Project Euler - What you've learned

[Edgardo]

So I registered on Project Euler (http://projecteuler.net/) and I must say: it's quite addictive!
I didn't expect it to be that much fun.

I'd like to share what things I've learned from solving the problems:


- When solving a problem you get access to a thread. I was impressed
how different the approaches are. People use "crazy languages" like K where
the code looks like this: */_*a@&1000=+/'a:b,'(_sqrt+/)'b*b:,/n,/:'1+!:'n:!1000

- Memoization (http://en.wikipedia.org/wiki/Memoization): Use results that you've calculated before

- Using psyco (http://psyco.sourceforge.net/) (a Python compiler) improved the Python performance of one my programs by a factor of 4.

- C++ is much faster than Python. But Python can handle really big numbers.
I will probably have to learn using GMP (http://gmplib.org/).

- Java is fast. But the BigInteger datatype is slow (maybe I used it in the wrong way?) and hard to use :frown:


What did you learn (for example what kind of techniques boosted the speed
of your programs)?
What programming languages do you use and prefer?

[chandubaba]

Hi, I registered in spoj.pl , www.topcoder.com , www.codechef.com, and many acm judges ... i am chandubaba everywhere. Read the book "introduction to algorithms" by cormen, awesome! Very good for all these sites. But i am not there in project euler ... i am not familiar with gcj stuff. Also we need to do a lot of math (learn it fun way, ie not preparing for exam but for fun)... dynamic programming and graph theory seems to be everywhere at the elementary level.

[Amanheis]

I learned:
C can handle bigger numbers than C++
Many algorithms are already better implemented than I ever could, especially in Mathematica
Thinking about a problem first can reduce the time needed to get the answer significantly