"In Open Transactions, nodes have to actively perform transactions, so you can't as easily have a public state of the network. So because the state of the [Ripple] network is public, if I want to make a payment to you, I can look at the exchange offers, I can look at the available liquidity in the network, and I can very easily find the cheapest path to make a payment to you.
In Open Transactions, I would have to go around to different servers and get quotes, and then the transactions can't really be atomic, they're working on it but they're not quite there yet, so if part of the transaction succeeds, and the second half fails, I can wind up with something that I was going to trade to pay you, but now I can't trade it, so I wind up with some asset that may be completely valueless... payments are atomic in Ripple, meaning the entire payment either succeeds or fails."
Today, Open Transactions developer FellowTraveler claims to have found "the Holy Grail," which is to use Bitmessage to allow for discovery across federated servers, server-to-server wiring of funds, escrow based conversion of currencies across federated OT servers, and the ability to send any currency p2p, all of this done without the issuance of credit or using a "pre-mined" currency.
Well, I am not a developer, so, the question is, have they done it? Is it now just a matter of who brings their product to market first?