nx_idiomcensus.nx

nx_idiomcensus.nx -- WHAT COUNTER IDIOM DOES THE SKIP-IDIOM FAMILY ACTUALLY USE? nx_gate_dry_apply SKIPs a large family because it cannot recognise how the gate counts pass/total, and the remedy proposed for it has been "write a recogniser" -- for a shape distribution nobody has measured. ★DIAGNOSE THE POPULATION BEFORE BUILDING THE GENERAL MACHINE. A recogniser written against the one shape somebody happened to eyeball is a recogniser for n=1. ★SUBJECT MATCHES CLAIM. It censuses EXACTLY the gates dry_apply classed SKIP-IDIOM, read from the NEWEST generation of knowledge/status/drysweep.log -- not "all gates", which is a different population wearing the same name. drysweep.log is APPEND-ONLY across runs: reading every generation double-counts (the identical defect already found and fixed in nx_adoptsweep), so the newest block is isolated first. ★UNKNOWN IS ITS OWN BUCKET and the partition is PRINTED AND CHECKED TO SUM. A shape that falls into a known bucket because nothing matched is the number somebody then plans against. ★MARKER TALLIES ARE NON-EXCLUSIVE AND LABELLED AS SUCH, separately from the exclusive partition -- two different questions ("how many gates contain X" vs "what IS this gate") must never share a counter. nx_idiomcensus [max] [logpath] [dir] exit: 0 censused | 2 usage | 3 NO-CONCLUSION (log unreadable / no SKIP-IDIOM rows / partition disagrees) license_tier: ORIGINAL No hw writes (Rule 26).