In a recent publication, Haestier [J. Appl. Cryst. (2009), 42, 798-809] has proposed a method to take care of the unit-cell-parameter uncertainties in the calculation of geometric quantities such as interatomic distances and bond angles by modifying the uncertainties of the atomic coordinates. This problem is addressed here with a different approach, which gives additional insight. For the cell-edge uncertainties, Haestier's results are confirmed and their importance is easily appreciated. However, for the cell-angle uncertainties, it is proved that there exists no simple solution short of calculating the derivatives of the quantity of interest with respect to the angles. Simple rules of thumb are presented for guessing the importance of edge-length uncertainties.