Post Archive
› September 23, 2003
Screen Reader Visibility
Web-Graphics now uses this technique to hide accessibility related features from graphical browsers.
Why? The short answer: Screen readers apparently can not "see" content that is hidden using the CSS display:none style attribute. Given that we (along with many other sites) have been using that to hide content specifically geared towards screen readers (e.g. jump-to links), it is not acceptable that such content would be unread by screen readers. The new technique provides the same effect, but without the drawbacks. I was happy to find the switchover to be a minimal change to the CSS file, it only took a few minutes.
Comments
1. September 24, 2003 02:42 AM
2. September 24, 2003 05:41 AM
3. September 24, 2003 08:05 AM
Nate Posted…
Yikes, thanks Phil - ok that didn't work in IE 5.2 for Mac at all, I've gone back to the old method for the moment. Liorean: what you're saying makes sense, but I would still guess that people who use screen readers will want to hear the "skip navigation" links? How well does a screen reader represent the layout of the screen?
4. September 24, 2003 03:57 PM
liorean Posted…
It varies. Most read out content boxes or columns as they appear from left to right, top to bottom. Some read them top to bottom, left to right (note that I'm talking about the content boxes, not the actual text here). Most have some problem with the order of non-rectangular non-symmetric layouts where the text is not clearly a running text. I've not used that many, but a friend of mine with an eye disease that means he's slowly losing his sight has given me the tour of those he's used.
5. September 24, 2003 05:04 PM
Russ Weakley Posted…
Nate, you mentioned that yu use this technique:
.off-left {
position: absolute;
left: -999px;
width: 990px;
}
Have you tried the version mentioned by Simon Willison?
http://simon.incutio.com/archive/2003/09/13/screenReaders
.skip {
height: 0;
width: 0;
overflow: hidden;
position: absolute; /* for the benefit of IE5 Mac */
}
6. September 24, 2003 05:42 PM
Nate Posted…
Liorean - I'm still having trouble understanding how having the "skip nav" links unreadable by screen readers is ok, maybe in concept screen readers are supposed to present the screen as is, but since audio is linear and websites are spacial, I can't imagine it would work for sites that are more than one column. It's quite possible I'm just missing your points.
Russ that sounds like a good idea, but I noticed that in the comments, the willison method isn't as compatible with screen readers. I wonder if I just didn't implement the 999 method properly? I noticed that the wiki page specifically says that it works with IE 5.2 for mac... hmmm...
liorean Posted…
Hmm, I've had this discussion before... the thing is, a screen reader (as opposed to an aural rendering engine/browser) is supposed to represent the screen rendering, not the aural rendering. It should read out the contents of the actual screen as rendered and in the order rendered, not the contents of the document and in the order of the document. A screen reader that aspires to be an aural browser, such as JAWS, have some os the capabilities of an aural rendering engine, but not all necessary to fully be one. (Note that several general screen readers only reads out the contents of the screen, and doesn't even ask the browser for the most basic accessibility features, such as the alt attribute on images)