Focuses on Universality and Irreversibility

PoxHash is a block hashing algorithm created for performative and educational means and can be used for various tasks. It was originally developed for one of my active C| | projects as a checksum for HTTP body data. But I decided to take it out, polish it, make it more intricate, and implement it in C, Rust, Go, Nim, Python and JavaScript. Along with the implementations in every language, I have also written a runner code.

Length of the hash is 64 bits. It is comprised of 4 unsigned 16-bit integers. Size of each block is 64 as well. The size of hexadecimal digest of the hash is 16. Besides base 16, the implementations offer non-decimal digests in bases 60, 20, 14, 12, 8, 6 and 2. The 4 16-bit factors are also offered as 8 unsigned 8-bit integers, 2 unsigned 32-bit integers and a single unsigned 64-bit integer.

• Learn how to make use of the libpoxh.* files -> USAGE.md
• Learn how to use the runners and poxh.sh script -> RUNNERS.md
• Learn more about the hashing algorithm in general -> SPEC.md
• Learn about ways to contribute to this project -> CONTRIB.md

To demonstate the universality of PoxHash, look at the table below. The strings are encoded with UTF-8 and passed to the hasher function.

• Message A: HTML for www.example.com
• Message A': Message A, with changes to`
• Message B: contents of https://www.ietf.org/rfc/rfc2616.txt
• Message B': Message B, with [Page 116] changed to [Page 161]

• Message A; [0b00100010, 0b01100101, 0b10110101, 0b10110101, 0b01011101, 0b1111110, 0b1111101]
• Message B: [0b00100010, 0b01100100, 0b10110101, 0b10110101, 0b01011101, 0b1111110, 0b1111101]
• Message C: [0b00100010, 0b01100101, 0b10110101, 0b10110101, 0b01011101, 0b1101110, 0b1111101]
• Message D: [0b00100010, 0b01100101, 0b10110101, 0b10111101, 0b01011101, 0b1111110, 0b1111101]

Althoug I cannot guarantee that PoxHash is ever-unbreakable, what I know is, it relies on several serially accumulative operations to scramle the message. These operations apply average, median, product, sum (and the weighted counterparts of these) to the 4 word factors during the operations. Walking backwards from this without knowing the original message is a hard task. You can notice some patterns with the digest, but these patterns are fairly faint and untracable. I still refuse to make a promise.

If you manage to break the universality or irreversibility of PoxHash, please let me know so I can relase a stronger version.

The following table shows the speed with which the implementations hash a message of 20KBs.

I wish to make Pox into a series of cryptographical tools and algorithms. I wish to make a PoxCipher, a PoxECC (Elliptic Curve Ciper), and my next target is PoxChain, a BlockChain implementation in Go using libp2p. If you find this project useful, please consider a small donation my way. It will go a long way!

If you have any questions or requests, my contact details are inside my Github profile. Thank you.

PoxHash remains untested and its methodology are experimental. It's not recommended for use in production for checksums or for security purposes even at the lowest stakes. If you decide to use PoxHash for any other purpose than education and recreation, I cannot be held reliable. I recommend BLAKEx family of hashes and SHA-x family of hashes for both cryptography and checksums and MD5 for lower-stake checksums.