The SA Fallacy: Open Knowledge Foundation Gets It Wrong

So, the OKF claim to define the open in open content. I’ve come to terms with the fact that few people actually read what I write. But that’s still no excuse for people coming along six years later (in 2004), co-opting my terminology, and then getting the definition wrong by 180 degrees. If they want to define the open in open knowledge, that’s their business. However, the definition of the open in open content is available at

The OKF blog is currently featuring a post called “Why Share-Alike Licenses are Open but Non-Commercial Ones Aren’t.” I’ve written about this ad nauseum (literally – I’m sick of it) in the past. But as I said above, apparently no one cares. So, for my personal therapeutic good feeling, I’ll make a brief commentary on the post.

“Share-alike or attribution requirements are allowed within the definition precisely because they do not break this interoperability.”

This is obviously, demonstrably, and 100% false. The share-alike requirements of the BY-SA and GFDL licenses are exactly what prohibit material shared under these licenses from interoperating. The interoperability-breaking nature of share-alike requirements is what forced Wikipedia to engage in what might be the largest relicensing of content in the history of humanity – changing its millions of pages of content from the GFDL to a CC license. This had to happen because share-alike requirements destroy interoperability with content licensed under similar (but different) licenses with share-alike requirements. Read that sentence again. I’ll wait for you. If share-alike requirements promote interoperability, I’ll eat hay with a horse. Share-alike requirements only promote interoperability of content using exactly the same license. Is that really anything to be proud of or write home about? Here I state the SA Fallacy:

  1. Share-alike requirements supposedly promote interoperability.
  2. Only content licensed under exactly the same share-alike requirement-bearing license is interoperable.
  3. Any two pieces of content licensed under exactly the same license would be interoperable even if that license didn’t have a share-alike requirement.
  4. THEREFORE, a share-alike requirement in a license does absolutely nothing to promote interoperability.

What is it, then, that share-alike requirements do? Nothing more than prevent a down-stream user from making her own choices about how to license the derivative works she creates. When authors adopt a share-alike license, they are saying: we value the freedom of content over the freedom of people. They are saying: we prefer (1) that all derivatives of our content remain “free” over the option of (2) giving people the “freedom” to choose how to license the derivative works they create. Share-alike requirements give authors a way to privilege bits and bytes above people.

“To reiterate: it is a mistake to view the set of licenses as some continuous spectrum of ‘openness’ with PD at one end and full rights reserved at the other — with the implication that all licenses in between are more or less open.”

No, this is not a mistake – it is a demonstrably accurate view of the real world. PD places no legal restrictions on would be users. CC BY places one restriction on users. CC BY-SA places two restrictions on users. CC BY-NC-SA places three restrictions on users. &c. In fact, the only reason to create more than one license is to mandate different degrees of restriction on people’s use of content.

The OKF argue that because they have defined open to mean one particular constellation of restrictions (including the restriction that prohibits content interoperability), there cannot be degrees of conformance with their mandate. You are not allowed to think about the world as it really is (many licenses falling along a wide spectrum of openness), you must conform your thinking to their alternate reality. However, their standard appears arbitrary (why are these restrictions acceptable while others aren’t?) and their standard does not promote the interoperability of content. And it certainly does not define the open in open content.