[Ietf-caldav] Re: RE : [Ietf-calsify] iTIP: persisting SEQUENCE number

Bernard Desruisseaux bernard.desruisseaux at oracle.com
Wed Oct 3 10:24:38 PDT 2007


Hi Arnaud,

Arnaud Quillaud wrote:
> 
> Cyrus Daboo wrote:
 >
>> Again, my feeling on this is as follows: iTIP makes it clear that the 
>> ORGANIZER'S CUA has to persist on a per-ATTENDEE basis, the 
>> SEQUENCE/DTSTAMP values (section 2.1.5). What is important is 
>> that the CUA 
>> track what the next expected SEQUENCE/DTSTAMP reply for any 
>> one attendee is 
>> going to be.
>>
> 
> In the case of a CalDAV based CUA, how is the client supposed to persist
> this per ATTENDEE SEQUENCE information on the server (e.g. in the case
> where the organizer is using 2 cal clients) ? If CalDAV scheduling is
> used, the CUA may try to guess the right SEQUENCE by looking at the
> outbox history but that is not guaranteed to work as servers are not
> required to keep the outbox history.

Currently, the CalDAV calendar-schedule draft only defines two new
iCalendar parameters RECEIVED-SEQUENCE and RECEIVED-DTSTAMP to
"persist" the value of the SEQUENCE and DTSTAMP properties that was
specified in the most recent reply received from a given Attendee.
Example:

   ATTENDEE;CN="Cyrus Daboo";PARTSTAT=ACCEPTED;RECEIVED-SEQUENCE=2;
    RECEIVED-DTSTAMP=20070214T123456Z:mailto:cyrus at example.com

The purpose of these parameters is only to allow the handling of
scheduling messages that arrive in an unexpected order (i.e., not
at all what you are asking for)

We are still debating whether there is a need for the Organizer to
"persist" on a per Attendee basis the last SEQUENCE number sent to
the Attendee.

I believe we need to answer the following questions:

1- Is there a use case where the Organizer bumps the SEQUENCE
    number but doesn't send an update to any Attendee?

2- Is there a use case where the Organizer bumps the SEQUENCE
    number and only sends an update to some Attendees?

    [ The case of an Organizer adding new Attendees and only
      sending an update to those new Attendees would be an
      example *if* the Organizer (his CUA really) also bump
      the SEQUENCE number.  According to iCalendar (though
      it should be "according to iTIP") the Organizer is not
      required to bump the SEQUENCE number here. If the
      SEQUENCE number is bumped, I would say that all the
      Attendees should receive an update... ]

3- Is there a use case where the Organizer would consider
    an Attendee reply that specifies a SEQUENCE number less
    than the current SEQUENCE number?

Cheers,
Bernard


More information about the Ietf-caldav mailing list