3

Example: Is it possible to recreate a file using only its hash

Here the question is clear. OP wants to "recreate" a file. Using only its hash and an algorithm.

Logical Flaw: Wants to ignore the fact that there can be hash collisions.

The accepted answer begins with "Theoretically yes", and goes on to explain why it is "technically it is quite impossible".

I find both the question and the answer problematic (not outright incorrect), though I can rationlize both of them (as in "I think I can see where they're coming from").

To someone browsing or referring to SSE for good answers, I think it gives a completely wrong impression of how good our community is.

Am I right? What is the best thing to do under the circs?

PS: I can see my question itself is an example of my dilemma - especially my use of "mislead". :(

  • I cant see from this post what the problem is. That answer says it is theoretically possible but technically impossible. What is the problem with that? It doesn't mislead, and in fact is completely explicit. The question, however, was less good :-) – Rory Alsop Jul 27 '17 at 12:29
  • Well strictly speaking, even without the additional assumptions ("if we know the file size, we can make an educated guess"), if we have a rainbow table for 2^64 bits, we do have the original file somewhere in there, though we don't know which one it is; and the OP's question is satisfied. We did "recreate" the file. But then that's also my point. An answer that is not really wrong - but gives the idea that it is just a matter of figuring out a few collisions (e.g., M1 ... M10). – Sas3 Jul 27 '17 at 13:13
  • 1
    I saw that question and stared at it a minute myself trying to figure out what to make of it. To be clear, everything there is 100% wrong. It is not theoretically possible. Hashes are specifically non-reversible mathematical functions, not just practically but also theoretically. It's ironic because the answer says "Theoretically yes even though they are non-reversible functions". That sentence is gibberish. There is nothing theoretical here. They are not reversible. That bugged me to no end. – Conor Mancone Aug 3 '17 at 14:39
2

In French, there is saying "Avec des si, on mettrait Paris en bouteille" (litteraly "With ifs, we could put Paris in a bottle") which according to my Google-fu, has the following English equivalent: "if ifs and ands were pots and pans, there’d be no work for tinker’s hands".

Personally, I think that there is already enough work and a lot yet to be done to address current, concrete risks. I used to vote to close questions which ask concrete ways to address purely speculative risks abiding by arbitrary and usually unrealistic limitations and also using arbitrary chosen means. IMHO, such questions are more the realms of WorldBuilding.

1

Your best action is a downvote on any post you think deserves one, and you can vote to close any question you think is not a real question, too broad, opinion-based etc.

  • Thanks. I was hoping for something more, but I guess that when something is not against the rules/guidelines, this is the only option left. – Sas3 Jul 28 '17 at 6:50

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .