What Can Robotics Research Learn from Computer Vision Research?

The fields of computer vision and robotics are both children of the artificial intelligence program that was spawned by the Dartmouth Conference in 1956. In recent decades the fields have diverged in terms of conferences and journals, research methodology and research rate. From a robotics perspective it seems that computer vision is in the fast lane while robotics is stuck in the slow lane. Roboticists hold a fundamental belief in the importance of experimentation but could it be that experiments are actually holding us back? Or is it that we are doing experiments poorly?.