W3C home > Mailing lists > Public > public-html@w3.org > January 2010

Re: Taking another round at @summary

From: Jonas Sicking <jonas@sicking.cc>
Date: Tue, 5 Jan 2010 15:44:46 -0800
Message-ID: <63df84f1001051544p7a5abe01n5d14362d70379627@mail.gmail.com>
To: Denis Boudreau <dboudreau@webconforme.com>
Cc: HTML Accessibility Task Force <public-html-a11y@w3.org>, HTML WG Public List <public-html@w3.org>
On Tue, Jan 5, 2010 at 3:12 PM, Denis Boudreau
<dboudreau@webconforme.com> wrote:
> Hi Jonas,
> On 2010-01-05, at 4:54 PM, Jonas Sicking wrote:
> On Tue, Jan 5, 2010 at 10:24 AM, Denis Boudreau
> <dboudreau@webconforme.com> wrote:
> Even if another mechanism was provided to replace @summary, we're still
> removing something that is useful to a lot of AT users today.
> Really? Do you have data showing that is the case? All the data I have
> seen indicates that @summary is used extremely rarely, and when it is,
> it's often used in the wrong way.
> What kind of data would you need? I'd be happy to provide.
> For every screen reader user that I put in front of two similar complex data
> tables, one accessible and one that's not, I'm pretty sure 100% would come
> to realize that the accessible version, thanks in part to the description
> provided through @summary, would actually be much easier to understand.
> How many names do I have to come up with for this data to be receivable? ;p

My understanding was that your argument was that if we remove @summary
from the spec, this would hurt accessibility for users. However this
seems to only be the case if any of the following happens:

1. Tools stop supporting @summary on the pages that already use @summary.
2. @summary is removed from existing pages without a replacement being
added instead
3. New pages are authored where authors choose not to add @summary,
nor add any equally accessible alternative

1 I believe goes against the HTML5 spec. IIRC the spec still asks
consumers of HTML5 to expose the @summary information if it is

I don't see a reason why 2 would happen. Additionally, the data I have
seen indicated that extremely few pages used @summary, and the ones
that did so, often did it incorrectly. Thus even the worst possible
scenario here won't affect most users.

Likewise I don't see why 3 would happen. If we extrapolate based on
the previous data, it also doesn't seem like the worst case scenario
here will affect very many people, given how few pages could be
expected to get a proper @summary.

The data I was thinking about in my previous mail was showing that 2
or 3 would be more likely, or that it would be more bad than the data
I have seen so far indicates.

> As a responsible member of this working group, why should I sign a
> blank check on something as important as this when I have no garantee that
> this new mechanism - that we have yet to envision - will do any
> better than @summary already does?
> It seems to me that the WAI already has come up with a better solution:
> aria-describedby. This attribute has two advantages:
> 1. It encourages the description to be visual to all users, not just users
> of AT tools. While still allowing the description to be only visual to AT
> users when that is desired by the page author. 2. It can be applied to all
> elements, not just tables.
> 2. This attribute is already supported in HTML5.
> Good solid point, thank you. OK, so maybe theoretically @aria-describedby is
> the new-and-improved @summary I was calling for. But it seems to me the main
> argument for getting rid of @summary in the first place was that people
> weren't using it properly, no?
> I may be wrong (please correct me if I am), but with @aria-describedby, we
> seem to be stuck with a description for the table that's now available for
> everyone as content, including 99% of the people who don't need, nor require
> it to understand what the table represents.

You are wrong. As mentioned in point 1 above, while aria-describedby
encourages the summary to be exposed to all users, it allows it to
only be exposed to AT users:

Exposed to all users:
<table aria-describedby="tableDesc">....</table>
<p id="tableDesc">Description here</p>

Exposed only to AT users:
<table aria-describedby="tableDesc">....</table>
<p hidden id="tableDesc">Description here</p>

> What garantee do we have that authors would provide a better, more suitable
> description for content associated with
> aria-describedby="table-description" referenced somewhere else in the page
> with <div
> id="table-description">This-table-presents-blah-blah-blah...</div> than they
> already do for a simple description with
> <table summary="this-table-presents-blah-blah-blah..."></table>?

The obvious counter question is: What guarantee do we have that we
will be more successful in getting people to use @summary more in the
future, than we have been in the past 10 years?

The amount of success @summary has seen in the past 10 years does not
IMHO speak for a "stay the course" strategy.

/ Jonas
Received on Tuesday, 5 January 2010 23:45:40 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:45:06 UTC