ghorn / classy-dvda Goto Github PK

Right now, a force acting on a rigid body is silently assumed to be acting on the body's centre of mass.

A lazy user might want to specify the force to be acting on another point, without doing the maths of splitting in an equivalent force+torque pair.

nice package, btw!

The following code seems to expose a bug in simplification logic. The output contains + --b'(t)

import Classy
import VectorMath

n = newtonianFrame

a = coord "a"
b = coord "b"
f1 = rotX n a "f1"
f2 = rotY f1 b "f2"

x = coord "x"
y = coord "y"
z = coord "z"

p1 = RelativePoint N0 (xyzVec (x,y,z) f1)

run = do
  print $ ddtF (subtractPoints p1 N0) f2

Some basis dot products are left symbolic, like Ax> . Cy> when the relationship is not explicit.

Since the frame hierarchy should be know, this relationship should be computable from #2

Also see http://hackage.haskell.org/package/lca-0.2.4

The following code does not work as expected:

import Classy
let n = newtonianFrame
(xyzVec (1,0,0) n) - (xyzVec (1,2,3) n )

Works totally fine for something else then the exact zero vector...

I always tend to think of frames as things that are related to each other by means of both a rotation and a translation.
The current notion of Frame does not seem to allow for translation.

Translation only gets into play with RigidBody, where I think it should already be present in Frame

Supposing the dependencies of all frames have a tree like structure, we could do with a function that gives the latest common ancestor of given two frames. A first implementation could just lie and return the NewtonianFrame.