Re: Properties applicable to root?

>From: Todd Fahrner <todd@verso.com>
>3. For HTML explicitly, the BODY (and HEAD) elements should accept all
>normal box properties, just like any other blocks. The HTML element itself
>should take most CSS properties excepting those outlined in (2).

I'm not sure what you mean by the HEAD element accepting box
properties.  The question of rendering the HEAD element is left
somewhat open in the HTML 4.0 specs, but I tend to think it is not
advisable.  The only renderable (not SCRIPT or STYLE elements) content
of the HEAD element is usually the TITLE element.  (I can't think of
anything else.)  I guess if one wants to render the TITLE as part of
the page this could be useful.  Therefore the default user-agent (UA)
value for HEAD would be display:none, and the default UA value for
TITLE would be display: block.  (If such a change were made to the
specs, this should be added to the sample UA style sheet.)

A problem with such an implementation is that it could encourage people
to put BODY content within HEAD.  Furthermore, I think the display of
the TITLE could *almost* be accomplished within the current CSS2 spec
in the following way (and I think it would be a better way):

BODY:before {
	display: block;  /* or float could go here, except float is
		not allowed on :before, for reasons I don't really understand.
		Maybe display should be allowed as well, but I'm not so sure
		about this one.*/
	content:  /* Here's the "almost".  There is no way of referring
		to the TITLE, although the spec hints that there might be
		such methods in later versions */
	}

There is one other side issue, and that is the meaning of CSS
properties on the FRAME and FRAMESET elements.  I don't want it to
sidetrack Todd's proposal, but it is something to think about.  If my
memory is correct, Netscape treats attributes on FRAMESET as if they
were on BODY.  There ought to be a way to duplicate the HTML attribute
FRAMEBORDER (on FRAME) and the Netscapism (the only non-HTML attribute
that I use) BORDERCOLOR (on FRAMESET) using CSS, and possibly do some
other things with frames.  I seem to be in a minority among supporters
of standards in believing that frames are a good thing, because they
preserve bandwidth and simplify development.  (I just think there
should be a standard way of encoding specific framesets in URI's.)

David Baron

Received on Wednesday, 30 September 1998 17:41:28 UTC