Re: [rtcweb] Preserving stream isolation when traversing the network

[Tim Panton <[e-mail]>]

On 7 Mar 2014, at 09:21, Martin Thomson <[e-mail]> wrote:

> On 6 March 2014 15:03, Matthew Kaufman (SKYPE)
> <[e-mail]> wrote:
>> Would be good to think about whether the default should be isolated (with a special way for sites to ask the browser to relax the restriction) or not isolated (with a way to ask for isolation). The traditional way for "the web" is to do the latter, but I think by now we've seen why we might have wished otherwise.
> 
> We've talked about this in the past.  There are two aspects that we've
> considered: whether to prompt for access to isolated streams (we
> decided that this could be considered creepy), and whether to default
> to isolation.  I don't think that we can realistically default to
> isolation at this point.

I'm skeptical about the usefulness of isolation. If I understand it correctly, it would disable the ability
of a user interface to do any webGL or webAudio processing on the audio or video signals.
Many of the use cases I see use these APIs to do things like add titlebars, mix audio streams play audio prompts etc,
isolation reduces webRTC to  door-intercom like functionality, whilst there are good usages of that, I suspect they are in the 
minority.

So I'd oppose any move to make streams isolated by default.

T.