29 December 2009
This is quite beautiful. Nothing beats asymetrical information in markets, especially if the other person doesn't realize it.
Computational complexity and Derivatives. For those of you too lazy to read the article and are willing to believe me, basically it is somewhat easy for me to construct a derivative which, with a computer, it can be quite hard for you to evaluate. There are whole classes of problems which can be hard to evaluate in any sensible length of time regardless of how much computational power you have. Note that 'sensible length of time' can easily be 'longer then civilization has existed on earth' lengths of time.
The paper is here
Another blog post is here
For those of you who the computational complexity idea seems bizzare, think of it this way. You want to pack your truck to move a long distance. You want to take as much as you can, you can only make one trip, and, the truck only holds so much. What's the optimial set of things you can put in the truck? The easy, and wrong, solution turns out to put the most important item in first. It might be a good enough solution, but, the right solution requires you to look at all of the combinations of items. It turns out this takes a long time. It also turns out that if I say that this combination of items is the best, it takes you just as long to verify that claim. This goes up by N! where N is the number of items. N! is N x (N-1) x (N-2) down to 1. This gets very big very fast. 1000! is 10 followed by 2566 zeros. So, if you can check 1000 solutions per second it takes you 2556 times the length of the age of the universe to find a solution.
Well, somewhat. For those of you who like your Christmas stores a bit darker and Wierder.
19 December 2009
Lisp and Graphics
L-system plant Drawing
08 December 2009
Yet again a long time since the last update.
Nice article on reverse square root and tracking who wrote it.
Interesting post on your looks and on-line dating.
A whole load of articles about Sergey Aleynikov. He's the guy accused of stealing code from Goldman Sachs and uploading it to a public site.
xp-dev is the site that he uploaded too.
Joel Reymont's take on the whole thing.
In the end it looks like much ado about nothing, ie, they couldn't build a case for some reason or another.
The actual complaint
One last word on Sergey
Last modified: Wed Jul 8 22:19:37 CEST 2009