Comments on ‘Formulating the tightest bounds on causal effects in studies with unmeasured confounders’ by M. Kuroki and Z. Cai, Statistics in Medicine 2008; 27:6597–6611