This is a computer-science tradition, I believe. When I was an undergrad at Rice in the 80s I saw an abstract posted for a talk by a professor Löof Lirpa showing P = NP, and mentioning, among other consequences, that henceforth all programmers would be out of work.
In (vaguely) related news, there is an email posted on Erik's office door from a businessman who is convinced he is NP-complete (because he is very good at Minesweeper), and wants to know how to use this skill to his advantage -- "What should I do if I am NP-complete?" Not an April fool.