## Engineering Design

Creative, innovative, patentable, engineering design requires more genius than ongoing adaptation to environmental changes.

Engineers skilled in design should be the ones to comment on the design seen in nature. One professional recognizes the work of another. I have known a number of very innovative engineers. One friend of mine had more than 100 patents.

This innovative engineer was contemptuous of “automatic design.” A programmer told me what it was like to try to show my friend contemporary computing methods (those of the mid 1960’s). The engineer wanted the roots of a fifth-order equation. The computer department assigned the programmer to find them by Newton’s method, a widely used homing method. In that method there is an initial guess at the solution. The engineer probably supplied the initial guess. Then the method refines the guess using the slope of the equation at the guess point to find a closer approximation. The program repeats the process until the root is found within some small error tolerance. The programmer wrote a FORTRAN program to apply Newton’s method. He had the program steps punched on data-processing cards. As soon as the programmer loaded the cards in the card reader, the engineer got out paper, pencil, and slide rule and went to work. Before the program was compiled and ready to run, the engineer had the root and was twitting the programmer about it. “How is it possible that I am faster than the machine?” he asked.

This made me realize that smarter mathematicians can sometimes replace computers. Humans make the intuitive leap, the creative conjecture. Has anyone ever succeeded in programming “common sense” so a machine can apply it to solving a problem?

Engineers skilled in design should be the ones to comment on the design seen in nature. One professional recognizes the work of another. I have known a number of very innovative engineers. One friend of mine had more than 100 patents.

This innovative engineer was contemptuous of “automatic design.” A programmer told me what it was like to try to show my friend contemporary computing methods (those of the mid 1960’s). The engineer wanted the roots of a fifth-order equation. The computer department assigned the programmer to find them by Newton’s method, a widely used homing method. In that method there is an initial guess at the solution. The engineer probably supplied the initial guess. Then the method refines the guess using the slope of the equation at the guess point to find a closer approximation. The program repeats the process until the root is found within some small error tolerance. The programmer wrote a FORTRAN program to apply Newton’s method. He had the program steps punched on data-processing cards. As soon as the programmer loaded the cards in the card reader, the engineer got out paper, pencil, and slide rule and went to work. Before the program was compiled and ready to run, the engineer had the root and was twitting the programmer about it. “How is it possible that I am faster than the machine?” he asked.

This made me realize that smarter mathematicians can sometimes replace computers. Humans make the intuitive leap, the creative conjecture. Has anyone ever succeeded in programming “common sense” so a machine can apply it to solving a problem?