ever tried crypo breaking on text caki? doesnt look like it
even though each of the characters change, you have a text file (in your example you stated a->z - perfect for example)
try making 'hello world i am a little bit of encrypted text with a bad algo'
all the h's are the same code, etc.. going by your example
so you begin this by breaking each portion, looking for repetitions..depending on the language (lets use english...) used in the text,
rules can be formed..such as i before e except after t, the chances of a word not having a vowel are slim, vowels will be the most
repeated characters, and space will be pretty common too...
so lets look at the example
firsly we do a character distribution matrix, calculating the highest used characters, from this we then form a dictionary attack to obtain the
best 'fit' for the word, and check that against the next text and so on
your crypt wouldn't last too long before its broken...
enigma is right joe, it was hard to crack because it was mechanical and each machine had a different cog setup... caki's example however
is not caesar, its a variant.. and a bad one at that...
enigma was also broken, not because of a code book, but because of the actual machine.. on the u-boat.. it was attacked and the
machine was obtained from it before it sank, then it was played with
frequency of words can be calculated, but its language dependant, like in english sentances will begin with 'the', 'i', 'im', this' and so on
from that, you can build a matrix.. and break it...
caesar text is typically rot-13, used by roman messengers to stop the 'secrets' being obtained if the messenger was attacked, it wasn't
really substitution, it was more or less adding the value of the letter in the alphabet to another sequence, looping if it overflowed
eg: a = m, thus m = a.. difference between them is 13, so u add or subtract 13 from the depending on where its situated in the alphabet
also called rot-13, and then taken to modern times with rot-18 where it does the same thing with numbers too..
any good cryptography will not do the crypt in the fashions described above, they will use mathmatics to do the end result
without knowing the formula and / or having the correct data/keys breaking it is tricky... some have flaws in the implimentation,
which is why they have been broken...
tea for example (tiny encryption algorythm) is used in safedisc, its a small fast routine (and easily portable to other chips/hardware)
however it can be brute forced, even though it is 128 bit... bruteforced in a few seconds actually... but thats another story