Knowledge Packaging (for Content)

Late in the afternoon at the Free Culture UK meetup back in April there was a discussion of the analogy between code and content in relation to open production models. This came up as an aside to the main discussion but it raised some very interesting points directly related to my ongoing consideration of the analogy between open source and open knowledge. Below is a highly impressionistic rendering of the relevant portion of the discussion:

John Buckman: Content is just different from code. If you lay down a guitar track you lay it down once. You don’t produce version 1, version 2, … version 7. And if it is going to be reused it gets reused a couple of times it isn’t like a code library.

Rob Myers: I think that is only true for ‘manufactured’ culture. For ‘folk’ culture, be it songs, myths etc they would get changed, and improved continuously.

Rufus Pollock: Pace Rob’s comment I think John’s got a good point there in relation to content but that’s because content is very highly coupled and is hard to componentize which is what is necessary if you want significant reuse. If you look at knowledge more generally I think there is a pretty good analogy, we just haven’t exploited much of the potential yet. For example if you look at databases the analogy is almost perfect: you’ve got bugfixing, versions, releases etc.

John Buckman: …. because on cc-mixter why that works is because [artist -name] went out and, for example, broke down all the guitar samples and labelled them with BPMs etc etc. Relatedly it would be much more useful if more people just contributed the raw track rather than adding in the distortion — you don’t really want that, it actually makes it harder to reuse.

Rufus Pollock: That’s fascinating. So he [the artist] went and **packaged** them [the guitar samples] — that’s really important. I think we’re going to see more and more of that. It’s just like at the start of software. It takes a whole bunch of effort to create a library suitable for reuse, to package it, document the APIs etc etc. At the start you just wrote your own code and that was it. It is only once there’s enough material — and that has only happened recently, flickr’s only 2 years old — that you start putting the effort into packaging, into creating the API — which in this case is the metadata.

Rob Myers: On the point about the raw track I had an argument with David Berry about this recently. I’d been saying that it now that it was possible you should make available the whole set of raw tracks for a recording and he said he couldn’t see the point.