| Popis: |
If the confinement of a trial wavefunction on a subspace of the full Hilbert space of a systems is taken as a “constraint” that prevents an approximate method to be labelled as a “fully quantum” one, then it can be shown that practically all known molecular approximate method are not rigorously quantal ones. Another weaker (and already known) criterion used by nuclear physicists for such a labelling is discussed and illustrated for the generator coordinate method (in this case a quantum approach) and Hartree—Fock method (a semiclassical approach. On the other hand, although Lathouwers' treatment of molecular spectra may be labelled “fully quantum” (for certain molecules at least) under this softer criterion, it will be remarked that the loss of the variational upper-bound property attenuates much of its theoretical attractiveness. |
