Re: New internal draft of HTML techniques

Joe Clark wrote:

> A "keyboard user" might be sighted and can simply read the text.
> 
> I would like to see results from actual testing proving that current 
> screen readers do not enunciate <optgroup> text. Absent that, it's a 
> vapourware objection.

A keyboard user could certainly be sighted, but if that person is using 
only the keyboard, that person cannot read the optgroup label in the 
current browsers that I am most familiar with.

It doesn't take an advanced test to realize this is the case. Just use 
your keyboard (I'm talking about Windows machines at the moment) to 
access a drop-down list that uses optgroup. Mozilla-based browsers, IE, 
and Opera only show you the options that you can actually choose. Since 
you can't actually choose an optgroup label, the browser skips over 
these items entirely. You can only see one item at a time in a Windows 
drop-down list when you use the keyboard, and since optgroup labels are 
"unselectable" they are never shown to the user. Of course, mouse users 
can see them just fine, because the full list pulls down. In other 
words, the browsers have two different kinds of behaviors: one for mouse 
access (showing the full drop down menu) and one for keyboard access 
(showing one item at a time).

 From what I remember about IE on Macs, the behavior is somewhat 
different, with the optgroup labels acting as a submenu, with flyout 
menus to the side. I don't have a Mac available to me at the moment (I 
am out of the office), so I can't say with authority whether the 
implementation is keyboard accessible or not for that browser on that 
platform.

However, from a screen reader perspective--or, to be more specific, from 
the perspective of using JAWS and Window Eyes--my own tests have shown 
that screen readers do not read optgroup. I assume that the reason for 
this is that the browser's behavior is to skip over the optgroup labels. 
Since the optgroup label never gets the focus of the keyboard (or indeed 
any kind of focus at all), it is never read.

If anyone else has experienced different results using screen readers, 
I'd like to hear about it, but my comments were not based on presumption 
("vapourware"), they were based on my own tests.

-- 
Paul Bohman
Project Coordinator
WebAIM (Web Accessibility in Mind)
www.webaim.org
Utah State University
www.usu.edu

Received on Wednesday, 28 July 2004 00:20:53 UTC