How to efficiently transform array of vectors, or why the homogeneous coordinates were not chosen to represent affine/projective transformations?

[babaq]

Hi,

I often need to affine transform a set of vertices, and have used the homogeneous coordinates to simply represent any composition of affine/projective transformation in a single matrix, and any transform is a matrix multiplications.

when i look at the package, it seems that i have to transform vertices one by one. and the affine implementation was a linearmap plus a translation.

Is there any suggestions to how use this package to efficiently transform a set of vectors, or explain the merits of current approach vs. homogeneous coordinates approach that have been widely used in computer graphics, such as OpenGL and Direct3D.

Thanks,
Alex