- From: Stefan Håkansson LK <stefan.lk.hakansson@ericsson.com>
- Date: Wed, 10 Apr 2013 14:17:08 +0200
- To: Martin Thomson <martin.thomson@gmail.com>
- CC: "public-media-capture@w3.org" <public-media-capture@w3.org>
On 4/9/13 6:30 PM, Martin Thomson wrote: > On 9 April 2013 07:43, Stefan Håkansson LK > <stefan.lk.hakansson@ericsson.com> wrote: >> My preference would be *1*. I don't really see why we would do any changes >> if we do not allow the same track to be part of more than one MediaStream - >> the current text says that constructing a MediaStream, or doing addTrack, >> clones the tracks. > > Not taking any position here, but I can't parse your rationale here. > > > Just to be perfectly clear, the reason we are discussing this is: > Tracks are mutable, their settings can change. Tracks can be grouped > in multiple MediaStreams. Certain properties of tracks were perceived > as only being relevant in the context of a MediaStream. > > I'm speaking here of enabled/disabled specifically, the other settings > are less problematic. Changing enabled/disabled only has meaning in > the context of the MediaStream. As a consequence, a change to a Track > has surprising side effects if that Track is used in two places. Yes, it can have surprising effects, but it can also be a feature. But my point was, that if we do not want to allow that, then I see little need to change the current spec (currently it is specified that addTrack as well as constructing a MediaStream always clones). > > ... An alternative that hasn't been considered (at least publicly) is > to have enabled/disabled as a Track-indexed property of a MediaStream. > Then Tracks don't need this property. This is an option we never considered. There was for some time a discussion if we should remove enable/disable from tracks, and instead do it on the consumer (Recorder, PeerConnection, media element), but a recent discussion concluded we should not. > Two possible embodiments: > split Tracks into enabled and disabled sets, or have a map (something > like stream.enabled[track]). That's almost too ugly to contemplate > though. >
Received on Wednesday, 10 April 2013 12:17:26 UTC