[Wien] Total Energy

[Chiung-Yuan Lin]

Dear all,
  I  would like to ask a basic question that doesn't seem to have been
discussed before.
  In many calculation procedures like mini, optimize, or
spin-polarized calculation,
we are comparing the total energy :ENE for different configurations,
i.e. internal
structure for mini, lattice constants for optimize, or FM-vs-AFM for
runsp. To my
impression, the total energy (and also other energies) is in reference from the
average coulomb potential in the interstitial region. For surface
(slab) calculations,
it makes more sense to me that one should look at the total energy relative to
the vacuum rather than the interstitial average coulomb. In the UG and among
many mailing-list discussions, it always states that one should track
the evaluation
of the total energy relative to the interstitial average coulomb (the
:ENE from case.scf) rather than further subtracting from the vacuum
level.
  Is there anything wrong in this "popular" statement? Or it's just my
misunderstanding.
I hope it can be cleared out here.

Thanks much,
Chiung-Yuan