When they say "you can't separate" the two, that's absolutely not in reference to sidechains or other similar schemes.
In this paper, we argue that it is possible to simultaneously achieve these seemingly contradictory goals. The core observation is that
Bitcoin the blockchain is conceptually independent from bitcoin the asset
I totally believe that BlockStream's business as it standards will depend on that fact.
There is a risk of destroying Bitcoin as we know it if BlockStream's goal is achieved.
So I understand you are also against federated sidechains? Because they also create this exact seperation you are against.
it sounds like you are desperate, in all the noise you may have missed it.
Peter-R put it most succinctly - my emphasis in bold:
(continued) Furthermore, except for in this thread, there seems to be very little concern for the new dynamics that this change could entail. I highly respect Greg Maxwell (gmaxwell) and Andrew Poelstra (andytoshi), and I want to mention that they're pretty quick to point out the need for rigour in any proposed crypto-system. Andytoshi wrote what's become a highly-cited paper on the dangers of altcoins designed by "amateurs." New designs should be peer-reviewed by experts in the field, tested, etc., etc., etc.
But SPVP sidechains are not just a new cryptosystem. They represent a change to the economics, game theory, politics, and probably even legal aspects of bitcoin. For the same reason that amateur cryptographers may be blind to the weaknesses of their proposed cryptosystems, professional cryptographers may be blind to the change in incentives that their new cryptosystem entails. SPVP sidechains should not be viewed as simply a technical problem; it's a multidisciplinary problem and we need to explore the concerns raised through many different lenses.
In the meantime, there's already the potential for enormous growth ahead of us with bitcoin as it is.
One argument I would like to make is I don't consider SPV to be a "feature" like blacklisting but moreso an "upgrade" on a scheme that is already possible within the existing Bitcoin protocol (federated peg can be implemented right now as you have pointed out).
I think it is clearly a new feature (not that that's necessarily a bad thing--one could argue that Pay2ScriptHash was a new feature too). The fact that federated sidechains are already possible in no way means that adding OP_SIDECHAINPROOFVERIFY is an "upgrade."
The federated servers sit on top of the bitcoin protocol, whereas the SPV-proof-based sidechains would be integrated within the protocol. If you argue that SPV-sidechains would be an "upgrade," would you not also argue that native support for colored coins or Counterparty features would also be an "upgrade"? Like federated sidechains, these features are already possible on platforms that sit on top of bitcoin too.
brg444: Do you support a hard-fork to increase the max block size limit?
The first notion I agree with.
The second is irrelevant to the notion you are defending : that sidechains seperate the BTC unit from its blockchain. SPVproof is a different trust model for this mechanism and the fact that is it embedded in the protocol makes it potentially more secure and trustable. Yes federated models sit on top of the protocol by the separation is the same, only the way by which these units interoperate with the mainchain is different.