1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
|
= Event Recurrences =
Events defined by iCalendar objects may recur when
[[http://tools.ietf.org/html/rfc5545#section-3.8.5|recurrence component properties]]
such as `RDATE` and `RRULE` are employed. Each recurrence of an event may then
be referenced using
[[http://tools.ietf.org/html/rfc5545#section-3.8.4.4|recurrence identifiers]],
and such identifiers indicate the originally-specified start point of a
particular recurrence (either implicitly specified by `RRULE` properties or
explicitly specified by `RDATE` properties).
== Recurrence Identifier Stability ==
A recurrence retains the same identifier throughout its lifetime. Even if a
recurrence's start date or time changes, it will still retain the same
identifier in its `RECURRENCE-ID` property which will no longer reflect the
currently-specified start point of the recurrence. Such identifier stability
is intended to provide a means of identifying the original recurrence so that
it can be hidden from any calendar or event descriptions and replaced with the
modified version.
== Recurrences and Time Zones ==
Since recurrence identifiers may be defined using time zone information,
imip-agent normalises the specified recurrence identifiers to UTC-based
datetimes to minimise ambiguity. For example:
|| '''iCalendar Property''' || '''Normalised Value''' || '''UTC Datetime?''' ||
|| `RECURRENCE-ID:20141114` || `20141114` || No ||
|| `RECURRENCE-ID:20141114T000000` || `20141114T000000` || No ||
|| `RECURRENCE-ID;TZID=Europe/Oslo:20141114` || `20141113T230000Z` || Yes ||
|| `RECURRENCE-ID;TZID=Europe/Oslo:20141114T000000` || `20141113T230000Z` || Yes ||
|| `RECURRENCE-ID:20141114T000000Z` || `20141114T000000Z` || Yes ||
Identifiers without time zone information are not in themselves sufficient to
unambiguously define points in time, and thus additional time zone information
must be provided to obtain such time periods for such purposes as detecting
conflicts with other events. In the above examples, those normalised values
not providing a UTC datetime representation need further conversion to be
usable for period comparisons. Such further conversion would be done by
nominating a disambiguating time zone, such as the user's configured time
zone.
By normalising identifiers using any object-resident time zone information,
imip-agent can use the resulting values without needing to consult the object
providing any redefined recurrence, knowing that any time zone information has
already been taken into consideration. Thus, all UTC-based datetimes used as
recurrence identifiers are readily usable for comparison purposes, whereas any
floating date or datetime values used as recurrence identifiers must need
additional conversion using the user's time zone to be usable.
It might be thought that there would be correspondence between a recurrence
identifier and the time zone details employed by the original object
describing the redefined recurrence (such as the `TZID` attribute specified on
an object's `DTSTART` property), and so any unqualified recurrence identifier
might be converted to a UTC-based datetime using such time zone details.
However, an assumption could equally be made that the recurrence identifier
should inherit time zone details from the redefined recurrence instead. The
only reasonable choice to be made when confronted with such ambiguity is to
treat any unqualified identifier as a genuine floating date or datetime, and
the normalisation process facilitates this strategy.
== Recurrences and Free/Busy Information ==
Events employing recurrences on fixed occasions can be readily recorded in the
free/busy information for a calendar user. However, iCalendar also permits
recurrences that may potentially continue forever, and yet providing free/busy
information for arbitrary periods in the future may either result in
substantial computation or substantial demands on storage resources.
Consequently, free/busy information may only be generated for a period ending
at a certain point in the future defined in terms of days from the present.
Within this period scheduling would make sense, and attempts to schedule
events outside this period would succeed at the participant's own risk.
Such a period where participant availability is known must be necessarily
expanded as time progresses. One-off events, once recorded in the free/busy
records, will not contribute further to expansions of those records. Recurring
events, however, may provide additional periods of interest as the
availability window moves forward in time.
To determine which events contribute recurrences, a list of objects (initially
all objects known to a user that have not been cancelled) is consulted and
their recurrence properties inspected. With such knowledge of recurring
events, upon expanding the availability window, only these "known recurring"
events need to be inspected for further contributions to the free/busy
records, and those no longer contributing after a given point can be discarded
from the list for future expansion of the window. Meanwhile, new events would
need to be added to the list, at least if they were defined as providing
recurrences that may occur in future availability periods.
=== Updating Free/Busy Records ===
To update and thus expand availability information, it is suggested that a
regularly scheduled task be used to consult the events known (or thought) to
provide additional free/busy periods and to record such additional periods for
each user. This can be done using a system's `cron` daemon and a suitable
script in `/etc/cron.daily` or equivalent. Such a script is
[[../CronIntegration|provided]] in the imip-agent distribution along with a
program that can expand availability information for all known recipients of
calendar information.
|
