Re: Summary of Thursday's IRC conversation about @summary

Ian Hickson On 09-06-08 10.25:
> On Mon, 8 Jun 2009, Henri Sivonen wrote:
>>> (and again: "The wider web is not an example of good practice.")
>> Is there a less wide web where @summary actually serves it's purpose 
>> well?
> 
> This question was asked before, and a less wide web consisting of pages 
> written or maintained by authors with a high level of motivation to 
> correctly use summary="" were offered. However, a study of that data 
> actually showed that the use of summary="" on those pages was a net loss 
> to the accessibility of pages for people who did not use ATs. (Indeed, 

But in reality, there is no guarantee that what is currently 
inside @summary would automatically have been moved to a better 
place if @summary did not exist.

It might be that it would be a good advice for authors to go 
through their instances of @summary and see if the content there 
could actually be useful to _all_ user groups. But that does not 
justify the removal of @summary from HTML.


> those pages were a big part of the reason why I proposed using <caption> 
> instead of summary="", to make the table help text universally accessible 
> instead of making it only visible to certain users.)
> 
> A detailed analysis of this "less wide web" can be found in this e-mail:
> 
>    http://lists.w3.org/Archives/Public/public-html/2009Feb/0601.html
> 
> ...starting from the paragraph that reads "Thank you! Real data always 
> helps us make better decisions".

Here is, so far, a re-analysis of the first 4 examples.

First, I have identified 4 concepts that you look for in your 
analysis:

	D for Duplication of content
	H for Harmless for sighted users
	U for Unique info in the @summary that this way
           is "hidden" for sighted users.
         L for Layout table

http://www.fbi.gov/ucr/cius2007/data/table_01.html

This @summary do not support your claims.

You conclusion is U: the info would have been helpful for all 
users. However, the @summary only tells that one may download an 
Excel version of the table. That this is possible is easy to see 
without @summary. Thus the @summary here, while not used as I 
would expect, is geared toward the usergroup for which @summary is 
intended.


http://www.cdc.gov/asthma/nhis/04/table1-1.htm

This @summary do not support your claims.

You conclusions: D, H and U.

However, only the first part of @summary duplicate info. The last 
part goes like this: "as a Total and Broken Up by Age Group, Race, 
Residence Size, and Closeness to Poverty Level as Determined by 
the National Health Interview Survey, 2004". Here the @summary 
describes the header cell content  in the leftmost column of the 
table. Thus this is clearly structure information. Sighted readers 
can see for themselves that the columns covers those categories.

Would it be harmless if this info was in the caption? Well, may 
be. But, so what. One can understand that the author wanted to 
create a short caption, as expected of a caption.

I cannot really see that @summary contain info that else are not 
found one the page. For instance, the @summary say "in the General 
Population of the United States" - but readers already know tha 
this is about the US, I think.


http://www.eia.doe.gov/cneaf/electricity/epm/table1_13_a.html

Your conclusion is L - layout table. Yes, indeed. Also, the 
@summary repeats what the nested table says in its "caption" 
(which is only TH cell). How might the author have been thinking? 
Probably he did not consider that user agents would be able to 
apply heuristics - there is a lot of confusion w.r.t. 
accessibility and layout tables. But at least, we cannot claim 
that this table represent "best of the best" w.r.t. accessibility 
when it uses a table for layout purposes.


http://www.vrg.org/journal/vj2006issue2/vj2006issue2mealplans.htm

You conclusion is D and L. You say that a number of the duplicate 
existing captions. However, the page do not have any <caption>. We 
can speculate that focus on @summary caused that the author did 
not use <caption>, but that only becomes speculation. And the use 
of table for layout means that we cannot say that this represent 
"best of the best".
-- 
leif halvard sili

Received on Monday, 8 June 2009 12:07:25 UTC