Re: Should <video> buffer control be tri-state?

On Jan 3, 2010, at 7:34 PM, Chris Double wrote:

> On 01/04/2010 3:55 PM, Silvia Pfeiffer wrote:
>> Durations could be provided by the server using a "Content-Duration"
>> HTTP header with a HEAD request on the resource.
>
> Not all servers allow adding headers like this. Amazon S3 for  
> example. I don't think mandating that 'no buffering' means nothing  
> at all is read (thus no duration being available) is a good idea.
>
> When I see a video on a weblog and I want to decide whether to play  
> it or not I like to see the duration.
>
> Before tinyvid displayed the duration on videos I got many  
> complaints from users that this was not available.

There might be a use case for "no metadata". Let's say you're showing  
a very large number of thumbnails that have no controls but can  
preview in place, either on click or on mouseover. In that case, you  
may not need the duration until the user picks a video. However, two  
things to keep in mind: (A) we should actually compute what the likely  
bandwidth savings are in a typical scenario before we assume it's  
worth optimizing - if it's more like 10% than 50%, then it's likely  
not worth the cost of an extra mode. (B) This mode can be simulated  
with script, by using an <img> element, while the others cannot, so  
we'd have to think this was a very common and important need to make  
it worth supporting the extra mode explicitly.

Regards,
Maciej

Received on Monday, 4 January 2010 03:56:43 UTC