Post

(with apologies to Philip Greenspun) any sufficiently complicated description of an implementation by a Turing-machinist will eventually lapse into a half-coherent rediscovery of lambda-calculus