@Epimetheus I have a couple of questions:
have you done any more molecules? I have been running Millsian 2.0 and using it to calculate energies of "difficult" molecules that are not included in its internal library (e.g. cyclopropane, N2, nitrogen trifluoride, sulfur hexafluoride...). I am comparing the atomization energies ("bond energies") and heats of formation (two quantities which the Millsian SW prints out for every molecule it can solve) to NIST experimental data. This is in order to make my own version of the plot Mills published (shown below).
After I double-check my answers, I will also compare the experimental values to more accurate traditional ab-initio results, since Mills's plot on the right side (below) assumes HF theory with the 6-31G* basis set, which is old-school and notoriously bad for "difficult" molecules.
Have you used Millsian? What are your thoughts.