I'm reading Lamport's new book "A Science of Concurrent Programs" and am amused that sometimes he refers to algorithms as being "invented" (e.g., Paxos) and at other times refers to them as being "discovered" (e.g. an N-process mutual exclusion algorithm).
