Written: Sep. 29, 1995.
Black vs. Comely; Formal vs. Rigorous; Applied vs. Creative; Computational vs. Conceptual. Fair vs. Strong. These are only 5 of the infinite set of biases, past, present, and future, in which humans erroneously take pairs of sets to be disjoint, while they really have a large overlap.
In a very eloquent letter to the editor, on the Oct. 1995 Notices, p. 1118, Andrew Odlyzko pointed out Hugo Rossi's tacit erroneous assumption that multi-disciplinary research demands less creativity than pure research. But even Odlyzko did not go all the way. He said: `There are no applied results that compare to Fermat's last theorem in both the sophistication of the tools used in the proof and the definiteness of the results'. This is perhaps true if you take `sophistication' and `definiteness' in the very narrow sense. But in the deeper sense of creative novelty, Coding theory, for example, that was started out my engineers Hamming and Golay, is far more novel and requires much greater creativity than Wiles's amazing but technical tour-de-force.