The exponent for the radians in the unit for e.g. torque is arbitrary, but to me the choice made in this template seems unintuitive. The template now has no radians in the unit for rotational energy and have the exponents for the other quantities based on that. To me torque seems like a more basic concept than energy, so having no radians in its unit would make more sense. Torque is normally defined based on only a force and a lever arm length and so would not need radians in the unit. Rotational energy would then be torque times angle and have radians in the unit. Is there a deeper reason for the present choice? Ulflund (talk) 06:39, 10 November 2014 (UTC)

The radians make sense for the units of dimension Tx quantities (e.g. angular velocity), but not of dimension ML2Tx quantities (e.g. angular momentum). Note that the corresponding articles mention no radians in these units. At maximum we could write N m = J/rad. Petr Matas 16:25, 13 February 2016 (UTC)