[ODE] Maximum Contact Point Generation

[Donny Viszneki]

I had taken a number I knew would be much higher than possible to 
achieve in my simple environment (which I discussed in an earlier email 
"Help with LCP internal error.")

The number was 32. I noticed that I never get more than 4 contact 
points, though I knew that technically 6 is a more reasonable maximum 
value.

Changing 32 to 6 seems to have changed my simulation dramatically.

Note: if you read my last email, I should point out that I changed my 
cubes' sizes to 3.0 from 1.0.

Now my cubes will sometimes fly around like crazy when using 6 as a 
maximum contact point generation. Using 32 it seems to be more 
realistic (though with a bit of jittering, and of course, the 
vanishing.)

This started to get me thinking - even though 32 and 6 are still both 
above the number of contact points I'm generating, it's strange that it 
has such a dramatic effect on my simulation.

Can anyone help flesh out this quirk for me? Perhaps altering these 
values could help solve my vanishing problems?

Again, I'm working on a deadline, so please let me know ASAP if you 
have any help for me!

Thanks in advance.