I found the assignments to his class to be really weird and challenging. The first one had you fill in a few functions of a half-complete percolation simulation, but without having written the other part of it, it felt like having to reverse engineer and guess at the solution. There was no real way to measure your progress. Either it runs or it doesn't, and there's no incremental progression towards a solution.

I'm curious if anyone else experienced this or could point out where I went wrong. I really wanted to do the class and I liked the lectures, but i started it after finishing Gardner's Stanford algorithms Coursera class which had some of my favorite exercises ever. They had you write an algorithm in it's entirety, gave you sample data sets to check against, and let you write it in any language. Compared with that, Sedgwick's assignments were like coding with one hand tied behind my back.

Also on Safari (both the books and the classes)