In the fall of 1985 Jene Golovchenko from Bell Laboratories presented a staff colloquium at the National Bureau of Standards (now NIST) on scanning tunnelling microscopy. This was an exciting time in surface science, since the tunnelling microscope was yielding such fundamentally new insights that the Nobel Prize would soon be awarded for its invention and application. In his seminar Golovchenko made a key point that has stuck with me ever since. His argument went something like this: Investigations of surface processes had been largely empirical and had made little headway before the advent of scanning tunneling microscopy. The field simply could not progress beyond a certain point without a detailed knowledge at the scale of atomic dimensions. Tunnelling microscopy was unleashing a flood of information which was forming the basis of a new surface science.