Comment by didibus | Hacker News Reader

Comment by didibus | Hacker News Reader

Parent Full thread

The blockchain is just technology. It makes all this more convenient, and possibly scale better.

But what people really argue for is that it makes the system much less prone to corruption and manipulation, because the system of tracking is much more resilient to being tempered with.

For example, right now Joe Shmoe could fabricate a fake paper deed showing that his great grandma owned the Mona Lisa, and it is thus his. And then sue the French government saying they need to either buy it off him or return it to him. Now you'd need forensic experts and all to try and check the authenticity of that deed and all that.

Now this is a bit of a stupid example obviously. What if the deed was real? Would the government still say well oh crap here you go? I guess that will depend on the French government and their laws.

But say the laws of France were such that yes, it should indeed therefore be returned to Joe.

Well with blockchain, the idea is that this wouldn't happen. First the "deed" would have never been lost, the ledger tells you who owns what, there'd be no confusion. Second, Joe (if he was lying), wouldn't be able to create a "fake" entry in the ledger as easily as maybe he can fabricate a deed. Now that's up for debate, but in theory a decentralized ledger need you to temper with all the nodes, so it's possibly much more difficult to scam with.

And back to the example in the article, the Mona Lisa would be transfered from the government to Joe, there'd be a record of this. Now the question is no longer, is this record valid? We know this happened, the question becomes was this transaction initiated fradulently? Obviously the computer doesn't know why the transaction came to be, just that it for sure did, not tempered with, no fake claims, this did happen and we know exactly from where and to where, and exactly what was it that was transacted.

Now you need people to investigate and asses if the transaction was fraudulent. Until there's a way for machines to with 100% accuracy read intent, this has to be handled by people. But in theory, a decentralized blockchain might make that process easier and more resilient to scams.

> Now this is a bit of a stupid example obviously.

Obvously. For some reason, no one can come up with a real non-stupid example for blockchain that would work better than whatever we already have.

> But say the laws of France were such that

Let me give you an actual concrete example. Some time ago someone mentioned that someone decided to create a land registry in Afghanistan based on blockchain because reasons: https://unhabitat.org/un-habitat-oict-and-lto-network-releas...

I came up with a very simple use case:

- Ahmed is registered as the rightful owner of Kolula Pushta Road, 25, Kabul.

- A month later Mohammad appears and is also registered as the rightful owner of that same place.

- Half a year later Zahra, who has never even lived in Afghanistan, wins a court battle and is registered as the rightful owner of that same place.

Questions:

- Who is the rightful owner?

- How does blockchain help here?

To skip long discussions, all of the below are possible, and are not solved in any way, shape, or form by blockchain.

- Ahmed is the rightful owner. Mohammad bribes local council which enters him as the rightful owner. Zahra is Ahmed's daughter, and has the money to chalenge this at court.

- Both Ahmed and Mohammad bribed the local council to become the owners. Zahra inherited this place from a relative and had to win a court battle to prove that she is the rightful owner.

- All of them bribed local councils and courts.

- All of them are rightful owners, they sold or gifted the property to the next person. However, at one point local authorities intervened and didn't allow Mohammed to transfer the property to Zahra, and she had to go to court.

I can't speak to the Afghanistan example, but I can see some advantages to a land registry on a blockchain, with the ensuing difficulties as well.

For example, one of the places that had toyed with the idea of a blockchain land registry was the state of Andhra Pradesh in India: https://www.ledgerinsights.com/indian-blockchain-land-record...

As an illustrative example of the issues with the land registry in the region, a friend of mine wanted to settle some land he'd inherited in his village. He had the deed to several parcels of land owned by the family (registered in his deceased father's name, as the head of the family) in his possession, but it turned out an uncle who lived in the village had 'sold' possession of some parcels of land to somebody else. It took several trips (and bribes) to the district office to even be able to view the land records for the land - which had indeed been modified. Sure, with a legal battle he could have gotten it back, but it would take decades.

How might the blockchain help here? Prevent modification of the land records without the express wishes of the deed holder. Court cases involving land are severely backlogged in India, on the scale of decades.

I also agree that the blockchain would also create additional problems & burdens.

> I can't speak to the Afghanistan example

The example equally applies to any country

> He had the deed to several parcels

> but it turned out an uncle who lived in the village had 'sold' possession of some parcels of land to somebody else

How is this a) different from my examples and b) can be solved by blockchain?

Blockchain automagically makes selling of something impossible?

> How might the blockchain help here? Prevent modification of the land records without the express wishes of the deed holder.

You don't need "modification". All you need is the next entry: "the rightful owner of X is now Y". How is blockchain going to prevent anyone from doing this?

> Blockchain automagically makes selling of something impossible?

