El Martes, 13 de Octubre de 2009, Daniel-Constantin Mierla escribió:
SIMPLE presence specs are more laborious, allowing more complex services.
Yes, I'm very happy because within 25 years my dishwasher will know the advanced presence status of my microwave :)
There are couple of mistakes in SIP specs, could have been more strict in requirements, but what makes it very complex is the large set of functionalities it wants to solve. Making call between two users is easy,
changing presence between 2 users is easy as well (end-to-end).
Not so easy if two phones are registeres with same SIP account and both publish presence status ;) Using a presence server this means a NOTIFY with two <tuples>. However RFC's specs don't state which should be the behaviour of a watcher when reveiving a NOTIFY with two tuples.
Also, a user could publish <person> and/or <device> sections. What happens if phone_A (sip:alice@domain.org) publishes a <person> section containing "Display-Name = Alice YEAH" and phone_B (also sip:alice@domain.org) publishes a <person> section containing "Display-Name = SuperAlice"? RFC's say *nothing* about it. They "talk" about a presence composer (a node in the presence server which should "mix" the presentities but doesn't specify how. In this point OMA does specify it.
A very complex world... :)