nx_ecomat_domtrend_lib.nx

nx_ecomat_domtrend_lib.nx -- EXACT growth attribution from the per-domain ledger. This is the consumer that closes debt 1785514540. nx_ecomat_trend reads the AGGREGATE ledger and can only attribute a span whose domain count held still -- measured 406 permil of real movement. The per-domain ledger (nx_ecomat_domledger) records who held what, so the same question becomes arithmetic instead of inference: diff two beats BY NAME. ADVANCEMENT = sum over domains present in BOTH beats of (cur_last - cur_first) EXPANSION = the levels and bar carried in by domains present ONLY in the later beat RETIRED = levels carried out by domains present ONLY in the earlier beat Nothing is ambiguous, because nothing is inferred from a total. A domain that was declared at the bottom can no longer be mistaken for a domain that climbed, which was the entire defect: the headline sum_cur/sum_bar falls when you declare a frontier, so the ecosystem's own self-grade punished expansion and made honest growth unreadable. SNAPSHOT CHOICE: first beat vs last beat. Intermediate beats are counted and reported but not diffed pairwise -- the endpoints answer "where are we vs where we started", and reporting the beat count lets a reader see how much history backs that. license_tier: ORIGINAL No hw writes (Rule 26).