[sugar] perceived sugar performance

[Eben Eliason]

On Tue, Apr 29, 2008 at 1:58 PM, Marco Pesenti Gritti
<mpgritti at gmail.com> wrote:
>
> On Tue, Apr 29, 2008 at 7:53 PM, Tomeu Vizoso <tomeu at tomeuvizoso.net> wrote:
>  > On Tue, Apr 29, 2008 at 7:44 PM, Michael Stone <michael at laptop.org> wrote:
>  >  > On Tue, Apr 29, 2008 at 07:26:12PM +0200, Tomeu Vizoso wrote:
>  >  >  > We cannot presume that _all_ activities will be able to put a window
>  >  >  > in 0.1-0.5s,
>  >  >
>  >  >  I think we are better served by presuming that activities which fail to
>  >  >  start quickly are broken need to be fixed. For goodness' sake, we have a
>  >  >  processor clocked at over 400 MHz that can play full-screen video. Does
>  >  >  someone here actually wish to argue that it's acceptable for process
>  >  >  creation, X connection setup, window creation, and painting to take long
>  >  >  enough to require secondary feedback mechanisms?
>  >
>  >  In a perfect world, you would be right. But that doesn't seem to be
>  >  the world we are living in, because so many apps seem to need a banner
>  >  while they launch (openoffice, gimp, banshee, etc.).
>  >
>  >  I'm not 100% sure that we need such a strong feedback during
>  >  launching, but just saying that we'll make everything fast enough and
>  >  slow activities won't bite us is a bit courageous, at least.
>
>  We should ban banners, really :)
>
>  The two points that makes our feedback different from a banner are:
>
>  * It reinforces the zoom metaphor.
>  * It deals with the problem of children clicking on 2-3 activities at
>  the same time, which proved to be a real issue in the field (will
>  faster activities address this? not sure).
>
>  I'm still worried that the feedback might be too strong and
>  unfortunately it's something hard to test until we have activity
>  starting in 1.5 for real. If we had an acceptable form of feedback in
>  the current builds I'd propose to first make activities faster and
>  then play with feedback. Unfortunately, after the redesign, I don't
>  think that's the case.

For what it's worth, a number of activities in the latest joyride
appear to launch in a second or two at most when run in emulation.
I've been testing them during the creation of the launcher feedback
patch, and I still think the behavior is nice, myself.  As mentioned,
it would be even nicer if the activities painted in against white
instead of gray, but overall the speed of the launch doesn't seem to
detract from the feedback thus far.

I encourage others to try out what I have working so far.

- Eben