Hacker News Reader

Notarizing documents is a great example where some people think blockchain is a good use case, and it's never true.

Where would be the advantage of the scheme you describe to the current system? There is none. The system already works: people are given power to validate documents by the authority. The authority stores that centrally. The validation is mostly collected centrally. All stuff you can do perfectly fine without needing to run replicated servers storing an evergrowing immutable blockchain sipping authority from the real authority.

The current scheme allows things like secret notarization and revocation, stuff a blockchain is very bad at.

> it'll be up to the receiving agencies like England's government to decide "We trust the system the US uses to validate it's notiaries, so we'll add 'USRoxs' to our list of accepted Established Authorities (along with 'TheEnglish' of course)"

Countries already have systems in place to accept documents made in other countries. Works with official translators and stamps/seals of authenticity.

In better and funny: Blockchains Are a Bad Idea (James Mickens), https://www.youtube.com/watch?v=15RTC22Z2xI

Notary services are expensive, and some notaries are corrupt. I'm no fan of bitcoin, etc. OTOH this does seem like a good blockchain application, and the fact that it doesn't burn a stupid amount of electricity is nice too.

I'm curious about secret notarization and revocation. Why is that particularly desirable? I guess it would be a poor fit for a blockchain indeed...

Not my area, but think of things like a will. You being a billionaire maybe do not want to let your murderous family know that you made a new one writing them out of your will.

> Notary services are expensive, and some notaries are corrupt

A blockchain being the place where the act of bureaucracy ends up being stored changes nothing about that.

> OTOH this does seem like a good blockchain application

Exactly. It does at first, because it's about authoritatively storing information, and isn't this what a blockchain is good at? But then you think about how the system actually works and what the requirements are. A blockchain brings you immutable and decentralized storage of peer-validated information. None of these properties are of any value here. We just need a centralized storage that does not forget values by accident, which is basically every other solid information storing system. And most of those are a lot easier to run than a blockchain.

> A blockchain being the place where the act of bureaucracy ends up being stored changes nothing about that.

This is where a lot of blockchain concepts seem to break down. In most systems at some point you need to trust the data entered in by humans. But how can the members of the system be sure they can trust other members? Maybe certification or auditing by a trusted third-party? Oh, we're right back to where we started.

> think of things like a will. You being a billionaire maybe do not want to let your murderous family know that you made a new one writing them out of your will

Hash the will. Store the signed hash on the blockchain. Then, no one knows what the hash represents until your lawyer reveals the will after you've died.

What do you gain if the lawyer has to reveal it anyway? If you don't trust the lawyer[1], just give a hash of the document to the people who would be affected.

I don't see what role the blockchain plays here.

[1] and who's to say there aren't multiple signed hashes on the blockchain anyway?

The chances of a document making sense as a will and matching SHA and MD5 hashes placed in a blockchain must be infinitesimal.

The blockchain places the production of the hashes in time; in this case showing that - all else being equal - this will supersedes any previous wills.

Proving things existed at a particular date is important for contracts, wills, IPR, and such.

It doesn't solve everything, you could have a collusion between relatives and lawyers such that they created multiple blockchain entries. But the blockchain would show, for example, if they did it after the subject had died. The relatives would also need to buy off those running the public blockchain, and not just the deceased's law firm, in order to get a legal attestation -- and audits would show the public servants lied. Auditors can be bought off too, no system involving people is perfect.

> I don't see what role the blockchain plays here.

> [1] and who's to say there aren't multiple signed hashes on the blockchain anyway?

The timestamping of the blockchain is the use case. If multiple signed hashes exist, the latest one supersedes all the previous.

> What do you gain if the lawyer has to reveal it anyway? If you don't trust the lawyer[1], just give a hash of the document to the people who would be affected.

I'm afraid I don't understand. Who would hold the original text of the document in your plan?

