Re: Pause on exit from Cue Ranges

On Thu, 18 Sep 2008, Dave Singer wrote:
> <>
> We think that the "pause on exit" aspect of the cue range should be removed
> from the specification.
> Why?
> * It's only one of the functions that you might want to achieve at the end of
> a cue-range (e.g. loopback is another)
> * It gives the impression that it'll pause exactly at the end of the range,
> but this is not (necessarily) true.  If the cue-range fired 'late' (because
> the system is loaded), it might pause 'late' (unless the specification is
> updated to say that it is pause-and-reseek-to-on-end-of-cue-range).
> * Even if it is, it's easily achieved by a script that does a pause and seek.
> * Removing it simplifies the API at no loss of functionality.

This feature was added because it is in fact significantly harder to pause 
the video exactly on the last frame of a cue range when the pause is 
triggered from script than it is when the pause is triggered from the 
native code monitoring the video. I think that the points above far 
underestimate the potential difference in quality here. Because of the 
inherently single-threaded nature of Web page scripting, any script 
running for even a few dozen milliseconds could screw up the pausing, 
while the UA could in fact implement this at the same point of the code 
that does the actual playback, providing perfect stops every time.

So while I agree that a simple implementation of this in the UA isn't 
going to be much better than a good script implementaiton with care taken 
by the author to not run too many long scripts, I think the potential 
benefit is far more.

Note that looping around a range is already possible using the looping 
constructs (which also prevent the user from seeking out of the range).

Ian Hickson               U+1047E                )\._.,--....,'``.    fL       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Tuesday, 14 October 2008 01:01:37 UTC