9.26.2007

Ed Boyden, Roy Baumeister

Ed Boyden, a neuroscientist over at MIT/Technology Review, has started a general-interest science blog. I'm happy to see this, as Ed seems not only smart, but prone to write frankly and creatively about deeply relevant issues.

Another interesting, rather speculative piece I'd direct people toward is "Is There Anything Good About Men?" - American Psychological Association, Invited Address, 2007, Roy F. Baumeister.

I've not decided whether I agree with his conclusions, and since it's a conference address it's a little citation-lite, but it's a thoughtful and thought-provoking mix of fact, theory, and conjecture.

Back to paper writing.

9.19.2007

Citizendium: Choosing a Free Content License

Citizendium, the wiki encyclopedia I volunteer with, is holding a call for essays about which free content license we should choose for our content. It's an open call for essays, so even if you're not a member of Citizendium and have an opinion you're encouraged to write it up and send it in. Wikipedians and other indirect stakeholders welcome.

Here's my essay (i.e., personal opinion) on what and how things should be considered. As a complete position statement it may be a little rough around the edges, and it may move rather slow for those who live and breathe this stuff. Feel free to leave comments.

Citizendium's License
Mike Johnson
DRAFT version

There's a lot at stake in our choice of free content license, and we have lots of options. In this essay I'll run through options that have been discussed on the forums, some practical concerns, potential future developments which may be relevant, and my own recommendation. I'm not a lawyer, and I don't claim this essay to be a definitive analysis- but I think, at minimum, it maps out some options, issues, and opinions that should be considered.


===Part 1: Our Options===

The most common content licensing options discussed on the forums have been:

- GNU Free Document License (like Wikipedia): people may use our content freely, so long as they credit Citizendium and any improvements under the same conditions, and conform to certain (some some circumstances, rather unwieldy) regulations pertaining to including licensing information along with documents. It's very similar to CC-by-sa, but as the license wasn't designed with wikis in mind, it's a bit 'creaky' in various ways (it may be sacrilege, but I'd refer readers to the Wikipedia article on the GFDL for criticisms of the license). Wikipedia was founded before there was a mature ecosystem of free content licenses and chose this license by necessity.

- Creative Commons: Attribution, Share-alike (CC-by-sa for short, aka the "Creative Commons Wiki License", like the Encyclopedia of Earth): people may freely copy and improve our content, as long as they credit Citizendium and release any improvements they make under the same conditions.

- Creative Commons: Attribution, Share-alike, Non-commercial (i.e., CC-sa-nc, like many images on Flickr): people may freely copy and improve our content, as long as they credit Citizendium, release any improvements they make under the same conditions, and don't use our content to gain "commercial advantage or private monetary compensation".

- Dual-licensing our content under both the GFDL and CC-by-sa, since they essentially lay out the same abstract set of rights and obligations (there are certain technical legal uncertainties about the possibility of doing this, but the people who work on these licenses with whom I've interacted didn't shoot this down as a possibility). The benefit would be better outgoing compatibility, that websites using either the GFDL or CC-by-sa could integrate our content with their content. The main drawback would presumably be another layer of licensing complexity.

- Public Domain. No legal protections on our content whatsoever; people are free to do anything they please with our content (and importantly, repackage, change, and sell it under whatever conditions they prefer).

Of note, any license we choose for our fully homegrown content will exist side-by-side with the GFDL, which applies to articles with content originally sourced from Wikipedia. There's no way of changing this.



===Part 2: Practicalities===

Non-commercial clause. One of the big question marks in choosing a license is whether to add a clause prohibiting commercial or commercial-related use of our content. On paper, it sounds great: we're contributing to Citizendium for humanitarian reasons, and we don't want anyone making money off the sweat of our brows. Or if anyone does, it should be us.