Edit: Nevermind, I see what you mean, here. I see the issue as one of public record. If there was a revision to the will which no one in the family liked, and benefited some estranged cousin of yours who lived in another country, the family and the lawyer could conceivably hide the new will and the fact that you were dead from this cousin. Having it all on the blockchain prevents this.

We've solved the trusted-third-party problem by having a trusted-third-party (i.e. the lawyer).

All hail blockchain as the most amazing non-solution to the problem.

At least in the US generally notary services aren't expensive. Most states set upper limits for the per signature fee that can be charged. Of those that set a limit California is the highest at $15. Of the states that have a set limit the vast majority set it at $5 or less.

You can also often get free notarization at your credit union, public library, local government, church or through your employer.

>Notary services are expensive, and some notaries are corrupt. I'm no fan of bitcoin, etc. OTOH this does seem like a good blockchain application, and the fact that it doesn't burn a stupid amount of electricity is nice too.

Seriously? Most US banks and credit unions have free notaries for their members. And otherwise is around $5 elsewhere for a simple document.

Private notaries allowed in US?

Notaries work differently in different countries; they emphasize different facets of a notary system.

In the U.S. notaries are traditionally decentralized, and a notary often maintains possession of their journals for years if not in perpetuity.

The primary functions of the notary in Common Law jurisdictions is the ceremonial and evidentiary value, not administrative value as in Civil Law countries. (All three functions matter in both systems.) Ceremonial because we can assume that when someone executes a notarized signature we can impute greater attentiveness to the contract. Evidentiary because the notary himself and the required third-party witnesses are (or should be) readily identifiable and available for questioning regarding the circumstances of the execution of the signature, including the authenticity and state-of-mind of the executor.

Plain signatures operate much the same way--we use them for their ceremonial and evidentiary value. It doesn't matter that a signature is easy to forge; the signature is a device for anchoring the act of consent to a concrete environment where we can begin asking and investigating pertinent questions as to authenticity and intent. In practice, the validity of signatures is rarely challenged in court. Either everybody agrees it's obviously or plausibly forged, or everybody agrees it's valid. But disputes regarding the circumstances surrounding the signature are endlessly varied. (Civil Law countries traditionally handle the problem by making notarized signatures the final and only word--non-notarized signatures have no value, notarized signatures cannot be repudiated. They sacrifice fairness for administrative efficiency. The trade-off between fairness and administrative efficiency is at the heart of most distinctions between Common Law and Civil Law systems.)

Public ledgers are not suitable substitutes for either of the above functions, whether according to the calculus of the Common Law or Civil Law systems. Arguably they're a step backwards. They're great for replacing all the other technical machinations of transactions except the most important ones--the ones that allow us to adjudicate disputes regarding intent and similar complex questions.

> The notaries can then add a record on the block chain saying roughly "I do notarization act X on document Y" (It'll be the equivalent of what they enter in their log books today and there will be some corresponding stamp that they'll stamp the physical document with, like they do today). They could now even notarize digital documents by including a hash of the digitized document in their record.

> Finally, the agency receiving the notarized document (digital or electronic) can easily go onto the blockchain and verify that the document was indeed notarized as claimed and that the notarization is not a forgery.

Or, they could just sign it using their key using old school public key cryptography. If you want to check if a notary is real just use certificate chains, just like for TLS.

What is unique to blockchain is trustless timestamping but timestamping services work very well in the X.509 world and are used for example for code signing (so that even if your certificate expires the signature is valid).

Now class, please remember. You don’t need a decentralized ledger to store a record of a centralized entity. How many times do we have to go through this?

Of cause you could make a blockchain based decentralized SaaS cloud service using Crispr to disrupt the established dogmas! You know to secure funding for your sure to become a unicorn startup.

Bog standard blockchain antidote, feel free to steal and reuse as needed (or disregard as usual):

If participants are willing trust some subset of other participants within the system for certain privileged tasks that are required by the design of this system, why would they not be willing to trust another subset of participant to operate data storage for the system? Note the potential that most any kind of datastore can underlie systems providing replication, operation of copies by multiple independent parties, consistency guarantees, availability through redundancy.

