Re: Public feedback on HTML5 video

On Fri, Jan 1, 2010 at 4:02 AM, Philip Jägenstedt <philipj@opera.com> wrote:
> On Thu, 31 Dec 2009 17:45:29 +0100, Julian Reschke <julian.reschke@gmx.de>
> wrote:
>
>> Philip Jägenstedt wrote:
>>>
>>> ...
>>>>
>>>> Potentially because of servers without proper partial request (byte
>>>> range) support...
>>>
>>>  In such a case I think one can still decode the first frame and report
>>> the duration as +Inf (this is allowed per spec, and it is true if the
>>> resource is a live stream like a webcam).
>>> ...
>>
>> In that case you'd need still to do a complete GET, and then abort the
>> connection (unless I'm missing something). Not good..
>>
>> Best regards, Julian
>>
>
> Why is that not good, and how could the browser possibly know that the
> resource is not seekable without trying to GET it?
>
> In any case, I think we should always try to get the first frame and
> duration unless we spec an opt-out from that behavior, the implementation
> details are quite beside the point.

Yes, either the first frame or the poster frame. And since the
duration could be done through a Content-Duration HTTP header, it
would mean not downloading anything for the video, which is identical
to putting a img there with a javascript link to replacing the img
element with a video element. I like this approach.

It may well be that autobuffer isn't the right name for this. I used
the word "initialise" before, but I'm not sure that's a good word
either for setting up the decoding pipeline. There is definitely a
difference between "no download of video" - "decoding setup only" -
and "full resource download".

Regards,
Silvia.

Received on Friday, 1 January 2010 01:43:35 UTC