Notifications: uselessness of

[Jordan Mantha]

On Fri, Feb 27, 2009 at 1:52 AM, Mat Tomaszewski
<mat.tomaszewski at canonical.com> wrote:
> Bruce Cowan wrote:
>> On Thu, 2009-02-26 at 18:18 +0000, Matthew Paul Thomas wrote:
>>
>>> One pleasant side effect of our work on notifications is that it has
>>> given us the excuse to rip out some gratuitous notification bubbles,
>>> such as the one saying "Your laptop battery is now fully charged". And
>>> when an interactive notification is really necessary, an alert box or
>>> other window that opens in the background will usually be less
>>> distracting then a bubble that floats on top of your work.
>>>
>>
>> Replacing a "distracting ... bubble that floats on top of your work"
>> with a distracting window that appears behind your work doesn't seem to
>> be much of an improvement. Also, why is the window list considered a
>> second notification area now?
>>
> First of all, let me quickly welcome everyone and introduce myself. My
> name is Mat and I'm with Canonical's design team, being responsible for
> large part of the design of the new notification system.
>
> Now, to the point :)
>
> I can't quite see how the window that appears *behind* your other open
> windows, so that you *don't see it* until you close/minimise other
> windows, is as distracting as big, ugly yellow bubble that appears *on
> top of* your open windows, covering your work. What you're saying simply
> does not seem logical.

I can't quite see how you can't quite see it. :-)  The notification
popups allow you to quickly address whatever has come up and get back
to work on whatever you're doing. When you just have a flashing window
list item you *have* to switch context to even figure out if it's
important or not! That's a huge flaw in design logic, IMO. The first
thing the user wants to be able to do is prioritize actions, the
flashing window list does *not* allow for that.

One of the primary arguments against the old notifications seems to be
that they are big and yellow, something that can totally be changes
without fundamentally changing the way notifications are done. The
"they're ugly" argument is a bit of a red herring and totally
orthogonal to the discussion.

> I also find it very arficicial and unconvincing to distinguish between a
> bubble that contains actions and an open alert window as two completely
> different things. One of them contains text, a button to invoke an
> action, and another button (x) to close it. The other contains text, a
> button to invoke an action, and another button (x) to close it. The only
> difference being that the first has a shape of a speech bubble and is
> yellow and the other looks like an app window. What if we decide that
> all app windows should be yellow and look like speech bubbles? Is it
> going to magically turn them into notifications? Really, when there is
> no *functional and behavioural* difference, they are effectively the
> same thing.

This doesn't make a lot of sense. If they're the same thing then why
are you changing it? Obviously there is difference in function and
behavior or else you wouldn't be gitting rid of one in favor of the
other. Put another why, if they are no *functional and behavioral*
differences then why are you bothering to change it?

> The only issue regarding the U-M that has been raised and I agree with
> is the fact that the window is big and clunky and uses a lot of RAM. We
> will be investigating a possibility of replacing it with a simple,
> lightweight alert window with short text and 3 buttons: [install now],
> [details...] and [later] (exact wording TBC). I'm really struggling to
> see how this alert box appearing in the background would be more
> distracting and annoying than the yellow ugliness :)

That's really the only issue you think has been raised? And again, the
"yellow ugliness" bit. It's like not liking the color of the house so
you decide to remodel the whole thing.

> Again, I'd like to reiterate the main point: we have a good reason to
> believe that persistent indicators only work for some very specific
> cases (examples being network connection, volume, etc). We are now going
> through long and painful process of carefully defining these cases. It
> is early days, and there can be reconsiderations. So please be patient
> and forgiving :)

Again, I'd like to reiterate that the "trust us, we have our reasons"
is not going to be very convincing to many people. I keep getting this
sort of double-speak feeling when the same team is having to keep
pushing both "we know what were doing" and "we're just starting to
figure this out so bear with us". If you really do know what you're
doing, patches welcome. If you're still not sure yet, maybe you should
consider waiting until Karmic before making such huge changes.

I'd like to point out that I'm not against the notification work per
se. I like the new, prettier notifications I'm seeing, etc. I do
however think there's maybe too much assumption from the Dx team that
whatever they see as "better usability" *is* better usability. This is
why in the open source world we like to have heavy discussion about
the fundamental design, not just the implementation details.

-Jordan