I want to get dialogue going on off-chain transactions. I think I have helpful info on solving the block size issue which doesn't necessarily require off-chain transactions, but I want to discuss the concept.
I've said off-chain transactions should exist with Bitcoin. Primarily I had scalability in mind, but also think of the progression of money. People once traded physical gold. People then traded claim checks for the gold instead since that was more effective (easier to carry, not subject to shaving, etc.). Today we use credit cards, PayPal and other methods to trade digital claims on paper dollars since that's even more effective than trading the paper.
Why couldn't the same happen with Bitcoin? Why can't people trade digital claims on bitcoins instead of the actual bitcoins?
Member retep posted:
Fidelity-bonded banks: decentralized, auditable, private, off-chain payments.
While I'm still considering the viability of that model I've also had a far simpler transfer method in mind which I call Bitcoin Clearing Houses (BCH).
A BCH starts with an entity that can garner general trust, say MtGox, Blockchain.info, the Bitcoin Foundation, etc. The BCH sets up a server with website interface allowing users to register with email addresses and deposit or withdraw bitcoins. Those transfers take place on-chain. However, since the BCH is widely visible many other users and merchants could also have accounts there. Then coin transfers happen within the BCH, or between multiple ones. A BCH also provides an API so that eWallets like WalletBit or MtGox can pass on transfers from their members without the members having accounts at the BCH.
Using a BCH has obvious advantages. Transfer is instant, no waiting for confirmations. Transfer is free or very low cost (profit might come from ads or features). Finally, of course, is scalability is helped greatly. (people could also send to email addresses eschewing wallet addresses)
While I think this is a strong case people raised concerns that this tends to centralize things and provide a target for authorities, noting Bitcoin exchanges are probably the biggest weak point. However, I don't think that's a concern. A BCH centralizes things no more than exchanges do currently, that is, they can come and go anytime and anywhere. A BCH has even more flexibility because all info is digital whereas Bitcoin exchanges have to merge in part with the traditional regulated system. Also, a BCH has no ability to create bitcoins or prevent their transfer since users could revert to the old system.
The largest concern I see is security of coins held or possibly manipulation or dishonesty with claims of coins backing transfers. However, that can be mitigated by users storing minimum balances (usual amounts for transfer) and withdrawing their balance periodically, and maybe even using different BCHs.
Thoughts?
By creating centralized points of trust, you provide profitable points to attack, either by people who want to steal the bitcoins, or governments or companies like PayPal who might want to damage Bitcoin. Compare to peer-to-peer file sharing, the centralized ones were attacked successfully, the decentralized ones are nearly impossible to stop.
The reason Peter's proposal was more complicated was to reduce the profitability of an attack on any of the Chaum banks - the worst that can happen is that the account holders would not have access to their funds for some period of time, but they are *guaranteed* to get it back.
Having said that, I expect we'll see what you propose in the future when/if Bitcoin becomes very popular, because the majority of people prefer simplicity over security/privacy, and the majority will have come to Bitcoin because it is popular rather than because of any of its inherent advantages.