[simias]

>So specifically, blockchain offers a totally accurate, reliable, secure way to track and exchange data across a multitude of distinct businesses with little opportunity for corruption.

This is genuinely amusing to me. I feel like in these discussions you can replace "blockchain" with "philosopher's stone", like it's some kind of magical element that solves problems using the power of dreams and internet memes.

Let's imagine that we're in the real world for a second (I know, lame). Now we're in a company that deals with fruit imports. We have a ton of contractors all around the world and we use the Holy Blockchain to deal with that.

One of our contractors adds a transaction to our chain saying "we're sending you the container #12 containing bananas". Then almost at the same time an other contractor in the same area says "we're sending you the container #12 containing lemons".

How does the Holy Blockchain, Peace Be Upon It, solve this issue? Do we use PoW? Does the contractor with the most powerful computer decides who's right? If it turns out that the lemon contractor wins somehow and gets its transaction on the blockchain but it turns out they lied or were mistaken, do we tell the customers that "the code is law" and they're just eating extra-bitter bananas?

I mean seriously, it's not magic, it's a bunch of data chunks linked through a SHA-256, it's not going to bring world peace and it's not going to magically make everybody in the world agree about everything. As long as bananas won't grow directly on the Holy Blockchain (May Satoshi Grant Peace and Honor on It and Its Blocks) then nothing will ever be totally accurate, reliable or secure. Unfortunately I hear that digital bananas are not very nourishing.

[askmike]

> PoW

I don't know any private blockchain solution in the supplychain space that uses PoW. There are more ways to form consensus.

> This is genuinely amusing to me. I feel like in these discussions you can replace "blockchain" with "philosopher's stone", like it's some kind of magical element that solves problems using the power of dreams and internet memes.

You can use any type of software, as long as it has one clear picture of what the current status is (a database would work, 100 different databases would not work). A bulk of commodity trading is a done on paper right now, because no one wants to trust a single party to hold this database. I am not familiar with fruits but I work in soft and hard commodities (anywhere from grains to beans to oil to metals) and everyone is looking into distributed systems to solve this exact trust issue. But maybe all these companies should hire you to build a mysql database instead I guess?

> One of our contractors adds a transaction to our chain saying "we're sending you the container #12 containing bananas". Then almost at the same time an other contractor in the same area says "we're sending you the container #12 containing lemons".

Those are two different transactions, I don't know any type of software that wouldn't compute that you get 12 bananas and 12 lemons? I guess I'm not understanding what you are trying to say.

[sanderjd]

They are saying that one party says the container with ID #12 has bananas and a different party says that container #12 has lemons, not that there are two containers, one with 12 bananas and one with 12 lemons.

[askmike]

Ah, well there is no system ever that will solve that problem. Blockchains are simply systems that allow for the (as much as possible) trustless formation of consensus. So as soon as you have fraudulence data entering your system you know exactly where it came from and also that everyone is relying on the same data. This removes the need for everyone to double check everything multiple times.

I am involved in commodity trade finance (the financing of supply chain) and the current process is basically a huge number of different companies all checking the same data (documents & contracts). The banks actually check these documents multiple times (4 eyes principle).

You can build an API around a mysql database that does the checking. But no company would trust your checking for compliance, legal and trust reasons. If you offer them a full node (piece of software that doesn't trust all data thrown at it) in a blockchain system that they can run on their own infra it's a different story.

[sanderjd]

> Ah, well there is no system ever that will solve that problem.

That was the parent commenter's point: the really hard problems are not ones that technology can solve.

Having said that, while I agree with that point, your extremely informed comments in this thread on the usefulness of a private blockchain for a real use case have been the best argument for the the technology that I've seen. It's an interesting space; I just wish I saw more of your kind of perspective instead of the constant hype and "crypto" speculation.

[Drdrdrq]

You are taking the argument too far, but then you probably know it.

It's just a "tamper-proof" ink for a reasonable price, nothing more, nothing less. It won't grow bananas or lemons, and GIGO (garbage in, garbage out) principle still applies.

How about if instead of coming up with the fictional scenarios that won't work, you instead learn about technology, ignore hype (and greed) and come up with the some that will work? Hint: they exist.

[taco_emoji]

Please try to track the context of an argument. You're ignoring the part where simias was responding to a post which stated that blockchain is "totally accurate [and] reliable", which is in complete contradiction to your view that it's GIGO.

[Drdrdrq]

You are misquoting:

> So specifically, blockchain offers a totally accurate, reliable, secure way to track and exchange data across a multitude of distinct businesses with little opportunity for corruption.

There is nothing said there about contents of the data, so the GIGO principle still applies, as always. Attacking blockchain that it doesn't help if input is garbage is thus pointless.

I would suggest you read the quoted statement again. It is very precise and true (imho of course).

[michaelmrose]

If its tamper proof ink can't we solve that by signing the document and emailing it?

[Drdrdrq]

You could, depending on your requirements. If someone needed proof that this was indeed signed and sent at that time, it gets more difficult with e-mail.