getUserMedia implementation roadmap

[Anant Narayanan]

Hi all,

After talking with several folks over the last week, here is a proposal
for how we can go about landing a getUserMedia implementation (the first
step towards a full webrtc stack!) across various platforms supported by
Gecko: Android, B2G, Desktop (which includes Windows, Mac and Linux).

https://wiki.mozilla.org/Media/getUserMedia

The general idea is to build a simple abstraction layer for device
access, which the DOM component for navigator.getUserMedia will talk to.
In Phase 2, Each platform will implement the abstract interface, and
there will be a fallback implementation (that returns noise, for
example). The simple image capture API is a priority for Android and
B2G, even if it means we won't use the abstract interface to begin with,
so we can coincide with the impending Apps release.

We will leave out the UI bits for the initial iteration, as we work with
UX teams across the organization on mockups. Phase 3 targets building
the full UI for all platforms, at which point we can enforce a
permission model, and the pref for enabling getUserMedia access from
content can be turned on.

The roadmap does not include any hard dates yet as we are still early in
the process, and determining estimated workload will largely be a guess
anyway. A rough deliverable is to land *at-least* Phase 1 in time for
Firefox 15 (June 5th). Better estimates will be known when we're mostly
done with Phase 1. Hopefully we can do a lot more before the tree closes
in June!

All feedback is appreciated.

Thanks,
-Anant

[Anant Narayanan]

On 5/7/2012 9:07 PM, Anant Narayanan wrote:
> https://wiki.mozilla.org/Media/getUserMedia

To clarify, this plan can be considered a "magnification" of the
"getUserMedia backend" and "getUserMedia in m-c" bubbles that are part
of the larger WebRTC implementation plan Maire sent out earlier today.

-Anant