News for hackers. Stuff that matters.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by eeeeeeeeeeeee on Friday February 22, 2019 @11:59AM (#19226863)
Anyone have any good counter arguments to articles like this? That there is a use case that makes sense?

I’ve been trying to find one and can’t, but it hasn’t stopped major corporations like IBM, who I would expect to jump on the hype train, but even AWS is embracing it to a degree with their latest announcements.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by pdpi on Friday February 22, 2019 @12:14PM (#19227002)
Blockchain-the-datastructure is incredibly useful and widely deployed — e.g. Git histories are blockchains (well, a slightly extended version of a blockchain that allows you to represent a DAG, same difference).

Blockchain-the-trustless-consensus-protocol has a fundamental challenge that makes it completely and utterly pointless for most applications — the oracle problem. You build up a whole system around a secure trustless chain-of-custody, but then require trust in some sort of outside entity for adding resources into the system. If you're willing to trust that entity, there's many simpler ways to sort out your chain-of-custody problems.

Cryptocurrencies "work" precisely because they don't need an oracle. They don't track anything that exists outside the chain.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by UncleMeat on Friday February 22, 2019 @02:05PM (#19228105)
Only in a crazy loose sense. Git histories do not have enforced immutability or trustlessness. And they are really merkle trees, which have been around for decades. If blockchains (as described in this manner) are so amazing, why wasn't there hype for merkle trees in 1995?
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by pdpi on Friday February 22, 2019 @02:20PM (#19228230)
I think you misunderstood my point — The reason I distinguished between blockchain-the-data-structure and blockchain-the-trustless-consensus-protocol is precisely because the word used to mean specifically the former but got coopted to refer to the latter. The blockchain data structure has been around for decades, and a Merkle tree is precisely the tree analogue to a blockchain's linked list.

Trustlessness is a property of Nakamoto consensus, not of the blockchain data structure, but the combination of the two (or something that bears passing resemblance to that combination) is what usually gets called a "blockchain". Unfortunately this usage of the word is borderline useless because people keep loosening the requirements of the system so as to allow their harebrained setups to count as a "blockchain".

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by ajkjk on Friday February 22, 2019 @12:20PM (#19227066)
For blockchain to be valuable there needs to be an otherwise unsolvable coordination problem.

One way the problem can be unsolvable is not to be literally unsolvable but philosophically unsolvable, which is why bitcoin itself is probably a legitimately good application of the idea. The anarchic ideal is to not trust anyone in particular with controlling money, and any actor who says "no trust me, I'll do it right" is probably less trustworthy for having said that, so it seems unsolvable. So a 'self-organizing' coordination is somewhat necessary, essentially for the same reasons we like democracies more than autocracies.

(uh but proof of work is evil, so, just cause it's a good usecase doesn't mean it should exist.)

Besides that, most of the 'coordination problems' that people purport to want blockchains for seem to be simple cases of glaciation and incompetence. If you can't reconcile accounts across banks / supply chains / etc efficiently, your industry just sucks technologically and needs to get it together, and it'll still suck when you have a blockchain making your IT more expensive, but at least then you don't have to reach consensus on protocols so good for you I guess?

I would love to hear of other useful cases of otherwise-unsolvable coordination problems, but I really can't remember any right now.

On the other hand, ETH-like cases are completely legitimate uses of blockchain-like things ('running code in the ether' is not something you can do any other way, basically by definition, and is also just _cool_), but they are rather different and not directly comparable.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by doubletgl on Friday February 22, 2019 @02:25PM (#19228282)
The supply chain management use case from the article might not really work in the sense that the oracles/involved parties can tamper with the their input sensors, but it still offers other parties plausible deniability.

Let's say you get meat delivered by cooling trucks and this is mirrored in a public block chain. You publicly claim that all your meat is sufficiently cooled and every single delivery is tracked. If this data is in a public block chain, there is no doubt that you didn't input this data, it was in fact the delivery company, and if they cheated with their cooling sensors, the liability is on them. With a privately hosted database under control of the corporation (you), there would be suspicion that you manipulated it.

So it doesn't actually solve the problem, but you can legitimately claim that you did as much as you could (as far as using technology goes at least) to ensure the quality of the delivered meat.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by Vinnl on Friday February 22, 2019 @12:33PM (#19227191)
As I understand it, blockchain is only a solution if your problem is of the shape: I need to record statements, but I cannot/do not want to trust a central authority to not tamper with those statements.

So bitcoin solves the problem of not wanting governments being able to print extra money. The sibling comment [1] about timestamping the date of birth of veal is not a good use case, because it's perfectly possible and easier to set up a central authority that the veterinarian can report to. Which, according to a response [2], is exactly what happened, and solves the problem of timestamping veal just fine.

[1] https://news.ycombinator.com/item?id=19226935

[2] https://news.ycombinator.com/item?id=19226997

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by h1d on Friday February 22, 2019 @12:20PM (#19227062)
I think they always fail because they try to use it inside a known party where there's always a better centralized solution because they already have some trust built for it to work.

Cross parties or even industries may be a better target.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by homero on Friday February 22, 2019 @12:10PM (#19226963)
The only application i can think of is proving something existed at a certain date by putting the hash in the blockchain.

Op says it won't help authorship since earlier real author can claim he didn't use the blockchain but it's still helpful to prove something simply existed.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by lawn on Friday February 22, 2019 @02:07PM (#19228123)
For example:

You find a security vulnerability in some applications. You want to make an anonymous disclosure but you want to be able to prove you knew about it at a certain time.

So you create the message "I John Doe has found the vulnerability ..." and add the hash to the blockchain.

