Skip to content

Publication status and archival corrections

Working drafts are not submissions

The repository contains a complete research narrative and a condensed RNAAS draft. Neither a local PDF nor a successful software test certifies scientific peer review, journal acceptance, or submission readiness. The v1.3.0 software is archived separately at Zenodo; that software DOI does not imply a journal submission or acceptance.

The legacy publication_ready field in the validation run JSON means that the pipeline's artifact acceptance checks passed. It does not mean journal readiness, submission, acceptance, or external scientific validation.

Artifact Meaning
reports/research_report.md Working narrative of the research v1/v2 results, with editorial corrections
reports/rnaas_draft.md Working Research Notes draft requiring final author and venue review
output/pdf/sxs_preprint_v1.0.0.pdf Archived research PDF; not automatically synchronized with current Markdown
CITATION.cff Software release citation; research-report metadata is a separate reference

Public software releases v1.0.0, v1.1.0, v1.2.0, and v1.3.0 are not the same thing as the research labels “v1” (baseline) and “v2” (scaled benchmark/search). The research PDF remains version 1.0.0; versions 1.1.0 through 1.3.0 are software releases. A main container tag is a moving build, not a new numbered release.

Corrections to read with the archived PDF

The archived eight-page PDF is in English and retains its original content. Later Markdown corrections do not rewrite the PDF or published release assets:

  • The effective Kepler BLS duration grid is 1, 2, 4, and 8 hours. The YAML also requests 12 hours, but the implementation excludes durations not strictly shorter than the 0.5-day minimum period. Reported recovery values are unchanged.
  • The PDF retains historical numbered development milestones and the wording “submission-ready.” Treat it as a research draft requiring author review, not as a current publication-status certificate.
  • The 20,000 candidate-level FAP rows reuse null distributions for candidates sharing a target: there are 14 unique targets with 1,000 realizations each.
  • The “no confirmed exoplanet” conclusion concerns the new-search shortlist, not the previously catalog-confirmed planets used for recovery benchmarks.

The final result remains 0 strong candidates, 1 weak candidate, and 19 likely false positives. KIC 8300900-r1 remains weak and unconfirmed, with empirical FAP 0.01998. No data, model scores, or candidate categories were edited to make the documents agree.

The archived frozen shortlist retains old phase7 score-provenance wording and reports/phase8_candidates/ figure paths. Input comparison recognizes those two editorial aliases without rewriting the archive; changes to scientific values are rejected (with a 1e-12 relative/absolute numerical comparison tolerance).

RNAAS format check

The official Research Note Preparation Guidelines, checked on 30 August 2026, specify a required abstract of up to 150 words and a 1,500-word overall limit including title, headings, captions, and references, with no more than one figure or one table. The older instruction to omit the abstract was incorrect. The working draft now has an abstract and one table. Automated Markdown word counts are estimates, not the final publisher submission-system count.

Before submission, the author must review the science, authorship, references, current journal instructions, and publication-license terms. The journal's article license and the software's custom commercial license are distinct; this audit does not accept a publishing agreement or change software rights.

Editable AASTeX source, bibliography, an editorial-note draft, and a concrete submission checklist are in manuscript/. The AAS account holder must personally verify the metadata and accept the publication agreement. A prepared package is not a submitted manuscript.

Permanent software-archive evidence is recorded in the Zenodo DOI record.

Future PDF generation

Use scripts/build_preprint_pdf.py for a new working PDF. Its default output is build/sxs_preprint_draft.pdf, not the archived release file. A fresh PDF requires page-by-page rendering and visual review before distribution. No replacement PDF is generated by this documentation audit.

For exact historical evidence, use the relevant Git tag or archived file. For current installation and usage, use the documentation accompanying your checkout. Historical audit counts are not claims about the current test suite.