Re: Current state of the summary discussion

Martin Kliehm, Thu, 17 Dec 2009 20:25:18 +0100:
> On 17.12.2009, at 20:00, Leif Halvard Silli wrote:
>> Martin Kliehm, Thu, 17 Dec 2009 18:08:10 +0100:
>>> So it seems a main objection is that @summary metadata is hidden. It
>>> struck me that it is the same with the @datetime attribute of <ins>
>>> and <del> elements. The difference is that there's also the <time>
>>> element in case you prefer datetime to be visible.
>>> 
>>> So for consistency, why don't we keep @summary which is also
>>> backwards compatible, and add something redundant like a <summary>
>>> element for people who want to make it visible or change the
>>> visibility using CSS? Thus everybody should be happy.

> I believe since nobody objects @datetime and <time> in its two 
> manifestations, using it as an analogy may help to get the point 
> through.

It is easier, probably, for an developer to programmatically update the 
date than to do the same with @summary.
 
>> My idea is to join <summary> with Ian's current proposal to allow
>> <caption> to contain more than a caption is currently allowed. Thus
>> make <summary> a container for this additional info.
> 
> I understand the rationale, but do you think it is clear for authors 
> where the difference is between <caption> and <summary> if one is 
> contained within the other? They both suggest something similar.

If they don't understand, then how can we accept that the HTML 5 draft 
advices authors to put exactly that kind of additional information 
directly inside the <caption>?  At least, if they have to put such info 
inside a <summary> element then the "real" caption is separated from 
the "helpful text" part of the caption. This would also make the 
content of the <summary> programmatically detectable for AT users.

As for <summary> as direct child of <table>, if that was your idea, 
then this is just another variant of an invisible @summary, because one 
needs to stand on one's head in order to make it work at all, so a 
invisible element seems like the only option. (Webkit spits most 
non-table elements out of <table> - so I actually don't know if it can 
work - as things stand.)
-- 
leif halvard silli

Received on Thursday, 17 December 2009 20:43:45 UTC