Topic: The Oversight "Problem" (Read 507 times)
I think that a diversity of clients on various OS's written in a variety of languages is one way to help ensure this doesn't happen. (Assuming we are speaking of a runtime client compromise). Most big pool operators are watching things well too. I don't believe this is the case now though so this an interesting discussion.
Hi everybody,
I have noticed that there are a few instances in the Bitcoin protocol where human oversight and intervention is necessary to verify the network and insure that it is operating correctly. The obvious instance would be the fork earlier this year, where nodes had to manually revert their software to correct he problem.
A generalization would be this: Imagine that the network experiences a fork due to differences in protocol, I.e. Assume some portion of the network was affected by a virus that altered its client. How would this be handled?
It seems that the solution is to manually identify if your client is following protocol (which must be declared by a central development team) and then intervene. So there are two hold ups in the problem, where manual intervention is needed.
My question is whether or not there have been discussions or proposals on how this might be automated. I think that to automate this, there would have to be some way to decentralize "protocol determination", if that even makes sense.
I have noticed that there are a few instances in the Bitcoin protocol where human oversight and intervention is necessary to verify the network and insure that it is operating correctly. The obvious instance would be the fork earlier this year, where nodes had to manually revert their software to correct he problem.
A generalization would be this: Imagine that the network experiences a fork due to differences in protocol, I.e. Assume some portion of the network was affected by a virus that altered its client. How would this be handled?
It seems that the solution is to manually identify if your client is following protocol (which must be declared by a central development team) and then intervene. So there are two hold ups in the problem, where manual intervention is needed.
My question is whether or not there have been discussions or proposals on how this might be automated. I think that to automate this, there would have to be some way to decentralize "protocol determination", if that even makes sense.
Jump to: