[kde-edu]: Step and Summer of Code ?

Vladimir Kuznetsov ks.vladimir at gmail.com
Thu Mar 15 16:44:36 CET 2007


Hello Carsten !

On 3/15/07, Carsten Niehaus <cniehaus at gmx.de> wrote:
> Am Donnerstag, 15. März 2007 14:22 schrieb Vladimir Kuznetsov:
> > Hello All !
> >
> > I'm thinking about applying Step as a project for Summer of Code
>
> <snip>
>
> > StepCore (simulation):
> > - collision detection with friction (I hope this will be ready before
> > summer) - movement restrictions (joints) (it can be ready too)
> > - global error estimation
> > - molecular dynamics (2d gas)
> > - soft bodies and massive springs
>
>
> Mmh, I think it is easier for everybody to decide if you would add your
> suggestion here:
>
> http://techbase.kde.org/Projects/Summer_of_Code/2007/Ideas#KDE_Edu
>
> Of course, that page is for mentors suggesting SoCs not for students :-) But
> still I find that structure+layout much easier to read...
Yes, I've added it:
http://techbase.kde.org/Projects/Summer_of_Code/2007/Ideas#Step_interactive_physical_simulator

>
> The only issue I have is that we had a very similar project last year which
> didn't succeed.
I've read about that project. It is really very similar, but not
exactly the same - that one was about scripting, Step is about
interactive editing (by the way Harish mentioned that he is willing to
add scripting to Step when he has free time and I'll try to do my best
to help him).

I hope that my project will not fail. I have some experience in
physical simulation (actually it is my speciality at University, my
current academic work is related to gas simulation), I have experience
writing time estimation and than working according to it (at my job
which is about system programming for linux on arm and powerpc), and
I've already worked on Step for two month (so it is easier for me to
estimate required time). If I try to apply I'll write detailed time
estimation for my project (possibly adding/removing some items from
currently proposed) and include it in my proposal, and then try to
work according to it.

> Of course, Step is already much more mature and already
> works. Its GUI sucks and all that
I now it, I just trying to keep simulation functionality and GUI
advancing simultaneously. You can see that GUI has changed quite a lot
since the time I've first posted here. I've added GUI polishing as
another item on TODO.

> but I already really like Step and will
> probably show it around in my school soon.
Great news for me ! Looking forward to hear critical feedback ;-)

> Personally, I like "molecular dynamics (2d gas)" most and also suggest (well,
> that is basically the same!) Brownian motion.
I've added it to the list. Actually molecular dynamics will allow to
simulate it directly, but it also can be simulated separately for
better performance (but I need more time to think about it).

I've also thought before about simulation of the molecules. It is not
hard to implement with Step: add required effective potentials (Step
is quite flexible here, you can easily add tensor potentials, etc.)
and particles with desired characteristics. But I'm not sure how
usable it will be in 2d.

May be it can be a topic for another project: extend required subset
of StepCore to 3d (which is quite easy since required subset will not
include rigid bodies and collisions), simulate molecules with it and
visualize using some other tool (new molecule viewer from Kalzium ?)
providing control over simple parameters ... Just a raw sketch.

> Good screenies:
>
> http://upload.wikimedia.org/wikipedia/en/a/a7/Brownian_hierarchical.png
> http://upload.wikimedia.org/wikipedia/en/f/f8/Wiener_process_3d.png
>
> As a chemistry teacher both (2d and 3d, 2d more important) would be very nice
> to have!!
3d is quite more complex both for GUI and for simulation (I mean rigid
bodies and collisions), so I'm not going to start working on it until
2d will be actually usable.

Thanks a lot for your reply !

-- 
       Best Regards,
          Vladimir


More information about the kde-edu mailing list