A note on "Loophole-free Bell test for continuous variables via wave and particle correlations (Phys. Rev. Lett. 105, 170404 (2010))"

The inequalities proposed in [Phys. Rev. Lett. 105, 170404 (2010)] are not Bell's inequalities in the usual sense: there are local classical models that violate these inequalities. Thus, their violation demonstrates nonlocality only under assumptions about the physical implementation and moreover they cannot be used as a device-independent test of nonlocality.