A good mathematical proof is like a poem, not a telephone directory.