But why would you need a persistent shared ledger, and not just a plain old database?
The point of blockchain is to establish trust via technology when the participants are not trustworthy. Is DMV not trustworthy? If so why? I have a driver's license, and it never occurred to me not to trust DMV, and I never heard anyone else being worried about that either.
Devils advocate, because increasing isolation of the necessarily public information a governmental service manages from the government organization servicing it is a good thing.
The role of title transfers could itself be argued to be a form of governmental overreach. For the mere act of interpersonal buying and selling, why ought there be a database managing that?
If we agreed such a database must exist (and horses and carriages seem to have done fine without one), wouldn’t it be preferable for individuals to be able to directly interface with that system interpersonally, rather than through a governmental intermediary?
If we agree there must be a governmental intermediary, wouldn’t we prefer such an intermediary to solely pay for and maintain dumb infrastructure they are not expected to waste taxpayer time and money manipulating, rather than giving them control over the underlying transactions and wasting a great deal of both?
> The role of title transfers could itself be argued to be a form of governmental overreach. For the mere act of interpersonal buying and selling, why ought there be a database managing that?
So the gov’t can make money. When you sale the car the new owner must pay tax.
The advantages start popping up when multiple DMVs are using the same ledger; they don't have to trust each other, yet can still read and write records. The end game would be for state DMVs, international DMVs, and even private entities to be able to share information on a vehicle and its history - think CarFax, but much broader in scope and without some corporation acting as a middleman.
In this sense, I'm a bit skeptical of the value add of California's implementation, despite "put car titles on the blockchain" being something for which I've advocated for years now. This would've made more sense as a multi-state partnership.
> The advantages start popping up when multiple DMVs are using the same ledger; they don't have to trust each other, yet can still read and write records.
What do you mean different DMVs don’t have to trust each other? If California DMV can’t trust data written by Nevada DMV, what’s the point of writing it in the first place?
> The end game would be for state DMVs, international DMVs, and even private entities to be able to share information on a vehicle and its history - think CarFax, but much broader in scope and without some corporation acting as a middleman.
Okay, but who is responsible for verifying all of the data coming in from random 3rd parties? If you wanna get a good deal on a used car being sold, just submit a fake entry about it having been totaled in an accident. Instant discount!
And not to mention all of the PII that’s core to DMV’s operations that can’t be public, so it can’t be on a public blockchain anyway.
I think we're talking about different senses of the word "trust". I'm more referring to it in the sense of "I trust you to competently administer this database and not tamper with its data" rather than in the sense of "I believe the the data you're committing to be accurate".
The former is what a (public) distributed ledger solves; a central authority is problematic even in interstate contexts, let alone international. Blockchains don't directly solve the latter, but do provide tools for tackling it - namely, by making the submitted data trivial to audit. Using your example:
> If you wanna get a good deal on a used car being sold, just submit a fake entry about it having been totaled in an accident. Instant discount!
Setting aside the immediate footgun here (since you've nerfed whatever resale value you'd hope to get out of it), your single datapoint is less believable than the concurrent agreement of:
- A police department noting "a vehicle with this VIN was involved in an accident"
- An insurance company noting "we paid out an insurance claim for a vehicle with this VIN"
- A towing company noting "we towed a vehicle with this VIN to a repair shop"
- A repair shop noting "we serviced a vehicle with this VIN"
In short: there are a lot more opportunities for verifying information when it's on a public ledger than there are when it's stuck in centralized databases.
> And not to mention all of the PII that’s core to DMV’s operations that can’t be public, so it can’t be on a public blockchain anyway.
Nobody said literally everything the DMV does should be on a blockchain.
In this context of "CarFax but better", very little (if any) of that PII needs to be present in the ledger - and certainly not in plaintext. The core datum here is the VIN, and that's already far from private.
> I'm more referring to it in the sense of "I trust you to competently administer this database and not tamper with its data"
A well-administered non-tamperable database of garbage data is still quite useless.
With an unpermissioned public blockchain you can make up your own police department, insurance, towing company and repair shops. Make some wash entries to make them look legitimate and you’re back to square one of having to confirm everything off-chain to figure out what’s real.
Or drown a real report in a sea of fakes if you’re trying to offload a lemon, since there’s not much you can do about spam.
