Correct it is a freaking huge change of monumental proportions.
Its a high-risk high-reward gambit that expands the potential in all sorts of currently unforeseen ways..
It really can't be implemented without the change,
Well... some special cases things can be done, (with the federation/oracles) but the vast capabilities of validating arbitrary code on (and with) the Bitcoin block chain is what takes the fork.
It'd be interesting if you could expand on that.
My understanding is any sidechain scheme can be implemented by replacing SPVproof with federation/oracles. It was even suggested by the developers that this would be the first step of the implementation for most sidechains (bootstrapping on a federated peg model) then implemented through OP_SPV.
Is this not the case?
Any side chain can of course be implemented entirely without Bitcoin at all. Perhaps, some would not be viable as federated systems, such as those needing a true zero trust environment and wanting to also hold BTC in their chain.
Its not really a zero trust application if you are trusting an oracle. Those would be fairly esoteric though, probably most of the ones that are being seriously contemplated would be accepted with a federated system, with the exception of those that may be requiring extraordinary security. If it is something where folks don't trust that the federating entity isn't their enemy in disguise. It would be one trust removed to have the protocol change.
Without the Bitcoin network running the verification, why use Bitcoin at all though? Might as well use an Altcoin as the "separated currency".
It would be a better bootstrap to prove the case anyway.
Though some of the issues are insidious and won't manifest their effects until there are a lot of locked coin.
Federated systems work fine until the federated server (or the Bitcoin network) goes away.
Bitcoin works fine, until the Bitcoin network goes away.