innovation of bitcoin

[dzoni]

I'm trying to figure out what differentiates bitcoin from previous ecash technologies. If I understand correctly, b-money by Wei Dai and hashcach by Adam Back already were distributed and solved the byzantine problem (did I get that right?). So, what was the innovation of bitcoin? Or was there no innovation and it was simply the first to be widely adopted?

[DeathAndTaxes]

b-money didn't "solve" the byzantine general problem (prevent double spends) in any practical manner.  The first proposal requires that all nodes have unjammable perfect and synchronous communication with all other nodes.  This would be more analogous to Bitcoin only having 0-confirm txns.  If all Bitcoin nodes have synchronous perfect and unjammable communication with every single other node on the planet then double spends of even unconfirmed txns would not be possible.  As soon as a txn by one node the rest of the network would know about it and would reject any txn which spent the same coins forever.  It solves it in theory but is utterly unworkable even on a scale of hundreds of nodes much less millions.  As the original proposal was unworkable there was a second proposal which used a small number of servers in constant communication with all other servers to reach consensus on the validity of txns.  Users would submit txns and then query servers on the acceptance of the txn.  This would be similar (but with a lower trust level) than SPV clients in the Bitcoin network.  The use of "semi-trusted" servers makes the second proposal practical but it isn't decentralized or trustless.  It is closer to the security model used by ripple, where a few elite "super peers" act as guardians of the network.  If you can trust the super-peers well you can trust the currency.

The major innovation by Bitcoin was the use of blocks to form a timestamping mechanism to (partially*) solve the byzantine general problem.  This allows trustless verification of txns by all (full) nodes.  Bitcoin wasn't the first to use the concept of PoW but it was the first decentralized currency to use PoW as a solution for the byzantine general problem.  In b-money, the PoW was limited to the creation of new coins.  Beyond controlling the creation of money it wasn't used to secure the network.   Bitcoin combined the timestamping of txns with the creation of money in the form of solving blocks.  Doing so could be seen as a secondary (and very practical) form of innovation in that is solved a second problem as well.  There is a chicken and egg problem associated with how to pay for the security of the network during the bootstrapping phase.  By giving miners a subsidy in the form of newly minted coins the minting rate is controlled and the security of the network is subsidized to allow it to grow securely until such time that txn fees can allow the network to pay for itself. 

That being said it is very likely that Satoshi was aware of b-cash and was influenced by it to some degree when designing Bitcoin.


* Partially in the sense that the security model requires an attack to not have a majority of the computing power.

[itod]

That being said it is very likely that Satoshi was aware of b-cash and was influenced by it to some degree when designing Bitcoin.

Great analysis, as always DeathAndTaxes. Just to add to the sentence above - not 'very likely' but 100%. The first reference in the Bitcoin whitepaper is to W. Dai:

References
[1] W. Dai, "b-money," http://www.weidai.com/bmoney.txt, 1998.

[dzoni]

First, thank you for your knowledgable answer. You helped a lot.

So, is it safe to say that Bitcoin is first technology ever to use PoW for consensus making? Is that the most important innovation of Satoshi?

The major innovation by Bitcoin was the use of blocks to form a timestamping mechanism to (partially*) solve the byzantine general problem.

Didn't hashcash solve byzantine problem by usage of timestamping? I think I read something about that.

dzoni

[jonald_fyookball]

I think hashcash was the first kind of proof-of-work as timestamp.
So its a critical piece, but not the whole solution.