Re: summary attribute compromise proposal

Maciej Stachowiak On 09-08-05 00.20:

> On Aug 4, 2009, at 2:53 PM, Leif Halvard Silli wrote:  ....


[ skipped some interesting parts for some main points: ]

>>> That being said, by my compromise proposal it's perfectly acceptable 
>>> for authors to use the summary attribute if they find that, on 
>>> balance, the reasons to choose another alternative do not apply.
>>
>> May we say that you suggest authors use @summary when the summary
>> should be hidden. And else that they go for Ian's option?
> 
> I would put it the other way around. Authors should make the table 
> description visible to everyone, unless there's a clear reason not to in 
> a specific case, in which case one technique they could use is @summary. 
> That's pretty much what my proposal says.

I agree that even if we agree that there is a "hidden summary" 
usecase, we could still generally advice authors to consider twice 
if they really have such a use case, so as to create visible 
summaries as often as possible. So far so good.

However, for the "hidden summary" usecase, why should @summary be 
warned against, unless hidden summary container elements would 
also get the same warning? Any such warnings squarely aimed at 
@summary would fail to send the "visibility for all" message.

The problem, however, if we would like to warn against both kinds 
  of invisible summaries, is that the media information is hidden 
for the HTML validator, since the media applicability is part of 
CSS and since the media attribute is only allowed inside <style>, 
<link> and (perhaps?) <a>.

A possible workaround could be to allow the media="" attribute 
directly on the summary text container element(s), so as to make 
it/them visible or hidden to different media without the use of 
CSS: Imagine a hypothetical <summary media="*"> were omitting 
@media would be an *error*.

At the same time, for User Agents, whenever <summary> fails to 
contain the obligatory media attribute, then its media 
applicability would default to the same as for @summary - roughly 
"screenreader, braille, speech".

Thus, to make it visible to all users, it would be necessary for 
authors to explicitly set the media attribute to media="all". 
(Authoring tools could default to inserting media="all".) Hence, 
if an author explicitly types media="screenreader, braille, 
speech",  then validators should consider it a "conscious act" and 
not display any errors.

Thus, we would place @summary and "<summary> without @media" on 
equal footing. They would both be errors.

(The name(s) of the exact summary element, is not something I'm 
married to. An ARIA attribute that turns something into an summary 
element - and thus permits @media - is also thinkable.)
-- 
leif halvard silli

Received on Wednesday, 5 August 2009 02:44:41 UTC