GFDL suggestions

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search
See Proposal for a modified GFDL for past discussions.

Currently, almost all Wikimedia projects are licensed under the GFDL, version 1.2, a license created by the Free Software Foundation for documentation and reference works.

There is currently a draft for a new revision of the license up on the website and open for comments, in very preliminary stages of completion. The FSF is very interested in hearing comments from users of the licenses, and in particular they have always been supportive of Wikimedia and its goals and needs.

You are of course welcome to make your suggestions for the new draft on the FSF website itself! However, we would like to know what comments, criticisms, and suggestions come from the Wikimedia communities specifically so that we can go to the FSF and talk about changes that Wikimedia would like to see.

Please read the current version and the draft version before commenting on this page; you might find some of your concerns already answered! Questions rather than criticisms and suggestions might be better brought up on the talk page.

Additional discussion is being held on the English Wikipedia.


One proposal[edit]

Attribution requirements for collaborative works[edit]

Attribution is important for ethical and practical reasons. Attribution is ethically required because it is customary and reasonable for people to receive credit for their intellectual labors. Attribution is practically required because without it people with bad intentions can claim to be the author and cause problems, it's also important to helping others understand the history, source, and possibly the credibility of the work. Attribution is also a fundamental legal requirement in some jurisdictions.

One issue we have encountered is how does one correctly provide attribution when a work has a huge number of authors. GFDL requires that you name at least the five primary authors, but for huge collaborative works it may be hard to determine who those people are.

CC-by-*-2.5 and later simply allow a service provider to take attribution for themselves simply by merit of hosting the content (see "terms of service" in cc-by 4.b.), even for works which are not collaborative.

Proposed attribution text[edit]

Text adapted from various FSF and CC licenses:

Except as required elsewhere in this license, you must, in a manner reasonable to the medium or means of distribution, credit the Original Author or another party explicitly designated by the original author. For collaborative works with multiple authors, you must either
  1. credit at least five of the principal authors (or all of them, if it has fewer than five), OR, if and only if the principal authors can not be reasonably determined,
  2. credit any suitable source for the the document which must: conform with all terms of this license, be accessible online without restriction, provide the covered work in a preferred form for editing, provide the revision history of the covered work which shall include the names (or pseudonyms, at the author's discretion) of all authors and contributors and indicate their contributions, and provide the complete text of this licenses.
If applicable, this source must provide the complete attributed revision history as well as attribution to the sites of origin for works which have moved between multiple collaborative sites.
  • I like these terms. Some comments.
    • This is a lot like what most wikipedian's ask people to do with their work rather than actual GFDL conformance.. It preserves attribution in the non-collaborative case, so photographers should be happy.
    • "explicitly designated by the original author" needs to be made more strong or removed, as it stands it still presents a little too much "terms of service" risk. We want to allow people to ask that some other agency be credited... some people really want to do that. I don't think the license should make it hard for them, but it should make it impossible to take attribution out from under someone.

--Gmaxwell 20:35, 9 May 2007 (UTC)

