The ripple name was bought to exploit the positive impressions people had of the idea and was slapped onto something entirely different and unrelated.

https://bitcointalk.org/index.php?topic=144471.msg1547843#ms...

Bitcoin's lightning is in many ways the spiritual successor to the original Ripple idea-- a pathfinding network between pairwise relationships, but instead of IOUs, the parties use smart contracts to twiddle funds around in bitcoin payment channels without having to take the updates to the public network.

This fixes the big security weakness of Ripple while preserving the positive scalability properties. Unfortunately it also requires more complexity and limits flexibility somewhat.

I could be wrong, but I always understood the primary purpose of XRP as a mechanism to increase liquidity. Since Ripple is all about cross-border payments, you're potentially dealing with hundreds of currencies. So if you were a bank that needed to do global business, you'd be required to hold a subset of each traded currency in nostro/ vostro accounts. XRP was then meant to be the "standardized" currency that member-banks in the Ripple ecosystem could transaction with.

Since you as a bank would only need your own fiat currency + XRP, you could materially cut down on additional costs from holding all the other accounts. When it comes time to convert currencies, because Ripple (XRP) transactions settle in seconds, the period you're exposed to market volatility is virtually zero, so you can move in and out of a position in XRP in exchange for whatever other currency you need almost immediately.

I don't see how adding a currency makes this easier. Even if I don't want the currency you have, I want XRP less.

Adding a currency that everyone can agree upon as the "intermediary" prevents you from needing to hold huge sums of random currencies. For example - you're a bank in the UK so your fiat currency is GBP. If you want to manage transactions within just Europe, you would need to have Euros, Croatian Kunas, Danish Krones, Swiss francs, Polish zloty's, Russian rubles, Turkish liras, etc. Therefore you as a bank are subject to the ongoing valuations of these currencies, not to mention you've got hundred of thousands of (insert currency here) tied up in these accounts. As a bank, you have to correct for this inefficiency so you produce conversation rates that are not consumer friendly, but necessary to recoup the money you've lost setting aside all those accounts.

With XRP, you replace all those currencies with some amount of XRP. Because all member-banks use XRP when you need to transaction, say GBP to EUR (simple example), you would take your GBP, convert it to XRP, then use that XRP to buy EUR from another bank, that bank then can hold the XRP or chose to immediately convert it back to EUR through other banks. Like I mentioned previously, that whole transaction takes a few seconds, so you're effectively not subject to market volatility that could impact the transaction itself.

> you would take your GBP, convert it to XRP, then use that XRP to buy EUR from another bank

But I could more easily just use my GBP to directly buy Euros, with one less exchange rate and one less transaction in the process.

> that whole transaction takes a few seconds

For a bank, that's slow. But even if it wasn't, using an intermediate currency simply takes twice as long for each step.

> But I could more easily just use my GBP to directly buy Euros, with one less exchange rate and one less transaction in the process.

In this scenario, the transacting bank would need to have (or want) to have GBP for them to sell Euros. While I agree this is a poor example, think of the more obscure conversions. Swedish kronas to South African rands, etc.

Anyway, I won't pretend XRP is a perfect solution. Just seems to me that it solves the edge cases of liquidity while doing so more cost effectively, which in turn benefits the consumer.

Why wouldn't the bank just use dollars or Euros as the intermediate currency? Have a pile of dollars, a trading pair for everything against the dollar, and you're all set.

Currency roughly speaking doesn't move across borders, correspondent banks in each country have a ledger, you send money by decrementing the ledger on the send side and incrementing it on the receiver side. Someone else does the opposite. The exchange rate is set such that the ledger balance tends around 0.

No new currency needed.

I have no experience here, so I understand this in theory, but in practice, banks would somewhat indefinitely have some digital sum of a range of currencies in which they may not wish to have exposure to. How do you solve for that as a bank?

I suspect they directly trade the major currency pairs, and trade via USD for minor pairs.

I remember thinking it sounded awful. Automated repackaging and reselling of user-valued loans? The original idea was like the subprime mortgage racket, only with less opportunity for the participants to understand the risks.

So what destroyed or displaced the efforts of the original concept - was it the profit incentive associated with Bitcoin's structure, essentially maligning incentives?

Interesting. Do you have any links to some of the new IOU market maker projects? Cheers

Raiden is doing that for the ethereum network, if you count the fact that each node can be a mediator for multi-hop transfers. https://raiden.network

Raiden has fallen short on delivering a workable system and has had funding issues... most people people in the ethereum space are now banking on rollups filling these use cases (a competing approach for low-cost transactions)

Though certainly I'd be pleasantly surprised if Raiden still pulls through on their plans at some point.

Raiden has many issues, but funding is not one of them. Part of the hiring pitch from Brainbot was that the ICO got them funds to finance their operation for a decade at least.

Speaking as someone who worked on Raiden's core team last year and spent a reasonable part of this year working on a project received a grant from the Raiden Trust [0], the last release is workable, but still requires a lot of elbow grease and knowledge of the protocol to get it running. End users will certainly have a better shot when the light client gets parity to the reference implementation and gets a mainnet release.

In any case, I just wanted to point out that Raiden is a project that is related to what OP asked about and that is still alive and kicking.

[0]: https://hub20.io, a self-hosted payment gateway that integrates with Raiden and abstracts all its complexity from the users and node operators to enable fast and near-instant transactions.

So happy to see actual discussion of cryptocurrency concepts and not just visceral hate for all things crypto on hacker news. New line it looks like you guys have really turned a corner I'm proud of you hacker news.