[Standards-JIG] rfc3921bis, RC3

Joe Hildebrand hildjj at gmail.com
Thu Oct 12 22:12:01 UTC 2006


If the user is now offline, how do you distinguish that from a  
transient network failure?

Maybe this should be an IQ, with a definitely correlated result.

On Oct 12, 2006, at 5:56 PM, JD Conley wrote:

>> How is this expected to behave in case user sends unavailable  
>> presence
>> broadcast , and then sends directed presence ?
>> Notifying recipients of directed presence currently uses the
>> unavailable
>> presence trigger ...
>> This requirement does add significantly to the server resource usage
>> since there is no limitation to who all a user can send directed
>> presence to.
>
> Interesting use case. If we aren't going to store directed presence
> (which a bunch of people oppose for resource issues), I would say that
> instead of the last broadcast presence (in all cases) we could  
> create a
> protocol extension for probes in general. Some sort of protocol  
> when you
> just want to know the last presence you received is still valid.
>
> <presence
>     type='probe'
>     from='romeo at example.net/orchard'
>     to='benvolio at example.org'>
> 	<ping xmlns="probeping"/>
> </presence>
>
> <presence
>     from='benvolio at example.org/pda'>
> 	<pong xmlns="probeping"/>
> </presence>
>
> -JD




More information about the Standards mailing list