Lists Home |
Date Index |
Emmanuil Batsis (Manos) wrote:
> By far the best post of this thread.
> Currently, (X)HTML is mostly used for presentational purposes alone. I
> would agree that in 90% of cases, XML+XLink+CSS2+ECMAScript should be
> used to do what (X)HTML does today.
I disagree. The fact that HTML is a known vocabulary allows Google and
browsers to (for instance) detect titles reliably. Who knows what other
semantic markup they may use. Similarly, I suspect that vocalizing
browsers and text-based browsers depend on the semantics of the markup
rather than trying to "translate" the style. The more a tool knows about
teh semantics of the markup the more it can do something that wasn't
anticipated by the designer of the tool. Sure, your "unordered list" may
be styled in some cases as graphical bullets but Lynx knows that it is
really an unordered list so it does the right thing.
The trend is to put more and more semantics in CSS X (X>1) but that's
just a little bit of a slight of hand. Now CSS X becomes the
"architectural supertype" and your vocabulary becomes the architectural
subtype. CSS becomes both the mapping language and the target language
which is arguably not an appropriate separation of concerns.