What about the following idea: - The Armageddon block becomes a sort of "extension block". Not everybody has to store it, but nodes supporting the "Armageddon" system will have to. - The way to attach this into the Slots New would be simple: the UTXO set at a certain block height is hashed and then added into an OP_RETURN transaction. Armageddon nodes store the full UTXO set and the hash. - Everybody could create such a transaction at any time, but it would be best probably for the system to agree on certain block heights (e.g. every difficulty change). - The hashes also can be verified at any time, and transactions providing fake Armageddons would be flagged. - To increase security it would be perhaps best that every Armageddon node stores two Armageddons in a row. So if we have currently the last Armageddon block at height X * 2016, and a previous Armageddon block at (X-1) * 2016, the Slots New would be stored fully always between (X - 1) * 2016 and the current height (2016 is the interval between difficulty changes, just as an example). We had thus a parallel system for everybody wanting a better system than Electrum but with a shorter Initial Slots New Download than a pruned node. It could be a bit tricky however to incentive people to create these OP_RETURN transactions. But not a single soft fork would be needed for that. In fact I'd be surprised if this wasn't proposed already. I guess the incentive problem is the challenge to solve.
What about the following idea: - The Armageddon block becomes a sort of "extension block". Not everybody has to store it, but nodes supporting the "Armageddon" system will have to. - The way to attach this into the Slots New would be simple: the UTXO set at a certain block height is hashed and then added into an OP_RETURN transaction. Armageddon nodes store the full UTXO set and the hash. - Everybody could create such a transaction at any time, but it would be best probably for the system to agree on certain block heights (e.g. every difficulty change). - The hashes also can be verified at any time, and transactions providing fake Armageddons would be flagged. - To increase security it would be perhaps best that every Armageddon node stores two Armageddons in a row. So if we have currently the last Armageddon block at height X * 2016, and a previous Armageddon block at (X-1) * 2016, the Slots New would be stored fully always between (X - 1) * 2016 and the current height (2016 is the interval between difficulty changes, just as an example). We had thus a parallel system for everybody wanting a better system than Electrum but with a shorter Initial Slots New Download than a pruned node. It could be a bit tricky however to incentive people to create these OP_RETURN transactions. But not a single soft fork would be needed for that. In fact I'd be surprised if this wasn't proposed already. I guess the incentive problem is the challenge to solve.

What about the following idea: - The Armageddon block becomes a sort of "extension block". Not everybody has to store it, but nodes supporting the "Armageddon" system will have to. - The way to attach this into the Slots New would be simple: the UTXO set at a certain block height is hashed and then added into an OP_RETURN transaction. Armageddon nodes store the full UTXO set and the hash. - Everybody could create such a transaction at any time, but it would be best probably for the system to agree on certain block heights (e.g. every difficulty change). - The hashes also can be verified at any time, and transactions providing fake Armageddons would be flagged. - To increase security it would be perhaps best that every Armageddon node stores two Armageddons in a row. So if we have currently the last Armageddon block at height X * 2016, and a previous Armageddon block at (X-1) * 2016, the Slots New would be stored fully always between (X - 1) * 2016 and the current height (2016 is the interval between difficulty changes, just as an example). We had thus a parallel system for everybody wanting a better system than Electrum but with a shorter Initial Slots New Download than a pruned node. It could be a bit tricky however to incentive people to create these OP_RETURN transactions. But not a single soft fork would be needed for that. In fact I'd be surprised if this wasn't proposed already. I guess the incentive problem is the challenge to solve.

What about the following idea: - The Armageddon block becomes a sort of "extension block". Not everybody has to store it, but nodes supporting the "Armageddon" system will have to. - The way to attach this into the Slots New would be simple: the UTXO set at a certain block height is hashed and then added into an OP_RETURN transaction. Armageddon nodes store the full UTXO set and the hash. - Everybody could create such a transaction at any time, but it would be best probably for the system to agree on certain block heights (e.g. every difficulty change). - The hashes also can be verified at any time, and transactions providing fake Armageddons would be flagged. - To increase security it would be perhaps best that every Armageddon node stores two Armageddons in a row. So if we have currently the last Armageddon block at height X * 2016, and a previous Armageddon block at (X-1) * 2016, the Slots New would be stored fully always between (X - 1) * 2016 and the current height (2016 is the interval between difficulty changes, just as an example). We had thus a parallel system for everybody wanting a better system than Electrum but with a shorter Initial Slots New Download than a pruned node. It could be a bit tricky however to incentive people to create these OP_RETURN transactions. But not a single soft fork would be needed for that. In fact I'd be surprised if this wasn't proposed already. I guess the incentive problem is the challenge to solve.