Re: Draft text for summary attribute definition

Hi Leif,

On Mar 1, 2009, at 1:55 PM, Leif Halvard Silli wrote:

> Robert J Burns 2009-03-01 19.24:
>> Some are arguing with you that your proposal doesn't make sense. I  
>> cannot imagine what they would be thinking in that respect. If we  
>> were designing HTML from scratch I could easily see attaching the  
>> 'summary' attribute to the 'caption' element.
>> However, I think given that we already have legacy content and  
>> legacy UAs that support a 'summary' attribute on the 'table'  
>> element instead of the 'caption' element, it is more important to  
>> discuss what compels us to make a change now. That's the only part  
>> I'm having trouble understanding. Indeed there are similarities and  
>> differences between captions and summaries and the common confusion  
>> by authors between them needs to be addressed. However, I don't yet  
>> see how moving the summary from the table to the caption  
>> substantially addresses the confusions. Indeed I'm concerned it  
>> might introduce other confusion.
>
> I said this myself that when all things considered, table@summary  
> might be better.
>
> My answer in a nutshell: authors.
>
> Even this wg has trouble understanding the issue. I believe we had  
> lead more fruitful debate with caption@summary.
>
> If we don't change anything about @summary, then all we can do for  
> authors is writing a better description of it in the draft. Well, we  
> can also define how UA should display it and so on. There are  
> certainly many things one can do still.
>
> But if HTML5 goes for caption@summary then it can be portrayed this  
> way:
>
> Because authors did not see that the two has to have different  
> content, and because authors kept writing "layout tables" in  
> table@summary, something they are not known to use <caption> for,  
> @summary was moved to <caption> to help authors to discern between  
> the <caption> and @summary and to help them see that both are  
> visible metadata, except that @summary is only visible for the  
> unsighted. (Plus that it is easier to test via CSS etc.)

I think that is an excellent point. I would also add to that another  
point. While Gez earlier lamented the problem of adding an empty  
'caption' element simply to provide a summary of the table, I think in  
practice we would find it very rare to have a table so complex to  
require a summary, but not needing a caption. I cannot say I have  
every encountered such a table, but they might exist.

Also sometimes a change like this has the effect or raising awareness,  
where without the change the same poor practices might simply continue.

Take care,
Rob

Received on Sunday, 1 March 2009 19:10:26 UTC