> From each participant's perspective they cost far less, both in terms of implementation and ongoing maintenance, than attempting to build and run a centralized database oneself
This comparison isn’t valid: you’re using a shared blockchain service with high transaction costs but then saying the comparison can’t use a shared service.
> 2. They do the precise opposite of adding reliability problems; they're in fact notoriously difficult to bring down given the sheer degree of data replication.
So you’re saying I can process blockchain transactions without a reliable, high-speed network connection? I never have to worry about fee increases or high volume impacting my clearance time?
> The comment to which you replied describes at length how they add considerable degrees of trust.
It describes a number of distractions from the core problem but doesn’t solve any of the hard ones. The fundamental misunderstanding you and the poster are operating under is that this kind of system is based on anonymity. Since the real world is not, you don’t need anything which PKI doesn’t give you far more efficiently.
> > If my inspector is being paid off, he'll do the same thing for the blockchain record that he was doing with paper.
> Which would then be public and uncensorable evidence against that inspector.
Just like it is currently, except with hefty transaction fees. The problem here is that you need someone to figure out where things are being faked in-person. A blockchain can't solve that because the problem happens in the real world and if you setup an oracle to inject that information you're just renaming that existing trust relationship, not removing it.
> > If things are being tampered with in shipping, the same people will do the same things and you'll have the same dodgy goods arriving at their destination except they'd now have valid but inaccurate blockchain records.
> And that inaccuracy would - upon detection - become public and uncensorable evidence against those tamperers.
Just like it is currently, except with hefty transaction fees. Again, “public” isn't relevant — all of the parties already know each other and if it goes to court they're going to produce records — and “uncensorable” is just meaningless blockchain marketing fluff because that's not a relevant problem or one which a blockchain effectively prevent.
In all of the cases I outlined, the problem requires real world checks to find who is faking the records. Whether those records are paper, actions tracked on a website, PKI signed documents, or transactions recorded on a blockchain is a rounding error on the difficulty of setting up the real world legal system and monitoring which actually prevent cheating. If I have a problem with counterfeit goods showing up, everyone involved is going to say that their records were accurate and the problem must have been somewhere else – all a blockchain tells you is that you paid more to store the receipt, not that it was accurate.
> This comparison isn’t valid: you’re using a shared blockchain service with high transaction costs but then saying the comparison can’t use a shared service.
The parenthesized bit immediately following what you quoted (in addition to other parts of the comment and its grandparent) specifically explains why the comparison can't use a shared service: who's going to run it while being trustworthy for all users of it?
> So you’re saying I can process blockchain transactions without a reliable, high-speed network connection?
You need one for any other database, especially a shared one as you suggested above. In fact, it'd need to be more reliable and more high-speed for a non-blockchain solution; even full nodes don't require continuous uptime (that's only needed for mining or staking, neither of which is necessary to query and post transactions), and blockchain transactions tend to be pretty light on bandwidth.
> Just like it is currently, except with hefty transaction fees.
1. Citation needed on "hefty transaction fees". There are blockchains other than Ethereum, in case you weren't aware.
2. You're forgetting (or perhaps deliberately ignoring) that it's considerably easier to "figure out where things are being faked in-person" when the evidence of fakery is permanently in the public record. There's also far less room for plausible deniability on the faker's part.
> “uncensorable” is just meaningless blockchain marketing fluff
You might be surprised to learn that traditional database records, unlike those on blockchains, are trivial to destroy or otherwise tamper with. The inability to cover up fraud by editing transactions after-the-fact is nowhere near as meaningless as you assert.
> Whether those records are paper, actions tracked on a website, PKI signed documents, or transactions recorded on a blockchain is a rounding error on the difficulty of setting up the real world legal system and monitoring which actually prevent cheating.
And setting up said real world legal system and monitoring is considerably easier when the digital records are auditable by pretty much anyone and effectively impossible to modify.
> all a blockchain tells you is that you paid more to store the receipt, not that it was accurate
A blockchain tells you that the receipt was not and will never be modified. If there's an inaccuracy, that makes it much easier to detect it. If there's a pattern of inaccuracies, that makes it even easier to detect it and quantify it. What would've previously taken weeks or months worth of wrangling records from filing cabinets or arbitrary databases can instead be done in minutes or even seconds - freeing up time for the actually-hard parts of such investigations.
This comparison isn’t valid: you’re using a shared blockchain service with high transaction costs but then saying the comparison can’t use a shared service.
> 2. They do the precise opposite of adding reliability problems; they're in fact notoriously difficult to bring down given the sheer degree of data replication.
So you’re saying I can process blockchain transactions without a reliable, high-speed network connection? I never have to worry about fee increases or high volume impacting my clearance time?
> The comment to which you replied describes at length how they add considerable degrees of trust.
It describes a number of distractions from the core problem but doesn’t solve any of the hard ones. The fundamental misunderstanding you and the poster are operating under is that this kind of system is based on anonymity. Since the real world is not, you don’t need anything which PKI doesn’t give you far more efficiently.
> > If my inspector is being paid off, he'll do the same thing for the blockchain record that he was doing with paper.
> Which would then be public and uncensorable evidence against that inspector.
Just like it is currently, except with hefty transaction fees. The problem here is that you need someone to figure out where things are being faked in-person. A blockchain can't solve that because the problem happens in the real world and if you setup an oracle to inject that information you're just renaming that existing trust relationship, not removing it.
> > If things are being tampered with in shipping, the same people will do the same things and you'll have the same dodgy goods arriving at their destination except they'd now have valid but inaccurate blockchain records.
> And that inaccuracy would - upon detection - become public and uncensorable evidence against those tamperers.
Just like it is currently, except with hefty transaction fees. Again, “public” isn't relevant — all of the parties already know each other and if it goes to court they're going to produce records — and “uncensorable” is just meaningless blockchain marketing fluff because that's not a relevant problem or one which a blockchain effectively prevent.
In all of the cases I outlined, the problem requires real world checks to find who is faking the records. Whether those records are paper, actions tracked on a website, PKI signed documents, or transactions recorded on a blockchain is a rounding error on the difficulty of setting up the real world legal system and monitoring which actually prevent cheating. If I have a problem with counterfeit goods showing up, everyone involved is going to say that their records were accurate and the problem must have been somewhere else – all a blockchain tells you is that you paid more to store the receipt, not that it was accurate.