And as I see it this can only lead to one conclusion. Eth and other coins with "global" smart contracts will fail. Smart contracts can not be used in a system that can overrule the contract by forks or any other way.
This is the important lesson indeed, and this was the true lie of the ETH/DAO promotors: they explicitly said that this was not a thinkable scenario, while it is, of course, because of the consensus nature of a block chain. That said, and that is why, if there is one immoral and dishonest party in this whole affair, it is the ETH foundation, because it was THEIR initiative to call for a hard fork. Of course, they needed suffiicient weight from miners to be able to pull the trick, but if they had stated, from the start, that the essence of smart contracts, and the value proposition of a smart contract platform such as ETH, is exactly that the code is the law, including unintended behaviour, there wouldn't, most probably, NOT have been a hard fork, even if the potential desire for it existed.
So, as you say, the potential for the code being overruled by a hard fork is an inherent flaw of the whole system of smart contracts, but on top of that, the very founders of a smart contract platform pushed for that potential to become reality, and as such betrayed all their claims concerning "unstoppable" and "immutable" and "the code is the law" themselves.
For now I think smart contracts should only be used for agrements between two parties. Thank you Eth for the lesson.
I think the issue is not so much the number of participants, but rather the weight of the value of a contract with respect to the value of the whole chain. The fundamental flaw in the DAO was that it was "too big to fail" and the main / sole ETH project. In other words, in the same way as having all the mining power in the hands of a few, having a large chunk of the chain value in one single contract is a centralization problem that will lead to corruption of the rules. If there would have been 20 000 little DAO on ETH, a fork to save one of them would never even have occurred in the minds of anybody.
I think your two-party smart contract will suffer even more from that if you consider that the code has to be the judge. Each party can fork off and do as he pleases. In fact, your two-party smart contract is probably just a cryptographic way to sign and date actions on both sides, *with the idea of going to court if one of the parties thinks that the contract was broken and got screwed*. But that's not a smart contract, that's a secure logbook. The only thing you're probably doing is introducing non-repudiation signatures on actions, so that you can use this as proof in court. I think that's a useful way to use cryptography, but I wouldn't call it a smart contract.
But I fully agree with you: thank you, eth, for having screwed up so royally and to have shown the limits of smart contracts, namely people messing with history and hard forking over it, before people taking contracts seriously, lost money. Now, only gamblers lost money, and even got it back, which is not dramatic. So, thank you again for showing the uselessness of ETH before it did real damage.