Wikipedia could be changing licenses. How, and why?

On Monday, 3 November 2008, the Free Software Foundation announced a new version of its Free Documentation License (commonly called the GFDL, for Gnu Free Documentation License) . This new version was produced after the Wikimedia Foundation, stewards of the Wikipedia project, advocated for license compatibility with one of the Creative Commons licenses. The story of how this situation came to be is surprisingly complex.

History: Why does Wikipedia use the GFDL to begin with?

Wikipedia was started as a project in 2001. At its creation, there was a need to choose a license for its content that would permit reuse and, critically, make it practical to run a project in which the original copyrights belong to thousands of different, potentially anonymous, people.

At the time, this meant choosing between the GFDL and creating a new license. The Creative Commons family of licenses did not exist yet. Developing a new license can be a complicated, failure-prone process, and may prevent reuse of materials in other projects unless they adopt the same license. The project's creators chose to use the GFDL because it offered at least some degree of compatibility with other uses, and was reasonably well-defined.

The GFDL in a nutshell

The GFDL was developed by the Free Software Foundation as a complement to its Gnu Public License (commonly referred to as the GPL). The GPL provided for the sharing and reuse of software source code, with stipulations for things like providing credit to the previous authors. The GPL is best known (to the consternation of some at the Free Software Foundation) as the license used by the Linux operating system kernel. The GFDL was originally intended for use by software manuals for software products using the GPL.

The GFDL allows anyone to reuse any piece of licensed text, provided they adhere to several requirements. The most critical are that they must provide credit to the "upstream" author(s), all derivative works must also use the GFDL, and they must make the complete text of the GFDL available.

Why change licenses?

The GFDL was the most suitable license for a project like Wikipedia that was available in 2008, but that does not mean that it is well-suited to a project like Wikipedia. Several aspects of the license are either counter-productive for Wikipedia or have an uncertain meaning for such a project.

The GFDL requires that the complete license text be made available to recipients of a GFDL'ed text. For the web version of Wikipedia, this is handled by simply linking to an HTML page on Wikimedia's servers that contains the full text. However, it becomes more difficult to handle when users print things. Users who print, say, a Wikipedia article for their own reading do not have to worry about this clause. However, if they print a few dozen copies to hand out to friends, they arguably do need to provide the full license text to each person in order to adhere to this clause. If they create a derivative work from it and distribute that, then they certainly need to provide the license text. The GFDL is a fairly lengthy document (nearly 3500 words in the current version 1.3), and so this clause can create a hefty burden on individuals that simply want to redistribute a small piece of GFDL'ed text.

The GFDL contains language regarding "invariant sections" and "cover texts" whose meaning is not generally clear in the context of a project like Wikipedia. The Wikimedia Foundation's solution to this has been to explicitly state in its licensing terms that no part of its content should be held as an invariant section or cover text.

The GFDL makes requirements for the listing of authors that can also be difficult to adequately define or comply with in the context of a project like Wikipedia. In our above scenario of printing and distributing Wikipedia articles, it is not absolutely clear under what circumstances authorship must be explicitly stated, nor (in the case of a project where dozens of authors may have contributed to a single page) how many authors must be named.

There is a distinction in the GFDL between a "document" and a "collection of documents", and it has never been perfectly clear which of these Wikipedia (as a whole) qualifies as. The Wikimedia Foundation has argued that Wikipedia is a collection of documents (where each article or image covered by the GFDL would individually be treated as a document). This has implications for several aspects of the GFDL, including the above-mentioned case where major authors may need to be enumerated; would that be the major authors of a single article, or of the entire project.

There have been other legal issues related to the GFDL, such as the use of "with disclaimers" templates on images, which indicate that warranty disclaimers, in addition to the GFDL text, must be carried forward to derivative works. These wounds are, to a certain extent, self-inflicted; the GFDL does not by any means require the use of the "with disclaimers" language.

Why can Wikipedia now change licenses?

On 3 November 2008, the FSF released version 1.3 of the GFDL. This new version adds Section 11, which provide for the one-time relicensing of content in "Massive Multiauthor Collaboration Sites" (abbreviated MMC sites; sites that provide prominent facilities for visitors to edit content; essentially a legalistic way of saying "wikis"). This was the result of lobbying by the Wikimedia Foundation, which has increasingly wanted to move away from the GFDL.

There are some stipulations in this new clause. Content may be relicensed only the the Creative Commons Attribution Share-Alike 3.0 license (commonly written CC-BY-SA 3.0). This is the third major iteration of the Creative Commons Attribution Share-Alike license (CC-BY-SA). And the window for relicensing is open only until 1 August 2009; after that, content that is under the GFDL will stay under the GFDL.

There is an additional, slightly misunderstood requirement regarding content originally published under the GFDL somewhere other than the MMC site and then incorporated into that site. Such content may only be relicensed if it was added to the MMC site before 1 November 2008; the FSF has indicated that the choice of a date that had already passed was intentional, to prevent people from "laundering" GFDL content through a wiki to convert it to CC-BY-SA.

Why CC-BY-SA?

The CC-BY-SA license is very similar in spirit to the GFDL; it permits reuse and derivative works, but requires that derivative works provide attribution and adhere to the same license. Thus of the Creative Commons suite of licenses, it is the most compatible with the GFDL.

While the CC-BY-SA is principally in agreement with the GFDL, it is generally seen as better-suited to a project like Wikipedia. The GFDL carries a lot of baggage specific to reproduction in printed media, which is one of the major sources for the complications and confusion discussed above.

The CC-BY-SA lacks many of the more problematic clauses of the GFDL, especially the requirement of providing a full copy of the license text. It still has a fair amount of legalese that will require some interpretation (and at around 3100 words, it is nearly as lengthy as the GFDL), although like all Creative Commons licenses there is a greatly abbreviated "human readable" version of the license that cuts to the chase.

There is one more major reason for preferring a Creative Commons license. A number of other major wiki projects, including Wikinews and Memory Alpha, use some form of Creative Commons license. While this change would not bring Wikipedia's license into perfect harmony with those licenses, it would be a step closer.

Does this mean all Wikipedia content is now under the CC-BY-SA?

No. The Wikimedia Foundation will hold a public discussion on whether Wikipedia should make the jump before formally deciding whether to relicense. The Foundation has until 1 August 2009 to make its decision. Given that the Foundation was the real insitgator of this effort (although the Foundation might have preferred true, two-way compatibility between the GFDL and CC-BY-SA), it is reasonable to assume that the Foundation would like to make the transition. And when the effort (and Wikimedia's involvement) made headlines nearly a year ago, the response was overwhelmingly positive.

  • RSS feed StumbleUpon del.icio.us Digg Yahoo! My Web 2.0