Rereading it, I think it was mostly right, but partly wrong. The path to massive programming productivity isn't to rebuild everything from zero; it's to reuse only the right things.