> solves a very real problem of how to prevent bitrot in the long term storage of data for example.

Not really. At least, that's not the new part in cryptocurrencies. The knowledge of how to have a verifiable list of transactions which could not be tampered with existed long before bitcoin: the main challenge was to only have one which everyone could agree on just using the same set of rules without having to trust everyone follows the rules or the intent behind them (if you have a single you can trust or even just a set of parties in which you can trust the majority then it's pretty easy). And that's far more to do with designing an incentive structure which works to achieve that then it is about any fundamental cryptography.

Crypto currencies aren't necessary for blockchain as a technology to be useful as in the bit rot example.

However crypto currencies cannot exist without the blockchain.

"blockchain" is a term only used in the context of cryptocurrencies where a chain of transaction blocks is agreed upon in a distributed system. In any other case it's called a merkle tree. The core tech which distinguishes blockchain is not the series of signatures which provide proof of integrity (and perhaps more important in cryptocurrency: that they were derived from the previous part of the chain. This doesn't matter for data integrity per se, but it matters for consensus rules), but the consensus rules which say 'this chain is the one chain which counts, not the others'.

Hence it is and should be named DLT. The "Distributed" separates it from merkle tree systems.

Except that there are also "permissioned blockchains" which are non-distributed systems. Of course those are more an abuse of terminology to hype various services, but the term has appeared in the research literature. The consensus algorithm in such a blockchain is just "whatever blocks the controlling entity accepts into the canonical branch".

Permissioned in that case simply means private (you need a permission to participate). It does not mean non-distributed. It also doesn't mean any entity with permission can add anything and its accepted. That would just be a shared/synced database. A DTL still has rules and all participant collectively enforce them. Double spend is still prevented by the same technology. There must be some form of central authority to make it private but DLT limits the power of that authority, it can maybe add or remove participants and such stuff but it can not change the rules. Every participants would have to agree to run the new rules

This is certainly useful to connect entries with shared balanced like banks who cooperate. Every bank still wants to validate for themself and not blindly trust whatever another bank writes on the ledger. Moving balanced form someone else or double spend is not possible in such a system, even trough all have the same permissions.

A permissioned/private "blockchain" is still a DTL as long as the consensus enforcement is still distributed. If all participants have to trust all others to simply follow the rules then its not a DTL anymore.

This is where the conversation gets tricky, though. If, as the bit rot example would imply, blockchain is just a fancy word for hash chain, then that statement is uncontroversial. Hash chains their cousins have been around for something like half a century and are well established.

That's really hard to reconcile with the "blockchain is revolutionary" meme, though. I think you've got to include distributed consensus and proof of work in the definition in order to get one that characterizes blockchain as a new invention and not just marketerspeak. Unfortunately, that definition doesn't jibe very well with the usual lists of successful uses of blockchain outside of cryptocurrency, such as corporate supply chain management, because they tend to be using blockchain where a hash chain would work just as well. Perhaps they're actually using hash chains, and just using the word "blockchain" as marketerspeak. I don't have inside knowledge and couldn't evaluate such things, but it wouldn't surprise me at all if the more commonly touted examples were doing so because distributed consensus is not a desirable thing for the problem they're trying to solve.

In the abstract, the problem is that it gets really hard to understand what everyone is actually talking about when the conversation centers on a popular buzzword.

I suspect what you're calling a blockchain is actually a merkle tree. The blockchain is a merkle tree structure shared via p2p communication over a protocol designed to prevent double spending.

Merkle trees were invented in 1979. If companies looking to prevent data tampering didn't make use of them until they got rebranded as "blockchain" then those companies are just bad at what they do.

> The blockchain is a merkle tree structure shared via p2p communication over a protocol designed to prevent double spending.

I'd say "The blockchain is a Merkle tree structure with a consensus algorithm which selects a canonical path to a leaf node at any given time."

