Re: [rtcweb] SIP MUST NOT be used in browser?[was RE: Remote recording - RTC-Web client acting as SIPREC session recording client]

[Harald Alvestrand <[e-mail]>]

On 09/08/11 16:54, Igor Faynberg wrote:
> Well,  maybe we should take it off-line--I will be happy to explain a 
> few things in October. Just some in-line answers:
>
> On 9/8/2011 10:30 AM, Harald Alvestrand wrote:
>>> I thought I had already proposed a subset: INVITE, SUBSCRIBE/NOTIFY, 
>>> and REFER.
>> ...
>>
>> That's the commands. What are the parameters?
>
> My proposal is to populate the methods with parameters in the 
> JavaScript application.
It seems that writing up your proposal in enough detail that we can tell 
which parts you intend to have done in Javascript and which parts in 
browser code is an useful exercise, because I just don't seem to get the 
picture.
>>
>> As I've said before, unless I've misunderstood what SUBSCRIBE/NOTIFY 
>> are for (I think they are just for presence), 
>
> No, no, no! Not just for presence, although it can be used for 
> presence.  These precious methods came from PINT, how could you forget 
> this!  They allow the client to subscribe for an event at the server 
> and then receive notifications asynchrously.  This specific feature 
> cannot be implemented using HTTP (just as INVITE cannot).  We used 
> them in PINT pretty much to enable the client to learn when PSTN call 
> is disconnected, but the SIP group had found many other uses later.
If the issue is getting a signal from a Web server to a client, there's 
approximately 100 ways to get notifications from the server to the 
client using HTTP (hanging GET being one of them). Now that WS is 
getting standardized, there will be 101.
>
> Again, my objective is to ensure that SIP (with all its bell and 
> whistles) can be supported by a JS application.  I am just starting to 
> study the code kindly sent my way a few minuts ago, but when I am 
> convinced that this can be done based on WebSocket alone, I will stop 
> harping on this.
>
> Igor
>