It's not about ASCII, it's about where that ASCII is presented whether it's terminal emulator like mlterm or a text box in a sane OS like Ubunutu. RTL languages can be detected and viewed correctly without any additional information.

Addendum: I'm using 'ASCII' only because in this context we are talking about 'plain text' and not the actual 7-bit ASCII. UTF-8 encoding is a given when talking about RTL languages.