This segregated witness proposal is turning into a complete train wreck. What else have they overlooked?The way you'd know that the Blockstream Core developers were approaching this problem in a responsible way is if they started with a robust fraud proof system first before taking on anything else.It would be very similar to the concept of test-driven development. Build a system of making easily-verifiable proofs that a block is invalid, and all the infrastructure that's needed to propagate the proofs. Require blocks to include the information that's needed for the proofs to operate.Heavily test that system for 6-12 months to make sure all the corner cases have been exercised.Then, after we're absolutely sure we have the ability for light clients to reject invalid blocks based on well-documented and well-understood proofs,then it would be time to do something like separate out the cryptographic witnesses from the transaction data to produce storage and malleability gains.Throwing that into the protocol right away as a last minute faux scalability improvement is unimaginably reckless.