Re: Updated Presentation API WebIDL - call for review

Hi Wes, All,

On 03 Mar 2014, at 18:29, Wesley Johnston <wjohnston@mozilla.com> wrote:

> Just glancing at this briefly, there's still nothing in the spec to describe how we would actually facilitate communication between two different UA’s.

Wes - thanks for your feedback.

> i.e. UA1 calls postMessage() -> Magic happens -> UA2 receives onMessage
> 
> We'd really like to avoid that sort of magic, and there are already specs (as I mentioned earlier, PeerConnection?) that do handle it.

The “magic happens” part is an implementation detail. Under the hood, the implementation could, for example, re-use existing machinery of RTCPeerConnection, WebSockets, XHR depending on what works best for the specific case and device at hand.

The API normalizes connection technologies to secondary displays such as Miracast, Airplay, the protocol used by Chromecast, and others. We gain flexibility by not specifying the connection method. This also insulates the web developer from details such as device discovery, connection setup and such.

> In the same way, if we're tackling two UA systems here, I think we also need to define how discovery and the initial setup of a connection between two UA's is handled. I'd like to avoid a situation where that communication has to be defined differently for each different type of device.

>From the web developer's point of view, not exposing details such as discovery, connection setup, makes the API much easier to use. On the other hand, implementers need to do a bit more work.

I feel the current design is aligned with the established design principle for the Web, Priority of Constituencies:

  http://www.w3.org/TR/html-design-principles/#priority-of-constituencies

> I like the idea of using this as a way to make Network Discovery work, but I think we'd be hesitant about implementing anything until there were solutions to those problems.

I’d like to understand what are the concrete problems that would need to be addressed you refer to above, and from which use cases those requirements are derived from.

Thanks,

-Anssi

Received on Tuesday, 4 March 2014 09:35:09 UTC