the work can be studied, not its inner workings or the nuances of its making. Also, as pointed elsewhere, "information" is really sterile.

I would rephrase it:

  • The freedom to study the work and use the knowledge gained from the work: The licensee must be allowed to examine the work and to use the knowledge gained from the work in any way. The license may not, for example, restrict "reverse engineering".

--Antoine 01:29, 2 May 2006 (CEST)

Applying Knowledge

First freedom now is worded as "the freedom to study the work and to apply knowledge acquired from it". This may however be a problem when you think about it in context of scientific research. Copyright is covering just a publication, but there may be (and usually are) also included other "IP" law such as patents. I understand "applying knowledge" as a very wide term. I'm not aware of any exisiting license which may guarantee that patents will be not used. However this is a very good statement if we want to describe our ideals.JaroslawLipszyc 02:06, 2 May 2006 (CEST)

I think it is an absolute fact that free content cannot be impeded by patents. We need to work towards a clearer distinction, I think, between expectations we have from the works which are called free content, and from the licenses which are used for free content. I don't need to use a complex license that has clauses about patents if that is never relevant to me. Similarly, regarding earlier discussions about source code, I don't need anything in my license about source code if I'm talking about an essentially transparent work such as an essay. So I think we need to work towards emphasizing this distinction more clearly, and then formulate precisely what expectations there are from a work that is considered free content.--Erik Möller 02:35, 2 May 2006 (CEST)
And maybe "freedom to study" is good enough as a license requirement? JaroslawLipszyc 03:38, 2 May 2006 (CEST)
This is one of the cases where the absence of a clause restricting freedoms may already meet the requirements, I think. Generally, you're free to study and apply knowledge from a work, unless the terms under which you have done that restrict you from doing so. What I think is important, especially if we want this definition to be a superset of the free software definition, is that we make it clear that the work itself must be available in a format suitable for modification. Whether or not the license protects that state of transparency is optional.--Erik Möller 09:54, 2 May 2006 (CEST)

Non-commercial not free?

It seems very arbitrary to call popular non-commercial licenses such as CreativeCommons Attrbution-NonCommercial-ShareAlike non-free! I believe all current usages of the term free content or free license allow a non-commercial clause. Very little content will be free under this definition as few content creators allow other people to profit from their hard work without paying royalties. I see no reason why a license prohibiting the receiver from profiting from content he got for nothing can not be called free.

In fact, I think CC A-NC-SA is the model minimum free license, as it makes use of most of the restriction clauses permitted here plus NonCommercial. This definition should allow non-commercial licenses to fall under this umbrella. The interests of Wikimedia can still be satisfied by saying only Comercial Free Content and Fair Use were allowed. Prohibiting the reuser from making profit serves to protect the free status of the content, thus justifying its use as a restriction in the license. 06:55, 2 May 2006 (CEST)

You can find my thoughts on the Creative Commons NC licenses here. The idea that any kind of commercial venture is an "evil" from which people must be "protected" is a highly fallacious one; indeed, it is exactly the commercial use of free content and free software that has enabled some of the most exciting projects today, ranging from the adoption of free software in large sections of industry to the integration of Wikipedia content into search engines, CD-ROMs, and existing lookup tools. Especially in a collaborative context, it is essential to grant this freedom from the start, as it is virtually impossible to get permission from any contributor to a wiki for a particular commercial use.
It is perfectly alright for authors and artists not to reliniquish rights in accordance with this definition. But the resulting works should not be called "free". There can be no compromise on the core freedoms in this definition. If anything, they may need to be expanded.--Erik Möller 09:15, 2 May 2006 (CEST)
I think that the provision on free redistribution makes it quite impossible to simply "profit from others' hard work". If someone tried to derive unfair advantage from a work, then people are free to reject the offer and are very likely to find the work that the offer is based on under more reasonable terms. In fact, if a share-alike or copyleft clause is used, then you'd be likely to find whatever was added to the original work freely available too. NC clauses do nothing to prevent unfair usage, but will prevent many legit uses.--Kari Pahula 11:33, 2 May 2006 (CEST)
Not to allow commercial re-use is a critical limitation, as Erik aptly pointed out in his essays, e.g. OpenSource-Jahrbuch 2006. So this content simply is NOT free. That's why Stallman does not support CC anymore, because the different models get to easily thrown into a single bucket altogether. The importance about the commercial clause is that some value-additions are ONLY possible in a commercial context, when it is possible to integrate the new content e.g. in Websites with advertisements, or sell packaged DVDs etc. To protect the idea of free it is enough to add Copyleft provisions. 17:13, 2 May 2006 (CEST)
There is no such thing as simply profiting from others hard work. You can propagate, advertise and store their work, but these are all of benefit to the work's author (although this may require or at least be helped by the moral right of paternity). NC is a fundamental error, and the definition of NC is causing headaches. NC is another layer of permission culture, and is not Free. The FSF and DFSG freedom definitions require that work be allowed to be sold, so it is not true that they allow NC clauses. --Rob Myers 19:39, 2 May 2006 (CEST)
So - not surprisingly, I guess - I think that it is important to give some kind of recognition to those licenses that grant freedoms that are more flexible than an "all rights reserved" or "personal-use only" license. While I agree that it is not possible to accommodate those licenses within the current definition of "free" being debated here, I do think that these can and should be classified under a different brand of licenses that enable important freedoms that are palatable to different industries than would contemplate a "free" license. --Mia Garlick 18:09, 2 May 2006 (PDT)
The distinction between commercial and non-commercial also hinders the evolution of an open society where there is no commerial action anymore ;-) -- 08:54, 4 May 2006 (CEST)
Mia I think that the various 'by the grace of the author' permission licenses like NC and ND are only less-unfree. Their palatability to any industries is not an issue if we are concentrating on whatever we believe "true" freedom to be. We should simply ignore them here, possibly noting that there are other licenses that give some permissions but not real freedom and warning people away from them. That said the FSF do maintain a page of licenses that explains whether a license is free and if not why not, so maybe that would be a good model to imitate. --Rob Myers 20:09, 7 May 2006 (CEST)

