> As for <EM> and <STRONG> these are ok for some purposes, but comprise
> too small a set.
No, they comprise too *large* a set -- there's no semantic difference
between them.
As I've noted in other circumstances, most of English literature has
gotten along with a *very* small set of procedural markup -- basically
italics and left and right margins. Italics are used for basically two
purposes (though Fowler's _Modern English Usage_ cites 8 usages, these
can be grouped into just two major categories): emphasis and
alternate-usage (such as ship names, foreign words, book titles,
newspaper names, variable or type names in programming languages).
Sometimes capitalization, bold face, or underlining is used to convey
alternate-usage. (The only real use for margins is to indicate
quotations or excerpted material.) But what realling needs to be
communicated is not <B>, but rather <ALTERNATE-USAGE>, not <I>, but
<EMPHASIS>. There are no agreed-upon conventions as to when bold face
is used in a document -- so you can't tell what it means if someone
sends you a document that merely says <B>. That's the crucial
difference between descriptive and procedural markup.
Chapters, paragraphs, and sections are typically indicated in English
literature with with a mix of presentational and procedural markup, and
it makes some sense to include descriptive markup to handle these.
I think the real driving force behind the welter of markup now in HTML+
is the desire to make posters, or pictures, or whatever you might call
them -- non-textual art -- combined with the misapprehension that HTML+
is the only document formatting language we can use.
Bill