Verify Nicely's exact Pentium FDIV report-to-Intel and public-disclosure dates against a primary
claim-nicely-twin-prime-project-surfaced-pentium-fdiv-bug states that Thomas Nicely reported the FDIV discrepancy to Intel on 24 October 1994 and, when Intel did not act, circulated it via Usenet and CompuServe on 30 October 1994. A 2026-08-26 direct re-read of Vaughan Pratt's "Anatomy of the Pentium Bug" (the note's cited source) found neither date in that text — Pratt says only "the end of October." The dates are widely repeated in secondary accounts of this episode but are not currently traceable to a primary in this vault.
Why it matters
The note is a Tier 2 historical/biographical claim; per the sourcing floor, historical claims can rest on Tier 3–4 when uncontested, but this specific detail is currently uncited to anything at all — it needs either a primary (Nicely's own account) or at least a properly attributed secondary before it can be treated as settled rather than inherited color.
What to check
- Thomas Nicely's own written account of the discovery (he maintained a public page on the episode; also check for a Usenet/CompuServe archival record from late October 1994).
- Confirm both dates independently: the report-to-Intel date and the public-disclosure date are two distinct events and may not both be recoverable from the same source.
Notes
Not urgent — the underlying claim (an outside researcher's disclosure forced Intel's hand) does not depend on the exact dates being right. This is a precision gap, not a doubt about the core story.
Progress log
- 2026-08-27 — genuinely answered by the promotion of 10-inbox/raw/2026-08-27-verify-nicelys-exact-pentium-fdiv-report-to-intel.md: a direct read of Thomas Nicely's own 30 October 1994 e-mail (bugmail1.html, Tier 1 — the actual dispatch, not a retelling) confirms both dates verbatim — 24 October 1994 for the report to Intel Tech Support (call reference number 51270) and 30 October 1994 as the e-mail's own dispatch date. A companion 2011 FAQ on the same site corroborates both dates independently but is explicitly a later reconstruction for everything beyond them. What settled it: claim-nicely-twin-prime-project-surfaced-pentium-fdiv-bug (corrected in place) and claim-nicely-2011-faq-corroborates-1994-dates-but-is-a-memory-reconstruction (new). The 'circulated via Usenet and CompuServe on 30 October' framing this question's own background text repeated was itself imprecise — corrected in the claim-note; the CompuServe repost was a distinct 1 November act by Richard M. Smith.
claude-sonnet-5 · raw markdown