Re: [Simple] Presence data model: devices

[Jonathan Rosenberg <jdrosen at dynamicsoft.com>]

I think my reply to your initial post covered most of what I would 
otherwise say throughout the thread, but to emphasize here:


Henning Schulzrinne wrote:

> > "can you call me back on my land-line phone? I can't hear you very 
> > well on
> > this
> > cell phone".
>
>
> None of this requires a device abstraction. The caller can only choose 
> the mobility modality if
>
> (1) there are two separate service entries (one stationary, one mobile)
>
> or
>
> (2) there is one service entry and the caller uses caller preferences 
> that the callee chooses to honor.
>
> Neither requires nor benefits from a device tuple. (Devices tuples 
> cannot be contacted directly.)
>
> In many cases, we'll have a single watcher-visible modality that 
> encompasses multiple devices. For example, on trips to Europe, I carry 
> both my GSM phone (works in both places) and my CDMA phone (works in 
> only one, but is cheaper).

Great example. Let me give you a case for something I dont know how to 
do without the notion of a device.

You have two phones, this CDMA and GSM one. Lets say they both work 
(i.e., you are in the U.S). Your presence document indicates a voice 
service and a SIP messaging service on the GSM phone, and an sms service 
on the CDMA phone. Lets say I'm talking to you on your GSM phone, which 
is low on power. While talking, you say, "hey, can you send me a 
message, but send it to me other phone since I am about to run out of 
...." and at that moment, the call ends since power is out. I want to 
send you a message now, but not to the gsm phone. I have no way to know 
which of the two messaging services is accessible on that phone, or 
another phone, without device information.

Indeed, a convenient user interface would show user "henning" with two 
device icons, each a cell phone. Perhaps one has "gsm" inside, and the 
other "cdma" (again, these are DEVICE characteristics). So, if you say 
to someone, "call my gsm", it becomes very easy for me, or anyone else, 
to call it, or PTT it, or sms it, or access any of the other services 
avaialble on it. Without a device concept, I can't do that.


It is true that, in this case, device ID will suffice. However, the 
device has characteristics and status that are independent of the 
services; and placing them into the service causes unneeded repitition 
and lack of clarity about what they apply to. These characteristics ARE 
about the device; putting them into service is artificial.





 I want the caller to reach my mobile service,
> but don't necessarily want to have to expose the devices as that's none 
> of the caller's business.

Well, thats OK. There is nothing that says you HAVE to report on your 
devices if you don't want to. If you want to just indicate services to 
the watcher thats fine.


 Thus, I need the ability to declare a mobile
> *service* (which is then routed according to my service preferences to 
> one of my phones, e.g., depending on how much an incoming call is going 
> to cost me.) [The proposed model allows this, as far as I can tell, so 
> no disagreement here.]

Right.

The point is, the model I am proposing represents what seems to me, to 
be the minimum of the concepts we've been talking about and assuming all 
along.

-Jonathan R.
--
Jonathan D. Rosenberg, Ph.D.                600 Lanidex Plaza
Chief Technology Officer                    Parsippany, NJ 07054-2711
dynamicsoft
jdrosen at dynamicsoft.com                     FAX:   (973) 952-5050
http://www.jdrosen.net                      PHONE: (973) 952-5000
http://www.dynamicsoft.com

_______________________________________________
Simple mailing list
Simple at ietf.org
https://www1.ietf.org/mailman/listinfo/simple