Just read the BLAKE3 paper. I'm really happy to finally see a standardized tree-based hash function that is fast on all architectures. Existing hash functions either didn't come with a tree mode (everyone had to invent their own incompatible tree algorithm) or had too many different (and domain-separated) modes - making cross-referencing hashes across multiple protocols impossible.

BLAKE3 should allow one to publish a single file hash that could be used both for identification and incremental verification on across multiple (P2P) file sharing networks. I'm hoping it will get adopted.

Sign in to participate in the conversation

A lonely little town in the wider world of the fediverse.