Re: Draft text for summary attribute definition

Hi Leif,

On Feb 28, 2009, at 12:10 PM, Leif Halvard Silli wrote:

> Steven Faulkner 2009-02-27 13.18:
>> Hi all,
>> I have taken up Sam Rubys suggestion:
>> "As for me, what I would most like to see in the next draft is a RFC
>> 2119 compatible definition for table summaries, either in terms of a
>> HTML 4 compatible attribute or in terms of a suitable replacement.
>> Note that I am specifically saying "in a draft".  What I am *not*
>> looking for is a reply to this email on how such a topic might be
>> approached." [1]
>> and taken a stab at a RFC 2119 compatible definition for table  
>> summaries:
>> http://esw.w3.org/topic/HTML/SummaryForTABLE/SummarySpecification
>> If you have positive contributions to make to this definition please
>> add comments to the wiki page in the notes section
>
> What I am missing in that text is a focus on authors.

I've certainly tried to address authors in the text I wrote (actually  
some minor changes to Steve's original text; mine is not version B and  
his version A). I do feel more should be said and as Steve  
acknowledged any version requires some easy to follow examples.

> We need to challenge authors:
>
> * to choose EITHER caption OR summary OR both.
> * if they use identical captions.
> * if the summary looks like caption
> * if they seems to have written "this a layout table" info

I agree with this for the most part. However, I think there is  
something to be said for a summary that looks like a caption. That is  
if an author considers a caption somewhat desirable, but in the end  
decides to omit it, those are cases where the caption should probably  
be included in the summary. For many simple tables it is very obvious  
what the purpose of the table is for the visual user. It is very easy  
to see the homogeneity of dates down a column and the homogeneity of  
text in another column. For the non-visual user, such table  
consumption requires a careful listen to every cell. So where an  
author considers a caption but decides to omit it, I think we should  
advise such authors to place the caption in the summary attribute  
instead of omitting it entirely.

Certainly we should prohibit redundant summaries and captions (which I  
think the draft does) and also prohibit "layout table" summaries  
(which I think we have also done).

> In short, authors needs help juxtaposing the two things. In general,  
> what one could do for this are:
>
> * somehow linking summary to caption;
> * ways to let author see both simultaneously;

I think displaying this and other values more obviously in the chrome  
is an important step. Also encouraging general purpose rendering  
engines and browsers to also offer authoring modes where a different  
authoring default stylesheet makes alternate content visible would  
also be helpful.

> * ways to discover tables with caption but not summary;
> * ways to discover tables with summary but not caption;

I'm not clear what you mean by these two items. Could you explain this  
further?

> I will try to offer a RFC 2119 compatible text that seeks accomodate  
> for this.

I welcome wiki style changes to the version B I drafted[1]. Such edits  
should: 1) stick to the general thrust of a summary attribute and a  
caption element; 2) and any substantive changes should be added to the  
"Issues for discussion" section[2]. If, like Leif, anyone wants to add  
a different mechanism to address these use cases, then please add  
another version (currently both version A and version B involve a  
'caption' element and a 'summary' attribute).

Take care,
Rob

[1]: <http://esw.w3.org/topic/HTML/SummaryForTABLE/SummarySpecification#head-f7ae874dfb50ef6dfa17e222128fd04365ab691c 
 >
[2]: <http://esw.w3.org/topic/HTML/SummaryForTABLE/SummarySpecification#head-e24f017b97cef6950cb2f1a528e4c9a9ff5962c2 
 >

Received on Saturday, 28 February 2009 21:48:19 UTC