I thought I was the *only* ERCS advocate here...
Anyway, it would allow the above. More importantly, it allows Chinese
to be use in attribute values, and it also provide a common way for
defining the mapping of numeric character entities.
ERCS provides a framework for treating many languages
as equals, and while it is somewhat biased toward allowing
native-language markup (as in the above), it will also provide
benefits to fixed-tag cases like HTML. It is not perfect, but it's a
good 98% solution.
I believe you know all this though...
Anyway, I guess everyone here is tired of hyperbole. My revised paper
contains little hyperbole, much discussion, and some conclusions and
recommendations. I'd appreciate the chance to get it finished, and to
send it out, before we discuss this. It should provide a good, solid
base for rational discussion.