I managed to find myself in a discussion on Twitter (@mmcclimon) this afternoon about licensing. The discussion started out about open access, where a colleague of mine (@krisshaffer) wanted to talk about the distinction between open access (which he equated with CC-BY licenses) and open source (CC-BY-SA). His argument was the open access (CC-BY) “still allows for-profit hijacking of materials.”
I responded and said that according to GNU, open source is not equal to free software, and subsequently that I didn’t particularly care for the GNU licenses. Richard Stallman has a well-known arbitrary definition of freedom whereby “do whatever you want” is less free than “do whatever I tell you to.”
This is a bit long, so I’ll put it behind the jump.
[Quick aside for context.] The Free Software Foundation (FSF) was founded by Richard Stallman, who in the 80s started developing an operating system called GNU, which he envisioned as the world’s first “free operating system.” This system still doesn’t exist entirely, though Linux is basically the same (despite what Stallman might tell you). The GNU Public License (GPL) is a software license which says that any modifications to the software must be released also be GPL-licensed (commonly referred to as a copyleft license). The Creative Commons system of licenses is similar, but are more applicable to non-software works: books, articles, music, etc. (All of Wikipedia is available under a CC-BY-SA license.)
[Aside over.] Kris asked which licenses I liked better, and the answer to that is “nearly all of them.” My software is typically under the MIT license (or Artistic License, for Perl), but I also like the BSD licenses quite a bit, along with Apache, Mozilla, and of course the WTFPL. He responded that he sees copyleft as “helpful for maintaining communtiy ownership and preventing corporate hijacking,” which is of course true. He continued, “CC BY lets the immediate user do more, but the restrictions of SA (GPL), allow the next users to do more.” I couldn’t compress a response into 140 characters, so I’m writing this instead.
Kris (and the FSF) are of course right that works under the GPL/CC-BY-SA allow later users more freedom. My disagreement with these licenses (the GPL in particular) is purely a moral one. The ultimate goal of the GPL is to force all software to be open source. I realize that this is in actuality not more restrictive than copyright, but it goes against what I feel is the spirit of open source.
When I write software (or anything, for that matter), what matters to me is getting it out to the public. If other people find it useful, excellent. If they expand on it, great. My thoughts on the GPL are pretty well summed up by this article. Both the LGPL and CC-BY (along with the MIT, BSD, et al) licenses require that the initial license is included in the final work, but not that the derivative work also be open source. Because the original, open-source version is still available, my work is still open source and any later user can do whatever they like with it. I don’t think it’s my place to tell other people they can’t do what they want to with my work. That’s a moral decision, and either way is really ok with me.
The more interesting part of this discussion is the connection to academia. Open access is a great thing (and pays my salary, incidentally). The distinction between open access and open source in academics is much fuzzier than in software, however. (And I mean “open source” in Stallman’s terms now: access to source code. When Kris is talking about open source he is really talking about copyleft.)
What’s interesting to me is that academia has always been open source, but not open access. Essentially everything is CC-BY: if I want to use David Lewin’s work in my dissertation (which I’d hesitate to call a derivative work!), I can. I don’t have to get permission from the Lewin estate, or from Oxford University Press, I just have to say that they’re his ideas and tell people where to find them.
This is because the distinction between “source” and “content” is non-existent in the humanities in a way that is different from software (I’m certainly not the first person to have pointed this out, but I can’t recall having read it before). There is no secret “source code” to GMIT that is then run to generate the book; if there is, I’d certainly like a copy! Open source is useful in software because I can ship you a piece of software that is a black box: you can use it and have no idea how it works. That’s not the case in academia, though I have seen some people try.
The distinction then, hinges on “access,” and in the humanities I’d argue this distinction isn’t terribly important. Let’s say that there is a website that has the full text to Lewin’s Generalized Musical Intervals and Transformations available for free (open access). If I want to use this resource, it doesn’t matter to me whether it’s CC-BY or CC-BY-SA: if I’m a responsible scholar I’m going to cite it in either case.
The caveat is that CC-BY-SA does not prohibit commercial use (the relevant CC license is the NC provision). If I want to print off a copies of this mythical CC-BY-SA licensed GMIT and sell them to my friends, that’s perfectly fine (just as it’s perfectly fine to sell GPL-licensed code). In practice this doesn’t matter much: because the GPL requires me to include the license, this original version is always available. Users are much more likely to use the free version rather than the paid version, unless there is significant added value to the paid version (I’d gladly pay someone if they had fixed all the typos in GMIT).
I think open access is a great thing, and I think more journals/presses should be doing it, but I don’t think the distinction between open source and open access in the humanities is one worth spending a lot of time on.