Other Free Content Definitions

A took the time to compile a list of other free content definitions, feel free to update: -- Ricardo Gladwell 14:16, 2 May 2006 (CEST)

Edited & added fc-uk definitions --Rob Myers 19:34, 2 May 2006 (CEST)


I see license compatibility as a freedom, too. The more standarised they are the more works can be combined and reused. This is crucial for a free culture movement: we have to avoid balkanisation. I suggest changing reccomendation section to reflect that threat. I believe giving recommendations for CC-BY and CC-BY-SA licenses (they are most popular) for all publicatons and art, and GFDL (Wikipedia standard) for reference content and textbooks. JaroslawLipszyc 22:22, 3 May 2006 (CEST)

I'm not sure recommandations are our concern, apart from telling which licenses satisfy to the definition. I'd rather have a characterization of licenses and let people choose (see my proposal on the Talk:Licenses page). Also, popularity is not a sufficient metric IMHO ;-)) --Antoine 01:17, 4 May 2006 (CEST)
Popularity doesn't matter when we think about freedom, but once we settle with that it's all we should take care about :) I do not think that recommendations should be a basic part of definition, but hey, this what people will look for! Sorry for being pragmatic, but i think we should also think about "net-effect" which works under free licenses may (or may not) create.JaroslawLipszyc 00:27, 6 May 2006 (CEST)
I agree with Antoine, we should remain as license neutral as possible and simply list those licenses that fit with our definition of free content, rather than favour any license over another. We should especially not distinguish between "artistic" works and "technical" works: that path leads to non-free Invariant Sections. --Ricardo Gladwell 12:55, 7 May 2006 (CEST)

Moral right

This is a somewhat off-topic subthread about moral right. I put it here for reference, but we should find a way of putting offtopic or obsolete discussions elsewhere so as to keep a clean discussion page ;-)) --Antoine 01:03, 4 May 2006 (CEST)

Mmm... I'm not at all sure about your point about moral rights and modifying sculptures and paintings. I don't know a lot about moral rights law, but surely that would only apply if you publicly distributed the work or used the author's name in relation to the modified work? So are we talking about the freedom to modify or the freedom to distribute? --Ricardo Gladwell 15:11, 2 May 2006 (CEST)

No, not only. Moral right includes the right to the integrity of the work, so if you modify a work (even privately), you infringe on the moral right of the author. Moral right considers that the work is part of the personality of the author. (of course, in practice, the author cannot exercise his moral right if he doesn't learn about the modification or destruction of the work) --Antoine 15:23, 2 May 2006 (CEST)
Dear Antoine, you don't know moral rights: the author shall have the right to claim authorship of the work and to object to any distortion, mutilation or other modification of, or other derogatory action in relation to, the said work, which would be prejudicial to his honor or reputation. If you modify a work, you infringe economic rights. You infringe moral rights only if your modification prejudices honor or reputation of the author. --Francesco Baldini 22:44, 3 May 2006 (CEST)
Francesco, I do know moral rights ;) The Berne convention is a generic formulation of moral rights, but it does not dictate how states implement it in practice, and implementations are widely varying. In French right at least, it is traditionnally the author who decides what prejudices his honor or reputation. Thus, making any modification to a work, even privately, can be forbidden by moral right if the author decides so. One must understand that the moral right is viewed as a fundamental right of the author (at the same level as freedom of expression, for example). So courts usually interpret it in a manner very favorable to the author.
Anyway, this not really the main subject here. --Antoine 23:11, 3 May 2006 (CEST)
Daer Antoine, the Berne Covention isn't a generic formulation of moral right! :)))) In all the nations the author decides what prejudices his honor or reputation! :))) Ignorance in this matter is inexcusable: I study law and I think that you must understand that you don't know moral rights. You are confounding the right of integrity (a moral right) with the exclusive right to modification (an economic right). --Francesco Baldini 00:06, 4 May 2006 (CEST)
(rewritten answer) Ok, I think you misunderstood me. I didn't say that modification always infringes moral right; I said it potentially infringes it. That's exactly what your excerpt of the Berne Convention says, by the way. --Antoine 01:03, 4 May 2006 (CEST)
By the way, moral right is really off-topic in the original thread, so I think I'm gonna detach the whole sub-thread and put it elsewhere.
...oh, and please avoid appeal to authority ("I study law"), thanks.... --Antoine 00:55, 4 May 2006 (CEST)
Sorry Antoine, but I'm not a stupid. :)

