[Ietf-caldav] Comments on 3.2.3. Handling Scheduling Object
Resources
Arnaud Quillaud
Arnaud.Quillaud at Sun.COM
Fri Sep 26 07:42:08 PDT 2008
** Default value for SCHEDULE-AGENT **
Having the default value being SERVER still sounds wrong to me. If I'm
using a calendar-access only client and my server is suddenly migrated
to a schedule enabled server, I will suddenly start sending invitations
without even knowing it.
** Reloading **
<<
As a result clients SHOULD reload the calendar data from
the server as soon as it is stored in order to update to the new
server generated state information.
>>
Why SHOULD and not just MAY ? This is just a hint, no ?
** Restrictions for create and modify**
<<
The server MUST reject any attempt to set the "PARTSTAT"
iCalendar property parameter value of the "ATTENDEE" iCalendar
property of other users in the calendar object resource to a
value other than "NEEDS-ACTION" if the "SCHEDULE-AGENT" property
parameter value is not present or set to the value "SERVER".
>>
We already discussed this but just for the records:
The server may end up generating an iMIP message for a particular
attendee, hence the response from that attendee (and hence his PARTSTAT)
may come via iMIP. In that case, the client must be able to update the
attendee PARTSTAT to a value other than NEEDS-ACTION.
If I understood correctly, a solution would be to have a new
"server-request" value for the SCHEDULE-AGENT. This value could be set
only by the server (as the result of sending the invitation via iMIP).
Any attendee with such a value could have its PARSTAT changed by the client.
** Restriction for modify **
<<
The client MUST NOT change the "UID" iCalendar property parameter
value in the scheduling object resource.
>>
but this is true already from calendar-access so is there any particular
reason why this was re stated ?
** Attendee removing an event **
<<
If
the server is unable to deliver the scheduling message, the
remove action MUST fail, and an appropriate "SCHEDULE-STATUS"
iCalendar property parameter set on the "ORGANIZER" property in
the scheduling object resource stored by the server.
>>
For the remove action failing, you imply that the whole operation is
synchronous which is not the case in reality.
Arnaud Q
More information about the Ietf-caldav
mailing list