I try to get to R0ml's talks, and as one of the screeners, I bring my experience of "what's an OSCON". We like Perl people. Larry and family like R0ml talks. So do Alex and Anna (fellow Pythonistas -- she had the Google glasses this year).
This year he was talking about the reflex-conditioning around "secret source" and how it still impacted open source practices. Aren't "config files" just a holdover from not being allowed to compile in your favorite settings? It takes less code to say it in source, he showed in slides. And what about testing? If you're doing that because the source itself is closed to you, then you're a good little monkey, but studies prove more bugs are found by people reading code with comprehension.
On this last point I could feel audience unease and we had a hand go up. Defense of testing runs deep and R0ml was uttering deliberate heresies. Even though we knew that, it was hard to not formulate a come back, right there and then.
As background, I should come clean that I take the podium in favor of "test driven development" for money. It's a soap box, with examples, and I know in the business world that strict TDD was never king of the hill. What's TDD? You actually write the tests of your code before you write the code, and then your code is all about passing your own tests. It's like playing chess against yourself. The testing is looking for weaknesses. You do counter-intelligence against your developer self, and that makes your developer self develop, which is why TDD is a valuable plank in the "agile" platform.
The other talk I've been thinking about is the one on Race and Racism. She was good at making us endure a little awkwardness as the theory is Race is best confronted with such a willingness. Your first moves as a ballerina, as a wannabe, might make you feel awkward, given you're a blundering oaf. Shrek trying to play Tinkerbell.
However much I appreciate the Gibbsian / visceral approach to self analysis, I think we should also provide road maps, topologies for these topographies, ala Euler (I'm using these names with a Buckyesque spin). She was always asked "what are you" in terms of race and she was force fed the usual answer of "mixed" having had a "white" father and a "black" mother. The idea of "mixed" implies "pure" a kind of phenotype, and then we're back to the five, seven or eleven "races of man" depending on which Bible-informed Social Darwinism or other Anglophone pseudo-science we happen to be reading.
As I was saying at NPYM, I really don't trust the Anglophones to puzzle their way free of the Race concept, because they base so much on it. The investment is still there. Getting "beyond" race does not mean deconstructing the concept as anthropologists, but being "tolerant" of other races and such bull crap, buying into the basic premise, which is there's some essence or blood substance we're keeping track of, and fractional bookkeeping applies.
1/32nd Native American is supposed to make sense. And it does, after a fashion, when it comes to family inheritance, legal rights and so forth. Membership in a tribe may hinge on proving that 1/32nd is "real".
But this is not really genetics. The genetic code has no specific "race" gene and the kind of mixing that goes on is not captured by our feeble words like "white", "black", "yellow" and "brown" (or "blue" if you think aristocrats are divinely "blue bloods"). Those are kindergarten concepts, for politicians who choose to avoid science. But they're also very embedded in English, these concepts, which taints it and limits its utility.