This issue here wasn't about selling something - unfortunately in India when it comes to land, possession is nine tenths of the law. It doesn't matter if you're the true legal owner if somebody else claims physical possession and has a vaguely genuine looking piece of paper. Proving or disproving the veracity of that piece of paper takes many many years, and in this aspect India is different from many other places in the world.

> How is this a) different from my examples

It's not very different from your example, though my anecdote is meant to demonstrate some of the real life issues meant to be addressed. I'm not convinced that a blockchain is the correct solution, but I can understand the motivation.

> You don't need "modification". All you need is the next entry: "the rightful owner of X is now Y". How is blockchain going to prevent anyone from doing this?

Unless I'm misunderstanding how blockchains work, how can that next entry be created without being signed by my private key (and my consent)? I have no idea if any concrete suggestions have been made around the architecture of land record blockchains, but it could presumably be something like a multisignature wallet: each transaction has to necessarily be signed by the seller, buyer, & government authority.

Of course, like I stated in my original comment, I'm not convinced that blockchain is the correct solution here, or how certain issues could be realistically resolved.

> Unless I'm misunderstanding how blockchains work, how can that next entry be created without being signed by my private key (and my consent)?

Because this concept breaks for any number of use cases, see https://news.ycombinator.com/item?id=27217681

It even breaks for my example of "Ahmet illegaly obtained ownership in the very beginning".

One of my own personal misgivings around the blockchain for government services is the management of private keys. However, I don't see this as a conversation killer, and is something that could probably be resolved. Issues around concepts like Power of Attorney, co-ownership, & time limited ownership can be resolved by something like Multisignature wallets. Court orders changing ownership without the owner's consent could also be resolved the same way: the Sessions Court is a non-required signatory on all multisignature land registry wallets, with the District Magistrate & Sessions Judge keys sufficient to sign transactions.

> It even breaks for my example of "Ahmet illegaly obtained ownership in the very beginning".

I don't think any sort of land registry record reform can hope to resolve issues like this. This problem isn't resolved by digitizing records to a regular database either.

In India, the sort of issues my family & others have faced around land ownership claims is that the original records are largely on paper, nearly impossible to verify with the district administration, and all most people have is a faded paper receipt decades old. This is the weak point exploited by bad actors. I'll offer a personal anecdote this time. Some years ago we tried to transfer the ownership of our home to the name of our mother (it was registered in the name of my late father). The only paperwork we had received decades ago was a paper receipt. We hired a lawyer, appeared before the District Magistrate, gave affidavits, and the process chugged along for years. When it came time to verify the details of the registration, the relevant pages from the registry were 'missing'. The fact that we had paperwork proving ownership meant nothing since the master records were unavailable. We filed another petition which took even more years, till our lawyer suggested that a bribe in the right hands in the district records office might surface the original missing pages. Sure enough, a few months later the missing pages were "rediscovered". Then there were inspection agents who came to remind us that everybody along the greasy pole had to be sufficiently lubricated. The entire process took 10 years, a great deal of anguish for my elderly mother, and quite some money. This was in a large city, and the situation is worse in more rural areas.

Does a blockchain based land registry resolve every single one of these concerns? No, but neither would a database. A database would resolve many, but not all of the problems I faced. Similarly, a blockchain might resolve different issues, and likely raise new concerns. From my perspective, given how bad actors at the district administration operate, I'd be wary of a database they had control over. Sure, there are probably ways to address those concerns, but I can also see why people see the blockchain as a possible solution.

The problem with blockchain is that it's just a distributed append-only log. There's really nothing else inherent in it. There are very few, if any, applications for it.

However, too many conversations around it attribute magical properties to it. And any solutions built on top of it can, and have been, built more efficiently using almost literally anything else. Well, paper records are less efficient :)

And as far as trusting databases goes, in the end you need government to accept whatever records there are. And nothing can help there. And even in blockchain world hard forks and re-writing of history happen even without any governments.

Maybe I'm confused here, but the blockchain basically will say:

Kolula Pushta Road 25 Kabul is owned by Ahmed with Public Key XYZ.

For it to be transfered to someone else, you need the private key that can pass validation for public key XYZ. Local officials do not have this private key, only Ahmed does, stored in his personal wallet. Local officials could try to temper with their node of the blockchain ledger, but other nodes would catch it and penalize them.

If you need to deal with an edge case, say someone claims Ahmed is dead and they are his heir. Local officials can initiate a transfer where the ledger says "conditional Amhed is dead and thus key XYZ is missing, transferring to his heir Mohammad with key HJL. This could be something that other nodes accept as a transaction. There could be many things here built into the system. You could have the nodes notify some associated email with the current owner Amhed and grant them like 60 days to cancel the transaction. Or you could let Ahmed show back at any time in the future and reclaim the possession, etc.

These are the kind of things that a decentralized trusted network running a blockchain could in theory be designed to manage.

