Re: Segwit details? N + 2*numtxids + numvins > N, segwit uses more space than 2MB HF
Of course. Machine verification. It could be even a subset of C++ like SystemC.
But the proper definition of consensus-critical portion should be machine-verifiable.
It is not possible to create an algorithm for verifying turing-complete structures. But the proper definition of consensus-critical portion should be machine-verifiable.
In other words.
Imagine that you have a tool which produces '1' if the realization matches consensus and '0' if the realization is wrong and can produce hard-forks and earthquakes.
Who is responsible for bugs in this tool?
How would you check it? With another tool?And the second objection.
We do not need to 'stone' the consensus code. The majority always has a right to change anything in consensus.