Jeffrey, Aparna, Morgan and I have been trying to decode the mystery
behind all of the false positives popping into the NOW section. A few
developments I wanted to bring to the attention of the list include:
We had previously decided to pop every occurrence in a recurring
series to NOW if a 'global or this and future edit' has been made.
However, given that the byline doesn't currently give very much
information about what's been changed, this can be really
disorienting and not all that helpful. So I'm now wondering if we can
instead *only* pop the NEXT OCCURRENCE when there's been a global
edit. Examples of global / this and future edits include...
+ Changing the end date
+ Changing the recurrence rule
+ Title/Location/Notes field changes
+ Any Addressing stamp changes
When specific occurrences are deleted, nothing should pop to NOW.
This would be consistent with what we do when non-recurring items are
deleted.
Morgen, Jeffrey? does this sound reasonable to you?
I've logged a bug to track this: https://bugzilla.osafoundation.org/
show_bug.cgi?id=11661
It's still not clear however that global edits are the only reason
behind the algae bloom of 'old' items popping into NOW. Some are non-
recurring events. Others are items edited by *me*, but then get
marked Unread. Still others are message items, so there might be some
email layer thing going on.
I *am* happy to report however that after reloading my data into the
latest trunk build, I haven't seen any false positives in the NOW
section of our 'TestPopIntoNow' test-collection...so long as email
isn't involved.
I am now able to reliably reproduce a case where if you email an item
to yourself (doesn't have to be a recurring event), the item gets
marked as Unread and pops to the top of NOW when you sync Mail. This
is causing some fraction of the 'Pop to Now' bugs.
https://bugzilla.osafoundation.org/show_bug.cgi?id=11660 - Sending
email to yourself in Chandler causes false-positive 'Unread' status
and 'popping into Now.
bkirsch, does this make sense to you?
Mimi
Begin forwarded message:
From: Mimi Yin <[EMAIL PROTECTED]>
Date: December 13, 2007 6:21:09 AM PST
To: Morgen Sagen <[EMAIL PROTECTED]>
Cc: "Jeffrey Harris" <[EMAIL PROTECTED]>, "Aparna Kadakia"
<[EMAIL PROTECTED]>, "Philippe Bossut"
<[EMAIL PROTECTED]>
Subject: Re: 'Weird triage and sharing bugs to focus on' Redux
I'm not sure that global edits are the only thing that is making an
entire series pop to the top. My 'Summarize Design List' weekly,
anytime on Wednesday event series just popped to NOW and I think
it's because yesterday's occurrence was auto-triaged to NOW. I
haven't made any edits to it...and they all claim to be edited by
me. It's also not a message item.
As for global edits, could we: Only pop the 'next occurrence' to
NOW as a representative of the entire series, if there is any kind
of global edit.
We can probably get away with deletion not popping anything to NOW,
since we don't tell users about deletions of normal items either.
I will summarize and forward this to the design list.
On Dec 12, 2007, at 2:43 PM, Morgen Sagen wrote:
On 12/12/07, Jeffrey Harris <[EMAIL PROTECTED]> wrote:
Hi Morgen,
Today's fun "popping to NOW" adventure seems to because a master
event's recurrence rule spontaneously changed from
FREQ=WEEKLY;UNTIL=20070118T131500Z to
FREQ=WEEKLY;UNTIL=20070117T235959Z on the hub. Chatting with
Randy
about it now.
So it ends up that Aparna (because I asked her to) made a change
to this
event with the Web UI. Even though the Web UI gave her an
error, it did
seem to make a this-and-future change to the series. This means a
recurrence rule change on the master event (which both Mimi and
Aparna
received in their Chandlers), and so Chandler saw the change to the
master and popped the entire series to NOW.
So Jeffrey, It seems that perhaps Chandler does need to ignore
certain
changes on the master when it comes to popping the entire series
to NOW?
Sorry, I missed this thread till now.
It doesn't seem obvious to me that it's more desirable to ignore
recurrence-rule end date changes than to pop the whole series to
now.
In many circumstances, I want to know about that kind of change
to the
rule (what, you canceled this Monday's rehearsal? I didn't
know...).
I think here we're bumping into two issues:
- It's ugly to move the entire series to now, it would be much
cleaner
to have a master occurrence, perhaps it could be rendered
differently in
the detail view
- For all these pop to now problems, it's frustrating that the
end user
isn't able to answer the question, "What did the author of recent
changes actually change?" I think if we had UI to expose the log of
recent changes, it'd be much less annoying when we get seemingly
spurious popping to now. Of course this is a tall order.
Anyway, with all that in mind, if the use cases we're seeing make
ignoring certain rule changes (when choosing whether to pop to now)
desirable, that's fine. Which rule changes should we ignore?
End date
change? The creation of an end date when the rule originally ran
indefinitely? How about the deletion of a single occurrence?
All of those currently will pop the whole series to now.
/me waits for Mimi to answer Jeffrey's question about what should and
should not pop an entire series to Now.
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Open Source Applications Foundation "Design" mailing list
http://lists.osafoundation.org/mailman/listinfo/design
