In passing, I firmly believe that research should be offset by a certain amount of teaching, if only as a change from the agony of research. The trouble, however, I freely admit, is that in practice you get either no teaching, or else far too much. "The Mathematician's Art of Work" in Béla Bollobás (ed.) Littlewood's Miscellany, Cambridge: Cambridge University Press, 1986.
