[Cryptography] Solving Matt's hash problem

Natanael natanael.l at gmail.com
Mon Nov 14 21:04:45 EST 2022


Den tis 15 nov. 2022 02:51John Levine <johnl at iecc.com> skrev:

> In Matt Levine's long article about cryptocurrencies, he describes in
> accurate detail how hashing works, and mentions in an aside that he
> cannot include the hash of the article in the article itself, you (the
> reader) figure out why.
>
> I know this is a preimage problem, but is there anything about making
> the hash itself part of the input that would make it harder or easier
> to solve?
>
> R's,
> John
>

For a typical hash functions (built on pseudorandom permutations) it
doesn't matter much what the target value is.

"Hash quines" have been done indirectly with MD5 which is possible because
collision attacks are fast. I think they technically don't contain the
actual bitstring encoded directly, but they use multi-collisions to embed a
sequence of strings representing each character (like colliding code blocks
printing each character for the full alphabet, multiple times in a
sequence), then you calculate the full hash over that and select the
characters you wanted.

https://www.rogdham.net/2017/03/12/gif-md5-hashquine.en

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20221115/bdb366af/attachment.htm>


More information about the cryptography mailing list