Thanks for your input Sukrim! I'll try to cover all your points in line below...
Creating these "cryptobonds" (=fiat IOUs) is VERY likely a regulated activity in most countries and not suitable for anyone without an e-money issuance license.
fiat IOUs would certainly be regulated if they were regulateable. Vendors are anonymous, and all transactions are done person to person with no official businesses involved.
Also I don't really get the backing aspect - cryptobonds have BTC values attached to them (how do you ensure there's no copy of the private key at the issuer? trust?)
Cryptobonds have actual wallets INSIDE them, so it's actually more like gold than an IOU. But the question of the issuer not stealing the gold is valid, so every DEX client that has a cryptobond in it, will constantly watch those wallet's values to ensure that there is the proper amount of BTC
at that address.
It's not trust, it's verification... And the penalty on vendors for failing to verify is high.
that are dependend on the current BTC/USD valuation, e.g. today there are 100 USD for 1 BTC, so I attach ~0.01 BTC to a 1 USD note. Then BTC crashes to 10 USD per BTC - my note now still says 1 USD but is only backed by 0.1 USD in BTC.
I wrote a series of reasons in the whitepaper why Vendors would still redeem at a loss. The most important one of course will be the Rating, but you should also consider that many or even most DEX users will be buying cryptobonds to cash them out, making their lifetimes very short. This means that a very high percentage of cryptobonds will be created and redeemed within the same day or thereabouts, and therefore the price won't have much time to fall.
It's not that vendors will always expect this; but they won't have control over what happens to their cryptobonds out in the DEX network and therefore they had better be ready to sell 100 today, redeem 65 of them tomorrow, and redeem the rest randomly over the next few years.
The NashX project tried to circumvent this by requiring actually higher than 1:1 backing (e.g. 1 USD note needs to have 2 USD in BTC attached to it, or even more) to ensure that not allowing someone to redeem the USD also hurts the issuer more.
I had considered this option, but having the issuer put up so much money is unsustainable for business. It's nearly impossible to grow business that fast.
The fact that every cryptobond ONLY has $1 or so inside of it makes them seem Ridiculous to risk losing his rating over.
Still as BTC floats freely and there have been plunges from high to low values that were quite significant, it would be quite expensive to issue USD that way (requiring ~1:10 backing) and you open up other problems like people destroying bonds/keeping them indefinitely (already in your example the issuer risks 102 USD in BTC for issuing 100 USD - he wins if nobody redeems and BTC go down, so he is going short on BTC and he looses if people redeem when BTC go down, so he is also long USD).
That's what the bond aftermarket is there for. Whichever bonds the vendors feel are no longer desirable, they can dump at the bond aftermarket for their desired price. Meanwhile speculators are always able to buy them there, speculating that the market will go up.
All in all fully (or even "overfull" e.g. more than 1:1) backed USD bonds are going to be issued by bears, lower than 1:1 backed USD bonds would be issued by bulls.
They get paid a fee up front, and at redemption, to issue 1:1 bonds on any currency. They can trade it away in an aftermarket if they don't like the outcome, either way. I feel it's a suitable balance.
How does one verify that one encrypted piece of data attached to a bond actually contains the private key?
That's up to the original vendor to ensure. He will be labeled a scammer if he issues bad bonds, and lose his business instantly on the day someone tries to redeem one.
About "ratings will force issuers/merchants to keep honest": pirateat40s OTC profile was spotless, on SilkRoad afaik one merchant took off after being one of the highest rated and highest volume merchants, in the "real world" banks can fail...
It's true, this will happen from time to time, I'm sure.
But in DEX the system doesn't come crumbling down when some vendor takes off into the night. His bonds will become non-redeemable for now, but they don't disappear yet... There is still the chance that he could return one day and recollateralize them, so we don't destroy them, but instead demote them to be sold back to the bond aftermarket somehow. (Still thinking about the details of that part now actually.)
The main thing is that these bonds will all be mixed as they are traded around the web and no one user is likely to get stuck with hundreds of uncollateralized bonds at once unless he was stupid enough to go buy it from someone in the fiat marketplace with a bad rating.
I think we can set the fiat marketplace to not even show any vendors with a rating less than 9 out of 10 as a default, to keep noobs protected. There are other filters we can use too, and of course this sets the bar high for who will make a good business out of selling cryptobonds.
I don't believe in trust systems keeping people honest or that past performance does mean in any way that this entity is honest, doing good business or will pay out in the future.
You just haven't seen it done right yet.
Clearly I have more work to do showing how the vendors will be held accountable. Thanks for bringing up these points, I'll work on addressing them more.
Last but not least, I'd love to see a chapter about comparing your system to the current implementation of OpenCoin Ripple, where do you differ (e.g. attaching collateral to IOUs), what are your main benefits over their approach with an integrated native currency and so on.
Oy, doing that properly is not going to be easy... More so for us authors than for any 3rd party.
If someone would like to make the comparison on a chart I'll be happy to answer any questions about my system for them. That would be an awesome project for someone who has quite a lot of room in their head. (Way more than I do.)
To me the only real difference seems to be that you still want to do mining and attach collateral to IOUs.
There isn't a lot of mining in my system... Bitcoin miners are already mining to process safety for the traders in DEX, and vendors process safety for themselves... But I almost didn't use the word mining for that processing because afterall, they create their cryptobonds whenever they feel like it and have collateral to do so.
The most major difference between DEX and all other systems is clearly the incentive structure. Your grandma can get paid to simply download it and let it run.
The second major difference is that it'll be Apple-product easy for your grandma to buy her first bitcoin through it.
Third? Distribution. With anonymous vendors, no government could ever clamp down on any bottlenecks of DEX, and therefore Bitcoin and all alt-coins and all cryptosecurities!
Ripple has two gateways at present. Even if they had thousands though, governments can point a gun directly at every one of them. This is a tragic flaw that wasted ever last second of their work making ripple from the start.