Re: [rtcweb] URIs for rtcweb "calls"

[Ted Hardie <[e-mail]>]

On Fri, Aug 12, 2011 at 12:03 PM, Matthew Kaufman
<[e-mail]> wrote:
> On 8/12/2011 10:47 AM, Ted Hardie wrote:
>>
>> Do others think defining something like this would be useful?
>
> No. For the same reason I don't think that Gmail should be implemented by
> going to "imap:mail.google.com/inbox" and having the browser implement a
> mail reading user interface or it own (or even having a page that contains
> "<showIMAPinterface mailbox='inbox' server='mail.google.com'/>").
>

So, one of the aims here was to allow you to specify what the
signaling path was separate from whom you were calling.  In the case
where they were similar, you'd basically end up with:

rtw:///player23424232;token=234234234

(That is the authority/signaling path would be the same as the current origin).

I'm also not sure that having the ability to create a compact
representation of this equals baking a phone into the browser.  To
take a slightly different example, imagine that I'm using a gaming
site and I want to invite friends to come play the game with me.  A
URI that I could use to paste into a chat window or send in email
would be handy, as a simple way to give an out-of-band pointer to a
rtc-web context.  That would still fire up the full game context,
slotting the user into the game (rather than making a phone call).

Hope that explains the range of things I'd like to cover better.

regards,

Ted


> We should be providing components that allow for all manners for real-time
> communication applications to be built using HTML and JavaScript, not baking
> a phone into the browser.
>
> Matthew Kaufman
>