W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > April to June 2006

RE: FIELDSET and LEGEND - WCAG2 Conformance Example Files

From: Bailey, Bruce <Bruce.Bailey@ed.gov>
Date: Tue, 6 Jun 2006 09:04:35 -0400
Message-ID: <CCDBDCBFA650F74AA88830D4BACDBAB50B2D4DBD@wdcrobe2m02.ed.gov>
To: "Loretta Guarino Reid" <lguarino@adobe.com>, "Chris Ridpath" <chris.ridpath@utoronto.ca>, "John M Slatin" <john_slatin@austin.utexas.edu>, "Gregg Vanderheiden" <gv@trace.wisc.edu>
Cc: "WAI WCAG List" <w3c-wai-gl@w3.org>

I like the idea of such a failure but care will have be taken to construct a (counter) example where the proximity is visually obvious but still an actual accessibility problem.  The other Common Failures come from the real world.

It is possible for something to be accessible but still out of spec with the standards, especially as the screen readers have gotten better at compensating for broken code.

With simple tables, it is not really a accessibility barrier if header rows are marked up using TD instead of TH.  This violates WCAG 1.0 P1 Checkpoint 5.1.  In our testing at ED, we require developers to correct this "non problem" because the standard is clear.  I believe it is reasonable for WCAG 2.0 to require additional semantic mark up.  For the discussion in question, this would mean always using LABEL FOR or TITLE and always using FIELDSET and LEGEND.

With the example Chris created, programmatic association follows from everything being in a single form.  The prompt is actually outside the form container in a stand-alone paragraph.  Expanding the example with multiple paragraphs each followed by checkbox selections still would not lead a practical accessibility problem (assuming a single form element and submit button) but it would make it more clear that the prompts are not programmatically associated with the input elements.

With HTML, is having the source code reading order being a logical reading order a sufficient technique for 1.3.1 and form elements?

Is using a layout table and containing the label and field in the same TD a way to provide programmatic association?

Otherwise, then I think we are going to have to be comfortable that 1.3.1 *requires* LABEL FOR (or TITLE) *always* even when there is not an accessibility barrier in actual practice.  There is a precedent with this for WCAG 1.0.  There will a similar implication for FIELDSET and LEGEND.


> -----Original Message-----
> From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On
> Behalf Of Loretta Guarino Reid
> Sent: Monday, June 05, 2006 4:57 PM
> To: Chris Ridpath; John M Slatin; Gregg Vanderheiden; WAI WCAG List
> Subject: RE: FIELDSET and LEGEND - WCAG2 Conformance Example Files
> 
> Perhaps the potential failure coming out of this discussion is using
> physical or logical proximity to associate content without also using
> appropriate markup. We've wrestled with this in the context of links,
> too.
> 
> Loretta Guarino Reid
> lguarino@adobe.com
> Adobe Systems, Acrobat Engineering 
> 
> 
> > -----Original Message-----
> > From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-
> > request@w3.org] On Behalf Of Chris Ridpath
> > Sent: Monday, June 05, 2006 1:31 PM
> > To: John M Slatin; Gregg Vanderheiden; WAI WCAG List
> > Subject: Re: FIELDSET and LEGEND - WCAG2 Conformance
> > Example Files
> > 
> > Thanks John.
> > 
> > If fieldset and legend is the only way to create labels
> > that can be "programmatically determined" to be associated 
> > with a group of checkboxes then should it be a failure if 
> > you don't do it?
> > 
> > Chris
Received on Tuesday, 6 June 2006 13:05:11 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:46 GMT