* element IDs
* SRC attribute on LI, UL, HR, and others
* the <DIV> element
* CLASS attributes
While these are not as sexy as tables, centered text,
and stylesheets, they will be very useful for creating
robust and long-lived documents.
I'm wondering when it will be safe for authors to use, for
example, <UL SRC="greenball.gif"> instead of <DL> and
multiple copies of <DT><IMG SRC="greenball.gif"> or some
other form of tag abuse to get the effects they want; or
<DIV ID=foo> instead of an empty <A NAME=FOO> at the
beginning of a section; or <HR src="redline.gif"> instead
of <BR><IMG SRC="redline.gif" ALT="-------------------------"><BR>.
(The latter construct seems to be less popular lately,
thanks mostly to Netscape's <HR> enhancements; now that
authors have something there to play with they're less
tempted to say <IMG> when they mean <HR>. This is another
case where NHTML has improved rather than degraded the
average quality of the HTML on the Web.)
These features would be really easy to support. Just
look for element IDs in addition to <A>nchor NAMEs
when resolving URL fragments, and treat <HR> like <IMG>
if it has a SRC attribute. (<UL SRC=> and <LI SRC=>
may be a bit trickier, but it's gotta be easier than,
say, TABLEs, which many browsers can do now.)
Minimal support for CLASS and <DIV> is trivial:
just fail to barf when encountering them. So in a
sense, most browsers do support <DIV> and CLASS,
but it would be nice if <DIV ID=foo>, <DIV ALIGN=CENTER>,
etc., worked as expected.
--Joe English