just one comment: quaternions are more accurate than matrix
multiplications, and it is a common practice to use quaternions to
apply multiple rotations - so i would prefer if the internal
implementation of Rotation3D would stay as it is
On May 4, 2009, at 5:30 PM, Daniel Russel wrote:
I think we currently have too many way of representing matrices in
the IMP API.
- Rotation3D
- EulerAnglesZYZ
- and EulerMatrix
The last doesn't make any sense as there is no reason to tie
together two disparate representations. Given that one can create a
Rotation3D from Euler angles I don't see any reason to expose the
second. If we find rotational matrices are faster than quaternions
for actually performing the rotations, then we should cache the
matrix in Rotation3D rather than add another class.