[Wien] Characters of atoms in the fold2bloch bands.

[Oleg Rubel]

Thank you very much, Peter, for pointing the way!

case.normsoup/dn is exactly what we need. I just looked up the norms for 
the valence band edge of GaAs (x4 degenerate states at Gamma). Here they 
are from case.normsoup

0.907085541766E+00  0.433544748129E+00  0.499707182587E+00 
0.159662527518E+00

and case.normsodn

0.929144582341E-01  0.566455251871E+00  0.500292817413E+00 
0.840337472482E+00

Pairs add up to 1 as expected for a norm. Each spin channel adds up to 2 
indicating no dominant up/dn component (something I would expect for 
GaAs). But individual norms can be very different from 0.5, which needs 
an action.

> Eventually one could simply add the spectral weights, but probably one 
> should weight them by the spin-up/dn norms in the case.normsoup/dn 
> files. In that way one would get a proper representation (and eg. a 
> small spectral weight in spin-dn does not matter if also its norm is 
> small, i.e. if this eigenvalue is mainly spin-up.

This can certainly be done, but it would be convenient if fold2Bloch can 
do it "under the hood". Files case.vectorsoup, case.vectorsodn, 
case.normsoup can be read simultaneously producing a single output file.

Best regards
Oleg