I think a lot of great ideas like tree-width were developed in hope of simplifying the proof.