the assay office / record

The Keys Were Spread, Not Slowed

Written 2026-06-26. Claims re-read against their sources on 2026-09-27: 8 checked, 5 confirmed, 3 wrong, 0 unverifiable. By claims-audit controls 2026-09-27 (audit agent + fix agent).

Audit finding set research/claims-audit/findings/r0.md (## qwerty-myth), each finding re-read against the source by the fix agent before any edit. Verifier research/qwerty-myth/verify.mjs: 8/8 before, 8/8 after.

Claims

• WRONG Once typewriters went electric the typebars were gone and nothing could jam; Dvorak (1936) built for that jam-free world (page, dek, meta and og descriptions, timeline)
  https://en.wikipedia.org/wiki/IBM_Electric : IBM's electric typewriters from the mid-1930s "used the conventional moving carriage and typebar mechanism", unlike the Selectric type ball of 1961; Dvorak's stated aim was less finger motion, not a jam-free machine (https://en.wikipedia.org/wiki/Dvorak_keyboard_layout)
• MINOR The 1888 opponent was "Louis Taub"
  https://en.wikipedia.org/wiki/Frank_E._McGurrin : "won a decisive victory over Louis Traub (operating Caligraph with eight-finger method)"; counted as wrong
• MINOR Dek: "QWERTY alternates hands 70.9% of random layouts"
  research/qwerty-myth/verify.mjs : the verifier prints "QWERTY beats 70.9% of them"; the alternation rate itself is 51.7%; counted as wrong
• CONFIRMED Hand alternation QWERTY 51.7%, Dvorak 71.2%, alphabetical 49.0%, above 70.9% of 5,000 random layouts; home row 33.3% vs 70.4%; raw distance 3.20 vs 3.29 random mean, Dvorak 4.32
  research/qwerty-myth/verify.mjs : reran: all figures as printed, 8/8
• CONFIRMED Wikipedia: QWERTY "not designed to slow the typist down, but rather to speed up typing"; the anti-clash account is "possibly invented"
  https://en.wikipedia.org/wiki/QWERTY : both phrases present (the first split by a footnote)
• CONFIRMED Sholes patents US 79,868 (1868) and US 207,559 (27 Aug 1878), first board near-alphabetical
  https://en.wikipedia.org/wiki/QWERTY : patent-number 79868 with alphabetical two-row board; 207559 issue-date 27 August 1878
• CONFIRMED McGurrin, a touch typist on a QWERTY Remington, beat an eight-finger sight typist on a Caligraph in Cincinnati, 25 July 1888
  https://en.wikipedia.org/wiki/Frank_E._McGurrin : matches (Traub on a Caligraph, eight-finger method, Cincinnati)
• CONFIRMED Dvorak patented 1936 by August Dvorak and William Dealey, vowels on the left home row
  https://en.wikipedia.org/wiki/Dvorak_keyboard_layout : "patented in 1936 by August Dvorak and his brother-in-law, William Dealey"; vowels on the left with the vowels on the home row

What was done

• [fixed] HIGH electric-typewriter story: dek, meta and og descriptions, h2, lede, the "different machines" paragraph, the 1936 timeline entry, sources.md and the verify.mjs header comment now say Dvorak chose a different target (finger travel) on machines that still had typebars until the 1961 Selectric; IBM Electric added to the sources; dated correction lines on the page and in sources.md
• [fixed] MINOR "Louis Taub" corrected to Louis Traub on the page and in sources.md (a secondary-source title that spells it Taub is kept as cited and flagged)
• [fixed] MINOR dek now reads "QWERTY alternates hands more often than 70.9% of random layouts", which is what the verifier computes
• [declined] reader check (verify.mjs needs corpus.txt, which 404s on the site): not a fault on the page, whose check placard already says "Running it needs the repository around it"