W3C home > Mailing lists > Public > public-html-a11y@w3.org > March 2011

minutes: Canvas Subgroup of HTML A11y TF weekly meeting 2011-03-07 [draft]

From: Gregory J. Rosmaita <oedipus@hicom.net>
Date: Mon, 7 Mar 2011 20:57:55 +0000
To: public-canvas-api@w3.org
Cc: franko@microsoft.com, david.bolter@gmail.com, Charles Pritchard <chuck@jumis.com>, swarren@aisquared.com, tlalor@aisquared.com, dmazzoni@google.com, public-html-a11y@w3.org
Message-Id: <20110307205654.M16557@hicom.net>

minutes from the 7 march 2011 Canvas Subgroup call can be accessed as
hypertext from:


as an IRC log at:


and as plain text following this annoucement -- please log any errors,
corrections, clarifications, mis-attributions and the like by 
replying-to this announcement on-list...

IMPORTANT SCHEDULING NOTE: due to CSUN, the next Canvas Subgroup
call will be held on 28 March 2011 -- a canvas discussion is 
tentatively scheduled for the HTML A11y TF face2face meeting 
immediately following CSUN

for more information about Canvas subgroup meetings, please consult:



                               - DRAFT -

             HTML Accessibility Task Force Teleconference

07 Mar 2011


   See also: IRC log - http://www.w3.org/2011/03/07-html-a11y-irc


          Gregory_Rosmaita, Rich, Chuck_Pritchard




     * Topics
         1. Strategy for positioning fallback content relative to
            corresponding canvas UI drawing counterparts
         2. Chuck Pritchard update on working modification to
            simulator caret tracking
         3. Outstanding Bugs and Issues Being Monitored
     * Summary of Action Items

   <trackbot> Date: 07 March 2011

   <scribe> scribe: Gregory_Rosmaita

   <scribe> scribenick: oedipus

   <scribe> Meeting: Canvas Subgroup of HTML Accessibility Task Force



   <Downchuck> helpful article for me as I work with ios




Strategy for positioning fallback content relative to corresponding
canvas UI drawing counterparts


   RS: looked at Chuck's path in SVG thing

   CP: just talking about PATH element

   RS: suggesting what? PATH for drawing?

   CP: have to do manually in CANVAS
   ... can convert SVG path to series of canvas path commands or
   reverse -- example of what is available -- first thing IMAGEMAP, if
   not available can generate PATH -- just an example

   RS: bind an object to something being drawn on screen -- how would 1
   set dimensions for object on screen -- how to say "here are bounds
   of this object"
   ... want to say this area i am drawing to is associated with this
   element, and here is bounding element/clipping rectangle
   ... device context and clipping rectangle, you are done

   CP: for arbitrary paths, don't want to lose ability to represent
   arbitrary path -- curve that only takes op portion of bounding
   rectangle becomes less useful
   ... keep going back to SVG because in terms of usability, very
   ... have example of drawFocusRing
   ... 1 touch event short -- in canvas, if get mouse event, doesn't
   require pre-binding elements in DOM -- could change title, draw a
   focus event
   ... with iPhone and iPad 1 more layer of touch events that haven't
   been exposed

   RS: fit testing within a region?

   CP: yes

   RS: how establish region?

   CP: arbitrary -- 2 diff ways: paint to bitmap image and color of
   fill is mapped over and associated with function call -- quick and
   dirty to make clickable canvas area
   ... other: keep object with x,y bounding boxes
   ... event elegance -- this is a bounding box i am listening to and
   will now run through program
   ... point of standards is to give flexibility while having a

   RS: can't limit to just path?

   CP: drawFocusRing works, but is hit-testing based -- don't want to
   force to redefine every path -- could reuse drawFocusRing with diff
   name, but requires setting region for each clickable element

   RS: would that work for touch devices?

