Fair use and Wikipedia

Thinking about the FSF and the GFDL made me think quite a bit about my attitude on the "fair use" of unlicensed media on Wikipedia projects. And the more I think about it, the more I think there is no justification for using unlicensed media on Wikimedia projects.

The Free Software Foundation is about free, reusable code. They care more than the code is reusable than that it be good, in no small part because being reusable means that it will eventually become good (the reusability is a large part of what makes goodness happen when it actually does happen). So they don't compromise: there is no "fair use" in GPL software. Now, this is in part because it doesn't make sense; I've never heard of anyone making a "fair use" argument for code.

I am becoming more and more convinced that the compromise that Wikipedia makes to allow fair use content ultimately reduces the goodness of the encyclopedia, and even more importantly it clearly reduces the reusability of its content. Allowing the use of unlicensed content as fair use when reusable, freely licensed content could instead be developed discourages the development of the freely licensed content: why make the effort to develop something when there's an acceptable alternative already existing? There's far less incentive to go out and take a picture of something when there's already a good-quality image there.

Of course, there are the counterarguments: what if the article is about something that is copyrighted? In a lot of cases, the article in question will not really suffer if the nonfree content is eliminated. [[IBM]] isn't going to suddenly be a useless article if the corporate logo were suddenly removed, nor would all the multitudinous articles about record albums suddenly suck if the thumbnailed album covers were chopped out of the infoboxes. In my opinion, the marginal loss in the quality of these articles is more than offset by the increased freedom of the result.

The more difficult case is where the article is about a singular work of art which really does need an exact representation; for example, Picasso's [[Guernica]]. And here, this is where the Wikimedia Foundation needs to start negotiating. Would Picasso's estate really be that adverse to agreeing to license a 300x135 grayscale version of this image under a free license? They might. They might not. I bet nobody has ever asked them. The thing is, Wikipedia is now big enough that some content creators may be willing to give a bit, either out of altruism/charity or out of a desire to Be Included In Wikipedia. We've already seen that happen once, with that one pair of twin male models that Jimbo negotiated a GFDL-released image from. Content creators that refuse to negotiate won't have articles about them that are quite as nice; this might motivate them to be more forthcoming. And if this does work, it is a real win for everyone committed to free content.

Of course, one counterpoint that can be made is that RMS originally wrote almost all of the first GPL'd code on proprietary operating systems using proprietary tools; the open source software environment has never been "pure", so why should Wikipedia be pure either? There is a fundamental difference here. In almost every case, the Wikipedia article, stripped of its nonfree content, is still very useful as an encyclopedia article. The nonfree inclusions may add value but in most cases they are incidential to the main force of the article. This is totally unlike the first GPL'd software, which would have been simply unusable without a platform on which to run it. The lack of such a platform did eventually inspire people to write a free platform, but I fear free software wouldn't have caught on if Stallman had tried to mandate that it could not be used on computers with proprietary operating systems. Wikipedia has already caught on; there's no need to compromise principle for the success of the project.

There is, of course, the additional advantage that not allowing nonfree content ought to reduce legal hassles when people don't approve of the use of some content. No need to do a fair use analysis; just check to see if it's really released under the GFDL and if it's not, delete it. That, by itself, will probably improve public relations.

So it's my conclusion that the Wikimedia Foundation ought to, in recognition of its commitment to free content (and especially if it wishes to be known as a true champion of free content), set into motion a ban of all nonfree content from all Wikimedia projects whatsoever. I'm sure that doing this overnight is probably not practical, so here's my suggestion: I hereby challenge the Wikimedia Foundation to ban all "fair use" from all projects on or before Wikipedia Day (January 15th), 2008. Make it so!