First off, after encoding your message, you have to find a way to tell your recipient which code it's in. This could be accomplished in a number of ways. You could either blatantly spell out the code ("nakedlunch"), or each code could be assigned a number or symbol. (nakedlunch=code1).

Also, a number or symbol must be assigned for the language the code is in, in addition to the code system itself (nakedlunch:english) or (1/1).

Also, it is my proposal that you should assign a value to another variable, namely, the amount of times it it run through the machine. Since it is not done with paper and scissors, you can easily run a message through a code three or four times. To un-encode it, you would simply run it through the unencoder the same number of times. Example: to tell someone to unencode a message you've written in the "Interzone" code in English, run through the machine three times, you might write: (interzone:english:3).

It is also my suggestion that particular symbols would denote a paragraph break, such as the "/" sign. This character could be programmed to never change, so that paragraph breaks would never be lost in translation.

Anyone wishing to read a more detailed explantion of my coding proposal should click here.

I added another encoding device HERE/