Quote from: r0ach on September 17, 2015, 09:10:59 PM

Quote from: monsterer on September 17, 2015, 08:25:26 PM

https://gist.github.com/oleganza/8cc921e48f396515c6d6

You can't use Bitcoin itself as an example of Byzantine consensus in an effort to justify it's own existence. That page is moving the goal posts all around and adding a bunch of new variables that aren't even in the original problem. All that page is doing is saying, Bitcoin works, therefore, the solution Bitcoin used is the answer. Circular reasoning.

Battle of the century of r0ach vs smooth regarding this issue. They call him "smooth" because it's like talking to Bill Clinton. You tell me who won:

<@smooth> The BGP as usually stated has a concept of identity ("Generals") which is specificaly not part of the problem definition in Bitcoin (which is what makes it sybil resistant). Bitcoin doesn't care
I made the arguement that byzantine generals is a ridiculous ivory tower example with too many open ended variables and the only real problem is sibil prevention
<@smooth> yes and for the millionth time bitcoin is totally sybil resistant
<@smooth> because identity doesn't matter
it's not sybil resistant, all pools can be owned by the same guy
<@smooth> pools are not actors in bitcoin. hash rate is
<@smooth> hash rate can't be sybil attacked, it is a physcal property
hash rate doesn't decide vote, it's delegated proof of work (bitcoin), only the pool owner does
what hash does is irrelevant
you're letting satoshi decide what you can criticize or not
instead of using your own logic
to figure it out
because the model that exists is nothing like the PDF
<@smooth> well if you are critizing bitcoin, you are criticizing somethign he defined
<@smooth> if you want to redefine it, and then criticize that, that's perfect valid science, just make a specific definition first
bitcoin does not function in the way his PDF describes at all, so when you cite satoshi, it's pretty much meaningless in that context
<@smooth> I disagree
<@smooth> the only portion that does not apply is the convergence proof
<@smooth> but that is because of hash rate concentration, not because of pools
<@smooth> even with pools (and I'll admit this is not a precise argument), if 50% of hash rate is honest, pools can't do anything because the hash rate will quickly flee a dishonest pool
<@smooth> Note this is not true if KnC Bitfury etc. is not honest, because their hash rate can't flee
<@smooth> even 1 cpu 1 vote is actually true still
<@smooth> again, cpus are a physical entity, can't be sybiled
it doesn't matter what the hell the cpus are doing since you're going through a 2nd layer of abstraction known as delegation (pool)
and the 2nd layer takes precedent over the 1st
<@smooth> i would argue the opposite
<@smooth> the 1st takes precendence over the 2nd, because is I said, you pull your hash rate
yes, i can pull my hash rate AFTEr the attack has occurred
that's fault recovery, not fault tolerance
this is known as the long con, I'm sure you've heard of it

Even if KnC/Bitfury/etc. were >50% and dishonest, the socioeconomic majority can flee their corrupted PoW by forking to something besides SHA256.

So smooth won, but for a reason not explained. Cheesy