You said:

1. "The Berne convention is a generic formulation of moral rights": false.

2. "In French right at least, it is traditionnally the author who decides what prejudices his honor or reputation": in all the countries!

3. "In practice, the author cannot exercise his moral right if he doesn't learn about the modification or destruction of the work": false

4. "if you modify a work (even privately), you infringe on the moral right of the author": false (where is "potentially"???)

I study law: this is only a fact. You can eliminate this topic and your errors, but you can eliminate your ignorance in matter. ;))) It seems that you don't accept the truth. Why?--Francesco Baldini 11:02, 4 May 2006 (CEST)

You are trying to split hairs.
1. Berne Convention is a generic formulation of moral rights (as well as other rights, of course - my wording was elliptic, but since you know the Berne Convention it seemed obvious that you were able to understand it)
2. If it's true in all countries, it's true in France, isn't it? (but it's not true in countries which don't have a moral right, you know)
3. This is a simple matter of logic, you cannot exercise your moral right if you don't know something has happened to your work. How could you sue for something you don't know has happened?
4. "Potentially" was implied by previous messages in the thread, where I explained that the author could sue for modification on the basis of moral right. To understand a message, it's often useful to read previous messages in the thread...
And why do you think I'm gonna eliminate the topic? That's ridiculous.
--Antoine 14:20, 4 May 2006 (CEST)

Antoine, you are an incredible sophist! :-D Clap clap clap! I think that you prefer to kill a children rather than admit your evident errors. :-D End of thread for me. You have the last word. ;-) --Francesco Baldini 14:54, 4 May 2006 (CEST)

I like that statement ("you prefer kill a children rather than admit your evident errors"), although I don't know if it applies to me - honestly I've never thought about killing a children, perhaps I should give it a try. Thanks for the good laugh :-))
On a more serious note, perhaps I should be wary of sophistry indeed. Hmmm. --Antoine 15:01, 4 May 2006 (CEST)


I have no problem with the idea of portals in some sort of unofficial capacity and on this wiki. I think we should encourage it. However, placing the link it in the definition, at least as it has been suggested/worded here, makes me a little uneasy. -- Benjamin Mako Hill 17:18, 4 May 2006 (CEST)

Could you be more specific about how it makes you uneasy, or just edit the description?--Erik Möller 21:35, 4 May 2006 (CEST)

God et al.

A couple of comments:

  • "Any original work of authorship is copyrighted. Under copyright law, authors are considered God-like "creators" and are given legal powers they can use against those who duplicate "their" content in altered or unaltered form."
    • This sounds off (the God bit -- which notion of a God), and the scare quotes on "their." I'd say something like "Copyright grants an author a monopoly on certain actions ..."
  • can the license require attribution be removed from any derived works? -- reagle

Scope of the rights covered

Hi. Thanks for your effort& initiative. I sense that this definition does not limit its coverage to copyrights. Or am I mistaken? I think that it is not necessarily bad that the definition deals with a broad range of rights (any use is allowed), but it comes with some consequences. Below is some exploration of the issue. But please be aware that I'm not a legal expert.

(Copyright-only licenses) When a license is copyright-only license, does it fail to meet the definition? If I remember correctly, GFDL is a copyright license (as opposed to a contract, which is governed by state laws in the U.S. legal context); the licensor gives copyright permission he has over a licensed work. Other rights he might have - such as right to privacy/publicity if it is his self-portrait, trademark, etc. are not covered. This, of course, would place a limit to legally acceptable use of the licensed material.

(Third party rights) The currently existing free licenses do not usually cover rights held by the third parties. For example, if I release a photo of a building with a passer-by, I give copyright permission regarding the photo's creative expression, but not the copyright subsisting in the building (as a work of an architecture), or the privacy right of the passer-by. I personally think that is a reasonable - though not necessarily novice-reuser-friendly - practice.

