I'll only make some brief notes on today's sessions. Hoping for a theme for today as well, I hoped to engage that side of me that spent a decade getting a doctorate in philosophy. You know, that liberal arts urge, at a technical conference. This year has been comparatively thin on sessions that concern social dynamics, legal issues, and histories of ideas. There has been some, but I miss seeing some of the scholars of intellectual property and freedom that have been at OSCon in previous years.
I saw an entertaining talk today by Robert Lefkowitz called Open Source as Liberal Art. I had run into Robert outside his session, and also saw his keynote last night. The keynote had been largely a standup-comedy routine that compared software development methodologies to Qintilian's Institutio Oratoria (i.e. a study of rhetoric; a field whose relative bad fortunes genuinely fills me with sadness). Not because Lefkowitz mentioned it, but just as a service to readers, I highly recommend Jay Heinrich's book Thank You For Arguing, a modern primer in rhetoric. As with all jokes, Lefkowitz' was a bit serious in his keynote. Today's talk, however, leaned slightly farther towards the serious and literal side of things. Lefkowitz presented a general classical opposition between techne and praxis (funny Greek words for... well, technology and practice). The former has been the traditional domain of the working classes, with the latter reserved for intellectuals and "leaders." So goes the tradition of "liberal education," in any case. In humorous fashion, Lefkowitz presented arguments for which side of things Free Software should fall on, tracing ideas of Knuth, Sussman, Marx, Babbage and a merry bunch of other thinkers who see code as a form of literature. Well, some do and some don't, but I think we should, and I think Lefkowitz thinks we should. If anyone wants to seriously injure their brain in this direction, a ponderous and profound philosophical work on this topic is Intellectual and Manual Labor by Alfred Sohn-Rethel. I don't actually think readers will read this classic of critical theory, nor even find a copy in print or at most libraries... but, hey, I'm not getting that decade back, so I better show something for it.
While it probably does not quite count as humanities, I always love topics related to machine extraction of knowledge from free form corpora. A talk entitled Machine Learning for Knowledge Extraction from Wikipedia & Other Semantically Weak Sources by Jamie Taylor, Colin Evans, and Toby Segaran (all from Metaweb) was quite interesting. Continuing today's tongue-in-cheek theme, Jamie Taylor dressed as a pirate, and the speakers presented the project they are working on called Freebase. Mostly what the speakers presented was various technical details of their scraping, slicing, dicing, and restructuring of Wikipedia as a large source of semi-structured knowledge. The underlying idea was to extract interesting connections and categorizations from the work of Wikipedia's millions of contributors. Their work sprinkled a bit of Bayesian magic around, some text processing, some use of semi-large backend databases. It's cool, and at least worth checking out their FOSS code. Or even more cool, they provide (in about 20 GB) a relational version of their filtered and processed Wikipedia content that lends itself to querying with regular SQL, and all the clever joins and filters that allows.