If you narrow "blockchain" simply to mean a signed ledger of transaction that maintain history then no, it's not really that much better, since it's still centrally managed and operated. It's only once you distributed things and let the network define rules of operations that it can become beneficial in my opinion.

Edit: Now that said, even if it was centralized, in theory, the ledger could only be tempered with by people with access to the code of ledger database or its servers. Sometimes you might be able to trust those more so than say any clerk or local official which can only use the software, because they might be better paid or there are less of them and hard to contact and thus bribe. So even in that case, it might be a little better, maybe the local official can enter the transfer in the ledger, but et least if Ahmed goes to court after there is proof that local official X did a transfer on day Y and that Ahmed was the prior owner. Or you could still enforce the need for Ahmed's private key, etc. So there would be more trail or recourse for Ahmed in theory. Then without a blockchain, there'd be no trail, or no need for a key, or it might be possible for users to modify the whole history, etc.

> This could be something... There could be many things... You could have the nodes... you could let... could in theory be designed

See how many hypotheticals and "could be"s you're piling on top of blockchain just to make it barely handle the simplest of cases?

Moreover, none of these require a blockchain, and can be implemented better, and more efficiently with literally anything else.

Even more moreover, your entire premise of "nothing can be done without Ahmed's private key" immediately breaks for any of these:

- death, or other incapacitation. Even loss of access to the wallet

- power of attorney

- co-ownership

- time-limited ownership

- court orders requiring change of ownership

> It's only once you distributed things and let the network define rules of operations that it can become beneficial in my opinion.

So far, the "benefit" is: let's have blockchain for the sake of the magical words, and build a tower of babel on top of it to barely manage even the simplest of cases.

It just depends if you need a higher level of trust or not.

You can implement all that on a central network, or a single machine. But if you have it running over a decentralized network there's more certainty that validation rules and data won't be tempered with.

I'm not arguing to build everything with blockchain. I know very much how business can be swayed by buzzwords, but those also shouldn't take away or detract from the potential of a technology.

In this case, the ideal network in my opinion would be that you've got some global blockchain network that can record arbitrary data and run contracts attached to the data where it could scale and not be power hungry.

If we had that, then you could start to record things like land ownership, and attach contracts to them that dictates the rules for transfer of ownership, time-limited ownership, death, loss of key, etc.

Unlike a paper contract signed in handwriting, this contract is signed with people's private keys, and the rules aren't enforced manually but automatically. And you can trust in the data and the enforcement of the rules because you can have trust in the network to do so.

> It just depends if you need a higher level of trust or not.

I need trust. And blockchain doesn't provide it.

> It just depends if you need a higher level of trust or not.

And yet, all you can come up with is "a bit of a stupid example obviously". The moment I propose to you actual real-world problems, your entire argument breaks down into piles of "could be"s and lots and lots of words how blockchain magically solves everything.

> you've got some global blockchain network that can record arbitrary data and run contracts attached to the data

Why?

> then you could start to record things like land ownership

- Ahmed pays local authority to record him as an owner. Where's your solution for that?

- Ahmed is an atual rightful owner, but then Mohammed pays local authority to register him as the rightful owner instead. Where's your solution for that?

> this contract is signed with people's private keys

- Mohammad is now registered as the owner. 5 years later, after a prolonged court battle, Zahra proves that the original record is incorrect, and the court rules that the property is now Zahra's. Mohammad doesn't give up his private keys. Where's your solution for that?

> dictates the rules for transfer of ownership, time-limited ownership, death, loss of key, etc.

Ah, yes. Instead of contracts written in plain language, we now have a complex programmable system written in an esotheric language that requires programmatic access and integration to dozens, if not hundreds of other such systems. And all these contracts are guaranteed to be correct, foolproof, without omissions, and to 100% correctly represent the interests of the parties that sign the contract. Because blockchain.

Well, except the time when they are hacked, and require a hard fork and re-writing of history. So much for "non-tempering": https://www.coindesk.com/understanding-dao-hack-journalists

> Unlike a paper contract signed in handwriting, this contract is signed with people's private keys and the rules aren't enforced manually but automatically.

Ah, yes. When a person dies, they automatically register their death from beyond the grave. Or when a person sells property, this is automatically entered into the system through the celestial sphere. Court orders are automatically applied to ownership simply because blockchain.

> And you can trust in the data and the enforcement of the rules because you can have trust in the network to do so.

Ah yes. The rules are enforced because some numbers in a highy ineffecient, slow distributed append-only log say something.

> The blockchain is just technology. It makes all this more convenient, and possibly scale better.

That’s an odd way to say “less convenient and slower”. Everything you described could be done faster, better, cheaper and in a distributed fashion using PKI – the important part is that a legal authority signed a particular document. The blockchain features aimed at working in environments where people don’t know or trust each other aren’t useful when the whole point is having a trusted third party verify claims.

Reply on news.ycombinator.com