[ODE] Particle/cloth/skeleton

[Adam D. Moss]

david@csworkbench.com wrote:
 > SIMD: I wasn't considering rewriting all of the code with SIMD
 > alternatives, but just rewriting Vector and Matrix multiplication and dot
 > product functions with a SIMD alternative.

Sounds good -- my concern was that this would end up in
some relatively volatile code, but the vector and matrix maths
is an ideal candidate.

 > Verlet:
 > You really could constrain more degrees of freedom with just the distance
 > and angle constraints if you wanted to. For example, to make a hinge, you
 > simply attach a given structure at 2 particles (see link page 4 for
 > graphics -- if they had used a 2 particle wide leg, the orientation
 > calculation becomes a cross product).

I understand this, those are very simple examples.  The complication to
me is setting joint-stops independantly on various degrees of freedom
simultaneously with dot-product and length constraints on a skeletal
system, e.g. allowing lateral movement while disallowing twist.

The solution I ended up with and which is looking good (though I'm
still trying it out for size) keeps some 'virtual' orientations on
the particles  themselves to which neighbours' are constrained,
since distances and inter-bone dot-products alone don't seem
expressive enough, at least for me, to manually construct
convincingly articulated skeletons with good joint-stops, and more
importantly to have a good notion of bone orientation without turning
the skeleton into a network of triangles and probably thus introducing
unwanted constraints (I stress how important it is to know a bone
and/or particle's orientation if you're rendering a weighted mesh
around a skeleton, though this is irrelevant information for cloth
or jelly).

I'm still experimenting myself after a few weeks.  The whole
sticks'n'particles iterative thing is a fun and simple technique.
I'm trying out various ways to gain expressiveness of simulation
(flexible joints, intuitive and useful joint-stops) without losing
very much simplicitly, but I'm not sure I've struck gold yet...

Regards,
--Adam
-- 
Adam D. Moss   . ,,^^   adam@gimp.org   http://www.foxbox.org/   co:3
... but his bosses didn't like him so they shot him into space.