IOTA isnt trustless, it has a central Coordinator, which can and has censored transactions before.

Why dont they just use any SQL database? No one would be buying their coins if they didnt lie and glitter their Coordinator around with "the tangle" and "muh quantum safe crypto hash function which is broken by rookie level cryptoanalysis".

That is the "genius" behind IOTA or even Ripple. Once you discover that you can publish just a symbol to trade no matter what is behind (beyond a curated story) it is very tempting to start your own cryptoasset, even if the crypto aspect is empty. Just imagine what can happen if you can list your symbol in the NYSE or Nasdaq just putting a database behind.

It sounds like in-game currency without the actual game.

> why not store in a write only sql database?

Probably because the project wouldn't have seen funding without "blockchain" in the proposal title.

> Can anyone imagine ANY uses for a centralized blockchain?

Sure. A centralized chain can be used for anything that a decentralized one can be used for (with one exception [1]), you just have to trust the provider.

[1] The exception is that a decentralized chain cannot be used to fairly distribute rewards to the participants in a decentralized protocol. That makes it harder to use a centralized chain as money because you need a separate mechanism for controlling the money supply than the one that maintains the ledger. In a distributed chain, the same mechanism serves both purposes.

If you can trust the provider, why use an expensive blockchain instead of an inexpensive database?

Because the amount of trust required is significantly reduced if the provider uses a block chain. Without a block chain you need to trust that the provider will not change transactions retroactively. With a block chain you only need to trust that the provider will not censor transactions.

If it is centralised, i.e. one entity controlling write operations, you don't need a blockchain. See e.g. the "when to use a blockchain" flowchart on page 41 (53 of PDF) in the "NIST: Blockchain Technology Overview"[0].

Note also that many systems are sometimes erroneously referred to as using blockchains, for whatever reason (misunderstanding, marketing, etc.). Leaving aside the debate about how decentralised it is, the project you mention has a distributed ledger, but uses a Directed Acyclic Graph (which they call a tangle) rather than a blockchain. Similarly "permissioned blockchains" or "private blockchains" aren't really blockchains in the strict sense because they don't need a consensus mechanism (indeed many now call themselves Distributed Ledger Technology systems).

[0] https://nvlpubs.nist.gov/nistpubs/ir/2018/NIST.IR.8202.pdf

I can point you to a few examples of centralised blockchain designs that make sense, but one that is (imho) easy to understand is what Loom Networks uses.

Essentially, users money is kept in the public Ethereum blockchain so all the cryptoeconomic security properties are assured but the games are developed (and played) in a centralised side-chain using a technique called Plasma that doesn't charge users per tx and has very fast block times (~1 second iirc).

Plasma is a way of extending (for scalability or any other reason) a main-chain such as Ethereum with a new app-specific or industry-specific blockchain, in such a way that the users can always trust their money without having to fully trust the side-chain owner's best intentions.

Haber and Stornetta used one to hash and timestamp documents to prove they said whatever at a given time for legal purposes. Some of the hashes are published in the NYT. Satoshi built on their work in coming up with Bitcoin.

It turns out there are many instances where interacting parties don't have a good way to share transactional data. There's no actual technical reason they don't, they just don't (think...banks FTPing files containing EBCIDIC encoded emulated punches card records to each other+). For these cases, blockchains that are managed (PoA) look like a pretty great solution.

+I've no idea if this actually happens, just trying to paint a general picture of crazy band aid stuff that happens.

You’re describing a need for a standardized protocol not a blockchain.

> Can anyone imagine ANY uses for a centralized blockchain?

I haven't seen any compelling use for either a centralized blockchain (i.e. one entity controls committing the blocks) or a closed blockchain (i.e. access is limited to a group all of which have known identities). Only in the open, peer to peer, case does it seem to have any value at all.

I think there are a bunch of use cases, for example verifying authenticity of high-value products. For example, if you buy a bag from Louis Vuitton from some third party reseller or ebay, you could verify it against LV's centralized product blockchain by scanning/checking some code on the label in the bag.

> "you could verify it against LV's centralized product blockchain"

You can replace the word "blockchain" with "database" in this sentence and it'll still work, i.e. "you could verify it against LV's centralized product database".

Fair enough

I can do the same with my MacBook’s serial number at apple.com. You don’t need a blockchain for that.

I think you hit the nail on the head. The vast majority of blockchain ideas basically sub a blockchain in the role of a traditional database.

The "killer app" for blockchain technology is a currency.

The real question is whether it's good for anything else.