Test-wise addition for / about nitrogen.  Tested on the example of
pyridine by openbabel from the SMILES string (`c1ccncc1`), and
diethylamine (`CCNCC`).

Tested for the examples of diethylether (`CCOCC`), THF (`C1COCC1`),
and propan-2-ol (`C(C)(O)(C)`) in .xyz files generated by obabel.

Tested with openbabel generated .xyz about diethyl thioether
(`CCSCC`) and furfural (`O=Cc1cccs1`).  However, contrasting to
e.g., the depiction with openbabel generated .pov, the C-S bonds
no longer are visible.

Tested with openbabel SMILES string of fluoroethane (`CCF`).  The
.pov/.ini and subsequent generation of .png seem fine.

Introduction of chlorine.  Tested with an .xyz about chloroethane
(`CCCl`) generated by openbabel, the chlorine atom is depicted as
sphere, separated from the backbone of the molecule.  Because of
a similar observation for compounds with C-S bonds, and because the
analogue CCF was rendered well, perhaps the assignment where to
draw a bond faces an obstacle here.

The previous approach requires four lines each for every atom type
as an instance of class Atom.  I think an approach which uses the
element symbol as key to a the corresponding values of mass,
radiosity, and rgb coordinates in a dictionary is - in comparison
to this one - easier to read, and to maintain.  Lacking any better
inspiration, atoms with atom types not yet explicitly defined in
said dictionary default to hydrogen.  In such an instance, the
conversion will not stop for good, however is going to issue one
warning for each atom not fully defined (yet).