The goal of this paper is to provide a critical review of the well-known resolved-acceleration technique for the tracking control problem of robot manipulators in the task space. Various control schemes are surveyed and classified according to the type of end-effector orientation error; namely, those based on Euler angles feedback, quaternion feedback, and angle/axis feedback. In addition to the assessed schemes in the literature, an alternative Euler angles feedback scheme is proposed which shows an advantage in terms of avoidance of representation singularities. An insight into the features of each scheme is given, with special concern to the stability properties of those schemes leading to nonlinear closed-loop dynamic equations. A comparison is carried out in terms of computational burden. Experiments on an industrial robot with open control architecture have been carried out, and the tracking performance of the resolved-acceleration control schemes in a case study involving the occurrence of a representation singularity is evaluated. The pros and cons of each scheme are evidenced in a final discussion focused on practical implementation issues.