On 17 Sep, 2007, at 14:53, Mimi Yin wrote:
The second EIM - end-user modeling issue has to do with
relationships between attributes.
2. There are dependencies between end-user attributes that aren't
reflected in EIM. One example of where this is a problem is:
+ User A changes the end-date on a recurring event which results in
the deletion of the last 3 occurrences.
+ User B syncs and sees a pending change to 'delete' the last 3
occurrences.
+ If/when user B discards that pending change, the 3 occurrences
are simply left as 'orphan' events.
There's currently no way to 'rewind' so-to-speak and 'undo' the end-
date change and restore the last 3 occurrences as members of the
recurring series.
Morgen, Jeffrey, Grant? Do I have this right? Are there other
examples?
That sounds right to me. Another example of this kind of dependence,
which is an attribute affecting the existence of entire items,
changing the recurrence interval.
I sometimes wonder, in cases like these, if we should just make
RDATEs for any to-be-removed events that contain user changes (i.e.
are modifications, in our recurrence parlance).
--Grant
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "Design" mailing list
http://lists.osafoundation.org/mailman/listinfo/design