Earlier I made a comment as to the problem with the term HTML semantics claiming that none presentational elements would make the term redundant. I also posited elements such as div and span which are considered non- semantics my are in fact semantics in the sense of meaningful.

OASIS's Robin Cover collected well over 400 XML vocabularies before giving up. No one document description could sensibly organize the breadth of human sense of meaning. IMO we must work with HTML DNA's tortured history of gene splicing, congenital defects, and ambiguity.

The term "content engineering" offers a way to thread this tenuous jungle. It recognizes the separation of concerns in page design, hence the appropriateness of elements for particular intent (to represent different domains: layout vs affordances vs the message proper).