Fair. Python's str.splitlines() treats \x1c \x1d \x1e as line breaks -- I assume due to a Unicode classification. I'm sure other string routines in other languages do as well. But it is not a string: libraries such as c0data-py process it as bytes. Nonetheless, I've amended that line and I've updated the spec to make that clear. They are control codes, so they have to be read as bytes. But the Unicode Control Pictures allow them to be usable as strings. Our editor plugins are intended to make that seamless.