goal of this group:
try out some of the techniques with a very small system (h2 molecule)

1: optimize the wavefunction
2: diagonalize the wavefunction (including the virtual state)
   and write out the orbitals, densities
3: calculate some properties:
    localized orbitals
    wannier centers
    dipole moment
    project wavefunction to atomic pseudowavefunction
    population analysis
4: compare different methods to optimize the wavefunction.
5: run geometry optimization (from scratch).

part 2 is MD:

the goal of this exercise is to demonstrate CP-MD
and the proper selection of the timestep and emass.

6: the 5au run diverges.
4au and 3au yield practically identical trajectories
and energies (if corrected for the timestep difference).

the use of CP-MD with fixed atom positions and electron
annealing for wfopt is revisited to demonstrate the CP-method
(to be compared with the Remler-Madden paper and others).

7. uses annealing for both geo- and wf-opt.

8. stabilizes the 5au timestep by raising
the fictitious electron mass (with much increased dragging effect)

9. runs a BO-MD off the same restart. one can see, that the
timestep is much larger (20au) yet the movement of the ions
still conserves the energy (reasonably). note the increased
convergence parameter.
when plotting EKS vs. the 4au/3au EKS one can get an empirical
scaling factor, that shows the dragging (about 25% due to the 
small weight of the hydrogen), yet the sampling of the potential 
energy surface is the same.