But from a practical standpoint, I personally don't like the non-commercial clause, because it
1. Adds a level of ambiguity to our license: commercial acts don't necessarily have clear boundaries. And if people have to ask whether a use will be okay, most won't.
2. Answers a danger (other people making money off our content) which may be rather trivial. Someone may sell our content if we license it under CC-by-sa-- but they could just get it for free from citizendium.org. I can't dismiss this as a concern, since in the future businesses may get more clever and obnoxious in their attempts to wring money out of free content-- but it hasn't been a major problem for Wikipedia yet. That Wikipedia is freely available online (and that the GFDL protects Wikipedia's content from people trying to claim ownership and economic rights over it) takes a lot of wind out of the sails of those trying to monetize it.
3. Puts our content at a competitive disadvantage in the quest for eyeballs. Content without this clause will simply get used more. I think getting our wonderful content out there to be read by as many people as possible is really very important. Relatedly, it
4. Walls our content off from being combined with other sorts of free content (one can't mix-and-match CC-by-sa and CC-sa-nc content unless one can clearly show it's not for a commercial use- a risk many people will not take.

So, I think it's cleaner and better for reaching more people with our content to forego a non-commercial clause.

That said, the clause would increase our options for fundraising via selling businesses the right to use our work in commercial settings. We may or may not want to do this, but options are generally good.

Whether or not to go with CC-sa-nc is probably the most consequential choice on our plate. However, other relevant factors among the licensing options are:

Compatibility with Wikipedia. This is a huge factor in choosing a license, and one that's been debated on the forums: some are strongly for compatibility, some strongly against. Personally, I see full and seamless compatibility as very desirable in the long-term: Wikipedia's going to be around for a while, and if we set things up such that we can seamlessly share content between projects, I think it'll help us both out in obvious (less duplicated effort, less license administration overhead) and non-obvious ways. And philosophically, it's very much in the spirit of free content to work toward a licensing ecosystem where omnidirectional sharing is easy.

But in the short-term, things are more mixed. I don't think any of us want Wikipedians to swoop in and copy some of our competitive advantage away (and I assure you, from reading certain blogs, some Wikipedians are waiting to try to do so). Honestly, for a variety of reasons, I'm not all that worried about this-- but I do strongly believe we need to avoid the appearance that the really good stuff from Citizendium ends up on Wikipedia anyway during our first few years of existence.

Now, to be honest, I feel a little dirty about bringing this up. We've used some of Wikipedia's articles as jumping-off-points for some of our articles, and it seems natural that we should return the favor. And it just seems right to allow content to freely pass between our projects. I think it's very important to be a good neighbor to Wikipedia (as well as to other free content projects). And I think choosing incompatibility for incompatibility's sake, to wall off our homegrown content from Wikipedia for competitive advantage, would simply be unacceptable and inconsistent with our mission of improving the state of free content.

But I believe that in choosing a license there are many things to weight, and we should also keep in mind that, though there are external stakeholders in this decision whose wishes we should consider, we're not beholden to any one of them: we are a sovereign community and our most fundamental responsibilities in choosing a license are to our community, its philosophical and practical goals, and its health.

So I think we need to take this situation seriously and, if we do choose to be compatible or incompatible with Wikipedia, we should make our choice with respect to the likely practical consequences to our community in addition to philosophical ideals. I don't know what these practical consequences would be with respect to compatibility/incompatibility with Wikipedia- I think we'd do just fine in either scenario, and perhaps this issue is a tempest in a teapot. This is quite possible. But I view talking about these issues surrounding compatibility with Wikipedia as 1. part of due diligence in writing this essay, and 2. a reminder to the Erik Möllers of Wikipedia that we are a sovereign community (I do encourage Mr. Möller and others who hold an opinion to submit an essay. Wikipedians are among the stakeholders in this decision, and our decision will be poorer for not hearing from them).

As a digression, I think it's in Wikipedia's best interests to not swoop in and copy our content wholesale, from both the standpoint of having pride in their own content (if Wikipedia copies wholesale from us, that's a pretty harsh statement about the health of and confidence in their processes), and that they should take care to avoid anything that might stifle competition, because competition is valuable to the health of their community. But although there are many thoughtful people at Wikipedia, I don't think we should count on Wikipedia per se to keep these things in mind, since it'll be individual contributors acting under many different assumptions that would be doing the copying. I think at this point we should assume that, depending on internal politics at Wikipedia, they may copy from us very freely if we legally allow it. And we couldn't really cry foul, since they'd be fully within their rights.

So, though my personal stance is to weight long-term seamless compatibility with Wikipedia as a significant positive (albeit among other important factors), and short-term seamless compatibility as a minor positive, I do suggest that people think very carefully and honestly about this issue. It's complicated and potentially important, and I doubt I'm doing it full justice.

