Hacker News Reader

Comment by sdiacom | Hacker News Reader

Per the Wikipedia article:

> User content in SSB is organized as an append-only sequence of immutable messages, where messages cryptographically sign adjacent messages for the purpose of guaranteeing unforgeabilitity of the sequences as they are replicated to other peers.

That's pretty much a blockchain. Of course, they don't call it a blockchain themselves, because the Scuttlebutt people are not running a grift, they're creating a communication protocol that happens to use cryptographic primitives similarly to how blockchains use them.

View on HN

That's using a specific definition of a blockchain that differs significantly from e.g. Bitcoin and Ethereum which are two of the most popular ones and the ones commonly referred to, along with their decentralization properties. The Schneier post I linked explains fairly how these are not novel and have nothing to do with the current hype beyond the confusing naming.

Concretely, if Scuttlebutt uses a blockchain, then Git also uses a blockchain by a similar reasoning. I don't think the original post claiming that "there are no use cases for blockchains" was implicitly claiming that Git (or Scuttlebut) is not relevant software, just that they don't use a blockchain, because indeed they don't for any interesting definition of "blockchain".

I agree with you completely, the "blockchain" term is confusing, and by the definition I'm using, Git would also be a sort of blockchain. The more that blockchain (the grift) expands, the more that the meaning of blockchain (the technology) is diluted.

A key difference to me between Git and Scuttlebutt, which in my (most likely flawed) view makes Scuttlebutt "blockchain-er", is the distributed nature of it. You add a friend on Scuttlebutt, and this will fetch their content from any node, and all their friends' contents. It has an "automatic replication across nodes" aspect which Git does not have.

[Disclaimer: I haven't used Scuttlebutt in a long while, so I may be getting the technical details of the friends-of-friends-replication thing wrong.]

> That's pretty much a blockchain

Only if you remove all of the parts of a blockchain that make it a blockchain. The definition from wikipedia isn't a blockchain, it's a Merkle Tree. They're not the same thing - a blockchain _uses_ a merkle tree. The bits that are bolted onto the merkle tree are what make it a blockchain, and are the bits that don't provide value.

Do the messages require computationally expensive proofs?

Thankfully, no. I don't think that's a requirement for a blockchain, though.

Reply on news.ycombinator.com