I don't think preventing double spending is a meaningful concept for non-cryptocurrency blockchains, since those have nothing to spend. Whether we accept that such things are blockchains determines whether this distinction makes any sense. Though, if all blockchains are cryptocurrencies then there's not really any need to separate the two terms.

Yes you're right, that's a better generalised description of them. Of course that's what I meant by "double spending" (essentially, the list becoming a graph, ie forking), but precision in communication is important.

I think blockchain is a great example of interesting tech with very few real applications outside of crypto currency. Others have pointed out that there's many cheaper ways to prevent bit rot. Almost anything you can do with a blockchain can be accomplished with a centralized ledger.

Bitcoin is a great example of interesting tech that overcomplicates something everyone understands to fix problems most people don't have. Should we ever find a major vulnerability in a common implementation, it could be game over.

> it solves a very real problem of how to prevent bitrot in the long term storage of data for example.

OK, I'll bite - how does blockchain technology help prevent bit rot in a way that can't be helped more efficiently by a simpler technology?

I've worked in blockchain for 8 years across the crypto and enterprise space and I can tell you with utter certainty that it is stupid. And I'm stupid too for not seeing it earlier.

Just to provide slightly more info:

For crypto (read blockchain with PoW, PoS), other than censorship resistance - which most people don't actually care about because they are lucky enough not to need it - it doesn't really offer anything useful. Most crypto projects are Rube Goldberg machines whereby the operators of these companies are re-hashing all the same things our we've done with money before. Of course they are doing well at the moment because it's morphed into a get rich quick scheme. If there was no easy money to be made there would be little interest in it, just like there was back in 2010 when I started looking at Bitcoin.

Enterprise blockchain struggles because there's no actual need for blockchain elements like chains of provenance and smart contracts if the participants are fully identified and already have business relationships. "Blockchain" is a direct substitute for trusted third parties and good legal agreements and should only ever be used in circumstances where all other options are exhausted because the technology comes with so many bad trade-offs. Doesn't scale, very complicated, difficult to manage upgrades and versions, customers find it hard to understand, etc...

> it solves a very real problem of how to prevent bitrot in the long term storage of data for example

That's the first time I've seen "blockchain" and "bitrot" in the same sentence. And I fail to see the connection, unless you reduce "blockchain" to mean "copying data over and over", at which point there are many other, better solutions.

Data storage is in fact not a property of a blockchain. There is nothing that prevents data in a blockchain to be lost other than the fact that many people store it, which is not guaranteed at all.

DLTs (blockchains) solve the double spending problem in a decentral (trustless) way. Thats "the definition" and what separates it from other similar tech like git or synced databases.

He really did write "The answer is nobody knows"? Blockchains are not that hard to understand. So he plainly admits he doesn't know what he is talking about?

What's hard to understand about a really slow append-only database where you have to set a barrel of crude oil on fire in order to make a commit?

(edit: yes yes, 'proof of stake' blah blah)

DLTs (blockchains) solve the double spending problem in a decentral (trustless) way. Thats "the definition" and what separates it from other similar tech like git, a synced databases or a write-only database. PoW/PoS are just rather stupid ways to implement this its not needed at all. Lookup federated byzantine agreement (FBA) for DLT systems without it.

You know, if you could come up with a better way to implement it, you could become a billionaire. But maybe you already are, everything is possible on HN.

Go to XRPL.org its literally that and it exists since like 2013, It whole goal back then was to make a better bitcoin.

Plenty other projects have adapted the concept or even forked the code. The core team behind this back when it was named opencoin are without a doubt millionaires now and some even billionaires. So you are right but it way way to late now. Not to says that there could not be a even better solutions. But its already several thousand times better if you compare raw metrics like speed/TPS/energy usage etc. an even better solution would likely not make that much of a difference again.

Ripple has issues, iirc. It is not "the" solution yet.

Ripple is a company. The XRPL is fully functional since many years.

> Blockchain is useful and afaik does have a working definition

Speaking of unsupported assertions...