That's actually my point. Python supports Unicode code points and UTF. If you get the output encoding in UTF-8 it would actually be variable length chars. What's important is your coding output, not the internal code point representation.