[sugar] On improving Sugar performance

[Tomeu Vizoso]

On Mon, Apr 28, 2008 at 7:21 PM, Gary C Martin <gary at garycmartin.com> wrote:
> > Sure, but rather than a useful tool, I would call measuring as the
> > only possible base on which decide actual work that needs to be done.
> > We could be refactoring and recoding for years and don't get any
> > noticeable improvement, if we don't measure.
>
>  Not that I want to start a CS grad war here, but with more than one foot in
> the UI camp, I just wanted to request that more than just 'clock watching'
> is done when selecting/implementing optomisations – though 'clock watching'
> is a very good place to start.
>
>         http://www.codinghorror.com/blog/archives/001058.html
>
>  Catering for subjective human response is tricky, but once you've passed
> some minimum threshold for utility** you can often win more hearts and minds
> going for the subjective. A concrete example I'd point to is smooth
> animation; activity switching with no nasty redraw flicker; pulsing icons
> with no visible strobe; and frame/notification transitions that glide
> smoothly onto the display giving the illusion of effortlessness.

I agree with this. We need to choose which code paths need to be
optimized based on (subjective) user reports. But once we know there's
a point where we are not giving enough feedback or where things just
should be done faster, we should measure in order to set goals and
choose the best approach.

I don't think we have enough resources to rewrite things and find out
later that we haven't improved user experience at all.

>  **how many hours must I have spent as a kid, excitedly waiting for some
> software to load up off a magnetic tape, only to have it fail a fair portion
> of the time and have to start the tape over again... and some folks here
> moan about a ~6sec launch time for an activity. Though it's true to argue
> that those old machines did have instant on, even if you were left to ponder
> a BASIC prompt and spend the next 5min trying to tune your TV in to get a
> clear signal :-)

;)

Thanks,

Tomeu