W3C home > Mailing lists > Public > public-media-capture@w3.org > April 2013

Re: MediaStream: multiple consumer case

From: Harald Alvestrand <harald@alvestrand.no>
Date: Wed, 03 Apr 2013 14:56:34 +0200
Message-ID: <515C2702.2080306@alvestrand.no>
To: public-media-capture@w3.org
On 04/03/2013 10:04 AM, Stefan HÃ¥kansson LK wrote:
> On 4/2/13 6:36 PM, Anne van Kesteren wrote:
>> Hi,
>>
>> http://dev.w3.org/2011/webrtc/editor/getusermedia.html#mediastream
>> does not seem particularly clear on the subject of allowing multiple
>> consumers. It's been suggested to me that this may be clearer
>> elsewhere in the draft, but I think it would be good to define aspects
>> of the MediaStream primitive where that primitive is introduced.
>
> I agree, and in particular I have always found the picture in section 
> 4.1 as confusing/misleading in that it indicates that the MediaStream, 
> with two consumers, would have to be defined with two (input/)outputs. 
> I think it should be changed in a way where several consumers are 
> attached to the single output of a MediaStream.

Yes, the two right hand arrows need to look more like this:

    +--->
  --+
    +--->

than what they currently do.
I also think the figure should come 4 paragraphs further up, under the 
paragraph that says "A MediaStream object has an input and an output", 
since that's what it's illustrating.

>
>>
>> Also, I wonder, if we introduce a Stream object for XMLHttpRequest,
>> should it similarly support such a multiple consumer case? And if so,
>> how would the read semantics of such a thing be defined?
>
> Are you referring to section 4 of the "Media Capture and Streams" 
> document (the name of the section is "Stream API"; it should be 
> MediaStream API I think) or to the "Streams API" document by the Web 
> Applications WG?

In the case of a MediaStream, the semantics are pretty clear - each 
consumer gets the same content (for some definition of "same" - below 
the covers, it may be adapted to the particular consumer, but if they're 
identically configured and are running under identical conditions, two 
consumers should get exactly the same content).

A consumer does not need to worry about whether or not there are other 
consumers - that's strictly the UA's business to deal with.

With XMLHttpRequest, do you want that semantic (all consumers get their 
own copy)?
Received on Wednesday, 3 April 2013 12:57:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:24:40 UTC