RE: table-summary argument

I wrote:
>> The proposal is to remove an accessibility-related attribute, without
>> offering a replacement for its use, with no explanation except they 
>> looked at the Google index and figured they could axe it. 

to which Sam replied:
> That seems to be misrepresenting the position that has resulted in the 
> summary attribute not being in the HTML5 draft.

I agree that it might be considered hyperbole, but in fact the
"replacement" offered was not a replacement "best practice", but
at best a conjecture that some new mechanism could be created which
would have an associated best practice ("I would provide a mechanism to 
associate this text with the table."). The explanations offered
might be justifications against adding a new feature, but hardly
an argument for removing an existing feature.  The only basis
for the conclusions drawn were, in fact, the data gathered from
the Google index. 

Allow me to restate the last phrase:

"with no explanations except those based on looking at the
 Google index and making assumptions about future best 
 practices based on  past usage."

Note that the current document makes many recommendations
about future best practice completely counter to past usage,
when it is suitable.

> I'm late to the party, but reading the following link, it does appear to 
> be a misrepresentation.  In general, drawing conclusions and inferring 
> motivation without doing the proper research isn't the best way to start 
> an open discussion.

> See http://lists.w3.org/Archives/Public/public-html/2008Dec/0175.html  
> for a summary of various issues.

I can't see anything in that message or any of the ones linked
from it that isn't also in the more coherent summary I did
provide a link to, viz:
      http://esw.w3.org/topic/HTML/SummaryForTABLE

> Scanning both links, it seems relatively likely to me that the HTML5 
> working group considered the points raised on the PFWG link;

I see no evidence that anyone considered the points. I can
find no rebuttal of any of the points. There is no evidence
that anyone other Ian posted saying they had reviewed the
points and found them unconvincing. Rather, there is evidence
that the points were dismissed as "opinions and assertions"
rather than "objective analysis and logical argumentation".

> and it 
> seems less likely to me that the PFWG considered the points raised in 
> the HTML5 link, but that could be my ignorance.

I was presenting a position that is shared by Adobe Dreamweaver a
and the WAI coordination group based on their recent discussion, 
so no, your guess here is wrong. There is ample evidence that a
large group of people examined the points raised in the HTML5 
link.

I was adding a new point:  that table@summary is implemented
in a popular authoring tool, some of whose users would complain
if it was removed.

Since the HTML5 specification currently bends over backward to
satisfy backward compatibility with a wide range of junky
web pages and vagaries of backward compatibility with old
browsers, it would seem reasonable to ask that consideration
also be given to users of existing authoring tools.

> Even without fully understanding the issue, it is clear that this issue 
> has been going on for quite some time, and I infer from that that it is 
> likely that it won't be resolved quickly.

I brought this up at this time because the minutes for the HTML 
working group telcon 22 Jan 2009 said:

  <trackbot> ACTION-76 -- Chris Wilson to create poll on issue-32,
     based on Joshue's page from action-66 -- due 2009-01-31 -- OPEN
   <trackbot> [24]http://www.w3.org/html/wg/tracker/actions/76 
     [24] http://www.w3.org/html/wg/tracker/actions/76
   <pimpbot> Title: ACTION-76 - HTML Weekly Tracker (at www.w3.org)

This is Issue-32. If there is going to be a poll on the topic,
it seemed timely to raise another point previously not listed
in the summary.

> Larry: it might be useful to write up a short statement (ideally a 
> balanced one that covers both positions, if that's humanly possible) to 
> be included in the draft for now?

This might be moot depending on the result of the poll on
the topic, but here goes:

table@summary: "A survey of existing web content showed that this
attribute was rarely used, and, when it was, it was frequently
misused, leading to the conclusion that it should be removed from
HTML 5.  On the other hand, the attribute has been in HTML for
since HTML 4, there are some sites that use it correctly, and
the WAI community has recommended its continued inclusion 
as a best practice in the hopes of increasing correct
deployment. This issue is not yet decided."

I think the general policy issue I s whether HTML5 should
recommend as best practice something that isn't exactly
current practice. Currently the document rarely does so
except in a few notable places, primarily where new
features have been added.

I think that's at the root of this issue.

Larry
-- 
http://larry.masinter.net

Received on Tuesday, 3 February 2009 23:51:32 UTC