@EquityDiamonds@shruggr@grok@CsTominaga Spend to R-puzzle output, the R-solution being stored in trust structure. Or just a private key held by trust that you destroy your copy of. Why put that burden on the miners? It's not necessarily their business. Miners validate transactions and make blocks.
@EquityDiamonds@shruggr@grok@CsTominaga See the longer answer I gave to you in another thread. Yes, the original client had a mempool. There was no eviction or TTL in the original client, but non-final transactions would be lost if the node was restarted. I think they added more mempool management later.
My lack of concise communication and cognition, clearly. I've also not really studied the old node in very much depth. Looking at the alpha source, the transaction is invalid until all inputs are final and any nLockTime restrictions have been met. It looks like miners will relay the not-yet valid transactions. You can send another with a different or no lock. The original would sit in the mempool until being superseded by the new one. The two transactions needed to have the same set and order of inputs. Then it would be relayed to other nodes that would consider it newer with the same validation rules. Finality is checked when it's pulled into a block. It changed years later on BTC.
nSequence is setup in the transaction-in (CTxIn) structure, not the transaction-out (CTxOut), so the sequences are set by the transaction creator, not on the outputs. Same as nLockTime. Lock time gates the transaction validity anchored in time, Sequence gates each input individually, I assume to support interactive processes with third-parties or other methods that one would want a sequence to anchor the flow of data and coordination back and forth.
@EquityDiamonds@shruggr@grok@CsTominaga They can't mess with Satoshi's outputs. Satoshi *could* have stored some sort of custom transactions already pre-written from years ago, with lock mechanics, that would need to be re-written before broadcasting, but that's only if transactions are broken (no locks).
@EquityDiamonds@shruggr@grok@CsTominaga Outputs aren't locked. Another transaction without nLocktime can spend one right away. nLocktime only enforced if at least one input is non-final (nsequence ~= 0xFFFFFFFF). The locks are transaction-based, not output-based.
@EquityDiamonds@LightBSV@grok@CsTominaga Miners are not keeping around peoples transactions in order to broadcast them when the lock matures. You might have 100 different versions of transactions which spend the UTXO, and only 1 of them will be broadcasted, based on the fulfillment of off-chain agreements.
Tufty the Cat getting smacked around by Ira Kleiman over simple truths wasn't on my 2026 bingo card, but I'll take it. The enemies of Bitcoin are starting to fight each other. Incredible discordance on display and in action.
@EquityDiamonds Model out two branches. One waits until nlocktime >= now, the other doesn't. It's an interactive process, could take years... This is what Craig writes about when he says that all computational branches can be modeled with transactions for each step. Only spend 1. Could have more
@EquityDiamonds You hand a time locked transaction to someone else, already signed from your side. The other side keeps it until it's appropriate timing. Could also drop it if another branch in the interactive process was followed. A way to model out all paths via transaction branches, p2p.
@EquityDiamonds They either weren't sent, or the locktime passed and the transactions were spent, or, more likely, the software just stopped processing them when the functionality was disabled and those transactions were no longer valid.
nLockTime works best for interactive, IP-to-IP flows, IMO.