Re: [SVGMobile12] STZLC3-4 animation of xlink:href of a media

Dear Doug,

As I said earlier today, I believe this answer is more relevant to STZLC3-3
about the definition of the intrinsic duration of a media.
As an answer to STZLC3-3, it satisfies my concerns, thanks. (I do not have
to like it :) )

For STZLC3-4, your message answers only a small part of our comment.
It does not answer the questions in my email and in Cyril's answer about
whether the timeline is restarted.

quote:
- at 0s, <video> renders "vid1.3gp"
- at 1s, <video> renders "vid2.3gp". Is "vid2.3gp" started at 1s into
the media, or at its beginning ?

The above is independant from syncBehavior in my opinion. I do not see a
reason to pick one or the other.
The SVG WG needs to be choose because animating the xlink:href is an SVG
behavior if I remember correctly.

Probably, the same is true for:
Should the 'active end' be reevaluated at each change of the xlink:href
attribute ?

I apologize for missing these two unresolved points this afternoon.
Best regards
JC

On 3/3/06, Doug Schepers <doug@schepers.cc> wrote:
>
> Hi, Jean-Claude-
>
> We agree that this is a serious issue, and since it is outside the scope
> of
> the SVG WG to define this behavior, we will coordinate with the SYMM WG to
> address these and related issues. In the meantime, this should not affect
> the ability to conformantly implement SVGTiny1.2, since we normatively
> refer
> to SMIL2.1 for this range of behaviors, and any changes to SMIL will only
> be
> relevant to later SVG Specifications.
>
> In the interim, UAs should define their own behavior, even if it might not
> currently be interoperable. It is anticipated that further coordination
> with
> the SYMM WG based on feedback from implementor experience will yeild an
> acceptable and interoperable solution for future specifications.
>
> Please tell us within two weeks if this response does not address these
> issues.
>
> Regards-
> Doug, on behalf of the SVG WG
>
> Jean-Claude Dufourd wrote:
> |
> |
> | We are wondering about the consequences of animating the xlink:href
> | attribute of a running media.
> |
> | Ex. :
> | <video xlink:href="vid1.3gp" begin="0s" dur="media">
> |   <set attributeName="xlink:href" to="vid2.3gp" begin="1s" dur="2s"/>
> | </video>
> |
> | - at 0s, <video> renders "vid1.3gp"
> | - at 1s, <video> renders "vid2.3gp". Is "vid2.3gp" started at 1s into
> | the media, or at its beginning ?
> | Should "vid1.3gp" be kept active in background ?
> | - at 3s, <video> renders "vid1.3gp" again. Does it restart at the
> | beginning or at 3s ?
> | Should "vid2.3gp" be kept active in background ?
> |
> | What is the effect on the timing of <video> if the implicit
> | durations of
> | "vid1.3gp" and "vid2.3gp" are different ?
> | Should the 'active end' be reevaluated at each change of the
> | xlink:href
> | attribute ?
> | What is the consequence of the connexion and buffering phase which
> | appears at the change between "vid1.3gp" and "vid2.3gp" ?
> |
> | We do not think the list of questions ends here. We believe this is a
> | Pandora's box.
> | One possible solution is to adopt text to the effect that changes to
> | xlink:href are only taken into account when the media element
> | is not active.
> |
> | Thanks
> |
> | Nicolas Pierre, Erwann Gouesbet, Julien Tranchant and
> | Jean-Claude Dufourd
> |
>
>


--
Jean-Claude Dufourd
Founder, Chief Scientist, Streamezzo
83 Bd du Montparnasse
75006 Paris, France   Tel:+33 1 53 63 28 47