John Hasler wrote:
> The Natural Philosopher writes:
>> So if it changes n anyway, you re bout 99% sure that the next time you
>> run MD5sum against it, you will get a different answer.
>
> More like 99.99999999999999999999999% sure.
I wont quibble, but the wiki article suggests that there is a crack to
allow corrupted fies to be padded out with the right bit pattern to pass
a test..