I discussed the open transactions server stuff a bit with several members here on the forums. I didn't really like how it worked. It's not how you'd think it would work, all decentralized like bitcoin itself. It was more like a network of exchanges (each server = it's own exchange) that could trade assets around. Even the trading of the assets was not what I thought, as a single issue could not leave a single server. (if you have sheep on server A, and goats on server B, sheep could not move A to B. You have to trade sheep for goats, -or- the issuer has to issue sheep on both servers.) I could accomplish nearly the same thing if I added a secure cross-site trading plugin to LTC-GLOBAL, then open sourced the code.
If an issuer issues sheep onto servers A, B and C, then the "sheep" currency is now available on three servers.
If you redeem a cheque that was drawn from an account on server A, then your wallet will naturally redeem it at server A,
just the same as a web browser trying to display an image from webserver A, would download that image from webserver A.
(In both cases, the client software just accesses the appropriate server based on what's in the URL.)
There are several ways to move your "sheep" from Server A to Server B:
1. Anyone who has a "sheep" account on both servers, can perform this transfer for you.
2. Including, obviously, the issuer himself. He can take sheep from you on Server A, and give you sheep on Server B.
(In fact, anyone with accounts on both servers can do this.)
3. You could sell the Sheep on the exchange in return for Bitcoins. Then withdraw the Bitcoins from the server directly, and move them onto server B, where you use the BTC to purchase Sheep again.
4. A protocol could be devised between servers that allows them to open accounts with each other, and thus perform "wire transfers" on behalf of their users.
There are single point of failure issues with OT because all the assets go poof if the server they are on goes offline.
This is not correct. The whole idea with OT is that the user stores his
last signed receipt, and is thus able to prove his balance, as well as which instruments are valid, simply via that last signed receipt. In other words, the receipt
is the account, and both parties have a copy of it.
Problem also emerges that you might have 100 people all running their own "exchange", making it all that much harder for the trustworthy operators to rise above the noise.
Remember that the OT server cannot change balances or forge receipts. (Your signed request appears on every server-signed receipt. So the server cannot forge a receipt because it cannot forge your signature, because it does not have your private key.)
And as long as the issuer has an auditing protocol in place that meets his standards, then the server cannot commit inflation either.
Therefore, whether there are 100 servers, or 1000 servers, they cannot defraud the issuers or users.
(OT guys feel free to correct me if I'm wrong!)
I don't want to start a flame war or anything. I admire the work the OT guys are doing. Just saying that based on what was described to me, it's not yet ready for prime-time. It has some design flaws that need worked out so that it works more like bitcoin itself, totally distributed and independent of any single point of control or failure.
Cheers.
Thanks for the opportunity :-)
