>Perhaps this is naive, but could someone explain to me how one can hack/decrypt even a relatively simple encoding algorithm? For example - suppose my algorithm is to add 1,2,3 respectively to each character's ascii code of the string I'm encrypting. Therefore the word "secret" becomes "tgfsgw". Now other than just flat out guessing (and the clue that it's a 6 letter word), how could smeone determine that the original word was "Secret" given ONLY "tgfsqw"?

The hacker may not know that "tgfsqw" = "Secret" (initially); however, he can pump his own "clear text" strings into the algorithm and retrieve the "cipher text" that pops out the other end (into a table, communications line, whatever).

Then, he can take his cipher texts to determine how the algorithm encrypted his clear texts in order to figure out "your" clear text (ie. "Secret").

The more complex the algoritm, the harder the "reversing" process is. I like to use more that 1 variable; eg. the clear text password, and say, the user Id as a seed.