Partly in response to my recent post on open source anthropology, Rex over at Savage Minds posted a nice summary of the problems with open publishing models in anthropology and the recent failure of Anthrosource. Getting beyond publishing, however, I was hoping we could address some bigger issues. Below is my comment to Rex's post:

Talking about open publishing models is important, but I don't think we should be stuck with the narrow view of anthropology or ethnography as simply a product rather than a process. The Halloween memo points out that process is the thing that open-source software does really differently than its proprietary competitors (e.g. Microsoft). And once we start talking about process there are many more opportunities for open source anthropology than open publication models and projects like AnthroSource provide.

I am working on a research team studying kids' informal learning with digital technology. Like Rex we're also worried about our work not being ready for primetime, but we're committed to open sourcing our process as much as possible in order to develop a community of 'developers' and 'consumers' around our work. So we're doing it in stages:

  • We use site-specific wikis to share fieldnotes and other research documents within the various working groups.
  • We use a private blog to share research memos and links, and to discuss relevant literature amongst the larger research team.
  • We're talking about sharing informal reports on our findings and developing themes on a weekly basis via our website (not yet up…).

Why shouldn't more anthropologists be opening up their process the way we're trying to – and the way the FOSS community has? In The Cathedral and the Bazaar Eric Raymond laid out many of the principles that make open source work – principles that could work for anthropology as well. Let's drawn an analogy, for example, between debugging code and analyzing ethnographic data ('debugging culture?'). Stallman says 'parallel debugging' (thousands of developers simultaneously working to identify and fix problems in the code) is essential to the open source process because, as Linus Torvalds said, every problem is bound to be transparent to somebody. I'd say the same for the task of analyzing ethnographic data. Sometimes we miss the forest for the trees. Sometimes we can't see the obvious because we're too entrenched in the data. Sometimes a colleague or an informant has just the right knowledge and context to provide that 'ah hah' moment.

I'm not saying it's the sort of thing that all anthropologists ought be doing, but rather that it's the type of thing that would truly constitute open source anthropology. In a nutshell it's this: adopting open access publishing models is about the product, but debugging culture is about the process, which is where the open source model really hits its stride. Could anthropology be open to this? Yes, I'd like to think so, but maybe not anytime soon. (See previous argument)

Update: Thanks to Joe for pointing out both that this post had no title and that I mixed up Richard Stallman and Eric Raymond. That'll teach me to post on the fly!