First of all this en/decryption method is inspired by the symmetrical encryption methods like DES, AES etc. But the algorithm for the encryption is a little other.
For understanding, how the encryption is working, I will make a simple example. Let say we have a given plain text in 2 Bits: 01 10 01 11 10 01 11 00
The idea of the encryption is to take always two of the given plain text code (in this example the code has the length of 2 Bits). We will see later, what this means.
To can encryption we need a function where the input will give something else in the output (this is called a s-box, but with this algorithm it will be a normal function). So let say we have an input x and a function f(x). When we apply some input to the function f, it will give an output y. It means, that y = f(x). This function must be something special. The following property should be true: f^-1(f(x)) = x. The f^-1 is the inverse function of f. This means that when you get an output y with applying f and x, then you should get x with applying f^-1 and y.
To get a feeling with this, let have a look at the following table:
| x | f(x) | | --- | --- | Because our plaintext has 2 bits per character, we will need a function which will take 4 bits. | 0000 | 1100 | | 0001 | 0000 | | 0010 | 0101 | | 0011 | 1111 | | 0100 | 1011 | | 0111 | 1000 | | 0110 | 1101 | | 0101 | 1010 | | 1010 | 0100 | | 1011 | 1001 | | 1000 | 0111 | | 1011 | 0011 | | 1110 | 0010 | | 1101 | 1110 | | 1110 | 0110 | | 1111 | 0001 |