The aim of the definition is to advocate something other than the monopoly that is currently granted to the auther of a work, i.e. to change the idea that authorship is a good reason to take away rights from the user. Other barriers to re-use can be discussed but most of the ones you cite as examples seem reasonable to me. I don't think we want to negate the people's right to privacy, self-image, etc.
On the other hand we could make the point that a free content work (or part of it) should not be subject to a trademark, because trademarks are the same basic idea of saying authorship (not exactly authorship, but...) is a good reason to take away rights from the user. This can probably be subject to discussion ;-)) --Antoine 15:21, 7 May 2006 (CEST)
Hopefully Creative Commons will tackle trademarks in the future. At this point in time Trademarks haven't really been considered, but they need to be considered as they affect artists probably as often as copyrights. Free Software seems happy to allow Trademarks to be used to protect the identity of software projects, but with strengthening Trademark protection law in the US this does not sit well with more general cultural freedom and the need to be able to depict and comment on trademarks as part of cultural criticism or atistic creativity. I do not know (and I do not believe anyone knows) how Freedom interacts with privacy and other rights. I think we should focus on defining creative freedom and then see how it might interact with other freedoms or rights later. --Rob Myers 21:26, 7 May 2006 (CEST)
Are trademarks a serious problem? Works that include protected trademarks are non-free, but if they're under a free software/content license it's trivial to strip all references to the trademark and make the work free? --Ricardo Gladwell 11:29, 9 May 2006 (CEST)

(Copyright in previous works by the same author) Perhaps a minor issue, but there is a kind of counter part to the third party rights. Some work might contain copyright works (or a part of it) from the same author. Imagine I develop a very detailed world & character settings for (and through) a fantasy novel or a sit-com. When I make use of the settings and create a sequel, and release only that sequel in an open content license, should the world setting etc. be also "freed"? When a serial cartoon strips has only one episode as free content, which part of the work should be free? All of the work? Or just the part that is new to that episode, but not the part shared with previous works? I think free licenses are not specific regarding this issue.

(Defamation) I think free licenses do not allow a licensed text to be modified as to defame the licensor. Suppose I write a short bio of myself, and someone modifies it to harm my social reputation with some false information. Is the use within the scope of "use" permitted in the license? More importantly, should a free license permit such use? I suppose a widely-acceptable answer to this question would be "no".

I think it is acceptable indeed ;-)) (it is the same problem as above really: free content does not aim to change the norm on other subjects than the problem of making works a priori re-usable by everyone) --Antoine 15:21, 7 May 2006 (CEST)
It is acceptible under the idea of freedom, as defamation is simply a modification of the work (which freedom requires to be allowed). It may not be morally acceptible though. And a copyright license cannot undo the libel or slander laws of a country. Which is of particular concern in the UK, where out libel laws are quite draconian. :-) --Rob Myers 21:26, 7 May 2006 (CEST)

But another type of defamation is arguably permitted by a license. (I am not sure..) Moral rights defined by the Berne Convention include author's right to reputation (or right of not used his work to harm his reputation). For example, even when I am not making a false harmful assertion about the author, my use of his work might harm his reputation in thise Berne Convention's sense. Is this type of use permitted by free content licenses? Should that be a requirement?

It is permitted a priori. However, the moral right of the author cannot be waived nor selled and it is eternal. Thus, a license or contract can absolutely nothing against it. An author may decide to object to a modification that has been done to a free work of his. Of course, it will harm his trustability in a free content ecosystem. --Antoine 15:21, 7 May 2006 (CEST)
Moral rights (there are several) can be waived in the UK and Canada, and are very weak in the USA. They are inalienable in most of the EU, and often include rights not covered by Berne. --Rob Myers 21:26, 7 May 2006 (CEST)

CC licenses include a requirement that licensee remove attribution from a derivative/collective work using the licensed work, upon licensor's request.

CC licenses (those who are free) are a particular implementation of free content licenses, and some of their clauses can be surprising w.r.t. the Definition ;-) The clause you are mentioning looks like a bizarre way to introduce a part of moral right in the license, for those cases where an author cannot rely on a real moral right (because they are in the wrong countries). We shouldn't follow such a path in our Definition; we should stick to precise but legally agnostic principles.
One of the Free Art License authors remarked that that clause could lead to two branches of modifications with different attribution requirements, and that merging the two branches would stumble on the question of whether or not to mention authorship for someone who asked to be removed from attribution in only one of the two branches. --Antoine 15:21, 7 May 2006 (CEST)

Sorry for just raising issues without going into drafting!

Tomos 10:05, 7 May 2006 (CEST)

Non-discrimination requirement

It might be too obvious, but I think it is not a bad idea to include a requirement that the license is for all members of the public, not for select group of people, without excluding anyone.

I got the idea from Open Knowledge Foundation's definition.

Some licenses might terminate when a licensee violates the term, and the licensee is not eligible anymore to use the work. I think a license like that is still free.

