David Woolley wrote:

>>1) <div> and <span> aren't sufficient, because I can't nest
>>    a <div> inside a span.  In practice, I'd want to be able
>>    to nest two or more divs ionside a single span, in order
>>    to be able to typeset in columns.
> 
> 
> That's abusing HTML as a page description language.

I really don't understand that assertion.  Why should
a page description language be predicated on a single
column model and force users to use floated elements in
order to achieve multi-column layouts ?  Surely the poorly-
supported "display:inline-block" model is tacit recognition
that HTML and early CSS were deficient in this very area ?

> That's why XML is called *X*.  You can define additional elements, as
> long as you maintain the well-formedness (which exists to support this,
> not to avoid confusing people with implicit tags).  You can use namespaces
> to avoid conflicts.

Yes, of course I can author in XML : but if I do so,
I cannot serve the resulting documents across the
web with more than the faintest hope that anyone will
be able to render them.  (X)HTML is the /lingua franca/
of the web, but it is closed (inextensible); what I am
arguing is that the /lingua franca/ needs to be extensible,
not that one should be forced to abandon it simply in order
to gain access to a richer extensible tag set.

 > [X]HTML is very much a lowest common denominator.  The original concept
 > was similar to the current Wiki concept, i.e. that it provided something
 > that any intelligent person could create and edit.

Exactly as was TeX : yet Knuth had the foresight to realise
that no closed finite set of tags could ever satisfy the
needs of "any intelligent person" so he deliberately made
the language extensible; the author(s)/creator(s) of HTML
seem totally unaware of this fundamental need for extensibility.

Philip Taylor