   CP: hit-testing and methods are passed -- would work with touch
   device if events available
   ... canvas on iOS -- click on canvas, thinks is image -- will expose
   title of image -- have to perform awkward steps -- don't want to
   pass touch events into canvas DOM until user affirms this is element
   the user wants
   ... slide finger over link -- give name of link -- set virtual focus
   on link -- double-tap to activate

   RS: if move finger over can tell if in clickable area?

   CP: only when activated -- before, just a canvas blob
   ... once double-tap element, starts processing events

   RS: bound to corresponding element, but what are you using as
   parameters -- how to glean bounding rectangle

   CP: with clicakable use same semantics as drawFocusRing and bind
   current path to element -- run for each element

   RS: seems like a reasonable approach

   CP: requires that all clickable areas defined ahead of time rather
   than after an event -- text clickability can get weird
   ... double-tap on a word, but don't want ot send extra word
   boundaries if don't have to
   ... if receive events, ARIA handles with activedescendant
   ... awkward on iOS getting events to fire -- brought up on www-dom
   list fact that what apple introduced is "innovative" and unique --
   like a touch hover event
   ... did same thing -- touch hover event not bound or standardized --
   if was, could have bounded into hit testing
   ... setClickableArea falls apart if lot of complexity (lots of text)
   ... if dealing with complex app, not as desireable

   RS: want to say "take current path to calculate"

   <Downchuck> setClickArea(element) ; clearClickableAreas();

   <Downchuck> [sic] setClickableAreas

   RS: diff btw text and objects -- text has font info so can compute
   bounding rectangle for text
   ... might be able to deal with that

   CP: more innovative user interfaces as touch and canvas implemented
   more broadly and robustly

   RS: apply to text? such as word processor -- do on each piece of

   CP: if apply ClickableArea apply to shape -- circle with text make
   outer circle clickable

   RS: but not on each piece of text in circle?

   CP: possibility -- a11y and ARIA standardizes things for UI -- use
   of it will depend on application
   ... might want to allow someone to double-tap to start reading with
   "you are currently over this word" -- how to do quickly?
   ... circular text box moseOver first thing to tell is word currently
   over, then speak text block

   RS: know in document area, but do hit-testing at word and character

   CP: when click on RTE, need to know where to put caret

   RS: does application rather than system do all the work

   CP: depends on application
   ... hit testing is going to be less memory intense -- asynchronous
   based model works better

   RS: would be good if could bind hit testing to an element
   ... more than a11y impact

   CP: make things easier to program

   RS: need to float this concept to WG -- setClickableRegion

   CP: RS float that, i will float "touch hover" concept


Chuck Pritchard update on working modification to simulator caret


   CP: hasn't progressed -- still at 1 character right now -- grabbing
   focus for ZoomText

Outstanding Bugs and Issues Being Monitored

   RS: nothing moved on any?

   CP: no -- nearly got WebKit to consider resolution of pixels





   Issue 131: http://www.w3.org/html/wg/tracker/issues/131

   RS: submitted as editorial issue

   CP: should be addressed; just a bit of pushback from webkit
   community about shadow DOM, but not too hard a pushback

   RS: thought webkit ok with response we gave them

   CP: need to let it get into codebase

   RS: if not moving, need to ping apple people to find out why

   CP: give it 2 days -- webkit collective discussed as group last week

   RS: regular meetings?

   CP: no -- converse over bug reports

   RS: conncected to WHATWG?

   CP: no directly with webkit bug tracker

   RS: travelling to CSUN -- could discuss canvas a11y at
   satruday/sunday f2f

   CP: not yet gotten tickets

   RS: face2face meeting in San Diego
   ... first time i can host a canvas call will be 28 March 2011

   CP: apple is 1 developer who have made the leap so far -- only full
   screen touch interface widely available

   RS: will float by maciej


Summary of Action Items

   [End of minutes]
Received on Monday, 7 March 2011 20:58:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:55:53 UTC