RE: animations and intervals

Hi -

I will jump in since I worked on the SMIL Timing stuff. Yes, you should keep
a list of historical intervals until you are reset (assuming you support
time containment at some point).  Hyperlink navigation is an entirely valid
means of interaction and should not be dismissed as "just seeking".  The
ability to seek backwards can be crucial to authors in some cases, and this
is why it was specifically included in SMIL 2 timing. The combination of
hyperlinking and indeterminate timing (e.g. event based timing) is somewhat
complex, but it is consistent. For the SVG 1 case without time containers,
it is actually fairly simple.

Climbing out on a limb, I will comment that if you are considering not
keeping the historical intervals, you may be treading close to some more
serious problems in the implementation. If you are just considering a minor
optimization to toss out a few small objects, fine. If you are considering
departing from the model described in the spec, you are likely headed for
trouble.

If you have other questions about the SMIL timing model, I would happy to
help clarify the spec. I know it is not entirely obvious - it is a spec and
not a tutorial. In some places, we attempted to clarify precisely what we
meant, but the detail can be bewildering on the first read.

Good luck with it - Patrick

> -----Original Message-----
> From: www-smil-request@w3.org [mailto:www-smil-request@w3.org]On Behalf
> Of Sigurd Lerstad
> Sent: Tuesday, April 29, 2003 1:14 AM
> To: www-svg@w3.org; www-smil@w3.org
> Subject: animations and intervals
>
>
>
> Hello,
>
> Reading up on intervals and current interval etc. in the SMIL
> timing model.
>
> Is it a good idea to keep a history of all the intervals created for an
> animation element? I'm looking into seeking backwards, and
> though the rules are not completely clear to me yet, is keeping a list of
> already happened intervals a good idea (or even necessery) to
> seek backwards? It is my understanding that this is not necessary when
> playing normally and when seeking forwards.
>
> As Thomas E Deweese said, it's a good thing to refer to the specs
> and saying
> why I think the spec isn't clear when asking a question, however, all this
> is very fuzzy at the moment, and I don't have a complete understanding of
> the spec yet, I'm simply asking for implementation advise on
> wether keeping
> a list of history intervals for the sake of seeking backwards is
> a good idea
> or not, a simple yes or no from someone with implementation
> experience will
> do, so I know I'm on the right track.
>
> thanks,
>
> --
> Sigurd Lerstad
>

Received on Tuesday, 29 April 2003 12:54:37 UTC