No sociologist should think himself too good, even in his old age, to make tens of thousands of quite trivial computations in his head and perhaps for months at a time.