> HTML 2.0 seems like a pretty stable specification of existing
> practice mid-94 with a few relatively non-controversial extensions. I'd
> rather leave it that way and get it out the door.
I tend to agree, and really didn't expect to see this stuff
in HTML 2.0. I *would* like to see them in a stable,
published DTD before mid-95, though, whether that's called
"2.1", "3.0", "HTML '95" or (heaven forfend it takes that long)
HTML 2.0.
[...]
> CLASS is one of the examples where ignoring unknown attributes
> for upward compatibility makes a lot of sense. I tend to think
> we should give people some clues about the semantics before
> we do too much with it, or we may see mutually incompatiple
> implementations doing different thinks with it.
CLASS is an extension mechanism whereby users can encode
their *own* semantics. By semantics I mean logical roles,
not browser behavior -- my understanding is that
browsers should not do anything drastically different with
an element based on its CLASS attribute at all; mutually
incompatible implementations would be basically illegal
to begin with.
--Joe English
joe@trystero.art.com