On 13/01/14 10:22, romain.pontida at gmail.com wrote: > againstthe reference implementation, but how do you prove that > the reference implementation is correct? I thought the reference implementation was correct by definition.