[phyzome]

The article doesn't make it totally explicit, but it sure sounds like each new hash incorporated the previous one.

(It's also possible that each published hash is just a brand new hash of the aggregate, but this seems somewhat less likely based on the wording.)

ETA: Looks like it's a chain, or rather a Merkle tree. From the horse's mouth: https://news.ycombinator.com/item?id=17865168

[bittcto]

A Merkle root including previous Merkle roots may have been innovative when Surety started, but while its required in bitcoin its not a key feature of what defines a Blockchain.

[phyzome]

Given that "blockchain" is a generalization of Bitcoin, not formally defined theoretically beforehand, I suspect we would not agree on what a blockchain even is, and would go 'round and 'round in circles.

But I think we can agree that a Merkle chain like this would be sufficient for most of the things people are trying to use blockchains for recently, which certainly says something.

[Groxx]

Yea, I think most would lean towards requiring one or both of:

    - it's a distributed consensus system (often around "blocks", but e.g. iota is not)
    - it's a chain of hashes to prevent mutation (around "chains")

And I doubt we'll ever get everyone to agree on what subset (including null) of those are "a blockchain".

Very (very) few things benefit from distributed consensus IMO. And not all altcoins use blocks at all. So I tend to lean towards hash-chains. Publicly-verifiable immutability is useful in quite a lot of real-world scenarios.