Archives for October 2006

graph isomorphism

Graph isomorphism is a weird sort of central problem in CS theory. An efficient algorithm would be quite nice. It would also be an achievement of some note – quite a few people have been stymied by this problem over the years. It also has the interesting property that if graph isomorphism is NP-complete, then [...]

wiki

So why haven’t I used a wiki for my idea bin? I think the main reason is that I want it to auto-link files. I’m toying with the idea of stealing the tiddlywiki front end and combining it with the ability to add comments, openid authentication for comments and for the editor, RSS feeds with [...]

bill gates

Theory: Bill Gates is autistic enough that he views other people as objects, (as do many vaguely nerdy folk – this tendency is much more common than people think, but nobody likes talking about it) but has overcome this natural inclination through hard work. As a side effect, people in other lands are just as [...]

flamewars

What causes flamewars? How can we tone them down (or cause more and more)? I propose a study that creates a big matrix of features and then searches for fora that have each combination of features, and then calculating the percentage of messages that are flames. After doing this, it should be easy to tell [...]

erasing

How powerful are turing machines that can’t erase? A fun link that I can’t track down explained that the reason that graduate students are forced to do research and write papers is best explained by analogy to computational complexity: DFAs, by themselves, can’t do much. But turing machines can do almost anything – and a [...]

brands

I think brands and certifications can provide all the niceness of unions without a lot of the baggage. When things are certified organic, that means something to people. People can then choose whether or not to support organic farms, but they at least have a semi-informed choice. We’re seeing similar progress with fair trade goods, [...]