[OpenSIPS-Devel] Idea: Presence composer module

[Anca Vamanu]

Hi Inaki,

The truth is that indeed there is no definition in RFC on how the 
documents should be composed. What the presence module does at the 
moment is to keep the first person node and add all the tuples nodes. 
But this is not good enough in all cases, one being your example. So I 
believe that your module proposition might be useful.
I understand that what you propose is to have a module that deals with 
aggregation and when aggregation is needed to call function from there. 
Also I suppose that the aggregation itself should be configurable and 
the rules dynamically defined. And also the aggregation functions should 
be per event.

regards,
Anca

Iñaki Baz Castillo wrote:
> Hi, I'm thinking that perhaps a feature/module called "presence composer" 
> could be useful for SIP environments with rich presence usage.
>
> Example:
>
> - A user publishes a presentity <person> section using XCAP for static data 
> (display name, photo icon...). In this way, if I susbscribe to the presentity 
> of this AoR, I could also receive those data (as in MSN network).
> - But what about if the device also publishes a presentity <person> section 
> (with also a <tuple> section) using normal SIP PUBLISH? In that case OpenSIPS 
> presence agent would append it to the presentity published via XCAP. Most 
> probably, the watcher would discard one of these <person> sections and loose 
> useful info about display name, icon and so on.
> - It would be useful if a new "presence composer" module could mix both 
> <person> sections giving priority to the section published via SIP PUBLISH but 
> keeping the data of the XCAP presentity not present in the SIP PUBLISH. This 
> is: if the <person> section in the PUBLISH contains just display-name info 
> then this would replace the display name in the XCAP presentity, but the icon 
> of the XCAP presentity would be present in the composed <person> section.
>
> This is just an example, I could imagine more cases in with such feature would 
> be useful. Also, the concept of "presence watcher" exists in presence related 
> RFC's (but of course is not defined at all).
>
> What do you think?
>
> Regards.
>
>
>
>