Speaking of diagrams, I once read a short (probably no more than 3-4 pages) paper with one theorem and one diagram. The diagram was essential for me to understa

Speaking of diagrams, I once read a short (probably no more than 3-4 pages) paper with one theorem and one diagram. The diagram was essential for me to understand the proof. The problem was the diagram was a diagram of the proof.