I'm reading papers about tree diff techniques, to try to improve my structural diff tool. Most papers focus on diffing XML. Their techniques are clever and relevant, but crikey some of the intro quotes really haven't aged well.