[Pw_forum] Best strategy to calculate band gap

[Yury Vishnevskiy]

Dear QE developers and users,
I am calculating a metallic-like system with gaussian smearing and want 
to estimate the band gap. Now I've faced the problem that my highest 
formally occupied band is in fact not uniformly occupied. It is occupied 
in its minimums on my plot but in the maximum (which is the formal place 
of band gap) the occupation number is 0.0. What would be the most 
scientifically grounded way to calculate the band gap in my case?

Sincerely,
Yura Vishnevskiy
Uni-Bielefeld, Germany