why? cause sorting around the global driver pool would basically use more cpu power then just copy data sequences into the right order And sorting data references costs 0 performance at all