Hacker News Reader

Comment by jeffbr13 | Hacker News Reader

Blockchains are necessary for Bitcoin, but Bitcoin is not necessary for blockchains.

"The Blockchain" (big B) is the original blockchain (small b) used for Bitcoin, but it's only one possible implementation. It's like saying Postgres, MySQL, etc. aren't 'true' SQL databases only Codd's original implementation is -- now, obviously the metaphor doesn't quite hold because of the distributed nature of blockchains, but it's disingenuous to say that there's only one canonical implementation of a technology.

The Blockchain and Bitcoin are valuable proof-of-concept implementations of blockchains, but there's value to be gained from separating the general ideas from specific instances.

View on HN

This is only partially true. While you can have a blockchain without bitcoin, you cannot have a blockchain without some sort of native digital asset giving value to the network supporters (in bitcoin terminology, the miners). I think many people don't realize this, and think they can somehow spin up a new blockchain without needing to also create a separate currency and convince enough people to start giving it value.

Realistically, most "blockchain" tech these days is actually using the bitcoin blockchain, because it is the most secure and widely-supported blockchain available.

Correct. The specific innovation brought to us by blockchains is "objective truth" over subjective truth. We've had journaled, auditable, shared, and logged databases for years. Even PGP signed SMTP was a sufficient mechanism for transmitting shared data. The innovation of Blockchains are their proof of work mechanism by which truth is objectively determinable.

Oh - and of the thousands of blockchains out there, only one is worth a damn. The others are barely burning more than $10k per day in anchoring their objective truth.

It is absolutely not an objective truth! It's an expensive consensus.

The mental image of 'burning' increasing amounts of wealth for a financial instrument of dubious value gets more strange the more you think about it, like a funeral pyre or a hugely destructive potlatch.

It's certainly an objective truth - it's attested to by the energy spent in producing it. Attesting to alternate truth is often possible (we have orphan blocks all the time) but - the older the truth, the more objective it is.

The intrinsic value of bitcoin is censorship resistant storage. The speculative value of bitcoin is somewhere above 0.

If you don't need censorship resistance - don't use it. But if you think no-one needs censorship resistance, then you are empirically wrong, as demonstrated by the hundreds of thousands of transactions per day in Bitcoin.

As for burning increasing amounts of wealth - well, the network is pretty well calibrated. It's not a chain reaction that will take over the world, it'll scale to precisely the size that the market demands.

Uh, no, you do not need to create a currency. You might need that if you want to attract random outside miners, but that isn't necessary to have a blockchain.

How will your proposed blockchain protect against double spends or sybil attacks without a native token? This was the entire innovation of bitcoin.

I remember designs for DNS-like services on alternative blockchains where you were supposed to pay the miners in Bitcoins after the fact. Or the miners could be paid for processed blocks by a central organizer (e.g. the central bank if we think about a banking application people seem to be desperately searching for some reason).

There has to be an incentive for miners to work with the network, but it doesn't have to be intrinsic to the network.

Like Namecoin (supported by a separate currency) or Blockstore (uses the Bitcoin blockchain)?

Even if you have a central organizer/bank, you are simply using their currency as the native token. You could use a native token which has external value _if_ you have the authority to issue that token (e.g. a fiat currency if a central bank decided to create their own blockchain), but that doesn't remove the need for the network to have a native token.

Aren't the miner incentives are only necessary if it's an open blockchain that requires public hashing power?

You can have a relatively smaller number of actors incentivized by other means on a network - e.g. organizations in an industry that validate transactions amongst each other.

True, but decentralized databases (where all nodes are trusted) were already invented long before bitcoin came along. You do not need a blockchain if all actors are trusted.

For perfect trust, sure, but amongst competitors in a same market space you would think that it's more 'imperfect' trust, even with contracts.

If the transactions are small and numerous, it would seem easier to maintain a block chain ledger than managing multiple write points and masters for a decentralized database. I'm not saying it's concrete, but it would be an interesting experiment at least.

How do you determine who gets to participate in this separate blockchain? Can anyone join? Or are there barriers to entry? Because if there are barriers to entry, then you've now created a single centralized authority that issues these keys, and you thus don't really have a decentralized system at all.

If each participant individually judges the merits of other participants, then everything will quickly diverge as some blocks are accepted by some participants but not by others, and the whole network collapses into forks.

There are barriers to entry, for the miners - that's the whole point of having a closed block chain. Nodes and addresses could be open depending on the hypothetical.

Mass decentralization is only a selling point for Bitcoin, but even then the Bitcoin organization (who controls the code) has authority to push a hard fork on the miners. They won't do it obviously (e.g. increase block size), because it would alienate their user base.

Miner validation can be controlled by the same organization in charge of the code, which isn't aligned with any one competitor in the industry.

The only barrier to entry for a Bitcoin miner is that you have to have an Internet connection and a computer. That is completely decentralized; anyone can buy a computer and Internet access. I don't understand your point. Surely any system you were proposing would require that as well. Your comment was talking about a blockchain amongst "organizations in an industry that validate transactions amongst each other". That seems like it would require a barrier to entry to me, namely, that you would have to validate to someone that you were actually so-and-so organization in so-and-so industry. Can you more fully flesh out your proposal? I'm seeing additional barriers to entry, requiring centralization, that Bitcoin does not have.

In practice you probably want a special purpose computer that is GOOD at calculating SHA256 hashes. But anyone can acquire these without limitation or restriction.

Hard forks push changes to the protocol from a centralized location, namely the bitcoin organization. You know how hard forks work, right?

It's not a "vs. bitcoin" comparison... just an example of how organizations in an industry would use block chain to store numerous small transactions, instead of a decentralized database.

You keep deflecting and not answering my questions and it's getting really annoying. I'm not going to keep responding to all of your tangents when you can't even address a single one of my points. Please respond to what I've been saying in a relevant manner, then I'll respond to your other stuff.

I've been telling you that you're making the wrong assumptions about blockchain applications because you're assuming it has to be based on bitcoin designs. Again, a closed blockchain doesn't require it to be open to the public for miners.

If you're getting frustrated and think that is deflecting then it's possible you have a cognitive bias.

You don't have to have forks. When there is no agreement, the network can just stop working. Some might think that a bit useless, but when you're trying to reach agreement it is in many respects a lot better than a fork...

But then anyone can trivially shut down the network by submitting a single invalid block. The very first bit of abuse and everything is shut down and stops working indefinitely. How could this design serve a useful purpose? It seems a lot less useful than the way Bitcoin does it.

But who's going to abuse it? Why would it shut down indefinitely? This is a group of people or entities that has decided to do this. Presumably they all feel it's in their interests to make it work. Same as with any mutual agreement of this kind, if there's some kind of a falling-out, they'll probably just figure it out somehow.

(As for invalid blocks - in the absence of a bug in the code that runs the system, all validators will say no to the invalid block(s), so there's no disagreement, and things carry on.)

So now we're back to having trusted parties only, and you thus have centralization, and don't need a blockchain at all. You just need signed transactions among trusted peers. We keep going round in circles.

It's only centralized from the point of view of somebody who isn't involved. Those who are involved most likely won't see it as such - since if they did, they'd have chosen some other mechanism. They don't need to trust one another.

Are you familiar with Bitmessage?

The mining incentive can just as easily be usage of service provided by the network which uses the given blockchain.

Even if the miners weren't interested in consuming the provided services on their own, the network could rely on miners being rewarded with services which they then sell access to.

Reply on news.ycombinator.com