Iñaki Baz Castillo wrote:
> El Miércoles, 10 de Junio de 2009, Paul Kyzivat escribió:
> 
>> Actually, *in theory*, you can put Accept-Language:es into your INVITE,
>> and then the text messages in the response ought to be written in Spanish.
>>
>> (And of course everybody's SIP implementation does that. :-)
> 
> Yes, I was to say the same XD
> 
> 
> 
>> Practically speaking, if you want the caller to display some indication
>> with a specific semantic then you do need some signaling that coveys
>> that semantic. The 182, ignoring the reason phrase, comes *close* -
>> close enough that IMO it would be reasonable for a UA to display some
>> message that the user would reasonably interpret as call waiting.
> 
> Ok, I like this, more than what you suggest now:
> 
> 
>> To get more specific than that, I have, for quite some time, been
>> promoting the concept of using Alert-Info with a URN, and defining URNs
>> with specific semantics. Having one with a "call waiting" semantic would
>> be quite reasonable. Alert-Info in a response is already defined, so
>> this is not an abuse. The main problem preventing use of Alert-Info with
>> URIs has, IMO, been one of trust - you don't know whether you are going
>> to get something offensive or dangerous. Using URNs with agreed upon
>> semantics eliminates that issue.
> 
> Do you mean that the UAC receives a 1XX with Alert-Info header containing an 
> URN for "call waiting" so the UAC would contact the URN URI and so on?

Yes.

> Sincerelly I don't like this idea, why should it contact a remote server just 
> to understand the call-waiting concept?

The point is URN, not URL. The URN is a *name*, not a *location*, so you 
don't contact any remote server. You either understand the name, or you 
don't. You are potentially able to understand it because there is a 
registry of URNs (in IANA) that points to a definition of what it means. 
But you will have to teach your software the URNs you want it to understand.

So, it is really just a different extension mechanism. In this case it 
would be used as an extension mechanism for alerts.

> Also, being realistic, most of the phones ignore Alert-Info header, or at 
> least, ignore contacting its URI. Am I wrong?

No, you are right, I think. This is just a potential way forward for 
functionality of this sort, where there might be lots of different kinds 
of alerts that are useful, but in an environment where I don't want to 
fetch some URL, especially an unknown one that I might not trust.

Because alerts are advisory, its probably fine if a recipient that 
doesn't understand one just ignores it.

Alerts defined by URNs could have all sorts of different degrees of 
definition:
- It could denote a particular tone or audio clip
- It could denote a particular piece of music but not its rendition
- It could be some text that is intended to be localized for the
   receiver
- It could simply be a concept, with the rendition left open for
   the implementation (e.g. EMERGENCY)
- ...

I believe this approach is being adopted by 3gpp for some things, but I 
don't have the details.

Of course if you just want something you can use today, this won't help. 
It is simply a possible direction forward.

        Thanks,
        Paul
_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to