Tomos 08:12, 9 May 2006 (CEST)

Not a license

I think we need to make it much clearer upfront that the FCD is not a license, but a standard of freedom to measure both works and licenses. I fear that confused people will start putting their works "under the Free Content Definition" and end up not licensing it at all. As noted above, I also think we need to state that being under a free license is a necessary, but not sufficient condition for the work to be considered free. There may be other issues which preclude it from being free, such as the unavailability of a modifiable version, or the encumberance by patents.

I'll make some edits in the next few days, taking this and other comments on this page into account.--Erik Möller 03:24, 15 May 2006 (CEST)

I agree with your main point (not a license), but more importantly, I see the other point you make - importance of emphasizing "necessary but not sufficient" aspect of the definition.
As I suggested regarding fair use issue, some freedom might be incorporated in the "recommended" section. It is not the aim of this project to promote the maximum possible freedom, nor is it to preach people (esp. license designers) that existing licenses do not offer some of the freedoms that are seemingly basic (such as patent). But we can offer explanations on what restrictions users of a "free content/expression/work" are subject to. Tomos 12:56, 19 May 2006 (CEST)
Actually, I disagree: this project is very much about seperating those licenses that are free and non-free. It's not about telling people how to publish their licenses but it is about creating a basic standard of freedom that will be, by definition, exclusive. If we start making exceptions the definition will quickly become pointless.
I agree with Erik that this definition should focus on works, not licenses. A work is either free or non-free. Licenses are only free or non-free in the sense that they enable free works or not. --Ricardo Gladwell 14:33, 24 May 2006 (CEST)

Transparent copies

Allowing such a clause is an invitation to miserable license drafting. For example, the GFDL v1.2 says: "Opaque formats include proprietary formats that can be read and edited only by proprietary word processors..." The MS-Word format and its specification is proprietary, but many documents written in it can be edited by OpenOffice. So, is it okay to release a GFDL work as a Word document as long as OpenOffice can handle that particular document?

Rather then settle such issues, I think it's better to just forget the whole issue and disallow such requirements. The GPL v2, which has worked well for software, doesn't forbid you from releasing source in a language for which there is no free compiler (e.g. you could release C++ code even before g++ existed). The GFDL has a transparent-copies requirement, but the GFDL is not my model of an ideal license.

No this will be disastrous. If I can take your Free work, extend it and then encode it in a proprietary format that only I sell software to read, then your Freedom is removed and any cultural license is rendered ineffective. --Rob Myers 20:26, 14 June 2006 (CEST)

If someone releases a document in a non-transparent format, you can usually find someone to convert it to a reasonable format for you, e.g. Word to OpenDocument (with the new OpenDocument plugin for Word). It has to be freed only once.

A similar issue arises over whether people should be able to charge for the binary (back to the GPL and software here). Sure, let them. If they charge too much, several people can pool and buy it and then release it for free. So the market will take care of such cases. Similarly, why not let the 'format' market take care of non-transparent formats?

If people worry about format balkanization (e.g. TeX and Word versions of a textbook running in parallel), that's true but it's independent of transparent versus non-transparent formats. With the GPL and software, there can be Pascal (god help us) and C versions of a program developed in parallel. This doesn't seem to cause a problem and may be beneficial. Many but not all physicists use TeX; others use Word. To welcome everyone into the community contributing to a physics textbook, it's useful to have a Word version -- and people working on the TeX version can study it, get ideas.

But if I write a Pascal module it can be linked to a C module. If I have a piece of work that cannot be used by people without them using my proprietary software, that is much more of a problem. Being allowed to look at a Word document is not "free enough". --Rob Myers 20:26, 14 June 2006 (CEST)

A related issue is fonts (maybe should be a separate topic?). Can a document made with non-free fonts be free? Yes, because as long as you have source code you can make a version with similar but free fonts. No, because you cannot regenerate the document with free tools. I don't know the answer.

The text is free, the typography isn't. These will have separate copyrights anyway (in English law at least...) --Rob Myers 20:26, 14 June 2006 (CEST)

Sorry for the bit of a ramble. I'm trying to draft a free license for my physics textbook so I'm thinking aloud here... --Sanjoy Mahajan, 13:22, 16 May 2006 (EDT)

