Re: handling focus

we still need to deal with caret position. The mag has to be able to track
the caret.

Rich Schwerdtfeger
Distinguished Engineer, SWG Accessibility Architect/Strategist


                                                                           
             James Craig                                                   
             <jcraig@apple.com                                             
             >                                                          To 
             Sent by:                  David Bolter                        
             public-canvas-api         <david.bolter@gmail.com>            
             -request@w3.org                                            cc 
                                       public-canvas-api@w3.org            
                                                                   Subject 
             11/19/2009 12:30          Re: handling focus                  
             PM                                                            
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           





On Nov 19, 2009, at 8:41 AM, David Bolter wrote:
> On 18/11/09 9:26 PM, James Craig wrote:
>> On Nov 17, 2009, at 2:56 AM, Steven Faulkner wrote:
>>> How do AT such as screen magnifiers provide focus highlighting of
interactive parts of the canvas if native focus is not provided?
>>> How are they able to follow and bring currently focused elements into
the viewport if there  focus is not programmatically exposed provided?
>>>
>> That's can be done with standard CSS positioning on the interactive
elements in the shadow DOM. I'll update the proof-of-concept to include
that use case. I can't promise it before the U.S. Thanksgiving holiday
though.
>
> This is a creative idea and reminds me of a very similar technique we had
used for Dojo widgets, whereby we put a real, transparent input control
overtop of the pretty dojo control, and the user (unknowingly) interacted
with the real input. The pretty dojo control responded to the real input
state and as such was really just a 'rendering' of that was going on. I
think this mirrors what you want to try for canvas here.

Precisely. The main difference is that the keyboard user would be
interacting with the real form element *behind* the visible representation,
as opposed to your example, where the user was interacting with an
invisible form element layered on top of the visible representation.

James