Javadoc
CTfile specification is ambiguous as to how parity values should be written
for implicit hydrogens. Old applications (Symyx Draw) seem to push any
hydrogen to (implied) the last position but newer applications
(Accelrys/BioVia Draw) only do so for implicit hydrogens (makes more sense).
To avoid the ambiguity for those who read 0D stereo (bad anyways) we
actually do push all hydrogens atoms to the back of the atom list giving
them highest value (4) when writing parity values.