Superpose3d

hardened against non-sensical user input. improved unit tests. added more silly badges.

Callers can supply an optional argument, q which will store the quaternion corresponding to rotation R. From it, the rotation angle and rotation axis can easily be determined.

It turns out that dividing the coordinates by the size of each point cloud (*Rgf* and *Rgm*) before superposition does *not* make the algorithm more numerically stable after all. So I removed the unnecessary and confusing code related to this bad idea.

RMSD is now calculated correctly when allow_rescale==True. (Equivalently, when c≠1. Note: The optimal rotation and translation are not effected. They continue to be calculated correctly. However the estimate of the RMSD was incorrect. Now RMSD is reported correctly.)

I made some minor changes to the "long_description" argument in *setup.py* in order to make the equations for RMSD more clear. (I've been experimenting with how to convey these equations. These changes do not effect the behavior of the "Superpose3D()" function. They just change the way the equations appear at pypi.org. Each time I change something, I have to upload the changes to github and pypi to see how they appear online. That's why the last few releases have been so trivial.)

Some embarrassing mistakes in the equations from README.md and setup.py were corrected.
These changes should not effect behavior of **Superpose3d()**.