Topic: Ripple or Bitcoin - page 53. (Read 34134 times)

Everything about ripple is a great idea except the stupid currency

It just screams "hey i designed this to make myself rich!!!!"

Essentially, Ripple solves the BG problem by having an intermediate "unable to agree" state and designing the system such that nothing terrible happens if some nodes are unable to agree. If a node isn't confident that it has an agreement with the majority of other participating nodes, then it doesn't report transactions to clients as validated.

but it will be the gateways that are exposed to all the legal and AML stuff I would have thought.

This would give a lot of power to the gateway, since it could create fake accounts.

I would have guessed you either have to buy XRPs to activate an account, or you get free XRPs but you must provide ID. Probably that will work through a gateway, to which you have to connect anyway.

Well, to be fair it might indeed not matter much.  Lots of ripple proponents advocate for it in opposition to bitcoin, which they see as a speculative asset.  They also seem to think that hoarding is bad, so a currency should lose its value with time.  Therefore people might not be trying too hard to acquire ripples, so the amount of "cheating" might be low.

I has to be tried anyway.

That's probably true, despite our best efforts. But really, how much does it matter? It will just mean the giveaway won't be quite as effective as it could have been because it won't get as many users. But we know people are participating in the giveaways just to sell the XRP anyway. It won't make Ripple any less usable as a payment system.

Good point.  It would also have solved the problem of initial distribution, which I still don't know how OpenCoin intends of solving.  Just saying "we'll give ripples to everyone" is not very serious, for if XRP gets any monetary value, lots of people will fake identities in order to get as many ripples as possible.   Instead of the CPU rule there will be a "cheating" rule.

Or OpenCoin could just bluntly sell ripples to the public, without giving any of them for free or at least not systematically.  But such a strategy would not seem compatible with an open-source project, as anyone could sell a similar product as soon as the source code is released.

I think it's a conundrum.  It's too bad because I really appreciate what they are doing.  We really need a distributed IOU market as it permits the decentralization of the exchange.   The software seems very smart and well written but there are a few aspects that seem to not fit together very well.

Let's not be intellectually dishonest. If there was any real motivation, Ripple could have been designed so that XRPs were produced as the result of proof of work. The XRP production schedule could be set to create 100 billion XRP over the course of 10 years at either a fixed or a decelerating rate, with the target difficulty adjusted to maintain the rate of production.

Of course, you guys objected to that because of the "wasted electricity" but unlike Bitcoin, Ripple's XRP production via proof of work would eventually end, with total power consumption dropping to zero (no more mining). After the source is released, anyone can make this change. The total power requirements for using proof of work to distribute XRPs using a stochastic process would be finite.

XRP Exchanges identical to Bitcoin exchanges (but hopefully not identical in terms of MtGox lag) would crop up that let you buy and sell these mined XRPs so that people can do things like open up accounts.

Now the drawback of this method is that there is no obvious way to fund the development of the software. I rather like that OpenCoin has a profit motive and a path to generate revenue. It ensures that we get good solid source code with lots of unit tests and juicy wiki / API specs / documentation. Plus, Ripple gets guidance via CEO and a board of directors who are doing the legwork to make the system successful, instead of a bunch of neckbeards whose idea of satisfying customers is implementing "M of N signature transactions".

Not any way that I know of. There's another thread where I solicited suggestions for a way to make that work and nobody came up with any.

The primary problem is that there's no known way to take Bitcoins off the Bitcoin chain such that they can be returned to it that doesn't require a central authority or have other significant drawbacks. There are a variety of other issues that were discussed in the thread where that was first proposed

That's not the only reason, but that's a very good reason by itself. Getting mass adoption is far from a sure thing. Being able to make it free for as many people as possible for as long as possible significantly decreases the chances that all the develop effort will be for nothing. Not to mention, XRP funds the development and ensures there's a healthy company standing behind the network.

Who would the BTC be paid to? Would those transactions have to take place in the Bitcoin chain? How would you ensure that someone couldn't game the system be being on both sides of that BTC transaction?

