Let me conclude by acknowledging John Tukey. The main theme of this talk sprung to mind after reading an old article that he wrote in 1965. It was the keynote that he gave at a conference, and the article is contained in the book of his collected works on graphics edited by Bill Cleveland.
Tukey argued, as I have argued, that the general purposes of analysis and experiment, match point-by-point.
Let me quote from his article: