I have never been particularly comfortable with the idea that you can simply assume that if the hashes are the same, then the data must be the same. The fact t

I have never been particularly comfortable with the idea that you can simply assume that if the hashes are the same, then the data must be the same. The fact that collisions happen so rarely (and sometimes only after the a hash function has become compromised) makes this even more dangerous. It's like a couple of decades of strong hash functions has made people forget what hashing actually is.