Conversation

@lain This one again. I still haven't read the paper, but my first thought was of a crank who used to post on sci.crypt a lot about how all functions in security applications need to be injective (indeed, bijective - he really had a thing about permutations). My second thoughts are in the directions of numeric precision and thermodynamics. Time-reversal symmetry of physics doesn't mean you can unscramble an egg.