The paper by Scott Rubin1 in the March 2019 (p. 150) issue of The Physics Teacher acknowledges problems with regard to the physical accuracy of its assumptions; here I hope to clarify the nature of these deficiencies more fully.
Firstly, since a rolling sphere would become non-spherical as it rolls and accumulates snow, it would seem better to consider a rolling log, as in earlier treatments.2,3 However, the difficulty remains that, in detail, neither a rolling log or sphere is strictly cylindrical or spherical as it accumulates snow. This nominally small issue turns out to have the consequence that one finds different equations of motion if one uses a force-and-torque analysis, as in the paper of Rubin, or an energy analysis. In effect, the problem is not well enough defined under this assumption for a unique analytic solution—different, apparently reasonable, approaches give different solutions to such variable mass...