Certainly necessary, but very far from sufficient. But even that isn't quite accurate. Every one of my published papers has some kind of error in it. When somet

Certainly necessary, but very far from sufficient. But even that isn't quite accurate. Every one of my published papers has some kind of error in it. When something is sufficiently interesting and complex, this is essentially inevitable given how the human mind works. Hence the late Voevodksy's attempt to come up with a framework for making such ideas machine checkable.