Not to mention the fundamental problems around private key management.
I don't trust the DMV. It's not that I think they're out to get me, it's that some crappy system on their end always screws something up and locks me out of whatever it is I need.
Blockchains that do not have a built-in motivation for all member nodes to comply with the protocol in order to keep the data on it secure and reliable, these are the ones that require motivation in the form of a lottery token assigned some speculative value[0]. This is the case for all intentional (and unintentional) cryptocurrencies and cash-redeemable blockchain tokens.
But if there is a motivation inherent to the value of the data itself, for example in the case where a common ground needs to be constructed from a number of otherwise-unaffiliated private or semi-public parcels of terrain, and the usefulness all of the parcels increases through the construction and maintenance of this common ground, then no simple financial motivation is necessary for all of the participants to agree that none of the participants should be the sole or majority custodian of this common ground.
In this case a blockchain can actually make some sense, although it's arguable that it is still a needlessly wasteful way to achieve this goal since blockchain security is derived from triggering a competitive resource arms race between all members, which collectively makes tampering by a non-member impractically expensive to attempt and highly likely to destroy the whatever value or purpose the outsider wishes to usurp even if they succeed.
That said, the fact that this record is being established as an asset on top of a garden variety pyramid scheme lottery token blockchain, tezos, rather than on a club blockchain whose tokens have no presence on any exchange, undermines even this hypothetical and charitable interpretation of this project.
[0] disclosure: "assigned" and "value" in this space mean very different things to a blockchain believer and a skeptic. My username would have the word Block prepended to it if HN allowed a few more characters. So this opinion is coming from someone decidedly on the skeptic side of this issue.
They can and they have bungled non-trivial subsets of the data and records they are entrusted with in the past. Having a public non-DMV source of truth would really help the consumer from getting the shaft in these cases. Of course, doing that in a useful way would likely violate people's privacy to the extent it becomes a non-starter.
That said, I think this is gonna wind up being a lateral move. Best, best, best case "blockchain" was just the magic marketing word they had to use in order to finally get management buy-in to shit-can some problem vendor or team responsible for the systems "blockchain" is gonna replace.
You can use your car as collateral for a loan. If you don't repay it will automatically claim ownership of title. Basically the same thing that happens now but with no middleman or paperwork.
On new IT initiatives? California DMV? The one that in the laat twenty-five years has launched, spent tens of millions on, and then abandoned as hopelessly flawed an effort to replace/update its core systems. Twice?
A distributed ledger doesn't solve that problem. Just sign the updates to a central database with the ID of the updater. We (people who aren't crypto scammers or scammees) have known how to make audit trails for decades now.
From an European perspective the idea of a government registry tracking vehicle ownership seems downright bizarre. Our DMVs generally just track keepers and not owners.
Vehicle registration has nothing to do with ownership, you don’t have to own a car to register it.
In Europe we don’t have titles for cars like in the US. You don’t have to perform a weird ritual of signing over the title to sell a car, you don’t need a written contract at all.
19 states have electronic titling without a blockchain.
As long as you have a central authority, you don’t need the blockchain. You do need robust identity verification and controls systems (IAM and audit logs), but the title and liens can still be records in a database. Want to look up if a title is legit? Offer a public page that accepts a VIN and spits out title details, no different than public real estate property records.
Solution looking for a problem. This is someone’s pet project. Someone commits fraud? You have an audit log and verbose identity and transaction documentation when you need to go to court or reverse transactions.
> 19 states have electronic titling without a blockchain.
Which means 19 states have their own databases, and exchanging data between them entails 19! (at minimum) separate ad-hoc connections doing a bunch of ETL work to massage between subtly-different database implementations.
> As long as you have a central authority
And who is that central authority? Nevada is not a central authority over Wyoming.
The federal government could be that authority, but that just kicks the can down the road; the US (despite its best efforts) is not a central authority over Canada or Mexico or wherever else. Do we have the UN establish a global DMV?
A distributed ledger sidesteps all this. Wyoming and Sweden don't have to trust each other and/or some third party when they can just post their records to the same chain.
The point of blockchain is to establish trust via technology when the participants are not trustworthy. Is DMV not trustworthy? If so why? I have a driver's license, and it never occurred to me not to trust DMV, and I never heard anyone else being worried about that either.