I heard that as the value of "num_iter(tag in wannier 90)" is higher, spread of Wannier function(=WF) is gradually lower in wannier 90.
If so, is this procedure that minimize the spread of WF meaningful physically??
If I use a lot of num_iter for any projected orbital, the spread will be minimized.
In this case, when different people draw the maximally minimized wannier functions for the same material, is there a standard by which they can be compared?
Or is there a standard value for minimizing the spread of WF?
I would like to know the criteria for WF being minimized.
Thank you.