- From: Joe D'Andrea <jdandrea@gmail.com>
- Date: Fri, 4 May 2007 12:13:42 -0400
- To: "M. Jackson Wilkinson" <jackw-w3c@jounce.net>
- Cc: public-html@w3.org
On 4/30/07, M. Jackson Wilkinson <jackw-w3c@jounce.net> suggested
using li inside dl:
> <dl>
> <li>
> <dt>Item 1</dt>
> <dd>Item 1's associated content</dd>
> </li>
>
> [ ... ]
>
> </dl>
Wonderful! I've often found myself wanting to group two or more
related dt and dd elements.
At first glance I really like the leveraging of li here. Sure, there's
the allure of "another CSS hook" but it also demarcs a group of
terms/definitions.
However ... :)
Consider this statement: "Any HTML4 document should render correctly
as an HTML5 document."
Fine. Let's try a quick acid test. From a DTD perspective (and I
realize there are strong POVs on the suitability of DTDs for
conformance checking, so bear with me), dt and dd would be permitted
in _any_ list that uses li. I suspect we don't want this:
<ul>
<li>
<dt></dt>
<dd></dd>
</li>
</ul>
Or this (?):
<ul>
<dt></dt>
<dd></dd>
<li>
<dt></dt>
<dd></dd>
</li>
</ul>
OK, how about _di_ as a definition item? Wait - scratch that - I sense
a slippery slope (dt becomes t, dd becomes d) and then we have the
potential for:
<ul>
<dt></dt>
<dd></dd>
<di>
<t></t>
<d></d>
</di>
</ul>
I am intrigued! I'm just not sure how to introduce it sanely,
especially if HTML5 is intended to be a superset of HTML, compatible
with previous iterations and/or existing web markup.
Put another way, the "definition list bed" has been made. Now we must
lie in it. Would be nice to freshen up the pillows though. Hmm ...
--
Joe D'Andrea
www.joesapt.net
Received on Friday, 4 May 2007 16:13:44 UTC