Native photometric states, deblending and exposure context resolve missing quality semantics
6 Oct 2026 · embeddings
Can bounded native metadata resolve upper-limit versus unmeasurable states and retain all existing infrared candidates without assumptions about physical ownership?
Register the same six unfiltered E156 sky rectangles and 70 native columns, adding27quality/designation/deblending/frame-count/exposure fields. Six corrected GETs,2MiB each/8MiB total,12MiB combined science/5GiB reserve, no retries/redirects/uploads/images. Join by exact cntr only with all45old native primitives unchanged; retain original54export fields and explicit measurement/time/count conflicts. No model, reference readout or candidate changes.
Nine contract/state controls pass in original and corrected namespaces. Independent XML/Decimal replay checks141624native primitives/16703nulls,88515unchanged repeated primitives,70812measurement/time/count states, every exact ID/schema/unit/bound/row/hash/join and field summary. Corrected workers observed terminal0 and44scientificbindings accepted. Preserve terminal1original DNS failure with zero received bytes and pre-registration reservation failure; corrected plans/URLs are exact.
All1967original sources gain27native metadata fields and catalogue designations.4664band entries are detected,3201are reported95percent magnitude upper limits and3unmeasurable; no Z/unknown or photometry/time/count conflicts.264sources have multiple fitted components and77active deblending. Complete band exposure ranges exist for1966W1/1966W2/1940W3/1959W4 entries. Six public query responses total4680511bytes; prior native measurements, candidate ranks and reference matches unchanged.
| Metric | Value | 95% CI | n |
|---|---|---|---|
| original sources with native quality metadata | 1967 | — | — |
| native reported upper-limit band magnitudes | 3201 | — | — |
| native unmeasurable bands | 3 | — | — |
| multiple jointly fitted components | 264 | — | — |
| actively deblended sources | 77 | — | — |
| confirmed physical births | 0 | — | — |
Next: Register a contamination-aware signed-flux/shape likelihood with explicit native measurement states and matched raw noise-aware/magnitude/brightness baselines; freeze model and review allocation before readout. Preserve uncertain/flagged sources for follow-up and cold far-infrared/optical tiers. Obtain representative independent calibration/negative truth and fresh transfer before robust survey claims, then refine physical ownership/membership and temporal birth-region evidence. Full goal active;E121zero, no images/list uploads/Git rewrite. External update queued behind pending separate remote readback approval.
Native flags do not remove artifacts or establish Gaussian flux errors, physical class or ownership. E159quality loss persists: only8/302Orion sources pass its conservative two-band diagnostic, no W3/W4 band there passes. A reported confidence upper bound is not an independent random censoring event and cannot be multiplied with the same flux measurement. Exposure means are not automatically astrometric fit epochs; deblending/times remain incomplete. No improvement in population precision/recall, classifier calibration, physical stars/dust, membership/age or birth established.