update on active screen locker

[Thomas Pfeiffer]

On Thursday 16 February 2012 12:37:04 Aaron J. Seigo wrote:
> > - After the timer reaches zero, the slider goes back to the start and the
> > text reads "Sleeping in 5 seconds" again. Instead, it should not go back
> > to
> > the start until after resuming and should just read "Sleep" after the
> > timer
> > goes to zero.
> 
> the text should indeed read just "Sleep". that's a straight-out bug, in
> fact.
> 
> the slider position is more difficult and something that i'm not sure how to
> fix properly. the problem is that we don't have any notification for when
> sleeping is starting or we have come out of sleep. (kernel / power
> management issue; so deep in the stack.) so the UI just resets itself to
> what you will see when the device unsleeps eventually.
> 
> i'll fix the text up, not sure how i can deal with the slider position.

The slider position isn's such a big usability problem, I guess. Maybe you can 
just delay the resetting a bit in hopes that the screen goes blank before it 
happens? Or would we risk delaying it until after resuming then?

> > - If I tap the slider while it's moving, it returns to the
> > starting position instead of remaining where I tapped it. This feels
> > strange when I want to "accelerate" the countdown.
> 
> hm.. indeed. ok, this will require a little more logic in the QML, but it is
> doable.

That would be great! It might make sense to postpone any work on that until we 
have decided on the widget and mechanics to be used there, though.

> > Now we only need to remember to integrate a reboot function in our update
> > manager so updates actually take effect (otherwise they would never do for
> > users who never shut their device down).
> 
> fortunately very few updates require a full power cycle, but yes, this is a
> good point indeed and should be done ...

Sure, but many updates would at least require a re-login, or restarting of 
applications, to take effect, right? Ideally, of course, the updater knows 
when only a re-login is necessary and when a reboot is necessary and trigger 
the appropriate action.

> > I just conducted some "extensive" user testing (n=1 ;) ). The result was
> > that the participant understood what happened and was able to shut the
> > device down using the slider. He said he got the clue when the slider was
> > moving automatically. However, as expected, he was confused when the text
> > read "Sleeping in 5 seconds" after resuming. But that's a bug that should
> > be easy to fix, right? ;)
> 
> yes :) thanks for testing. glad it basically works ...

I found these micro-usability-tests (often called "Hausfrauentest" in Germany) 
to be really helpful, especially on a mobile device: You just hand it to 
somebody, give them a simple task and see if they can do it. Takes less than a 
minute and even though generalizeability is limited with just one person, you 
can identify things that don't work for _anybody_. 
I will do them regularly for things about which we are unsure from now on.
 
> > And by the way: Can we suppress the notification "The screen is being
> > locked"?
> 
> yes; and easy to do as well .. should probably not be there on the desktop
> either.

True. Actually, I never found them useful on the desktop either, they're just 
less annoying there.

Great to see my suggestions welcomed! :)