Re: Request to re-open issue 131 -USE CASES, USE CASES, USE CASES

On Fri, Jan 13, 2012 at 11:39 PM, Frank Olivier
<Frank.Olivier@microsoft.com> wrote:
>
> I’ve created a more normative change at http://www.w3.org/wiki/Canvas_hit_testing. I think the actual spec change that is required for this is relatively small, as the underlying concept (associate a path; forward all events) is simple. Please let me know if you have any questions or feedback.

Seems generally good. :)

You've interspersed bits of rationale with proposed text. I think it
would be clearer to keep rationale not intended to be included in the
spec itself in the "Rationale" section of the change proposal.

You say "User agents should use the information set by
setElementPath() to create accessible user experiences", but beyond
saying a path is associated with an element, you don't formally state
the meaning of that association.

With respect to the key normative requirement - "[w]hen the user
interacts with the canvas, the user agent should forward the
associated events to the fallback element" - it would help to provide
or link to definitions of key phrases such as "interacts with the
canvas", "associated events", and "forward the … events".

Examples of questions that spec text should answer include:

  * When we say an event is forwarded to the element, does this mean
in DOM4 terms that the event is dispatched on the element instead of
on the <canvas>? ("Forwarding" is not a concept mentioned in the DOM4
or HTML specs AFAICT.)

  * Especially, bearing in mind that <canvas> objects could be nested
and that events could be forwarded outside of the <canvas>, is there
any way of determining what <canvas> element received the forwarded
event? (Might it be worth adding a "forwardedFromTarget" property or
similar to the Event interface?)

  * What exactly happens when the user interacts with the canvas when
the drawing path is not closed?

--
Benjamin Hawkes-Lewis

Received on Sunday, 15 January 2012 15:10:52 UTC