- From: Christopher R. Maden <crism@yomu.com>
- Date: Sun, 02 Jul 2000 15:42:24 -0700
- To: w3c-wai-ig@w3.org
At 08:41 2-07-2000 -0400, you wrote:
>The example there, as a fragment of XHTML stripped of accesskeys and other
>features, is something like
>
><fieldset>
> <legend>Choice of role</legend>
> <input type="radio" name="choices" value="1" />Option 1
> <input type="radio" name="choices" value="2" />Option 2
></fieldset>
>
>An HTML browser that doesn't recognise this will present the text "Choice of
>role" as a paragraph on its own before the radio buttons, so most assistive
>technologies should render it as text even if they don't associate it in any
>more meaningful way.
That's not quite accurate. A browser that doesn't recognize this will
ignore any unknown markup:
Choice of role
<input type="radio" name="choices" value="1" />Option 1
<input type="radio" name="choices" value="2" />Option 2
This will display the text on the same line as the radio buttons.
That's why I said that I would have preferred that <legend> allow block
content, for fallback, but it's too late to debate that now.
>Using an XML browser that can expand/collapse the tree (most of those I have
>seen) the legend will appear as one of the things in the fieldset (the
>others being radio buttons, and possibly labels as well), and it can opf
>course be controlled by CSS as well.
Most *browsers* don't expand and collapse the tree; just editors. MSIE,
confronted with an XML document *and no stylesheet*, will present a tree
view, but this isn't the intended usual mode of operation.
-Chris
--
Christopher R. Maden, Solutions Architect
Yomu: <URL:http://www.yomu.com/>
One Embarcadero Center, Ste. 2405
San Francisco, CA 94111
Received on Sunday, 2 July 2000 18:50:32 UTC