W3C home > Mailing lists > Public > wai-xtech@w3.org > February 2007

Re: Example of using required, invalid and alert

From: David Poehlman <poehlman1@comcast.net>
Date: Thu, 22 Feb 2007 18:20:28 -0500
Message-Id: <2BC4AE20-D836-4F41-A2A0-A2B27ED4DD03@comcast.net>
Cc: Becky Gibson <Becky_Gibson@notesdev.ibm.com>, wai-xtech@w3.org
To: David Bolter <david.bolter@utoronto.ca>

The problem with tab is that not enough information is given through  
the use of it when interacting with forms to make an informed  
decision as to what to do.

On Feb 22, 2007, at 4:05 PM, David Bolter wrote:

Becky all examples work for me.  I see three levels of annoyance. The  
one were you trapped me in the text field until a valid entry was  
made was definitely annoying. The others were fine but I think it  
depends on circumstance.  If I was using a screen reader and want to  
explore things via tab, the first example would also annoy me I  
think.  Do you think that is a common use case?  To explore a form  
via tab?

BTW if we go with a tooltip alert approach we discussed off-list  
(sorry all) we are going to have similar questions I think.


Becky Gibson wrote:
> I have updated my examples a bit and gotten the questions  answered.
>> Do we expect that there can be more than one state - both required  
>> and invalid?  Should I remove the required field when I set the  
>> invalid role?
> Since required and invalid are properties, both can be set on the  
> same element.
> I also updated the example  to catch the onblur event when a field  
> is losing focus and forcing focus back to that field when it is  
> invalid. This can be quite annoying so I'm looking for feedback on  
> how folks feel about that.
> I've added an example where I monitor the onkeydown event and catch  
> the tab key and enter key and check for errors.  When the tab key  
> is pressed and there is an error, I make the error message visible  
> and force focus back into the field.  If the user presses tab a  
> second time, I allow focus to move to the next control. The error  
> message is still displayed but at least the user is not "trapped"  
> in the field until the error is fixed. Does this seem like a  
> worthwhile convention to encourage?
> I've added a bit more text for each of the three examples to  
> explain the behavior of each.
> This is an XHTML file with a content-type of application/xhtml+xml  
> so it can only be opened properly  in Firefox or Opera, although I  
> did not test in Opera.
> Becky Gibson
> Web Accessibility Architect
>                                                        IBM Emerging  
> Internet Technologies
> 5 Technology Park Drive
> Westford, MA 01886
> Voice: 978 399-6101; t/l 333-6101
> Email: gibsonb@us.ibm.com
Received on Thursday, 22 February 2007 23:20:44 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:51:32 UTC