Citation ledger

Bentley's own column, in his own words, introducing the ninety-ninety rule — not a summary of it by someone else.

A third-party YouTube retrospective, not the developers' own account — kept as this week's illustrative case, not as authoritative evidence for the claim.

NTCA is the rural telecom industry's own trade association, publishing its own economic analysis of rural broadband subsidy costs.

The Council of Graduate Schools is the primary national body tracking PhD completion, publishing its own completion-project data.

A customer-feedback vendor's marketing blog citing an uncredited survey stat with no named underlying study — kept as an illustrative figure, not authoritative evidence.

TLIE is a malpractice insurer publishing its own resource on the risk it underwrites and pays out on.

Original legal-scholarship research by its named authors, built on 1,000 hours mining Florida's own malpractice-claims database — law-review editing isn't blind peer review, but this is the same class of citable original academic work.

The Florida Bar is a state bar association — the same category as the ABA example in CLAUDE.md's sourcing rule — reporting via its own news channel.

Reputable, non-collaborative tech journalism conducting its own citation-tracing investigation into Wayne and Bossavit's research, not merely repeating someone else's essay — CLAUDE.md's fourth, 'similarly citable material' category.

NIST is the US government's own standards body, publishing its own economic-impact report.

The book itself, by the researcher who did the citation-tracing work — a primary account, not journalism about it.

An original essay by its actual author, a named software-process authority, making a specific argued case.

ASQ's own page returned HTTP 403 to every automated check (not confirmed bot-friendly), so this quotes and attributes the figure to ASQ one step removed rather than first-hand — flagged honestly as secondary, not upgraded to look primary.

The peer-reviewed paper itself, not a summary of it — IEEE Computer Society's own published record.

The original peer-reviewed investigation, IEEE's own published record of it.

The European Space Agency's own account of its own Inquiry Board's findings.

NASA's own Lessons Learned Information System hosting its own investigation board's report.

The US House committee's own final report from its own 18-month investigation.

A peer-reviewed conference paper from the American Society for Engineering Education, formalising the prevention/appraisal/failure cost-of-quality model.

Toyota's own official global site describing its own production system in its own words.

NASA's own official mishap investigation board report into its own mission failure — CLAUDE.md's own named example of a primary organisation.

CB Insights is the market-research firm that compiled and published this venture-funnel data itself, not a third party summarising someone else's study.

Procore is the construction industry's most widely used project-management platform; this is its own practitioner-authored explainer of how occupancy inspections and gating actually work, used here for procedural facts rather than a contested statistic.

Boeing's own official annual accident-statistics publication — CLAUDE.md's own named example of a primary organisation.

Gabriel's own essay, in his own words, making and later reflecting on his own argument — the primary text this week is built around.

A peer-reviewed meta-analysis published in a Nature-family journal.