Re: Grouping in Definition Lists

On Tue, May 18, 2010 at 3:09 PM, Dean Leigh <dean.leigh@deanleigh.co.uk> wrote:
> On 05 May 2010 17:00 Tab Atkins Jr. Wrote:
>>> XHTML2 is a dead spec.
>>> The W3C has explicitly closed down the working group tasked with it,
>>> and it will not be developed any further. You shouldn't look to it for any guidance.
>
> For guidance I agree, but I would argue it is a valid as a point of reference for debate.
> However, I understand your perspective and in future will try to avoid reference to XHTML2.

You don't necessarily have to avoid reference to it.  ^_^  I'm just
saying that anything done in XHTML2 has no bearing on what is done in
HTML5.

>>> All of the legacy browsers that were around in 2007 are still around.
>>> They have somewhat less market share, but are still significant in
>>> terms of dictating what we can do with parsing.
>
> Firstly, is there a "cut off point" for supporting older browser versions, even those that were quite deliberatly non-compliant?

It doesn't matter how deliberate their non-compliance is if they have
significant market share.  ^_^  "Significant" is somewhat subjective,
but anything more than a percent or two is usually significant.


> Secondly, does grouping DTs and DDs in something like DIs really cause parsing issues, in older browsers, surely they would ignore the tag?

Test it out!  I recommend looking up previous threads on this mailing
list about using <dt>/<dd> in <figure> and <details>, to see what sort
of problems we had there, and see if you can find any similar parsing
problems.


>>> HTML5 does not define a <di> element, since the grouping is completely
>>> expressed in the current markup.
>
> Many posts have argued that single DTs could/should exist without a DD.

Can you point to any?  I confess that I don't understand what a lone
<dt> should mean.


> This combined with valid multiple DTs means that the counting DTs following DDs is unreliable as an indicator for grouping.
> Therefore I would argue that the grouping is not "completely expressed in the current markup"

It works fine as long as you avoid trying to mark up lone <dt>s and
<dd>s.  ^_^  Like I said, though, I don't understand what a lone <dt>
or <dd> would mean.


>>> If there is a lack in the CSS side (which I agree there is),
>>> it needs to be taken up with the CSS group (www-style@w3.org).
>
> Even if this were not the case surely CSS still requires a tag to reference.

Not necessarily.  There have been proposals in the past for a ::di
pseudoelement.


> In summary:
> DL. DT. DD. Seem to be a natural way to present common content/presentation combinations such as index pages consisting of:
>
> 1. A clickable thumbnail image (DT)
> 2. A clickable tect link (DT)
> 3. A description/summary in plain text (DD)
>
> We can group in UL and OL by nesting but they only have one type of child element, LI.
> We group related but different form elements in the Fieldset element.
> The DL element is unique by not allowing grouping where it natually occurs in both data and language.
>
> In the abscence of new HTML tags (which I would assume we are trying to keep to a minimum) we could make very good use of an existing one for the sake of a small adjustment.
> I propose again the DI or DG (definition group).

I agree that being able to visually group together <dt>s and <dd>s for
CSS purposes would be great.  I have run into problems trying to use
<dl>s to mark things up precisely due to that lack, and had to change
my markup structure (either using a <ul> or <table) so I could style
things the way I wanted.  But that's purely a styling issue.  I've
never had content that I wanted to put in a <dl> that I was unable to
because of some *structural* problem that would be solved by a <di>
element.

Thus, this is a CSS issue, and should be handled in CSS.

~TJ

Received on Thursday, 20 May 2010 23:26:20 UTC