To be World-Wide, the Web must be usable by people who don't speak the language of its developers. In general, it makes little sense to develop a format for the exchange of information that is only usuable for one or two human languages. Thus, if a specification allows human readable text anywhere, it must allow text in any language. That text must allow all characters from the Unicode repository and it probably should allow multiple ways of encoding them, including at least UTF-8. ASCII is not forbidden, but it cannot be the only encoding.
Don't assume that text is always horizontal, left to right and that there are spaces between the words. Not to everybody does 7/1/92 mean January 7, '92, and in many languages, 3,141 is close to pi and 3.141 is a thousand times larger.
Also, when you write a specification for the W3C, remember that W3C specifications are written in English, but mostly read by people for whom English is only a second language.