So I was looking at this answer and was amazed by the compactness and unreadability, when I suddenly thought: "That are an awful lot of non standard characters!" So I get that coding golf languages utilize single characters for extended functionality, but as I remember it, a "byte" are 8 bit, so only 256 one-character-commands could possibly be used. And then I found out that the language in question uses more. So, are you really that lenient with encodings? Can one just select which parts of the command list are to be taken? But then that would be the same as counting any character as 1 byte. So shouldn't languages that are okay here only use up to 256 bytes?
(I understand that single characters can have an encoding bigger than 255 for display reasons, I just kinda disagree that more than 256 characters are allowed in general.)