What potential byzantine faults exist in the design of the system such that providing fault tolerance benefits the system? Can those be eliminated by a better design? Can the risks posed by faults be absorbed or mitigated by the administrative/business/legal infrastructure for the system (which probably will have to exist, regardless)?

If byzantine fault tolerance is required, why is a ledger + immutable chain of history objects + Proof-of-* system the most suitable byzantine fault tolerance solution for the system, given inefficiencies (Cost of Proof-of-*, constraints on throughput), inconveniences (Immutability), and introduction of complexity which is auxiliary to the intended purpose of the system (a ledger of currency)?

Do participants have insufficient incentive to participate in the system such that providing them a balance of a currency is required? Can the system be redesigned to better incentivize those participants? If the system provides little incentive to participate in it from its primary purposes, is there any utility in creating the system at all? Does attracting participation from participants uninterested in the primary and ancillary functions of the system to speculate in the value of an auxillary currency pose a benefit to or liability upon the system?

If participants are motivated primarily due to the fashionableness of blockchain technology, do they realize it's not 2017 anymore, what will happen to the system when they catch up and get bored? Is there a way to use an "AI" instead?

The benefit of the model I proposed is that from the perspective of the agency receiving a notarized document, they only have to make a single API call to validate any notarized document. Their system becomes much easier to validate and trust.

Otherwise, they'll have to have one system for US, one system for England, one for India, etc.

How does blockchain provide that property to the system? Why couldn't Stripe launch an API for a notary in multiple markets and why couldn't India chose bitnotary while Pakistan chooses dogenotary?

If you mean "blockchain technology" to mean cryptographic ledgers than yes.

If you mean blockchain technology as defined by the original bitcoin paper, then why? What use does the system gain by allowing anyone to participate anonymously (e.g. money).

Blockchains solve specific problems. Cryptographic ledgers, like we already use for things like CAs, are the better choice almost always. Certainly they could be made more like blockchains in their openness, transparency, and participation. But going full blockchain isn't required.

However some people have been using "blockchain technology" to include the ideas of cryptographic ledgers, so I want to be clear.

Certificate Transparency implements similar aspects to what be considered a "blockchain" (to money hungry VC people), and it's a similar use case to being a notary.

I'm not sure what the blockchain part is for? Digital signatures + if you want a public list Certificate-Transparency-style logs (alternatively potentially timestamping services) seems like they'd do the same?

This is a reasonable point. The main benefit of this approach is that it centralizes the whole process (meaning, gives it a single Api) without making any one entity dependent on another entity for business critical functionality.

The issue with digital signatures is that if you're using it to verify physical goods then you have to have to store the digital signature somehow, either email it, print it out, etc and deliver it to the one who's accepting the notarized document. That person will still need to somehow get that signature, obtain the notary's public key, and verify the siganture. And if there are more than one agencies who's documents an organization accepts (think of embassies wanting notarized photocopies of passports from various countries) then that's a lot of systems you have to integrate with.

This blockchain approach simplifies that without requiring all these embassies/organizations to use a service offered by a single company.

Any idiot can create a blockchain though, so which one do you trust?

And you’re back to square one

As some one who usually hates on the blockchain, this actually sounds like a very reasonable use for it. It's a solid step up from the current system of embossed seals, the only potential pitfall it adds is some additional complexity (but I can very easily imagine platforms taking that on, ala DocuSign but for Notaries). My only concern is how this would fare out in an a court system that is constantly behind the times.

The only thing using a blockchain is being used for there is timestamping.

This can be done by multiple trusted third parties just fine.

This is also possible with SSL. The state/government being the CA, and you being a signed entity. Public crypto fits this mold really, really, well.

I'm starting to feel like I'm shilling Estonian eID (I'm not, just a regular citizen), but I think that system is the very least a cool proof of concept (if not a state of the art example) how public crypto can be used to sign and encrypt.

Don't you have to trust the state then?