We need to follow the procedures used by our colleagues in physics, chemistry, biology, etc. in order to buildup a theory of mathematics education…. We need to start with extensive, careful, empirical observations of mathematics teaching and mathematics learning. Any regularities noted in these observations will lead to the formulation of hypotheses. These hypotheses can then be checked against further observations, and refined and sharpened, and so on. To slight either the empirical observations or the theory building would be folly. They must be intertwined at all times. (p. 242)