Huh, I've always understood that quote very differently, with emphasis on "premature" ... not as in, "don't optimize" but more as in "don't optimize before you've understood the problem" ... or, as a CS professor of mine said "Make it work first, THEN make it work fast" ...