Comparing algorithm

I'm working on a program that basically compares strings to find similarities in the way the have been generated, and so it finds the encryption key. More are the strings, better are the results...basically, with this you can create any keygen ( not only for piracy, but also for testing serial code algorithms )...It's simply to do I know, but I want to optimize the code to have more accuracy in the results. I wrote a bit of it in C ( the final version will be wrote in Python ) but I'll post it later, because it's too buggy for now and also I need suggestion about optimizing the algorithm...any idea about improving the speed? I use recursive functions and for cycles for doing the comparitions, and for now a bit of inline assembly for memory handling.
I'm also working with a different version of the program to decode tripcodes, or at least for finding the key to decode them all....I know their algorithm is very weak and they are already been decoded, but I want to do something better decoding also secure tripcodes. Ideas?

OK, tripcodes and many registration keys are generated by a crypto hash function, typically MD5/SHA.

To find the key to decode a tripcode, you have to break a hash function -- that is, given T and H (tripcode and hash funct.), you have to find T=H(S). Finding S given H and T is extremely hard -- it's called preimage resistance, which is a guiding property of hash functions.

What does this all mean? Well, it's simple, really...

If you think you can break a hash given a large number of inputs, you don't know enough about mathematics and computer science to do jack shit. GTFO

Your entire premise is fucked. Please LEARN2HASH and try again.