I disagree that transparent copies necesarily cause license problems: do you have any specific examples of such problems. I do agree, the GFDL is poorly worded here (and elsewhere) but that doesn't mean that a workable, universal transparency clause is impossible.
Transparency addresses the legitimate issue that, by allowing distribution in proprietary formats, vendor lock-in may force a levy on the ability to read and modify free content and thus make works non-free. At the moment we can rely on the ability for developers to legally reverse engineer properiatary formats and create free software tools, but legal frameworks may not support this in the future.
I would also not read too much into the lack of a transparency clause in the GPL: the absence may be down to the simple fact that the GPL must maintain backwards compatability with previous versions. I believe RMS has decried the use of proprietary software tools to create free softare (for example, Java).
GPL 3 adresses transparency. --Rob Myers 20:26, 14 June 2006 (CEST)
I suspected as much but could not find confirmation. I thought the transaparency in the GPL only applied to "Non-Source" distribution? --Ricardo Gladwell 18:18, 15 June 2006 (CEST)
In your example about non-free fonts, I would note that the content (or text) of the document is free, but the document itself is not. --Ricardo Gladwell 19:37, 16 May 2006 (CEST)
At first I agreed, but now I am not sure. One issue is how free the fonts have to be. If they are licensed merely for use within a document, but not for modification of the glyphs, then the font is not free according to the most free-software definitions; but surely an otherwise free document using such a font (e.g. Palatino?) would be free?
Also, what about trademarks? The issues there are similar to free fonts. Linux and Debian are trademarks, so one cannot blithely take a Debian CD, delete nine-tenths of it, and then redistribute it as Debian. Similarly, one cannot take the Linux source code and distribute five files from it and call it Linux. Surely the Linux executable is free software even though you have to rename what you redistribute. So I don't know how to distinguish between a document and its contents in a general way. It seems to be a special-case solution for printed textual matter, which means it'll be a nightmare to draft a covering clause for all documents, let alone works of art.
One way that might work: "A recipient must be able to recreate the work from the source using (if any software is needed) only free software and/or free content." Ack, it gets a bit circular. -
I'm not really sure what you are trying to say, especially wrt fonts. True, there are trademarks for Linux and Debian but I believe both trademarks are licensed under separate trademark licenses[1] so I believe you can completely modify and redistribute with both. Even if you could not, it is a trivial obstacle to redistribute these works with the trademarks removed.
You can think of a digital document as a piece of software to display and print a textual work, and the content of the document being the actual words: you can export the words of a document to raw text and still read the work, albeit at reduced quality and at a loss of information (emphasis, formatting, etc).
But I think we are getting confused again: Transparency, as defined in the FDL, does not mean that you must supply all components of a work, only that you provide the source of a work in an open, standard, clear format. I think this new requirement we are discussing is separate, but also important. Not sure what we should call it. --Ricardo Gladwell 14:28, 24 May 2006 (CEST)

Sorry to be so slow. I find wiki editing misreable compared to reading and replying to email threads in emacs. But you're right, I'm losing track of the distinctions. It might be easier with examples. So, what do you think about source code in the following formats:

1. Quark Xpress (.qxd): A proprietary binary format with half-done attempts at reverse engineering, and can be used only by a proprietary program.

2. XML for a proprietary word processor but with a public DTD: Perhaps Quark XML is such a beast, though I don't know whether the scheme or DTD is public or available only after signing an NDA.

3. ConTeXt: I don't know whether ConTeXt (a layer upon plain TeX, and an alternative to LaTeX) is publicly documented. The documentation is often years out of date and ConTeXt develops very fast, but the mailing list and developers solve almost all problems.

4. GIF for images: The format is publicly documented, but it has patent interferences.

5. Word 2000 file that cannot be imported cleanly by OpenOffice.

6. Word 2000 file that can be imported by OpenOffice 1 or 2.

7. Word 2000 file that can be imported by OpenOffice 2 but not OpenOffice 1.

I think 1, 4, and 5 are not free; 3 is free; and I'm not sure about 2, 6, and 7. 2 would be caught by a clause such as "Source must be suitable not only for non-free-software programs." 6 and 7 are borderline. DOC is not publicly documented, but if an available OpenOffice can accept it, then the source seems free (even if the format isn't in general). So a clause such requiring source in a "publicly documented format" might be too strong?

Another issue is the DFSG. Paragraph #3 says "The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software." Does that forbid forbidding people from converting say TeX source to Word and distributing it? Not that the DFSG should apply verbatim to documents, but it's a good start.

One issue, related to case 2 above, is what we mean by 'publicly documented format'. A narrow interpretation is that the source should be in something like ASCII or UTF8 (or tar.gz of that), since those formats are documented. But that the ASCII could represent formatting codes for a proprietary processor, e.g. for troff before it was free software as groff. A wider interpretation is that the formatting codes be for a publicly documented or free-software program, e.g. Inkscape. --Sanjoy Mahajan 11:53, 5 June 2006 (+0100)

Do you have an email address I can contact you on so we can continue this discussion off-line? TIA --Ricardo Gladwell 12:10, 9 June 2006 (CEST)
Just got back from holiday. It's --Sanjoy Mahajan, 23:50, 19 June 2006

Is anyone there?

