[Standards-JIG] RE: Timed presence events

Jean-Louis Seguineau jean-louis.seguineau at laposte.net
Thu Mar 17 19:54:21 UTC 2005


Peter,

If you browse to
http://www.cs.columbia.edu/sip/draft/rpid/draft-ietf-simple-rpid-05.html
you'll notice in section 3.1 that many SIMPLE 'rich presence' tags may be
qualified with a 'since' and/or 'until' date attribute. You are certainly
right when saying that the most probable use case is to have this values
generated automatically by a calendar application. And you can trust our
friends at MSFT to implement a SIMPLE version of JEP-0132 and claim a patent
for it ;)

I'll be coming back soon with a prototype JEP for this.

Jean-Louis

-----Original Message-----
Message: 1
Date: Wed, 16 Mar 2005 12:14:32 -0600
From: Peter Saint-Andre <stpeter at jabber.org>
Subject: Re: [Standards-JIG] RE: Timed presence events
To: Jabber protocol discussion list <standards-jig at jabber.org>
Message-ID: <20050316181432.GM11416 at jabber.org>
Content-Type: text/plain; charset=us-ascii

On Tue, Mar 08, 2005 at 12:02:36PM +0100, Ralph Meijer wrote:
> On Tue, Mar 08, 2005 at 11:28:29AM +0100, Jean-Louis Seguineau wrote:
> > [..]
> > 
> > My question was more targeted at the protocol level. There are other
> > contexts where having a way to indicate the time boundary of a status,
or an
> > activity, are precious. And this information may be applied to other
> > application than just the interactive IM.
> > 
> > As an additional example, we have a way of expressing a user's activity.
In
> > certain circumstances, we may want to add time related information:
> > 
> > <activity xmlns='http://jabber.org/protocol/activity'>
> >   <relaxing>
> >     <partying/>
> >     <valid xmlns='http://jabber.org/protocol/something'
> >        from='2002-09-10T20:41:07' 
> >        to='2002-09-10T23:41:07'/> 
> >   </relaxing>
> >   <text xml:lang='en'>My nurse's birthday!</text>
> > </activity>
> 
> Right, for a few Extended Presence features, this would be very useful. I
> will discuss with stpeter, how to best go about it, but your example is
> good candidate.
> 
> I don't believe we should want this for normal availability (using
> <presence/>). You availability should be set by the client, not
automagically
> at some time in the future. If I am not there, I am not there.

My understanding of the PIDF/SIMPLE extension (IIRC this was in early
versions of PIDF but seems to have been dropped before it went to RFC)
is that the time periods are hints. The main use case is to integrate
with a calendar -- if my calendar knows I will be in a meeting from 9:00
to 10:30, the calendar could set my show to be "dnd" or my status to be
"in a meeting", and when it updates my presence it could also specify
the expected time period for that presence. Of course, if my meeting
ends early or goes over, then I might manually update my presence since
I doubt that my calendar is smart enough to know that -- unless of course
it also implements JEP-0132. ;-)

The validity syntax looks fine to me (all times UTC, please!). Perhaps a
small JEP on this is in order? I could probably write that one with my
eyes closed. :-)

/psa




More information about the Standards mailing list