Re: [Simple] Presence data model: devices

[Paul Kyzivat <pkyzivat at cisco.com>]

hisham.khartabil at nokia.com wrote:
> > > You are making the assumption that tuples for the same service
> > > will get merged.
> >
> > Henning was enumerating the cases. This one case assumes the service 
> > forks to multiple devices. If there is indeed to be presence 
> > information 
> > from each of those then yes, presumably it is being merged.
> >
> > He covered the other cases as well.
> >
> > > I'm not making that assumption.
> > > My watchers will see 2 tuples for the same service if they get
> > > published by different PUAs. Each tuple will have a GRUU.
> >
> > Then by the terminology we are using here, they are two different 
> > services, not one.
>
> I don't agree here. They are the same service.
> You can communicate with me in exactly the same fassion.
> Are you telling me that the IM on one device is not the
> same service as IM on another?

Yes. You don't communicate in exactly the same fashion - you use a 
different address for one than the other.

Jonathan't data model explicitly states that the contact address serves 
as a UID for each service.

> A service does not equal a device.

Agreed.

As Jonathan has described, there can be a many:many relationship between 
services and devices. But when a service is associated with multiple 
devices it still must have a single address. Behind the scenes it may 
fork to one or the other of those devices, but to the watcher, to be a 
single service, it must have a single address.

If this were not the case, then either a tuple describing a service 
would require multiple contacts (not permitted) or else it would be 
necessary to have multiple tuples describing (aspects of) the same 
service. Jonathan's data model explicitly jetisons that approach.

	Paul


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