You can now definitely prove that you knew about it and had documented it at a certain timestamp before the disclosure.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by homero on Friday February 22, 2019 @02:42PM (#19228446)
Exactly, only thing a blockchain is good at and very hard to do without a blockchain. Even if you tell a lawyer or something, you can never prove it 100%.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by sp332 on Friday February 22, 2019 @01:40PM (#19227841)
Blockchain is useful in a narrow set of cases: when you need a distributed network with no trusted central party, and at the same time you need to prevent double-spending. Usually when you have something valuable and want to prevent double-spending, there is a central authority like a bank or a physical item like cash which will authoritatively answer who holds what. And usually for distributed systems like bittorrent, lying on the network has relatively little risk to other people.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by stale2002 on Friday February 22, 2019 @12:39PM (#19227242)
People are over complicating what a *Blockchain" really is.

All a Blockchain is, is a public, append only database. That's it.

And no, it does not require "proof of work". People have released lots of Blockchains that do not have proof of work. They might even use a central authority to sign their blockchain transactions.

Public append only databases have existed for a long time, and are quite useful.

Someone else gave the example of git. Yes git is a Blockchain. Really.

There are a lot of reasons to use a public append only database, as opposed to a private database, hidden behind a central authority.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by melkiaur on Friday February 22, 2019 @12:07PM (#19226935)
There's one I like: the farmer has a veal and introduces its date of birth on the blockchain.

The veterinary that comes to check it posts a validation mark on the blockchain. If you trust the veterinary, you should trust the birthdate.

The veal grows up to become a cow, and is sent for slaughter. The farmer stores an event on the blockchain and sends it to be cut into meat.

The slaughterhouse stores an event (with the weight) on the blockchain and sends it for distribution.

The butcher creates the packets of meat, with a QR-code which lets you check the whole list of events, with their dates and places, and of course, you can check that the same part of beef doesn't end up in two different shops. You have complete traceability of your meat despite all the actors not knowing each-others.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by danielmg on Friday February 22, 2019 @12:14PM (#19226997)
That's already a solved problem in the UK after the BSE crisis. From what I was told (by a friend who worked on it), it uses certificates to provide trust and verification. Uses a central database hosted by DEFRA. No blockchain needed.

https://secure.services.defra.gov.uk/wps/portal/ctso/

Everything you describe can be solved, in a better and more efficient way, by non-blockchain technologies.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by GauntletWizard on Friday February 22, 2019 @12:19PM (#19227058)
Part of this problem is that signed merkle trees are the solution to lots of problems, and block chain advocates say "Those are blockchains!" and then try to introduce the undesirable feature (forward looking hashes that require proofs), while crypto nuts have already been screaming about merkle trees forever.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by melkiaur on Friday February 22, 2019 @12:19PM (#19227050)
It could be solved, but is it really ?

Shall I remind you of this: https://en.wikipedia.org/wiki/2013_horse_meat_scandal#UK_Inv...

The blockchain doesn't offer a technological solution that couldn't be done with a centralized authority (I particularly like the way passports rfid reading involves certificates signed by each country). But it offers means to achieve that even if nobody safe enough wants to host the bloody service.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by danielmg on Friday February 22, 2019 @12:21PM (#19227071)
That wasn't about the provenance of the meat but the production of the end products - e.g. pies etc. They were just dumping in bad meat in place of good.

Blockchain isn't going to solve that - you could put the ingredients in the on it and have them signed etc but at the end of the day humans are mixing the food.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by dmitriid on Friday February 22, 2019 @12:23PM (#19227089)
Someone marks horse meat as cow meat on blockchain. How exactly blockchain solves that?
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by hnbroseph on Friday February 22, 2019 @01:20PM (#19227668)
ot, but even as i would probably be a little dismayed to learn my burger wasn't made of what i thought it was... it's kind of amusing that horse meat causes a stir while cow meat doesn't it's all just a matter of what's normalized i guess.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by richardhod on Friday February 22, 2019 @12:25PM (#19227106)
Agreed. Yet there is a case: I think the use case is for those who would rather do without centralised government and rather have a libertarian polity.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by Tomte on Friday February 22, 2019 @12:09PM (#19226955)
And you need to trust each and every one of these people not to lie.

That's the key issue: Bitcoin works, because Bitcoins are "born" on the blockchain.

Anytime you need to relate real-world concepts to the blockchain, you need trust again. As you said: "if you trust the veterinary".

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by melkiaur on Friday February 22, 2019 @12:15PM (#19227006)
No, I think you "only" need to trust the veterinary, and make sure that all animals are tracked (which is the case in Europe, which is why the Lasagna scandal came to be a scandal). Both things are already true here.

The anti-double-spend features of blockchains is what matters after that: your cow cannot be sent to slaughterhouse 1 and 2 at the same time. Slaugher house cannot ship it to butcher A and butcher B. And it's fairly easy to check that the butcher doesn't sell it twice.

Though as I said in another comment, this can all be solved very easily if you give the DB hosting to the veterinary authorities, and very likely to be much cheaper too.

Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by ajkjk on Friday February 22, 2019 @12:13PM (#19226988)
None of those seem to require a blockchain at all; a trusted authority database would accomplish the same thing at a fraction of the cost and complexity.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by h1d on Friday February 22, 2019 @12:17PM (#19227028)
The point is you don't need a central authority. And I'm sure running a central authority costs some too which could also gain power that no one in the industry can resist.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by ajkjk on Friday February 22, 2019 @12:22PM (#19227080)
Yes, that's the point: pay lots of money and make your system incredibly complicated, in exchange for not having a central authority when you literally already have one you could just use, for a worse and less-maintainable version of the same technical result.
Re:You Do Not Need Blockchain: Popular Use Cases and Why The... by danielmg on Friday February 22, 2019 @12:19PM (#19227052)
but in most countries there is a central authority in the shape of a government agency responsible for agriculture.