  • "I'm not likeing "if and only if the principal authors can not be reasonably determined" there are now two terms in there open to interpritation rather than the one in the old version.Geni 03:36, 16 May 2007 (UTC)
    • Can you suggest a better measure? You've point out what I dislike most about the above text, but I'm not quite sure how to fix it: The intention is to only leave the "alternative attribution" for collaborative works with lots of authors where the primary authors can't be easily determined. For things like single images with single authors the attribution requirements shouldn't allow someone to deny them the reasonable and customary credit.
    • Perhaps the license should allow the authors of the work to specify the (up to 5) primary authors, and if they do so you must attribute to those authors, if they fail to do so you do the alternative attribution? This removes the judgement call, but it creates the issue of needing rules for updating that list for modifyed versions. --Gmaxwell 15:01, 8 August 2007 (UTC)
      • I don't think the license should talk about "title pages" at all, in which case I don't see a need to figure out who the principal authors are. Crediting 5 authors and ignoring the rest doesn't seem to ever be appropriate. - Anthony 23:49, 3 December 2007 (UTC)
  • User:Jamesday pointed out to me a while back that the words "or another party explicitly designated by the original author" might create an opportunity for service providers to take attribution for themselves, something which both he and I regard as unethical. It was my hope that the the words 'explicitly designated' would be enough to avoid some service provider putting "All your work transmitted via our system will be attributed to EvilCorp(tm)" in their ToS, but he doesn't think so... and I guess I agree. The intention is to make it clear that the author has the freedom to assign attribution to someone else (perhaps a party which funded his efforts), but the license should be constructed in a way which makes it impossible for this to happen without the authors clear and considered intent. I don't know how to accomplish this goal with concise languages. --Gmaxwell 15:01, 8 August 2007 (UTC)
  • "shall include the names (or pseudonyms, at the author's discretion)" :
    1. It sounds like the user should (or should not) need to contact every authors individualy to find how to name them "at their discretion". Or at least, the ambiguity stands. And: would an author be allowed to ask for a change in the way he is named after you received and used (ie. printed) the work ? Remember that on Wikimedia's projects, users renaming occurs more often than not.
    2. Why do we have this precision in the second statement and only the second one ? (ie. the first statement, "credit at least five of the principal authors" doesn't explain how to name them, neither does the general introductory sentences): an user might wonder if it has the same opportunity (ie. to name by pseudonym) if he choose the first option.
    3. Would a transliteration be allowed ? I mean, if I receive a work from 鈴木清順, am I allowed to credit him as "Seijun Suzuki" ? This is important for printing.
I would like the attribution text to define clearly and globaly (ie. in a way applicable for the two statements/options) what is an "author name" (to credit). If I understand the intent, that would be something like this, along the lines (well, and after you translate this to proper native english ;) : "An "author name" means a name credited as an author, as you received it attached to this work, be it a real name or a pseudonym.". Benjamin Pineau 13:13, 24 December 2007 (UTC)

Another proposal[edit]


One proposal[edit]

Allow GFDL works to be combined with CC-BY-SA (or other similarly attributed copyleft) works, provided the combined work follows all of the restrictions of both licenses; where two conflict, use the more restrictive.

Another proposal[edit]

Print reuse[edit]

One proposal[edit]

  • Make it easier to use excerpts, short works, or photographs without having to reproduce the entire text of the license by allowing a credit to any site containing both a copy of the license and the full attributed change history?
    • Perhaps require also a tagline, if the full license will not be linked to? Something like "this work is free content and may be reused under the terms of the GNU Free Documention License"?



Debian has finally approved the GFDL. There was a project wide vote and it turned out the people complaining about the DRM clause in the GFDL were a tiny but vocal minority. The Debian-legal related complaints have almost always been cited behind any DRM related complaint that I've seen on Wikipedia.

Creative Commons has refused to water down the anti-drm provisions of their licenses ... surely the GFDL shouldn't be more permissive than by-sa in this regard.

I, personally, would be very unhappy if GFDL-1.2's DRM clause were weakened in any material way. I think the current FDLv2 draft's handling of DRM is poor compared to 1.2. It says that if you lock up a work behind DRM you can't stop anyone from breaking that DRM, but what if criminal charges are brought?

I think there are two primary DRM threat models a copyleft free content licenses needs to address, and which I think FDL-1.2 is already strong against:

  1. Parties circumventing the license requirements by locking the content up in DRM wrappers.
  2. Devices applying DRM against their operators' will (Microsoft Zune for example)

The first can largely be addressed by requiring non-drm copies be made available in parallel, or with a working "you can break" clause (which I do not believe FDL 2 draft 1 provides).

I think the second is especially dangerous, although the risks are not as obvious... The risk here is that it may generally become illegal to traffic in devices/software/etc which are willing to handle non-DRMed content. This would neatly close the analog hole that plagues most DRM systems. Most laymen regard this as unlikely, but it is well precedented: The DMCA requires all analog video recorders sold to be compatible with macrovision.

Already most video and audio exchanged on the Internet is in proprietary and patented formats... I suspect that if industry could stop being greedy for a bit and cooperate on a DRM standard which could painlessly DRM all content that we'd already be there.

Obviously mandatory DRM would have a devastating effect on free content, free software, and even free speech. It's too big a risk to simply ignore. Fortunately it shouldn't take much much friction to stop it... DRM provides almost no benefits to consumers, so simply making copylefted free content mildly incompatible with compulsory encryption which doesn't provide the recipient with clear-text access should be enough.