I'm not sure what happened but suddenly comments have dropped off dramatically. Is anyone still reading this Wiki? -Ricardo Gladwell 23:07, 18 May 2006 (CEST)

Benjamin and I most certainly are. It's not a bad thing if discussion drops off a bit; this will allow us, as initiators, to spend some time digesting all the comments and working on a new version of the definition. Benjamin is currently a bit busy, so I'm also waiting for him to catch up on some open questions, most importantly the naming question (I've had a long e-mail thread with Lessig behind the scenes about that).--Erik Möller 11:47, 19 May 2006 (CEST)
Should we not be keeping the discussion public and open on this topic? --Ricardo Gladwell 12:11, 19 May 2006 (CEST)
I'm not here everyday, but I am catching up the discussion at least once a week. I intend to do that for a while. From User talk:Antoine, I believe Antoine is around/ will be back. Tomos 13:00, 19 May 2006 (CEST)
I try to keep things as open as possible. However, sometimes it's simply easiest to have a phone call or exchange a couple of e-mails. I'll certainly document my reasoning for anything I do here on the wiki.--Erik Möller 13:36, 19 May 2006 (CEST)
Thanks for your responses here and elsewhere. My other concern is that, given the falling off of input into this project, how are we going to reach consenus to ammend and develop the FCD? --Ricardo Gladwell 12:15, 20 May 2006 (CEST)
We only need consensus among those who still participate. That's the beauty of wikis. ;-) --Erik Möller 01:05, 22 May 2006 (CEST)
My concern is that this is also the flaw of wikis: a truly universally accepted definition of free content requires the maximum input from all affected parties for it to work. With only a minimum number of people involved in this project how can we expect to get the same kind of acceptance as, for example, the FSF free software definition? --Ricardo Gladwell 14:13, 24 May 2006 (CEST)
The trick is to not worry too much about making mistakes. The definition has a version number for a reason. :-) --Erik Möller 15:00, 24 May 2006 (CEST)
I have to admit, there aren't any version numbers on the free software freedoms. They're simple, unambiguous and, because of this, universal (at least as applied to software). I don't see the same sort of thing happening here especially when people can pick and choose the version number of their liking. --Ricardo Gladwell 15:17, 24 May 2006 (CEST)
Unfortunately, copyright isn't simple or unambiguous. The FSF Definition works not because it is so accurate, but because there is also a shared consensus within the free software movement and open source community about what freedoms need to be granted (copyleft vs. BSD being the main point of disagreement). This is beginning to change a bit with the GPLv3, where people now disagree which licensing model is "more free" -- and indeed, both the FSF Definition and the Open Source Definition provide little guidance in resolving that question. Problems become even more apparent when you move outside the realm of free software, where questions like the "transparency" requirement haven't even been debated to any degree yet. The "pick and choose" effect cannot be avoided, whether you number your versions or not -- people can fork and create their own definitions. I'd rather see them use an official version number, so at least people will be able to clearly see the evolution and the underlying discussions.
If you feel that a particular group of people needs to participate more strongly, please be my guest and advertise the process and the definition. However, one thing I certainly do not want to do is debate in great detail here why, e.g., derivative works or commercial use should be permitted. That is why, for instance, I haven't advertised this definition with the Magnatune community (a CC-NC record label). To a large extent, I believe that the broad movement that respects the freedoms described here does not exist yet, so it is difficult to identify people who would agree, a priori, with certain assumptions while helping to work out the details. That's why I hope that, as this movement grows and defines itself, the definition, too, will evolve to meet different needs.--Erik Möller 16:32, 24 May 2006 (CEST)
I have to admit, I'm not aware of the disagreement over the free software definition or license models you are speaking of. I only see the normal debate and discussion that arises over any changes to a license. I also never said copyright was simple and unamiguous, only that the free software freedoms are (I also didn't say it was "accurate" either, but I'm not really sure how it is "inaccurate") which is, IMHO, why they are accepted across the free software and open source world. Sure, they're open to some interpretation but it's fairly obvious what is meant by the them.
Similarly, I have no problem with forking for the purposes of ironing out the draft definition but I think forking the actual definitions in general may lead to fragmentation where we are trying to create agreement.
I have to ask, if we're unconcerned with wider acceptance and if people can pick and choose the precise definition of free as they feel like it, was exactly is the purpose of this project? Given the current definition is close to the free software freedoms, I assumed this was an attempt to consolidate the various free/open content movements with a simple, unambiguous definition of freedom for all kinds of works, exactly as the FSF have done with software. A super-set of the free software freedoms. If I'm wrong please correct me and clearly define the objectives. --Ricardo Gladwell 16:47, 24 May 2006 (CEST)
What makes you think I am "unconcerned with wider acceptance"? Nothing I wrote above should have given you that impression. I am extremely concerned with acceptance, however, I believe that acceptance will come as part of a broader movement which has yet to evolve. This unified, defined movement exists for free software, and we can speculate all day and night about why it does and what role the FSD played, but it doesn't matter -- it does not yet, as an identifiable and recognizable movement, exist for free content as a whole. This definition aims to become part of the foundation of such a movement. We cannot prevent people from disagreeing with us. Of course we should try to find consensus where possible. I'm not sure what, exactly, you are expecting, however.--Erik Möller 17:40, 24 May 2006 (CEST)
I'm not sure I understand where the two of you disagree ;) I think we share the same intent Ricardo just stated ("an attempt to consolidate [etc.]"). Perhaps you just have different views on what magnitude of effort it implies.
Erik, as for why the Free Software Definition succeeded, it seems to me that there was no clear consensus prior to this definition. The FSF did actively advocate this consensus, partly by drafting the FSD (in short: free software did exist as a set of works and practices but not really as a concept). I'm not sure the GPLv3 threatens it that much. The question is not what is "more free" or "less free", but what is "free enough". That's why some rights and requirements are essential freedoms, while others are merely criteria for choosing a license, left to the personal taste of the author.
As for version numbers, I for sure hope we'll reach a stable version one day and settle on it for a long time. We can use separate pages if we need to expand on complicated topics (source code, fair use, legislation changes, etc.), so as to let the definition itself in a frozen state.
--Antoine 17:16, 24 May 2006 (CEST)
I think my main criticism of the FSD would be its essay-like structure, which is not very helpful to the reader. It is also normative on issues where I disagree (for instance, I personally prefer the public domain to all licenses, while the FSD advocates copyleft). I think it makes sense for our definition to be well-structured and agnostic on these internal variations. I do agree that we should strive for a stable core definition with dynamic pages about elements that are subject to change (new licenses, new legal precedents etc.).--Erik Möller 17:55, 24 May 2006 (CEST)

