Conway: Half a proof might be very long but not very hopeful.
Gries: Yes, and a whole proof might be wrong.