Re: summary="" in HTML5 ISSUE-32

Hi Ian,

On Feb 23, 2009, at 9:54 PM, Ian Hickson wrote:

> Before I jump in to the e-mails themselves, I want to make sure we all
> agree on the underlying goals that we are trying to accomplish.
>
> Problem statement: some users find navigating tables complicated, and
> would like a description of the table so that they can make better  
> use of
> the table. Such users might be blind, using an accessibility tool, or
> might have cognitive difficulties, or might just be unfamiliar with  
> the
> structure of particularly complicated tables.
>
> There are several things to consider when evaluating possible  
> solutions:
>
> 1. Whether the solution would actually solve the problem if used  
> right.
>
> 2. Whether the solution hurts any users, or fails to help users that
>    should be helped.
>
> 3. Whether the solution would be used correctly enough for users to
>    actually pay attention to it. A feature that is rarely used in
>    practice is better than a feature that is used wrongly, since the
>    latter will cause users to ignore the feature even when it is used
>    correctly.
>
> I think it's clear that summary="" could solve the problem if used  
> right.
> It seems like it would fail to help users that don't get access to it
> (like users of visual browsers).

You make a logical leap here that has no basis. There is nothing that  
I can see in the HTML5 draft that requires visual browsers to hide  
'summary' attribute values (or any attribute values for that matter)  
from users. So we cannot conclude that UAs will fail to provide access  
to other users to the table summaries. In fact authors and users can  
provide such access through the strategic CSS authoring[1] . Earlier I  
have suggested adding new norms for interactive UAs to display such  
information and you argued it was inappropriate for the spec to  
address such norms[2][3] (despite the fact that the spec already  
addresses such norms). However, if the spec cannot address such norms  
than it is also improper to assume UAs will not make such information  
available. So your repeated arguments that 'summary' is not available  
for all users is simply incorrect. Making such an argument would be  
like saying the 'title' element is of no use since it is not displayed  
by UA. Yes it is often displayed in the title bar, but that is a  
custom and not a requirement, recommendation or even suggestion of the  
HTML5 draft. The HTML5 draft does recommend (even though you say it is  
inappropriate) that UAs "use the document's title when referring to  
the document in their user interface", but it is still not displayed  
in the flow of the document by default which appears to be what you're  
saying about the 'summary' attribute.

More importantly, the claim that users are being short-changed because  
they cannot read content intended for the visually and cognitively  
impaired seems at best misguided and at worst malicious. I say  
misguided since table summary information is meant to describe  
features of a table that may be readily apparent to the visual and  
cognitively unimpaired users and is therefore redundant (e.g, an  
earlier example where a table might visually indicate its current sort/ 
collation state and such information would be redundant to repeat as  
text).

I say malicious since the continued repetition of the fallacious  
arguments seem directed at ensuring such information is not made  
available to visually and cognitively disabled users. Table summary  
information is information that often should not be displayed in the  
normal flow of a document. Authors should not be required to include  
the information in the normal flow of a document in order to include  
it within a document. However, by forcing authors to use caption or  
paragraphs or other visible by default facilities, means authors may  
be less inclined to provide table summaries within their documents.

Certainly other facilities might be added to HTML to replace the  
summary attribute. However, using facilities such as the 'details'  
element again requires the information be exposed in a way that may  
make authors reluctant to include the information. If the summary  
information is obvious to a visual user why would an author want to  
invite the user to click a disclosure button and read through a  
distracting summary to learn that the information in the 'details'  
element tells them what the user already knows? However allowing a  
summary attribute (or in the future a summary element within the  
caption might be suitable) permits authors to include the crucial  
information and direct it squarely at the users who need it.  
Simultaneously, these features allow authors to create the visual  
experience they prefer and not the visual experience preferred by you  
as the editor of the HMTL specification.

I do not understand how you think this WG would tolerate these types  
of impositions on the needs of authors and users, defended with the  
weakest arguments we've yet seen.

While the other concerns raised regarding the frequent misuse of  
summary are certainly valid, such misuse does not suggest the feature  
is substantially harming users. On the contrary, users can relatively  
quickly determine whether the summary is of any use or not and move on  
to table interrogation, to another section of the document or another  
document entirely. Even though it should be avoided, there is no great  
harm done in a poorly authored summary. Users might even become  
familiar with sites/authors who produce valuable table summaries and  
those that do not. The time lost on a poor summary is small when  
compared to the overall time needed to comprehend a complex table.

The existence of poor table summaries suggests that more needs to be  
done in the HTML5 draft, in other W3C WGs, and elsewhere to inform  
authors about how to author such summaries.

1) More should be conveyed to authors about the difference between a  
caption and a summary and how they should not be authored redundantly.
2) Authors should be told not to include information easily discerned  
through algorithm such as the number of columns and number of rows.
3) For simple tables authors should be advised that a summary may not  
even be necessary. And so on.

The point is that poorly authored table summaries do not indicate a  
need to remove facilities from HTML, but instead the need to convey to  
authors the importance of authoring effective table summaries.

Take care,
Rob

[1]: <http://malform.no/html5/summary+css>
[2]: <http://www.w3.org/Bugs/Public/show_bug.cgi?id=5756>
[3]: <http://esw.w3.org/topic/HTML/RicherUIAccessToHTMLData>

Received on Wednesday, 25 February 2009 06:48:28 UTC