Re: [SVGMobile12] STZLC3-6 overlapping display=none and media element

Hi JC,
This is the official response from the SVG WG from your email
(http://lists.w3.org/Archives/Public/www-svg/2005Dec/0100.html).

Key concepts:

* The execution of timing elements (e.g., video, set) over the course of
time (i.e., the clock runs) affects the current state of the timing
engine. The current state of the timing engine keeps track at each
moment of time which timed elements are executing (e.g., running an
animation or display a video).
* The current state of the 'display' property controls the current state
of the rendering tree (i.e., what things are rendered visually and
aurally).
* The timing engine is orthogonal to the rendering tree. Timing elements
(video, animate, etc.) are active no matter whether they are in the
rendering tree or not.
* A connection occurs between these two orthogonal systems when an
animation controls the 'display' property, which forces an update to
current state of the rendering tree. 
* The event target for 'begin' and 'end' events are the animation
elements (e.g., set, animate elements), not to the animation's target
element (i.e., the thing pointed to by xlink:href).

Therefore, here are answers to your questions:

Q: Is a begin event sent at t=1s or at 10s or not at all ?
A: There are four begin events, one for the video element and one for
each <set> element. The begin event for the video element occurs at 1s.
The begin events for the animation elements occur at 0s, 40s, and 55s.
The current state of the 'display' property has no affect on when begin
events are dispatched.

Q: At t=10s, we think the video at 9s should be presented, right ?
A: Yes, the user will see the frame at 9s into the video. (But note that
at t=10s, the first animation is in transition into an inactive state.
Thus, at t=9.9s, the video doesn't show, but at t=10.1, the video does
show.)

Q: Is an end event sent at 40s ?
A: No. There are four end events, one for the video element and one for
each <set> element. The end event for the video element occurs at 61s.
The end events for the animation elements occur at 10s, 90s, and 120s.

Q: Is a begin event sent at 50s ?
A: No. None of the animations begin at 50s.

Q: At t=50s, we think the video at 49s should be presented, right ?
A: Well, yes, except that the video is not part of the rendering tree at
that point because the second animation is in effect, so the user will
not see anything.

Q: Is an end event sent at 55s or at 60s or not at all ?
A: No. See previous answer about when the end events occur.

Q: Should the 'timing' part of a media element be considered as that of
a 
SMIL Animation element, independent of the A/V rendering, with events 
sent regardless ?
A: Yes.

If this response is not satisfactory, please respond within two weeks.

Thanks.
Jon Ferraiolo
Member SVG WG

------------------

From: Jean-Claude Dufourd <jcdufourd@gmail.com> 
Date: Mon, 19 Dec 2005 23:13:55 +0100
Message-ID: <43A730A3.6020501@gmail.com> 
To: www-svg@w3.org 

We are wondering what should happen in the following cases:

<video begin="1s" dur"60s" ... >
  <set attributeName="display" to="none" begin="0s" dur="10s"/>
  <set attributeName="display" to="none" begin="40s" dur="50s"/>
  <set attributeName="display" to="none" begin="55s" dur="65s"/>
</video>

Is a begin event sent at t=1s or at 10s or not at all ?
At t=10s, we think the video at 9s should be presented, right ?
Is an end event sent at 40s ?
Is a begin event sent at 50s ?
At t=50s, we think the video at 49s should be presented, right ?
Is an end event sent at 55s or at 60s or not at all ?

Should the 'timing' part of a media element be considered as that of a 
SMIL Animation element, independent of the A/V rendering, with events 
sent regardless ?

Thanks

Nicolas Pierre, Erwann Gouesbet, Julien Tranchant and
Jean-Claude Dufourd

Received on Wednesday, 1 March 2006 18:05:26 UTC