...
Yeah and I posit that @gmaxwell is going to be the spectator.
Side-chains are fundamentally flawed and can't ever work securely. Lightning Networks can't work decentralized.
Side chains are constrained by a fixed number of coins. As far as I can see, without an emission of new coins there is no practical way to secure them with one notable exception; namely
demurrage. The credit for this belongs to Freicoin
http://freico.in/I know this is controversial but it can work, particularly for a side chain whose purpose is transactions as opposed to wealth storage. A 1% per year demurrage rate is not a real issue if the velocity of money is say 2 weeks.
The problem is incentivizing mining, because merge mining is insecure, side-chains can't create coins, and transaction fees are not incentives compatible in PoW.
How do you envision demurrage incentivizing sufficient mining to secure a PoW side-chain?
Secondly merge-mining suffers from 50+% attacks if the chain being merge-mined doesnt have a majority of total hashing power... which kinda defeats the point if we're worried about miner scalability.
...
The problem is incentivizing mining, because merge mining is insecure, side-chains can't create coins, and transaction fees are not incentives compatible in PoW.
How do you envision demurrage incentivizing sufficient mining to secure a PoW side-chain?
...
A 1% per year demurrage rate can produce a 1% per year block reward while keeping the number of coins in the side chain constant. The demurrage amount and block reward are adjusted continuously to reflect the number of coins in the side chain. In Freicoin demurrage is used to compensate the miners while keeping the number of coins constant. Their demurrage rate I believe is higher.
1% is approximately the tail emission of Monero (0.6 XMR per 2 min block). This roughly corresponds to 3 XBT per 10 min block for Bitcoin.
Afaics, this is not secure for reasons:
1. The block reward is much smaller than the rewards on the main chain, thus unless the side-chain uses a different mining algorithm, then the main chains miners can 51% (or 33% selfish-mining) attack the side-chain. To incentivize sufficient security, then side-chains will not be competitive with coins kept on the main chain (or a different consensus algorithm which doesn't need to spend so much on security as PoW does, which is my yet unpublished design because of lower sunk costs as compared to PoW).
2. An attacker can yoyo the difficulty of the side-chain by moving many coins into and out of the side-chain. Difficulty attacks either leave a coin with periods of being very slow to confirm or enable the attacker to 51% attack during periods of very low difficulty.
There are likely other problems. Side-chains are simply a broken concept.