Re: Disposition: "MediaStream with worker" for video processing

Hi Harald,

Thanks for the response. We'll move forward as you suggest.



> On 20 Aug 2015, at 16:03, Harald Alvestrand <> wrote:
> Anssi,
> after discussion with the other chairs, we've concluded:
> - We think this is a valuable use case to explore.
> - We don't think anything related to it needs to hold up finishing the
> current Media Capture and Streams spec
> - We can offer a repository in the w3c github project to host this spec
> - just drop a note to Dom to tell us what name you want - it may want to
> start with "mediacapture-", but doesn't have to do that even.
> - We think it's best that the most interested parties gather together in
> a forum outside the WG/TF mailing list to hammer out a proposal -
> mailing list, bug tracking or however you want to do it
> - When we roll around to TPAC time, it should be an appropriate time to
> deal with plans for future work on media that doesn't necessarily move
> over networks - there are more issues here, as both our discussions with
> the Audio WG and the establishing of the Timed Media WG are signs of.
> The TF will continue to focus on finishing the Media Capture and Streams
> spec for now.
> Harald, for the chairs
> On 08/13/2015 08:33 PM, Kostiainen, Anssi wrote:
>> Hi Martin, Eric,
>>> On 13 Aug 2015, at 19:29, Martin Thomson <> wrote:
>>> On 13 August 2015 at 03:59, Kostiainen, Anssi
>>> <> wrote:
>>>> * Reuse the existing Media Capture Task Force (joint activity of the Device API WG and WebRTC WG) for mediacapture-worker [1] initially. The proposal was introduced to the TF already by CTai, and I don't recall hearing any concerns on adding it to the TF.
>>> Um, Eric raised concerns about this, specifically regarding the
>>> capacity of the group to complete new work when there is a significant
>>> amount of existing work outstanding.
>> Thanks for your feedback. I definitely don't want to slow down the existing work -- my expectation is we can handle that issue by being clever with tooling, and by getting new people to participate and contribute.
>> Tooling. This work would happen in its own GitHub repo to which people subscribe to if they want to. I'm also not expecting this work to steal any telco time, since I believe async participation is preferred. No extra traffic on the public-media-capture or public-webertc mailing lists.
>> People. I'd not expect this work to steal bandwidth from the same people working on existing deliverables, rather get new people to participate and help. I believe CTai has volunteered to edit, and we have a bunch of other people who have indicated interest to contribute, review.
>> While the Timed Media WG may be the long-term home, I'm wondering if we can find an interim solution that allows interested parties to collaborate already now. I'm seeing wider support among implementers and others and it would be pity to hold off this effort due to process reasons (i.e. WG creation takes time).
>> I'm open to suggestions. Eric proposed to form a mailing list for the Timed Media WG and wait. That might be one option. Any concerns with that? Others?
>> Thanks,
>> -Anssi

Received on Friday, 21 August 2015 08:12:50 UTC