May 22, 2007

Well, that sure was a hell of an update drought, wasn't it? I have a bit to show for it, though.

As always, more links are to be found on the Links page: Online Etymology Dictionary, f8d, Dresden Codak, Big Science, and Proverbs.

I've read a few more books: The Selfish Gene, The Blind Watchmaker, A Devil's Chaplain, The Satanic Rituals, and Age of Spiritual Machines.

I played through System Shock 2 again, and extracted quotes along the way. It's still a great game.

You should watch Beyond Belief 2006. Those guys are idea ninjas.

I watched the entire (first) season of Firefly along with Even, and now I've joined the ranks of the lamenting fans that Glench talks about. It truly is a jewel.

I've solved a bunch more Project Euler problems, recorded on Random.

When reading Dawkins' The Blind Watchmaker I was inspired by his description of a program that tries to generate the phrase "METHINKS IT IS LIKE A WEASEL" by both single-step chance (all tries are separate) and multi-step selection (if a try is slightly similar to the target phrase it's bred from in the next try), so I whipped up my own little test (it's immensely inferior; Dawkins' program reaches the phrase in around 40 generations, while mine does so in, on average, 10,000 generations). I might make it a Java applet, when I learn how, but if you want you can try to decipher it from the source. I'll try to explain how it works. First it generates a totally random string of equal length to the target phrase. It then breeds children from it (ten by default), and each child has a small chance of having one of its letters changed to a completely random character. The children are then rated, and the one that's most similar to the target phrase becomes the parent of the next generation. It goes on like this until it reaches the target phrase.

It could probably be made much faster by using sexual reproduction rather than asexual reproduction, by having mutation shift a letter slightly up or down the letter array instead of mutating it completely at random, and by tweaking the parameters (number of children, mutation rate).

09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0

<< | Previous entry (March 12, 2007) | Next entry (June 7, 2007) | >>
Back to Archive