W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > April to June 2012

RE: H65: Using the title attribute to identify form controls when the label element cannot be used

From: Phil Spencer <spencer_phil@hotmail.com>
Date: Thu, 10 May 2012 14:22:02 +0100
Message-ID: <DUB106-W3823B90113F7E6B3D2322DF1160@phx.gbl>
To: <listas@ramoncorominas.com>
CC: ax interest list <w3c-wai-ig@w3.org>

Hi Ramón,

Yes good point. 

I just created a little test page stripped back to just 2 radio buttons, one with the title attribute directly on the input element and one with the title attribute on the surrounding span (no table markup at all) and it does seem like the specific browser / AT combination I tested with treats each radio button the same. 

However as others have helped prove - this is not expected behaviour nor is it consistent across other browser / AT combinations.


> Date: Thu, 10 May 2012 13:49:47 +0200
> From: listas@ramoncorominas.com
> To: spencer_phil@hotmail.com
> CC: w3c-wai-ig@w3.org
> Subject: Re: H65: Using the title attribute to identify form controls when   the  label element cannot be used
> Hi, Phil and all,
> In these cases I would also consider if there exists a clear visual 
> identification of the input for sighted users, since the title attribute 
> is not keyboard-accessible (IMO this would be a failure of the SC 2.1.1 
> and SC 3.3.2).
> Regarding the span-title being announced by the screen reader, is it 
> really the title of the <span> what is read? Maybe it is the content of 
> the <th> elements... (just to be sure)
> Regards,
> Ramón.
> Phil wrote:
> > The actual case here is that of a grid of radio buttons with column and 
> > row headers but no individual <label> element for each input.
> > 
> > And just for total clarity, I am absolutely NOT considering the 
> > implemented technique (not my suggestion BTW, but implemented by a 
> > developer in a separate business), to be acceptable. I was surprised to 
> > find that it appeared to "work" with IE9 + Jaws, In fact I almost didn't 
> > bother to test it at all and just flag it as non conforming, but when it 
> > did appear to "work" it occurred to me that perhaps there was something 
> > in the HTML specification that I was unaware of which defines a form 
> > control without a title attribute inherits from it's parent element or 
> > something similar. Hence my question here to see if anyone knew if the 
> > behaviour I experienced was expected or atypical.
> > 
> > Steve F very kindly confirmed to me that it is not expected behaviour so 
> > the problem has been flagged!
Received on Thursday, 10 May 2012 13:22:37 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:44 UTC