Re: [rtcweb] New Version Notification for draft-jesup-rtcweb-data-01.txt

[Wolfgang Beck <[e-mail]>]

On Mon, Nov 14, 2011 at 1:29 PM, Hadriel Kaplan <[e-mail]> wrote:
>
> No only the psuedo-transport layer is being standardized - i.e., SCTP/DTLS/UDP.

That's what I hoped to hear.

> What goes on top of that (i.e., what the JS actually communicates) would be whatever the JS wants and thus yes of
> course it would only truly "work" if the same JS was used on both peers or someone went and defined some specific
> application protocols for it.  That might actually happen - like if someone went and defined an IM protocol to run on the
> data-channel they could create a spec for it for other web-apps to be able to interoperate.  (or a JS library doing
> something over the data channel could become so popular as to be a defacto spec for one)

In this case, the different JS clients will have to negotiate this
protocol. In the current ROAP model, this is done by the browser, not
the JS client. It will probably get ugly if I want to negotiate a
combined rtcweb-data and A/V session.

Thanks for the clarification.


Wolfgang Beck