- From: Shane Stephens <shans@google.com>
- Date: Thu, 17 Oct 2013 20:29:28 +1100
- To: Brian Birtles <bbirtles@mozilla.com>
- Cc: "public-fx@w3.org" <public-fx@w3.org>
- Message-ID: <CAGTfzwQd7NeuVpMoTHw0f9GbT_eRxxzNrNZ4_Gup9aDOfJfTVw@mail.gmail.com>
Hi Brian and list, I like option C best out of A-C, because I think it fits well with intuition and has nice properties with regards to reversal and play-on of appended animations. I don't think that the statefullness of this option is a drawback - to me it feels like Groups are the unit of synchronization, not Players. For example, to implement the use case you discussed, I'd recommend that you put the async resource in a Par Group along with the sync resource. This will ensure that the resources are permanently in lock-step. I'm afraid I don't really understand option D. Would you mind explaining it in more detail? Thanks, -Shane Stephens On Thu, Oct 17, 2013 at 5:54 PM, Brian Birtles <bbirtles@mozilla.com> wrote: > Hi again, > > I've thought about this[1] some more and another arrangement comes to mind. > > Option D: The perpetual record player with annotations > ==============================**======================== > > This is identical to the current arrangement where the current time of the > player keeps increasing indefinitely but the difference is we still add > "end-like" behaviour. > > Specifically, we still: > > * have an 'ended' attribute that gets set when current time > content end > time. > * can define 'end' events as needed > * can define a 'finish' method to jump to content end time > > In essence we act like the player ends but we don't try to fix the current > time when it does. > > On the up-side: > + there is no hidden "actual current time" (unlike option A) > + there are no issues with the timing of operations causing different > results (unlike options B & C) > + swapping the source content of players or "pre-seeking" a player works > as expected (unlike option B) > > However: > - reverse() is no longer a no-op when called twice > > I'm not sure what is better here, option C or D. Option D is much cleaner > to implement and spec and I'd suggest more internally consistent. But if > authors expect the current time to stop accumulating when a player fires > its end event then option C may be better. > > If anyone has any gut-feelings, author's intuition, etc. I'd be glad to > hear it. > > Thanks, > > Brian > > [1] "This" being the following post for those following this out of > context: http://lists.w3.org/Archives/**Public/public-fx/2013OctDec/** > 0059.html<http://lists.w3.org/Archives/Public/public-fx/2013OctDec/0059.html> > >
Received on Thursday, 17 October 2013 09:29:59 UTC