Re: minutes: Canvas sub-group HTML A11y Task Force telecon, 2010-02-15 [draft]

Rich Schwerdtfeger
Distinguished Engineer, SWG Accessibility Architect/Strategist

"Charles McCathieNevile" <chaals@opera.com> wrote on 02/15/2010 07:47:35
PM:

> "Charles McCathieNevile" <chaals@opera.com>
> 02/15/2010 07:47 PM
>
> To
>
> Richard Schwerdtfeger/Austin/IBM@IBMUS, public-canvas-api@w3.org,
> public-html-a11y@w3.org
>
> cc
>
> david.bolter@gmail.com, jcraig@apple.com, "David Singer"
> <singer@apple.com>, janina@rednote.net, cooper@w3.org.
>
> Subject
>
> Re: minutes: Canvas sub-group HTML A11y Task Force telecon, 2010-02-15
[draft],
>
> On Mon, 15 Feb 2010 23:59:30 +0100, Richard Schwerdtfeger
> <schwer@us.ibm.com> wrote:
>
> ...
> > 20:40:11 [richardschwerdtfe]
> > Authors MUST render visible focus of the ADOM subtree element in the
> > canvas area
> > Authors MUST render visible focus of the canvas subtree element in the
> > canvas area
> > Authors MUST render visible focus of the canvas subtree element on the
> > rendered canvas
>
> As far as I can tell, by using an image map to represent areas sensitive

> to user input, all of these things are automatically managed by the User

> Agent in pretty much every modern user agent (starting with lynx). There

> is a bug in some versions of Opera which prevents an imagemap inside the

> canvas from being recognised - this is fixed in current builds of Opera
> 10.5 and should remain fixed.
>

I think this is an implementation for some canvas renderings - sure. I
would not say all. For example, it would not work for Bespin.

> > Authors MUST render the keyboard caret insertion cursor of the canvas
> > subtree element on the rendered canvas
>
> I'm not sure about handling this one completely automagically, but the
use
> of either forms, or existing editable HTML elements may again provide
this
> easily.
>
It will and our intent is to allow HTML content to be included in the
subtree. The problem is
that the caret location, in say a contenteditable area does not match the
visual location on the screen
to allow a magnifier to track it. So, what we do is use a textfield or
contenteditable area in the canvas subtree and
define an API, like we did for visual focus, to drive the caret location on
the screen and map to the accessibility API coordinate
location. Steve Faulkner has an action item to address this with the canvas
API team.

> > 20:42:40 [cyns]
> > authored content SHOULD respond to system settings for font, color and
> > zoom.
> > Authors SHOULD should ensure that the rendered canvas responds to
> > system settings for font, color and zoom
> > Authors SHOULD should ensure that the rendered canvas responds to
> > system settings for font, color and zoom
> > Authors SHOULD should ensure that the rendered canvas responds to
> > system settings for font, color and zoom
> > 20:46:06 [richardschwerdtfe]
> > Define user settings as system, user agent, or web application settings
>
> This stuff is trickier. We need to describe how the author would find out

> what those settings are if they are using canvas (although I also need to

> read up on the canvas text API to make sure I'm not missing something
> basic).
>
To this we all agree. I am going to send out revised wording on this and
the other pieces based on Monday's call today. This is not trivial as CSS
deprecated the system settings feature in CSS3 due to platform
inconsistencies. A challenge is determining what we put in the spec. and
what we put in best practices. We just need to provide some guidance in the
HTML spec. on what to do.

Thanks for reviewing the proposal. A revision is coming to the group. I
want to have something solid for the HTML working group by Feb. 25. Some of
this will need to be done in different groups. For example, the canvas API
was separated out of HTML 5 and is on a separate track as I am sure you
aware.

> cheers
>
> Chaals
>
> --
> Charles McCathieNevile  Opera Software, Standards Group
>      je parle français -- hablo español -- jeg lærer norsk
> http://my.opera.com/chaals       Try Opera: http://www.opera.com