I think the community needs to develop a process ahead of time for determining how to deal with these issues. A good start might be: Can the protocol be made m

I think the community needs to develop a process ahead of time for determining how to deal with these issues. A good start might be: Can the protocol be made more robust against this class of error? Is the proposed protocol change victimless? (nobody has already spent forked funds, etc.) Is the protocol change elegant? (If you were to start over would the implementation have included this change to begin with? e.g. no hard-coded addresses with special code paths)