To what extent can derivative works be free?

One question that came up today on Wikimedia Commons, the Wikimedia media archive, which is meant to be exclusively free content, is whether images such as this one (a user-made photo of a Star Wars R2D2 toy), this one (a user-created artist's impression of Jar Jar Binks) or this one (user-made photo of a C3PO action figure) can be free content. Are photos of object representations of copyrighted characters from movies infringing? How about user-created drawings? We need someone with a strong knowledge of case law to give us some advice here, and I think the definition should include a few words on the matter when defining what a free work is (in addition to something about patent encumberance and practical modifiability). In doing so, I think we should be careful not to explicitly or implicitly embrace far-reaching copyright claims.

Mia, are you still listening? If I compile a few more concrete questions on this for you, would you be willing to give us some expertise?--Erik Möller 01:05, 22 May 2006 (CEST)

A somewhat disturbing example: if you make a photo of the Eiffel tower by night, distributing the photo infringes on the copyright of the lighting architect, because the nightly lighting of the Eiffel tower has been considered a work of art by a court. The tower itself is a work of art as well, but it is in the public domain since the architect is long dead. So taking a photo by day is safe, but it is not by night... --Antoine 15:48, 23 May 2006 (CEST)
I am interested in reading Mia's answer. And I am not a lawyer. But I have already discussed this type of issue (along with others) on this page. So let me summerize my assumptions in a form of Q&A.
Q. Are fictional characters such as a robot copyrightable?
A. Quite possibly yes.
Q. Are fictional characters performed by a human actor copyrightable? - Probably not how the actor's face looks, but his hair style, make-ups, costume, and other features combined might constitute a copyrightable work. I am not sure.
A. Are hand-drawings and other non-mechanical reproductions derivative works of the character? Yes, unless the drawing uses original character's "idea" only, and not its creative "expression". Where is the line between idea and expression? I don't know. But three-legged robot, for example, is an idea that alone does not make a drawing of a robot a derivative work of the star wars' R2-D2.
Q. Is it possible to release a derivative work of a copyrighted work under a free license like GFDL without the permission of the author of the original work?
A. I think so. But I am not completely sure. Some of you would remember that CC-1.0 contained representations and warranties which read in part like this:
1. Licensor has secured all rights in the Work necessary to grant the license rights hereunder and to permit the lawful exercise of the rights granted hereunder without You having any obligation to pay any royalties, compulsory license fees, residuals or any other payments;
2. The Work does not infringe the copyright, trademark, publicity rights, common law rights or any other right of any third party or constitute defamation, invasion of privacy or other tortious injury to any third party.
(taken from CC-BY 1.0 5a)
Subsequent generic CC licenses do not have these assurances, but have warranty disclaimer on these matters. So users cannot assume that a CC'd work is not violating others' copyrights.
But I think it is possible to combine fair use and free license in a manner that the licensor can say "use of the others' works in this drawing is fair use, and some rights I have regarding this drawing is available under a free license." A typical example is a Wikipedia's page containing a quote.
Q. Are those derivative works "free?" - It depends on how you define the term. It is under specific license that people widely recognize as free license (CC, GFDL). It is not freely usable because of the third-party copyrights, rights other than copyright held by the licensor or others. These rights not covered by a free license could become major hurdles for users who want to use a free work in certain context or for certain purposes.
How should we define "free" for free contents then? That's the issue I raised in the "#Scope_of_the_rights_covered" section above. Tomos 08:05, 28 May 2006 (CEST)