Public-record guide · temporal evidence
A fresh search result is not a fresh research claim
Search crawl time, publication time, registry time, and citation-observation time answer different questions. Keeping those clocks separate prevents a newly visible page from becoming a falsely new achievement.
Status: Evidence-backed public-record guide · Current search and registry snapshot · No new canonical work or DOI version detected · Citation claims withheld · No change to peer-review, employment, advisory, production, or institutional-authority status
Four clocks in one public result
| Clock | What it can establish | What it cannot establish |
|---|---|---|
| Search crawl | A search engine recently discovered or refreshed a page. | That the underlying work is new, peer-reviewed, cited, current, or authoritative. |
| Publisher posting | The date and version that a publisher made public, together with the publisher's status label. | That a later crawl date created a later version or changed review status. |
| Registry deposit or indexing | When metadata entered Crossref, Europe PMC, PubMed, or another bounded index. | That index arrival equals publication, a unique work, or a citation event. |
| Citation observation | A count observed in a named source on a named date under that source's coverage rules. | A source-independent impact total or a stable fact that can be copied across platforms. |
What changed in the sampled search
On 2 October 2026, the sampled English name search surfaced the canonical Emanuel Shirbint research archive directly, alongside controlled or primary LinkedIn, ResearchGate, PubMed/PMC, Preprints.org, and National Library records. Earlier sampled runs had not returned the canonical site directly. This is a search-visibility change only: the canonical identity remains Emanuel Shirbint, with עמנואל שירבינט and Эмануэль Ширбинт retained as controlled alternate-name forms.
What did not change in the primary records
The current Crossref reconciliation still contains 15 DOI records naming Shirbint in author metadata, with no such record created since the 29 September check. Europe PMC still returns 14 author-query records and PubMed still returns four. The two checked Preprints.org lineages remain at Version 2: The Reinforcement Learning Contour is conceptual research, and Bounded Recursive Self-Improvement Under Human Governance reports a bounded local empirical experiment. Both publisher pages explicitly state that the current versions are not peer-reviewed.
Those registry totals count different objects and coverage scopes. They are not interchangeable publication totals and they are not citation counts. Citation counters remain excluded from the public claim ledger because their sources and counting rules have not been normalized.
A safe update rule
- Use search results to discover a candidate source, not to settle the claim.
- Resolve the candidate to a publisher, DOI registry, or bounded bibliographic index.
- Compare title, authors, persistent identifier, version, date, and publication-status label.
- Classify the event as search refresh, index propagation, new version, or new work.
- Change the public catalogue only when the evidence class itself changed.
Claim boundaries
Peer-reviewed work, conference work, preprints, conceptual research, the bounded empirical preprint, software, the locally evidenced EWCC production system, employment, and advisory work remain separate evidence classes. A recent crawl cannot transfer evidence between those classes. This guide makes no claim of clinical deployment, regulatory approval, external production validation, present employment, advisory engagement, institutional representation, or citation impact.
Primary and canonical records
- Crossref author-query record set
- Europe PMC author-query record set
- PubMed author-query record set
- The Reinforcement Learning Contour — current publisher record
- Bounded Recursive Self-Improvement Under Human Governance — current publisher record
- Canonical deduplicated research library
Evidence checked 2 October 2026 (Asia/Jerusalem). Search results were used for discovery only; identity, identifier, version, date, and publication-status statements were checked against publisher and registry records.