Call for Papers/Talks/Ideas!

[Marco Pesenti Gritti]

On Sat, Mar 22, 2008 at 8:32 AM, John R. Hogerhuis <jhoger at pobox.com> wrote:
>  Based on observing my daughter, it's not much of an issue with
>  a USB mouse. But that's not the usual mode she works with the
>  laptop. She pulls it off her shelf and sits on the couch with
>  it on her lap. A mouse isn't an easy option.
>
>  If anyone asks and thinks they won't be redundant I will enter
>  a couple narrowly focused tickets. (popover frame + trackpad,
>  and "reading required.") I know the slow-activity-start
>  bug (5228) is already being tracked, though I am surprised it
>  wasn't considered high enough priority to be triaged for
>  update.1.


Hello,

first of all, this is very useful and constructive feedback, thanks!

I have not seen a good analysis of the frame problem yet. For example
I know that at some point the trackpad jumped to the corners very
often, and that obviously aggravated it. Also in my experience the
thing is nowhere so annoying on my work laptop (still using a
trackpad), why? Finally, when this happen, is the mouse left still in
the corner or do kids just touch the corner momentarily...

Eben, how do we go about solving this one? I think it should be a
priority for Update.2.
Some possibilities:

1 You and Christian do some simple usability tests, possibly with
kids. It should be trivial to figure out how they are triggering the
problem. Once we know that, it will be much easier to figure out
possible solutions or decide if we have to turn off corner activation
completely.
2 We try out a few possibilities (delay on activation, completely turn
off etc) as control panel preferences and we encourage people to try
it out and to provide feedback.

My feeling is that we should start with 1 and then do 2.


>  If I had to pick a single most important high profile problem
>  with the XO, the slow-activity-start bug would be it. It's
>  impossible not to stub your toe on this one within the first
>  15 minutes of use.


Tomeu managed to reduce startup up time to 2-3 seconds in the faster
branch using an approach similar to maemo launcher:

https://stage.maemo.org/svn/maemo/projects/haf/trunk/maemo-launcher/README

If we don't find any blocking problems with this approach, I think we
can make startup pretty much instantaneous.

>  As a programmer, the problem calls into
>  question the basic software architecture of the XO, in
>  particular the choice of heavy use of Python and
>  "abstraction layer"-type packages like Telepathy which
>  weren't purpose built for resource constrained machines like the
>  XO.

Python was chosen mainly because it's a good development tool for
kids. I think the few performance problem it's causing are all
solvable, it's just that no one had time to focus on it until now.

Marco