“A mathematical proof should resemble a simple and clear-cut constellation, not a scattered cluster in the Milky Way.”