I assume that's just adding up file sizes within the package? not counting cluster size overheads?

I work for the Department of Redundancy Department