The language in FDL-1.2 has been widely misunderstood and misconstrued, so the language could be clarified but I think the actual intention and behavior are desirable.--Gmaxwell 01:55, 13 May 2007 (UTC)

However whatever the initial author of a work not derived from a previous GFDL work uses must be considered a transparent format for that work otherwise we hit problems with scans of text and the like.Geni 03:34, 16 May 2007 (UTC)


  • Regardless of what people think of the substance of the GPLv3 (although most of the complaints were about old drafts, few dislike the current drafts)... Virtually everyone agrees that the style change was good. In particular GPLv3 avoids using words which have local legal significance, and instead defines its own terms. For example, rather than saying "distribute" the GPLv3 talks about propagation:

To "propagate" a work means to do (or cause others to do) anything with it that requires permission under applicable copyright law, except executing it on a computer or making modifications that you do not share. Propagation includes copying, distribution (with or without modification), making available to the public, and in some countries other activities as well. To "convey" a work means any kind of propagation that enables other parties to make or receive copies, excluding sublicensing. Mere interaction with a user through a computer network, with no transfer of a copy, is not conveying.

I think that changing to this style would remove a lot of ambiguity in the GFDL, and it would also improve the word wide enforcability of the license. --Gmaxwell 21:59, 9 May 2007 (UTC)

request contact author[edit]

It is requested, but not required, that you contact the authors of the Work well before conveying any large number of copies, to give them a chance to provide you with an updated version of the Work.

I see no value in this section. Either something is required or it is not. if it isn't it shouldn't be in the lisence.Geni 03:40, 16 May 2007 (UTC)

Hm. I suppose I agree that a license shouldn't include mere suggestions, as good as the suggestion may be! Kat Walsh (spill your mind?) 00:54, 30 June 2007 (UTC)

E. Preserve all the copyright notices of the Work.[edit]

This still leaves the problem of watermarked imaegs.Geni 03:42, 16 May 2007 (UTC)

That's interesting. Should it be explicit that you can remove a watermark so long as you keep the information elsewhere, or should it simply not say that you can't? Kat Walsh (spill your mind?) 00:52, 30 June 2007 (UTC)
Tricky. Something along the lines of "where the copyright notice is imbeded in the media it may be moved elsewhere in the document" might work but still has issues. Even that present problems with highly fancy copyrigt notices or when you want to shift formats (want to use a screenshot from a GFDL flim? what if there is a spoken copyright notice?).Geni 22:53, 4 August 2007 (UTC)

Still geared towards books[edit]

The wording of the license is still obviously geared to long texts such as books. The CC licenses have much better more generic wording. The wording should either be made generic or it should be made explicit that the wording only applies to certain types of works. e.g. What is a 'Title Page' in the context of an image?

Secondly the whole thing of 'Excerpts' clearly shows one is thinking about long texts... 20,000 characters could easily be an entire Wikipedia article. What kind of 'excerpt' is that? I would change it to something like 'the lesser of 20,000 characters, or 10% of the entire content'. And also explicitly state what an 'excerpt' is when the work is an image (is it a thumbnail? a crop? nothing at all?). It is really weird to say what it is for text, audio and video, but NOT images. pfctdayelise 13:25, 9 August 2007 (UTC)

GSFDL: clause 0a[edit]

My suggestion for this clause is simple. Remove it entirely. What is the point of acknowledging that the license can be used for all kinds of works, and then acting as if the work is a manual? pfctdayelise 13:25, 9 August 2007 (UTC)

'You need not include a copy of this license if...'[edit]

"You need not include a copy of this License in the Work if you have registered the work's license with a national agency that maintains a network server through which the general public can find out its license."

How is that going to work on the internet with massively collaboratively authored documents?? Oh right, it's not. :( pfctdayelise 13:25, 9 August 2007 (UTC)


Please see discussion of problems with the Excerpts section on the English Wikipedia. Kaldari 19:48, 17 October 2007 (UTC)


The section that contains the history needs to be more flexible with regards to its title since other language wikipedias don't necessarily call the history section "history". For example, on the Hebrew Wikipedia it is referred to as "previous versions". Yonatanh 15:01, 21 October 2007 (UTC)

Communication with the FSF[edit]

Can anyone tell me anything about how the Wikipedia community is communicating with the FSF? Do we have a designated liason or representative? Does the FSF have a person in charge of communicating with us? Is the FSF also communicating with the WikiMedia Foundation in any way? If so, how? Kaldari 19:52, 25 October 2007 (UTC)