I agree with you that your definition is what we should define a blockchain as, as Natamoto defined it. However, I'm afraid to say probably the majority of companies that claim they use blockchain probably use something very close to Git instead, or even just a digital signature or a hash function. Seriously, it's ridiculous.

Imagine if Google called certificate transparency blockchain? Journalists would lose their shit. Such an easy article to write. In fact, Google explicitly say on their website that they're kinda against the hype: http://www.certificate-transparency.org/general-transparency "These technologies are strongly related to the much-hyped blockchain. The reality, of course, is that there isn't a "the" blockchain, and that decentralisation is not always the answer. We are not making "the" blockchain, and we do not claim to support decentralisation."

I'll never forget my first job the business people wanted to investigate "the blockchain" and the yes men came up with a bunch of PoCs. I just asked "how is this different to Git?" Such an awkward moment.

> I agree with you that your definition is what we should define a blockchain as, as Natamoto defined it.

Funny enough, "blockchain" as a term wasn't defined in the Bitcoin white paper at all.

If you don't have decentralized production and validation of blocks, you don't have a block chain.

Why should it need to be decentralized?

A blockchain is a chain of blocks. Tautological, but the words mean something; it isn't just a name. If you hash blocks that contain hashes of previous blocks, you have a blockchain.

At most it might be argued that a centralized blockchain has no reason to exist, but I doubt that's true.

> A blockchain is a chain of blocks. Tautological, but the words mean something; it isn't just a name.

No, it doesn't work like that. "Blockchain" is understood to mean something and you can't decompose the compound word and say "well, here's a block. Oh, and here's a chain. I guess this is a block chain." Your description, e.g., applies to a Merkle tree [0], which git uses, and was published in 1979. But no one has ever or will ever call it a blockchain.

Is a Ford F-150 that's been lit on fire a firetruck? How much straw is in a strawberry?

[0] https://en.wikipedia.org/wiki/Merkle_tree

You're on the right side of the linguistic prescriptivism vs descriptivism debate, but I think the meaning of that word might not have settled enough yet. Personally, I think it makes sense to restrict "blockchain" to a chain (or DAG) of blocks with hashes, and refer to Bitcoin etc. technology as blockchain + POW (rather than having "blockchain" subsume the POW).

That might be a losing battle, but not sure the battle is over yet...

My go-to example for this is "putting a Hershey's bar in the microwave is not hot chocolate."

Because the term was coined by Nakamoto to describe the technology behind bitcoin. Decentralization, proof of work and other characteristics of bitcoin are intrinsic to making the system work. Otherwise you have an inefficient database that has no reason to exist.

> Because the term was coined by Nakamoto to describe the technology behind bitcoin.

It wasn't- there's one mention of a "chain of blocks" in the white paper. The term "blockchain" appears to have been coined by Hal Finney.

You don't really need the 'otherwise'.

You don't need decentralization for a blockchain.

All you need is 1) a reverse linked list of entries and 2) have the list be cryptographic in nature. Even git can be used for a blockchain since it uses a cryptographic hash for commits and can use GPG signing.

Validation, consensus and decentralizations are additional properties that you can add on top that make it useful for certain use cases.

Because the term was coined by Nakamoto to describe the technology behind bitcoin. Decentralization, proof of work and other characteristics of bitcoin are intrinsic to making the system work. Otherwise you have an inefficient database that has no reason to exist.

This is one of the problems with the generic use of “blockchain”. People think you just need Merkle roots including previous blocks, like you do, but that’s useless and pointless without the rest.

>Otherwise you have an inefficient database that has no reason to exist.

That doesn't mean it isn't a blockchain. A inefficient or poor implementation of something is still an implementation of that thing.

>People think you just need Merkle roots including previous blocks, like you do, but that’s useless and pointless without the rest.

I'm well aware of the additional properties required to get it to behave like bitcoin or have cryptographic safety. But those are to my knowledge not a must-have for a blockchain nor defining characteristics.

As the name implies, it's a chain of blocks, the less obvious part is you link them with something cryptographic.

And just because Blockchain was first used and defined in Bitcoin doesn't mean the definition or methodology cannot be separated from Bitcoin, in fact, I consider it heavily separated from Bitcoin by now.

And again, just because they are factors in making Bitcoin work doesn't mean you need them to make a Blockchain.

Except that's a merkle tree. Which we already have a word for. Git isn't a blockchain, but it is a merkle tree. Blockchains are merkle trees, but the reverse isn't true.

A reasonable definition might be that a blockchain is a decentralized merkle tree.

I think the most general thing you could say is monadic mutations of some global state. Blocks act as operators on this state. I don't necessarily think it has to be a consensus algorithm. It can be used as that though.

Blocks are taken as part of the blockchain once there is consensus amongst the nodes that it is valid, so the consensus algorithm (as well as the POW algorithm) are integral parts of what makes a blockchain a blockchain

That's just event sourcing. The innovation in the bitcoin whitepaper was the consensus algorithm.

So they'd have to do something like publish the hashes in a major newspaper?