Quote from: DumbFruit on September 24, 2013, 02:03:36 PM

Quote from: Sukrim on September 24, 2013, 10:06:42 AM

Your argument is now that it is hard to determine a proper amount to value this 1 BTC IOU. This is true however for any IOU and has been this way for millennia.

The difference is that for millennia the lender and borrower had a direct relationship and the lender could make an accurate determination of the risks he takes on when he lends by getting as much information as he can from the borrower. He can also determine the interest that would make it worth the trouble.
In ripple, this relationship is severed, making accurate determination impossible, rather than difficult.

Two consequences of this jump out at me;
1.) Trust lines will be determined by how much the lender is willing to lose, no questions asked, without notification, and without remuneration.
2.) People can and will abuse this system, and all of the most trusting/gullible will suffer the most from it.

As a direct result of this, trust lines -if they exist at all- will a combination of extraordinarily short and in tiny amounts.

You still have a direct relationship between your borrower (= your gateway) and the lender (= you). You have an account there, hopefully did due diligence in checking if they are actually paying out if you send them IOUs, checking if they are useful in your jurisdiction or your preferred currencies etc.

I am not sure about your terminology, with "borrower" I mean someone getting/holding an asset (e.g. a gold bar), and a "lender" is someone holding an IOU (e.g. a note saying "hand this note back at ... to receive 1 gold bar") for that asset that (s)he previously might have owned, but gave away to the borrower for later redemption.

Quote from: DumbFruit on September 24, 2013, 02:03:36 PM

Quote from: Sukrim on September 24, 2013, 10:06:42 AM

It is my belief that 1 BTC (on the blockchain, as an asset) has exactly 0 value ... BTC is not based on their inherent value but rather upon their usefulness as money and store of value.

Everything is valued based upon it's usefulness, not its "inherent value". When Austrian Economists talk about intrinsic value, they are only referring to the intrinsic properties of the item in question that has value, subjectively, to certain people.

Then why does the price of BTC change that much if not that much happens in the BTC-ecosystem? Why do bubbles happen if everything is valued upon usefulness alone?

Quote from: DumbFruit on September 24, 2013, 02:03:36 PM

Quote from: Sukrim on September 24, 2013, 10:06:42 AM

Your argument turned around would be the following: The inherent flaw of digital assets like XRP or BTC is that they cannot be priced by the owner, as they are worthless by themselves...

No, but a different phrasing of my argument would be;
IOU's cannot retain value as they are passed along a Ripple Web of Trust because the terms of the loan, which determines the value of the loan, is vaporized by the first Liquidity Provider.

I have terms only with my personal gateway and will only accept/receive IOUs denominated by them. Kinda like the bank that I have my account at. Even if someone from a bank whose name I can't even spell and whose currency I haven't even ever heard of sends me money, I will end up with some balance on my bank account, denominated in EUR, which means my bank now owes me a few EUR more.

I don't know your bank, I don't care about your bank and I certainly don't care about having any contracts with them. Still its IOUs might be worth something to some people. As long as you hand them to someone else at the same bank, they would likely value them at the same value as you do. You do nothing else when handing them to a liquidity provider, which is just a person who has 2 bank accounts - one at your bank, one at mine. He takes your balance and hands you a balance that I accept. That might be done for free or at a fee. Still there is nothing radically new or different about this, it has worked also for millennia like this when travelling to other countries that have different currencies for example.

Quote from: DumbFruit on September 24, 2013, 02:03:36 PM

Quote from: Sukrim on September 24, 2013, 10:06:42 AM

The solution to both your arguments about rating trust in IOUs and the opposite argument about uncertain pricing of assets is also known for millennia: an open market where people come up with prices and hopefully also match these from time to time based on both their needs/beliefs and other market participants.

That doesn't work because the mechanism by which lenders can make these determinations in the market do not exist in Ripple over Liquidity Providers.

As long as you are lending out e.g. BTC to only one borrower (e.g. Bitstamp) and trading only there, why would you care about anyone else? As soon as you need to pay in different IOUs (e.g. in CNY.RippleCN), you need to trade on the open market and that's where people come into play who take the risk or work of trusting more than 1 entity and can get a fee for that service. Still I could care less if someone only accepts 100% fraudulent BTC (e.g. by TradeFortress), as long as there is a way to trade mine for these, it will be found and the cheapest one will be used by me. I never get exposed to anything and the only person(s) that might get burned are the liquidity provider(s) inbetween that willingly trusted TradeFortress and the person that accepted them in the first place.

I am still unsure what your actual criticism is - maybe it would help if you scetch a small network with 2-3 gateways, users, liquidity providers etc. and point out where what according to your theory goes wrong. As it seems somehow you think gateways might be able to run fractional reserve: At least in the EU issuing e-money is a VERY regulated business and you are definitely NOT allowed to run fractional reserve at any point of time.

Quote from: DumbFruit on September 24, 2013, 03:58:33 PM

Imagine that the IOU was a written contract that is literally passed down line of people.

The contract says, "I [name] owe [name] [principle amount]".

No, the contract says: "I, [gateway_name] give anyone who hands me this note [amount] of [currency]"

On Ripple itself you only state "I [user_name] accept up to [amount] of notes in [currency] from [gateway_name]"

This is all you will end up with as a balance and the only IOUs you are concerned with, are the gateways you are trusting. I don't see how a liquidity provider (who is simply someone who is a customer at 2 gateways) would "break" anything in there.

Again, please provide a small sample network that clearly demonstrates your point of failure, as I fail to see it honestly.

Quote from: timeofmind on September 24, 2013, 04:28:18 PM

One major drawback of Ripple is that it relies on clients to maintain lists of nodes that are trusted not to collude. Bitcoin is not reliant on this assumption that nodes are not colluding. I could see a scenario where the NSA forces some of the most commonly used nodes to collude in order to comply with the law; even if users would suspect such nodes to have nothing to do with one another. In bitcoin, the NSA would have to take over the mining power. In Ripple, the NSA would just have to force a limited number of nodes to comply, since in order to ensure non-collusion between nodes, the client must choose from well-known nodes. If a client was to simply choose from a random set of nodes, there is no assurance of non-collusion. Ripple claims that this does not require trust, because we only need to be sure that nodes are not colluding, but it basically amounts to a dependence on a set of trusted nodes, because untrusted nodes cannot possibly be trusted to not collude. ie. Where are you getting your list of trusted nodes from?.... it has to be a trusted source. If you take your list of nodes from a public pool of nodes, then there is the potential for someone to contaminate this pool with a large number of colluding nodes.

This "limited number" is about 80% or even more of all Ripple nodes. This number is already RIGHT NOW higher than the number of servers you need to take over to severely mess with Bitcoin (which is about 2-3 pool servers). Also the NSA does not police the world - they would need to break into a lot of servers at the same time to pull this off. It would be easy for example to put only a handful of servers from the US into your UNL and mostly trust servers stationed in other jurisdictions. Also collusion is quite visible on the network too, if only a subset of servers always insists that foo_suspicious_transaction is included... You could for example have some "staging" area where you put NodeIDs of untrusted servers and watch them for signs of collusion if you want to download a huge list of random nodes from the net.

I agree that choosing a proper UNL is difficult, however this is far from impossible and if for example Bitstamp signs their NodeID with their issuing address, this is a VERY strong sign that this message is actually coming from them. Choosing a random subset of all known NodeIDs is of course not the smartest thing to do as you said, clustering them according to certain criteria and choosing some representative NodeIDs from these clusters however might already improve security a lot. Also bringing online a huge amount of nodes all at the same time might work once (when the source code gets released) - after that it would already raise a lot of suspicion too.