*old* BitShare Economic Theory 10 BTC bounty to prove me wrong... paid. - page 12.

I do not think that you structured this very clearly. And this is causing errors in your own beliefs.

First of all, the best way to characterize this is in fact a true open source distributed "ripple" based on the bitcoin codebase. Another definition would be a "native" colored coin blockchain. Since colored coins is frankly awkward, I have been considering a blockchain that accepts multiple currencies deeply. What is blocking me is frankly that I'd want bitcoins to somehow be a "native" part of the solution, not an external currency that requires a backer.

Point 1: You somehow think collateral makes your USD not need a gateway. Uh, No. There's really no nice way to say it but your thinking here is flawed. Essentially each individual is a mini-gateway, securing the genesis of your "crypto-USD". And BTW, nobody who really needs a loan can put up that kind of collateral... people put up cars and houses because they can simultaneously USE them. A loan is truly underwritten by FUTURE the earnings potential of the person who received the loan. Irrespective of your blue-sky scenarios, when black monday comes around, the people who received the $1 USD and essentially "created" the 1 crypto-USD have to make good on that and convert the crypto-USD back to USD. Or be sued to get it.

Fundamentally, there IS no crypto-USD unless its the US government issuing it. There is only someone's individual promise to pay. And as we learned in the 2008 mortgage crisis it can be REALLY BAD to "bundle" these promises together under the assumption that they are equivalent!

Ok, now that we've got that out of the way, lets structure the system.

Base:
cleaned up bitcoin codebase with a single "native" currency, let's call it "thecoin" (THC)

new transactions:
1. Coin-type genesis. A coin type is represented by a public key. URL and hash of a coin contract are included. The private key controls that coin type. Could require a nontrivial txn fee of THC to miners, not to just one miner but distributed across the next 1000 mined blocks or something to reduce coin type spam. However, just like there can be essentially an infinite # of bitcoin accounts, so with coin types. This is good because as I was saying above, you can't combine fiat promises together.
2. Coin genesis. Signed by the private key, this allows new currency to be minted. (THC txn fee)
3. Coin destruction. A transfer to the coin type's public key is essentially coin destruction because it gives control of the coin back to the issuer.

Transaction changes:
4. Of course, all txns can have multiple inputs and outputs that take different coin types (coin type public key is of course included in the TXO). This allows atomic exchange.
5. A new transaction type called "market-based exchange" that is essentially the exact same as a normal txn but just marked as an exchange between anonymous parties (see below).

Exchange:

For some reason you feel that bids/asks should be encoded in the blockchain. There are huge issues with that, including blockchain bloat and matching speed. But there is no need. Bids and asks (actually they are ALL bids, just offering different currencies if you see what I mean) do not need to be remembered forever.

These could be a special half-signed txn specifying what is offered and what is asked in return. These are propagated throughout the network. Someone trades by supplying and what is requested and signing, making a full txn. The issue here is the half-signed txn cannot easily be withdrawn. The only sure way to withdraw is to "double-spend" the underlying TXO so the $ the half-signed txn represent does not exist. So perhaps bids have an "expiration block number".

Another choice is an unsigned half-txn. The only issue here is your client needs to remain on-line to sign the txn if someone matches it.

Matched txns are committed to the blockchain, with a fee in THC paid by either/both parties. Multiple matches against the same half-txn are resolved when the next block is found -- whichever full txn makes it in that block wins. Note a higher THC fee would encourage more miners to use your trade (txn) as opposed to someone else's, which might make things very interesting...

Another issue would be bid flooding to attempt a DOS. This could be handled to some degree by miner's fees -- essentially clients would refuse to forward the bid if the TXin does not offer enough THC as a miner's fee. Clients can't tell the size of the bid because they don't know the value in USD of the coin-types being traded. So we don't know if a txn is "dust". So the miner's fee essentially needs be used to eliminate dust txns. Clients could also refuse to forward half-txns that are far outside the trading range of currency-types as known by looking at txns committed to the blockchain and marked as being an "market-based exchange".

Another possibility is proof-of-stake. You sign the bid half-txn with an account holding at least M THC, clients only hold N bids from a particular signature at one time.

Done! Easy! :-)

Topic: *old* BitShare Economic Theory 10 BTC bounty to prove me wrong... paid. - page 12. (Read 10026 times)

Topic: old BitShare Economic Theory 10 BTC bounty to prove me wrong... paid. - page 12. (Read 10026 times)