I have always attached great importance to the manner in which an experiment is set up and conducted … the experiment should be set up to open as many windows as possible on the unforeseen.