Everything is in the ledger, just like the blockchain.
I've already went through this with another Ripple supporter
here:
Since transactions in Ripple only operate on account balances, not on inputs like Bitcoin, and you can only branch at the top not "mine" on an earlier "block", there is no explicit need for keeping history, not even headers.
There is no explicit need for keeping history...
So there is no need for keeping a ledger, in other words.
Dude, seriously?!
That's your interpretation and definitely not the final conclusion! Ripple ledgers are similar to a signed UTXO set in Bitcoin terms. If something like this (also called "ultimate blockchain compression" in the dev forum here) was implemented in Bitcoin in the first place, there would also have been no real need for full nodes to download gigabytes of stale transactions and you could be up and mining from scratch after loading a few dozen MB.
It is not my fault that you confuse technical terms (Ripple ledger) with accounting terms (a book keeper's ledger --> full transaction history) or that the proper choice of terms for Ripple would have been "balance sheet" maybe. I did not design that system, I use and study it.
XRP are cryptographically accounted for since ledger #32570 and there are efforts to push this back to have the genesis ledger at #0. Still at 32570 there were less than 100 billion XRP already and their amount has been only going down since. There is also no transaction possible currently (and yes, you can check that in the code for rippled) that would allow issuing any more of them.
Anyways, why are we even talking about XRP in here so much?! They are a centralized currency and not the greatest one anyways. You can get more of them than you'll likely ever need for the price of a beer and be done with them. Anything more than that is pure speculation and there are easier markets for gambling your money away...
There is spam concern because there are transactional limits to everything. If there is no spam prevention, anyone can attack the network with say 100,000 false transactions and grind the network to a halt.
Okay, then use a cryptocurrency to prevent spam. In fact, you could use Bitcoin. The way to prevent spam is simple, which is attach a cost to an electronic action. Spammers are successful only when their cost to spam is negligible.
Actually that's a project that might be very interesting - to create a fork of Ripple that can use XRP and BTC or some other cryptocurrency for the purpose of spam prevention. They use similar enough crypto that it might even work, the only issue is that Bitcoin transactions are damn slow, but if someone is really disgusted by XRP, that won't stop them I guess. Another problem tmight be that validators then would need to watch a block chain too, creating more attack surface and another source for errors. It's definitely possible though.
Therefore, a valid transition between two ledgers can only maintain the number of XRPs in circulation or reduce it, and from that property, it derives that the only XRP that can ever be created need to be created in Ledger 0 (or genesis ledger).
Okay, if you say so... but what about this:
The oldest available ledger is #32570 (with about 100 entries) so this is currently the genesis one, the older ones are presumably lost or at least not available until JoelKatz finally publishes the transactions that took place in these first 1 1/2 weeks so these ledgers can be restored back to 0. Since then there is a unbroken hashed chain of headers (also cryptographically linked to the transactions and account state at each point of time). Nodes verify this by getting the most recent finalized ledger and continuing from there. Since transactions in Ripple only operate on account balances, not on inputs like Bitcoin, and you can only branch at the top not "mine" on an earlier "block", there is no explicit need for keeping history, not even headers.
Why is there this mysterious missing genesis ledger? In Bitcoin the creator disappeared, but it's no matter because we have and can verify the genesis block. With Ripple it's backwards, we have the creators but the verifiable origin of Ripple amount is a secret.
The genesis ledger is #32570 and it is publicly available with all its trust lines, XRP amounts and trade offers within.
It's easy to say that exchanges are just for exchanging and people should cash in and out quickly. The reality is that there are costs in terms of fees and inconvenience (captchas, email verifies, delays) that make it difficult to do so. So many people end up just leaving money on the exchanges. And get bitten when they fail. Is Ripple any different in this regard? Is it 0 cost in terms of fees and near zero in terms of inconvenience to cash out my bitcoin IOU for a real bitcoin?
Otherwise, I don't see how that is much different from any of the other exchanges.... entities that I put approx zero trust in.
You would rarely hold Bitcoin IOUs to begin with, Ripple for bitcoiners would work similar to Coinbase or BitPay - most if not all of your funds are with you and you only convert them to IOUs if you want to trade them (e.g. atm. on traditional exchanges) or buy something denominated in a different currency (e.g. you buy a computer game on humbe bundle via coinbase, who then trade on behalf of HB).
Edit: Unlike coinbase, Ripple would likely add other payment networks via bridges though, not just single merchants. Once someone strikes a deal with e.g. PayPal, you could pay with Bitcoin to any site that accepts PayPal for example.
There are currently fees (for sure in XRP and potentially in IOUs too) as well as the normal market problems: spread and slippage. Unlike for example a bank transfer to an account in a different denomination though, you can get active there too and just put in a different offer, that might get taken.
Sorry, but Mr. Popescu simply did not understand a few things, his example with the little daughter for one thing is flawed and can't happen that way. Feel free to try it out with a few accounts inside Ripple right now, it won't work. Unfortunately as far as I have heard from him, he seems like lomeone who rarely if at all changes opinions after he has made one... I wrote a longer piece about this blog post elsewhere, if you really want to, I can copy-paste it in here too.