Since I don't think that's technically feasible, it's definitely not one of the motivating factors. There are too many unsolved technical problems to making that work, and I don't see what the pitch for this network would be -- "It's just like OpenCoin's Ripple network except there's is free and you have to buy Bitcoins to use ours" doesn't sound very attractive to me.

Joel, can you clarify that this is not one of the motivating factors for not open sourcing the code for the servers yet?

Is Ripple's consensus algorithm another novel solution to the Byzantine General's problem? If so, how? If not, why doesn't it need to solve that problem?

It should make it easier for them to get funding, which should help with development, publicity and ecosystem building. It would also allow them to stabilise the exchange rate to a degree, which could help enormously. I've heard that's what they're doing, but I don't know if they've made any official statements about it.

Pardon my ignorance as I try to wrap my head around Ripple, but couldn't it just use BTC instead of XRP? How would it not work as well? Is the only reason for it not working as well with BTC is that Opencoin wouldn't have a stock of XRP to hand out to promote the system? If that's the only case it doesn't sound like a good reason. If so I would assume that as soon as Ripple becomes open source a fork would be created with people opting to prefer their "free-ish" transaction fee be denominated in BTC as opposed to XRP?

Well the ledger contains the state of the entire system, including every account balance, every offer in the every order book, every contract, and every trust line by every person. In short, the ledger contains everything with the exception of the labels that you give to your contacts (those are kept in the encrypted wallet blob).

But they are not heavily promoting the concept of self-issued credit networks between light weight peers. The vast majority of users will have exactly one trust line, to the gateway of their convenience. Extending trust to multiple gateways can create a situation where payments can be routed "through" you, causing your IOU balances to fluctuate without your explicit action (although the sum of your balances would remain the same).

There is no way for trust relationships to be kept private - the client software needs this information in order to calculate paths.

I've been playing around with Ripple over the past day. It's an amazing system and I think I see how it could help revolutionize payment and also create a decentralized exchange. But one concern I have is about privacy. If my wallet will eventually contain information about everyone I "trust" and how much I "trust" them, I don't want that becoming public. Who would have access to this information?

Once the server is open source, we expect a large number of people to start running the server. As they build trust networks and the system becomes decentralized, we will have to establish a consensus before we can make any change that affects how transactions are processed. By intentional design, the system, once distributed, is very hard to change so that people can rely on it doing what they expect it to do. Right now, we are adding new features rapidly and occasionally fixing bugs that affect transaction processing. Once the server is open sourced, the deployment schedule will have to slow down dramatically.

The design isolates those portions that do require consistency (how the set of accepted transactions affects the construction of the next ledger) from those portions that don't (such as interactions with clients, pathfinding, most of the consensus process) so that as much of the code as possible can be changed and improved without requiring universal agreement. This is much the same as the way Bitcoin works -- you can change the RPC commands you support, the way blocks are stored locally, how you form your own transactions, and so on, without needing community agreement. But any change on the affects of a transaction or any case where you accept as valid a transaction everyone else would reject requires a community consensus, deployment of the new code to everyone, selecting a point at which the feature is enabled, and so on.

The algorithms have been tested in simulation under various conditions including large networks with realistic path latencies. The implementation has been tested in test networks and is currently running the live network. It hasn't been tested on thousands of nodes on multiple continents, so there's always the possibility of growing pains. Having our implementation massively tested at no cost to us is another incentive for us to open source the server as soon as possible. Finding and fixing any growth issues sooner rather than later is also a priority.

Well, I guess I missed that part.
 

Sure, but has it been tested??

Not yet, but we have committed to open sourcing it. The reasons it has not been done yet have been discussed extensively.

It's documented in the Wiki and we've answered any number of detailed questions about how it works.

On ripple.com/bitcoiners I can read:


It's difficult for me to consider this as an honest statement.  The server of the source code has not been published, has it?   Then we are not even sure their method for decentralization actually works.