Starting discussion on a new version of the notification spec

Brian J. Tarricone bjt23 at
Mon Jun 15 11:47:53 PDT 2009

William Jon McCann wrote:
> Hey,
> On Mon, Jun 15, 2009 at 1:45 PM, Brian J. Tarricone<bjt23 at> wrote:
>> On 06/15/2009 08:10 AM, William Jon McCann wrote:
>>> Do you have a specific response to the problems that they describe at
>>> the following?
>>> I'd be very interested to see it.  I think that rationale is fairly
>>> compelling actually.
> ...
>> There's only one valid criticism of notifications in this section, and
>> it is valid regardless of whether or not you allow actions: the fact
>> that notifications block user interaction with stuff under the
>> notification.  Canonical has chosen to solve that problem in notify-osd
>> in a way that's incompatible with allowing actions.  Again, this has
>> nothing to do with whether or not actions are conceptually bad; it's
>> just a design tradeoff they've made.
> So, you don't actually have a response to the problem then.

I design interfaces as an amateur.  I'm not a professional/trained 
UI/HCI guy.  Just because I don't have a solution doesn't mean there 
isn't one.

>  I don't
> think anyone is saying that actions are bad - what does that even mean
> anyway?  However, the presence of actions implies a set of behaviors
> and expectations that I think Matthew has rightly identified as
> problematic.

Ok, so solve the problems.  Don't just throw away useful behavior.  (And 
what do you mean by "set of behaviors and expectations" [plural]?  I 
only see one behavior that's an issue.)

Personally, I never considered the issue of "stuff behind the 
notification" to be a big deal.  They're pretty small and unobtrusive, 
and disappear after a short time.  In the case of my implementation, you 
can drag the notification out of the way or simply click the close box 
(or anywhere in the notification) to dismiss it.  Yes, that's "more 
work" than just clicking "through" the notification, but I don't see it 
as a huge burden.

But hey, one man's "no big deal" is another's showstopper.  It's just a 
shame that a prominent player in the community seems to have an opinion 
very much opposite of that of the other spec implementers.

I'd even consider implementing click-through for notifications that 
don't have actions, but, as I said before, that creates two very 
different interaction models, which might be confusing.  Maybe making 
action notifications visually distinct in other ways could alleviate 
that and still make sense.


More information about the xdg mailing list