W3C home > Mailing lists > Public > public-html@w3.org > September 2009

Re: [Bug 7508] <dialog> needs a way to add non-speech related information

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Thu, 10 Sep 2009 13:52:04 +0200
Message-ID: <4AA8E864.2070306@xn--mlform-iua.no>
To: Toby Inkster <tai@g5n.co.uk>
CC: HTML WG <public-html@w3.org>
Toby Inkster On 09-09-09 23.45:

> In the case of, say, a block of stage directions inserted between two  
> lines of dialogue, I'd suggest the following markup:
> 
> 	</dialog><aside>...</aside><dialog>


That option is definitely there - regardless. A - or some - micro 
formats could define this in detail.

 
> The directions are not part of what is being spoken, so I think this  
> would be the correct way of handling them. It might be nice to have a  
> "follows-from" attribute containing an IDREF allowing a <dialog>  
> element (and indeed all the list elements) to be explicitly marked as  
> following on from a previous one. Its semantics would be to specify  
> that logically the two elements constitute the same list / piece of  
> dialogue, but it has been interrupted to allow some other text to  
> appear.

Seems like a good idea. OTOH, one could, I guess, also just place 
a heading - same level each time - between each  list?

How does this usecase differ from - say - an alphabetically sorted 
glossary with a heading for each letter?

<h1>Words on A</h1> <dl ... /dl>
<h1>Words on B</h1> <dl ... /dl>

> The case of e.g. timestamps appearing by a speaker's name is  
> different though. They're not an interruption from the dialogue, but  
> additional metadata about it.
> 
> To cover this, I'd suggest that the HTML5 draft be modified to,  
> instead of saying that <dt> indicates the speaker of the following  
> <dd>, say that it contains brief metadata about the following <dd>.  
> This metadata could include, but is not limited to, the name of the  
> speaker, the date or time it was spoken, and the manner it was spoken  
> in.

I think it is important to also decide what a dialog is. The 
<dialog> elements currently offers a narrow view of what a dialog 
is and contains.

A dialog is more than the words that are uttered. Joining/enter 
and leaving a dialog is part of the dialog. Muting is part of 
dialog. Closing your ears is part of dialog. Nothing of this 
belongs in headers or in an aside.

(Well, if you don't want to go into detail about it, you could of 
course perhaps lump such things into <aside> - you could keep 
<dialog> list inside <aside> with such details. ;-) But you could 
do the same with parts of the spoken dialog that you were not 
interested in presenting.)

Instruction that go into great detail should be placed in <aside> 
and the like. But references to an act in a dialog should not 
necessarily go there.

I have said that such info as "joined" or "left" should go into 
<dt>. It could be argue that it should go into <dd>, as well. If 
we bring <cite> into the picture:

<dl><dt>Leif <dd>Joined.
     <dt><cite>Leif</cite><dd>Hi, folks
     <dt>Leif <dd>Left</dl>

Here those <dd>' that aren't preceded by a <dt> containing a 
<cite> do not represent a spoken word.

OTOH, the dialog arguably becomes simpler to read and follow if 
"joined" and "left" is kept in <dt> since the spoken/said/uttered 
words after all is the main part of a dialog.
-- 
leif halvard silli
Received on Thursday, 10 September 2009 11:52:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:48 GMT