Re: H1 semantics

Daneel Pang Swee Chee (daneelp@ncb.gov.sg)
Fri, 13 Aug 1993 14:33:15 +0800 (WST)


:)Klaus Harbo said:
:) HOWEVER, I my view (as I think I've stated before, only then with
:) regard to P elements) HTML ought to be redesigned to have richer
:) structure, ie. using containers much more. The present HTML DTD makes
:) document instances flat, being mostly a sequence of elements.

I agree with this completely. Over on our side, we have "extended"
HTML a little for our needs, which, amongst other things, included the
ability to share information at a structure level (for example, my chapter
2 is actually your chapter 7 or even your section 5).
However, we wanted "compatibility" with HTML so that we can share the vast
amount of resources available there.
We had a hard time with <P> basically because it's not a container in
HTML but should be one in our case.

:) <!ENTITY % h1contents (h1title,(A|P|H2|%text;|....)+)>
:) <!-- this sketch only serves to demonstrate
:) that H2 elements would nest in H1 elements. -->

:) <!ENTITY % h1contents (title,(A|P|H2|%text;|....)+)>
:) <!-- with a generic title element -->
:) which is much nicer, which - again - illustrates the desirability of
:) using container rather than sequencing everything.

However, you'll loose the ability to utilize old HTML documents which
are already out there, wouldn't you? since your "H1" is now a container
for both the TITLE and the rest of the "sub-document" (similarly for
H2, H3....).

What we did was to define it this way:

<!ELEMENT HD1 - - ( H1,(%s.zz|HD2)*) -- part -->

-- 
[Daneel Pang]              | A well adjusted person is one who makes the same 
daneelp@ncb.gov.sg  	   | mistakes twice without geting nervous.
(65)772-0517               | mistakes twice without geting nervous.