Am 2017-05-10 16:45, also sprach Eriban:
>> Bonus features > flyingRocksDev

> Interesting to see the falling rocks sequence evolve. Is it all generated via
> custom collision detection and physics modelling,

yes. Custom java emitting SDL.
http://git.buckosoft.com/gitweb/pov.cgi?p=TteoacService.git;a=blob;f=src/main/java/com/buckosoft/povspew/tteoac/Fishtank/FishtankRunner.java;hb=HEAD


> or did you re-use
> off-the-shelf software for that? The final sequence looks impressive.

Thanks.

> Two comments though. The balls jiggle a bit too much at the bottom of the
> aquarium to be realistic.

Yeah, this bothers me too.  I *do* like the effect, somewhat, though. 
It looks almost molecular.  Perhaps the laws of physics on Planet Bucko 
allow for infinite hardness. :)  The problem is, the energy of a top 
rock is transferred to a bottom rock, who then has to transfer it back 
because the tank bottom doesn't move.  The opposite problem is, I want 
the rocks to spread out on the bottom of the tank and if I absorb all of 
the energy, so they stop jittering, they end up in a slumped pile in the 
corner.

I already had to cheat this. In order to get the full tank coverage in 
the time I allotted, I had to speed up the clock. (I think the sped up 
clock contributes to the static looking initial falling column)

> Also, while falling the column of balls remains a bit
> too static I feel.

Yes, especially the initial part. Static is a great word for it. I do 
like the little plume when they first hit the tank bottom.

> When falling in air, the balls would speed up, so their
> distance should increase as they are going down. In water the effect may be less
> pronounced.

Terminal velocity in water is much less than in air. The rocks speed up 
in the air, but the "water" is at the top of the tank and the rocks 
don't spend much time in the air.

> However, due to the extra friction and turbulence of the water, I
> would then expect a bit more sideways movement.

I expected more sideways movement too. But, a sphere in water at any 
vector is going to rapidly decay to straight down. There's no hydrofoil 
to channel the water.

> Maybe you can tweak your
> algorithm a bit too fudge this? (Realistic modelling of turbulence and friction
> of the objects might be a bit too difficult...)

I'm thinking on it.  A rock is radius .25. I thought about filling the 
tank with radius .01 water (just for calcs, not to display) so I can 
track the currents better.  I think to do that, I'd have to switch from 
Java to c++ and I don't know if I want to write the framework just to 
support that.

>
>


-- 
dik

Post a reply to this message