Den 12 feb 2015 15:53 skrev "Mike Hearn" <mike@plan99.net>:
>>
>> > So you're just arguing that a notary is different to a miner, without spelling out exactly why.
>
> I'm afraid I still don't understand why you think notaries would build long term businesses but miners wouldn't, in this model.
>
> I think you are saying because notaries have identity, brand awareness and because they have big up front bonds, that means they will be trustworthy.

Miners aren't contractors, they don't have to care about repeat business. Individual miners don't have enough impact to have a negative impact on their own capital investment. Zero-conf transactions also aren't that tied to the Bitcoin valuation.

Multisignature notaries need to convince people to select them. They want to know that even with collateral, their funds won't be temporarily locked up and unspendable for days at a time.

What services would miners provide here, do you think?

> Well, sure. It's the same model governments use and is why being a money transmitter in the USA is so difficult: you need to put up large sums of money as collateral and have your fingerprints taken 48 times. Then you can start advertising to get customers!

Obviously you need to have collateral to provide collateral. Can't make cryptographic verifiable guarantees if you don't have the resources to back them.

> The reason mining is such a nice model is it doesn't have these sorts of requirements.

And also can't make these assurances. Any minority miner can be overrun.

>> As notaries can be small operations ..... [snip] ...... (almost every large organization in the world have some unallocated funds somewhere).
>
> Which is it? Are notaries small operations or large operations? 

The operation itself is small. A few people maintaining a few servers.

The collateral needed depends on how many and how large simultaneous transactions they want to provide assurances for, so they can chose to be a small player for one niche market or large and global if they have the funds for it.

> I think exploring new consensus models with semi-trusted notaries is interesting, but it's not Bitcoin.

Methods for decentralized consensus that aren't PoW also aren't Bitcoin.

> Please don't try and apply this logic in the real world :( Rephrased:
>
> "That's a nice house. I noticed it's made of wood. I'm going to start fires until it burns down, because there is no guarantee your house won't burn down in future and it's important you understand that wooden houses aren't safe. Really I'm just doing you a favour."

Actually that IS often a bad idea. But fortunately the risk and threat is low, and mitigation is well understood.

> I'm really not a fan of Peter's approach, which is "hey let's try and cause as many problems as possible to try and prove a point, without having created any solutions". Replace-by-fee-scorched-earth doesn't work and isn't a solution. Miners can easily cut payment fraudsters in on the stolen money, and as they'd need to distribute custom double-spending wallets to make the scheme work it'd be very easy to do.

Security analysis requires having the mindset of an attacker. Sometimes that reveals suboptimal choices. Then you want them changed to more stable choices such that once the incentives change, the risk already is gone. Minimization of damage, simply put.

>> Your also ssume people will expect the Bitcoin network to keep zero-conf safe forever and that Bitcoin valuation is tied to that. Given the options available and current state of things, I'm assuming that's wrong.
>
> Why? You think ability to make payments in a few seconds is some irrelevant curiousity?

No. But you can't be certain it is secure without having a solid reliable mechanism to provide such a guarantee.

You want zero-conf to stay safe without involvement of servers? Then please, try to find a way to secure it. Right now you're assuming it can remain safe based on circumstances which can change and assumptions about market participant's valuations that likely aren't true.

> Let's put it this way. If BitPay's business model evaporates tomorrow, along with all the merchants they support, do you think that'd have any effect on Bitcoin's value? If not, why not?

It would. They'd tank. But you're assuming too much about the basis for valuation.