Compatibility with Creative Commons. Wikipedia is, of course, a major factor in compatibility, but Creative Commons content is growing very quickly as well. If we went with CC-by-sa we'd be compatible with e.g., the Encyclopedia of Earth. There are reasons to put compatibility with Wikipedia on a pedestal in relation to Creative Commons, but in the long run, perhaps not as many as one would think.

Ease of internal administration. Frankly, I think this is the strongest argument for going with the GFDL. It's simple, and simple is very good. If we have some content licensed differently than other content, it becomes a headache to administer very quickly. Going with the GFDL isn't necessarily a silver bullet, however, since we'd face problems with importing Creative Commons-licensed content. It would be worth examining Wikipedia's approach to importing CC-licensed content to see if they've figured out clever and low-friction ways to handle this.

Quality of License. Put simply, Creative Commons licenses are better laid-out and more suitable for wikis than the GFDL.

Customization. We should consider whether we want to accept any of these licensing options as-is or whether we want to (at risk of making our content less compatible with other free content) add in any additional terms to the broad-stroke license we choose.


===Part 3: Future Developments in Free Content Licensing===

- CC-by-sa has a 'compatibility framework' which attempts to bridge the gap between a Creative Commons-structured license and other free content licenses. The current revision of CC-by-sa isn't compatible with the GFDL (or any FSF license), but from personal correspondence with the FSF I believe it was a matter of timing and not substantive reasons.

- The current draft of the next revision of the GFDL includes a relicensing provision that
If the Work was previously published, with no Cover Texts, no Invariant Sections, and no Acknowledgements or Dedications or Endorsements section, in a system for massive public collaboration under version 1.2 of this License, and if all the material in the Work was either initially developed in that collaboration system or had been imported into it before 1 June 2006, then you may relicense the Work under the GNU Wiki License.
The unreleased GNU Wiki License will presumably be much more suitable for wikis than the GFDL, and include compatibility provisions matching up with CC-by-sa's compatibility framework.

So, as a bottom-line, I would say there's a good chance CC-by-sa and GFDL content will be compatible sometime in the future. This potential development would make choosing between CC-by-sa and the GFDL a lot less pressing (though, naturally, we shouldn't depend on this happening). I think it'd make CC-nc-sa a less desirable licensing choice because there wouldn't be that potential for radical simplification since the GFDL and CC-nc-sa simply can't be made compatible.


===Part 4: Other Issues===

Copyright sharing. I think it's desirable for contributors to share copyright with Citizendium, if and only if Citizendium enters into a legally binding contract with contributors to not use said power for certain aims. Or to use said power only for certain aims. The devil's in the details, but I see no reason why copyright sharing paired with such a contract couldn't provide all of the positives and none of the negatives put forth in these arguments. Of note, copyright sharing would only apply to homegrown content, which would reduce its utility.

One of the possible benefits of copyright sharing would be the ability to merge homegrown CC-licensed and GFDL-licensed articles under a GFDL article. There would like be many other such relatively minor legal maneuverings that would become possible with copyright sharing. However, once we start importing non-homegrown CC content, such merges would become more complicated. Copyright sharing isn't a silver bullet.


===Part 5: My Recommendation===
If you've been reading closely, you can probably guess what my recommendation will be: that we should license our content under CC-by-sa. It will probably give us eventual full compatibility with Wikipedia (good, in my mind), short-term incompatibility with Wikipedia (mixed, but slightly negative in my mind), full compatibility with Creative Commons sources, and a more sane wiki license than the GFDL. Until compatibility is reached, there'll be some administration headache in keeping licensing issues sorted out; hypothetically, after CC-GFDL compatibility is established, most of that should go away.

Alternatively, I think simply going with the GFDL is a good second choice (for many of the administrative reasons argued here), provided that we will indeed be able to relicense our content under the GNU Wiki License. If we choose this route, I think we should make a point of following up with the FSF for information and advice.

As a third choice, I would have few qualms about dual-licensing our content under both CC-by-sa and the GFDL. It would add some legal ambiguity (depending on whom one asks, the GFDL may not allow this) for what seem like rather ephemeral gains. Perhaps others have alternative perspectives on this.

There's no silver bullet licensing choice-- but conversely, for what it's worth, with the GFDL and CC-by-sa likely converging I think there are relatively few bad licensing choices.