W3C home > Mailing lists > Public > wai-xtech@w3.org > April 2008

Re: [html4all] New issue: IMG section of HTML5 draft contradicts WCAG 1 & WCAG 2 (draft)

From: Matt Morgan-May <mattmay@adobe.com>
Date: Fri, 11 Apr 2008 13:07:13 -0700
To: Dave Singer <singer@apple.com>, John Foliot <foliot@wats.ca>
CC: "'HTML4All'" <list@html4all.org>, <wai-xtech@w3.org>, "'HTML WG'" <public-html@w3.org>
Message-ID: <C4251301.6291%mattmay@adobe.com>

On 4/11/08 12:24 PM, "Dave Singer" <singer@apple.com> wrote:
> Consider an image that is 'part of the content'
> <img ... alt="an image">
> tells the user agent that there is a useful alt string that is worth
> displaying to the user, which is a lie (the string provided is not
> useful), and

I disagree. See below.

> <img ... alt="">
> tells the user-agent that the image is not 'part of the content',
> it's not worth describing, which in this hypothetical case is also a
> lie, whereas
> <img ... >
> tells the user agent the truth, that there is not a useful
> author-provided string.
> Lies are "worse" than the truth.

We know from over 10 years of experience that authors are going to lie. That
is something we can't control, and it's not in my mind sufficient reason to
dismiss the need for a required alt attribute.

However, for those who do validate, requiring alt is the only way to get the
author to signal his or her intent. To remove that barrier forces us to
assume that _all_ instances of <img> without alt present on the web simply
couldn't be expressed throws out the bathwater, the baby, the tub, the
pipes, and a chunk of the sewer line. For every one image being held up as
the edge case here (i.e., user-generated content), there are dozens if not
hundreds of others elsewhere on the web that have no such excuse for not
having a meaningful alt attribute.

The absence of semantics is not the semantics of absence.

You cannot turn one of the most common validity and accessibility failures
on its head simply because an empty alt attribute is unattractive. We
(including accessibility evaluation tools) have relied on this as a
statement that the author has evaluated the image and determined that it is
not worth an alternative. If they're lying, they will need to take that up
with a higher power than the W3C validator.

For that matter, I would argue that a site like Flickr _should_ apply
non-empty alt text for their images, given that those images are the main
content of the document. Even if there is no reasonable text equivalent,
there's nothing to say that a blind user wouldn't want to be informed of
that image. (And before you laugh at that, I know at least one blind person
with a Flickr stream. Probably more.)

Received on Friday, 11 April 2008 20:08:10 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:51:35 UTC