Re: Firing media events early for throttled downloads

John Foliot wrote:
> Much earlier in this thread, Ian Hickson wrote: "...we have an explicit 
> "autoplay" attribute which we should be encouraging authors to use 
> instead.", to which I responded that we shouldn't be _encouraging_ authors 
> to use autoplay - period. (we should instead be suggesting that in-page 
> media not autoplay, but leave the final decision to the author - providing 
> the UA's can over-ride the author's suggestion).

Ah, I see.

I guess in my proposal it doesn't matter much which one the page uses; 
they get treated identically.  The benefit of autoplay="" is that it 
works even with script disabled.

I think the right message is "in-page media should not autoplay, but if 
you must have it autoplay, use the declarative method of doing so".

>> The UI and interaction details would be up to the UA, of course; the
>> spec would allow such behavior, but not require it.
> 
> I thought you were talking about a client-side scripted feature[*], and was 
> saying instead that it needs to be a UA feature, which is what you just 
> said.  Yes, exactly, this is what we should have.
>
> [*] "... allowing the script _some_ way...", "... window.open restrictions 
> are heuristics that differ in different browsers..."

Oh, I see.  Yeah, there are both features here.  There's a client-side 
scripted feature: the ability to call play().  There's a UA feature: the 
ability to intercept such calls made in contexts where the user probably 
did not ask for them and take user preferences and possibly a user 
prompt into account when deciding whether to honor them.

The heuristic part comes in when making the "user probably did not ask 
for them" decision.

-Boris

Received on Saturday, 6 June 2009 22:42:09 UTC