> > [Joe English wrote:]
> > I think it would be a good idea to orthogonalize these attributes;
> > that is, leave CLASS attribute values entirely in the user's
> > namespace, and use ROLE for everything with predefined semantics.
>
> To get the best out of the subclassing idea, we will want to evolve
> some common class names. I am merely helping things get off to a
> good start with some root classes that should be useful for all
> browsers.
That's not a bad idea, as long as the predefined semantics
for classes don't cause a drastic change in the way
browsers present the elements.
For example, I want to avoid the situation where I make
a bunch of intra-document links to paragraphs tagged with
<NOTE ID=FNn CLASS=footnote>...</NOTE> at the end of a document,
and six months later they mysteriously turn into iconified
pop-up notes because of a change in the spec.
Hmm... looking at the 01-Mar-95 DTD, the only predefined
roles and classes that cause what I would consider a
"drastic change" are <NOTE role=footnote> and (possibly)
<DIV class=banner>. I believe it would be cleaner to
reintroduce <FN> instead of using an attribute for pop-up
notes anyway.
Only <NOTE> and <HTML> have ROLE attributes; everything
else uses CLASS. If there is to be no "user-reserved" namespace,
is there any essential difference between ROLE and CLASS?
If not, the two attributes ought to be unified.
--Joe English
joe@trystero.art.com