Re: Draft text for summary attribute definition

Gez Lemon 2009-03-01 11.19:
> 2009/3/1 Leif Halvard Silli <lhs@malform.no>:
>>> 1: The summary attribute isn't a property of the caption element, but a
>>> property of the table itself (its purpose is to describe how to read the
>>> table, not how to read the caption).
>> Could it become a real problem that authors would  think that
>> caption@summary is describing how to read the <caption>?
> 
> Yes, it would be a problem. 

Of course, yes. But this is not what I asked. So I answer myself: 
It is unlikely to happen that authors will think that @summary is 
a summary of the caption content. After all, what would a summary 
of a tittle be? A one word sentence?

> The summary attribute is a property of the
> table, and in no way related to the caption element. All it would do
> is add confusion to something that already seems confused. I don't see
> the point in making it more confusing.

Don't you see <caption> as a property of the table?

Perhaps we should say that it is meaningless to a have separate 
caption element. It would be better to have a @caption attribute 
of <table> so that authors understsand for certain that caption is 
related to the table.

>> It is exactly because authors needs to understand the difference between
>> titling and summarizing that they need to be close.
> 
> Relating unrelated concepts does not aid understanding.

@summary and <caption> are both related and unrelated.

>>> 2: There isn't a strong relationship between the caption element and the
>>> summary attribute; the caption element isn't required, but that doesn't
>>>  mean a summary shouldn't be provided.
>> Both <caption> and @summary are optional. So why not keep the optional meta
>> info in the same element?
> 
> Because making the summary attribute dependent on the caption element
> reduces opportunities where the summary attribute can be used - if a
> caption isn't provided, it's impossible to provide a summary with this
> proposal (unless you provide an empty caption element, but the caption
> element shouldn't be empty if it's provided). As the summary attribute
> is no way dependent on a caption, serves a completely different
> purpose, and is a valuable accessibility attribute, it doesn't make
> sense to reduce the opportunities to provide a summary attribute.

As you admit below, caption@summary doesn't reduce any opportunity 
where it can be used. For instance, no one would add - in David's 
words - "narrative" (aka @summary) to a layout table. And neither 
would they add a title (aka caption). So they two are *extremely* 
related.

>> In my proposal, caption will be needed to provide a summary, (as long as you
>> want to write undeprecated code). <caption> itself can be empty though. As
>> long as <caption> is emtpy, it will not caption any attention in visual user
>> agents.
> 
> An empty caption element is an ugly hack just to make the summary
> attribute dependent on the caption element.

We can discuss if it is a hack. But you have not hit the spot when 
you characterize why I propose this.

You are mistaking me for anothor ghost.

>> The only drawback I see here is that it will require more to add the
>> @summary since one needs to add the emtpy caption element first.
> 
> The fact the summary attribute is in no way related to the caption
> element is also a drawback.

This remdinds my about me and my wife. When I say "similar", she 
say "no, completely different".
-- 
leif halvard silli

Received on Sunday, 1 March 2009 16:01:02 UTC