WebAIM - Web Accessibility In Mind

E-mail List Archives

Re: When alt=""


From: Sven Jenzer
Date: Mar 23, 2009 2:00AM

Hi Cliff

Your idea doesn't work, because longdesc needs an URI, a simple link to
a html-document or also a textfile.
See spec: http://www.w3.org/TR/html401/struct/objects.html#adef-longdesc-IMG
My opinion and our experience here, with hundreds of accessibility-tests
each year, each one with participation of people with disabilites, is:
Images for decoration or layout-purposes, whitout content-relevance,
should have an empty alt-attribut and no longdesc-attribute.

Sven Jenzer
Swiss Foundation Access for all

Cliff Tyllick schrieb:
> When Jared announced the initial results of WebAIM's survey of people who rely on us to make sure they have access to the Web, I was struck by the strong and disparate positions on when an alt attribute should be empty. Lots of people had good reasons for wanting the photo next to a name or an image added to provide a splash of color to have an alt tag. In particular, I was struck by the responses along the lines of, "I have partial vision, so I can tell that *something* is there. Why can't I tell what it is on my own?"
> Out of the blue, a possible solution hit me: What if we always populate longdesc when alt is empty?
> For example, a key point of contention in the discussion that ran on back then was whether Jared's photo next to his name on his blog should have an empty alt tag--and, if not, what the content of the tag should be. If we were to adopt "populate longdesc when alt is empty" as a standard practice, then the coding could be:
> <img src="sourcepath/sourcefile.jpg" alt="" longdesc="author's photo" />
> Presumably, people who don't care to know what images with empty alt tags are could set their readers to ignore longdesc when alt="" and people who wanted to know everything could adjust their readers to always read longdesc to them. (I'll admit ignorance on what is and what isn't possible in the adjustment of screen readers. Still, it seems to me that vendors could easily make this possible if it isn't already.)
> Yes, it would be using an attribute for something other than its original purpose, but I think this sticks with the spirit of the purpose for having longdesc. And it would give the user complete control over what they hear.
> Am I missing something? Or does this make sense?
> Cliff Tyllick
> Web development coordinator
> Agency Communications Division
> Texas Commission on Environmental Quality
> 512-239-4516