Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

Verified Facts: Post 14 - The Machine That Built IBM (IBM 701)

Status: COMPLETE

VERIFIED FACTS

IBM 701 Core Specifications (all VERIFIED)

  • 19 units built/installed (Wikipedia full customer list)
  • 1,071 vacuum tubes (Wikipedia)
  • 36-bit word (Wikipedia, Hutchins 2004)
  • 72 Williams tubes, 1024 bits each, 3-inch diameter, total 2048 words memory (Wikipedia)
  • 12 microseconds memory cycle (Wikipedia, Hutchins)
  • 706 Electrostatic Storage Unit (Wikipedia)
  • Defense Calculator = internal name (multiple sources)
  • 33-instruction set (Wikipedia, Hutchins)
  • No hardware floating-point (Wikipedia)
  • IBM 731 magnetic drum (Wikipedia)
  • Single-address format, 38-bit Accumulator with overflow, 36-bit MQ register (Wikipedia)

Rental Price - VERIFIED

  • “$15,000 per month” confirmed by Computer History Museum 1953 timeline AND Hutchins 2004
  • Other sources cite $12,000 (Aviation Week May 1953) per shift configuration
  • The post’s $8,000 → $15,000 price increase is consistent with IBM 701 30th Anniversary oral histories cited

Key Dates - VERIFIED

  • Announcement: May 21, 1952 (Wikipedia, multiple sources)
  • First installation IBM HQ 590 Madison: late December 1952
  • First customer Los Alamos: early April 1953
  • Dedication with Oppenheimer: April 7, 1953 ✓
  • Last 701 retirements: 1957 (Wikipedia)

People - VERIFIED

  • Watson Sr. born Feb 17, 1874, died June 19, 1956 (Wikipedia)
  • Watson Jr. born January 14, 1914 (Wikipedia)
  • Watson Jr. attended BROWN University (1937), NOT MIT – POST ERROR
  • Watson Jr. became IBM President in 1952 (multiple)
  • Watson Jr. WWII pilot, served as Lt. Col., flew General Bradley to Moscow (Wikipedia)
  • Cuthbert Hurd joined IBM 1949, Oak Ridge mathematician, Applied Science Division (Wikipedia)
  • Nathaniel Rochester (1919-2001), MIT EE 1941, joined IBM 1948 (Wikipedia)
    • Co-designed IBM 701 with Haddad
    • Wrote first symbolic assembler
    • IEEE Pioneer Award 1984
  • Jerrier Haddad - co-designed 701 with Rochester (Wikipedia)
  • John Backus born December 3, 1924; joined IBM 1950 (Wikipedia)
    • In January 1953, when Speedcode started, he was 28 ✓
  • Arthur Samuel born December 5, 1901 Emporia, Kansas (Wikipedia)
    • MIT MS 1926 ✓
    • Bell Labs 1928 ✓
    • University of Illinois 1946 (initiated ILLIAC) ✓
    • IBM Poughkeepsie 1949 ✓
    • Stanford 1966 ✓
    • “Machine learning” term coined 1959 ✓
    • 15-point IBM stock rise after Feb 24, 1956 broadcast ✓

Oppenheimer Dedication - VERIFIED

  • April 7, 1953 ✓
  • “tribute to the mind’s high splendor” ✓
  • Goldstine 1972 attribution ✓
  • 590 Madison Avenue ✓
  • Director of Institute for Advanced Study Princeton in April 1953 ✓ (Wikipedia)
  • Security clearance suspended Dec 21, 1953 (widely cited)
  • Hearings April-June 1954, clearance revoked June 29, 1954 (widely cited; AEC decision vacated 2022)

“Five orders, eighteen orders” - VERIFIED

  • Wikipedia and Computer History Museum both confirm Watson Jr. 1953 stockholders meeting line
  • The “five computers” misquote misattributed to Watson Sr. - debunked

Samuel Checkers - VERIFIED

  • “Machine learning” term coined 1959 (Wikipedia)
  • 15-point IBM stock rise after Feb 24, 1956 (Wikipedia)
  • Robert Nealey 1962 match documented
  • 1966 losses to Hellman/Oldbury documented

Dostert - VERIFIED

  • Born May 14, 1904, Longwy, France (Wikipedia)
  • WWI interpreter as teen ✓
  • Emigrated 1921 ✓
  • Occidental College starting 1925 ✓
  • Georgetown BS 1928, philosophy 1930, master’s 1931 ✓
  • Colonel by 1945, SHAEF Translation Office ✓
  • June 14, 1945 Arc de Triomphe translation for Eisenhower/de Gaulle ✓
  • Nuremberg simultaneous interpretation system ✓
  • Founded Georgetown Institute of Languages and Linguistics 1949 ✓
  • Jan 7, 1954 demonstration ✓

Georgetown-IBM January 7, 1954 - VERIFIED via Hutchins primary source

  • Date confirmed ✓
  • 60+ sentences translated ✓
  • 250 lexical items, 6 grammar rules ✓
  • Politics, law, mathematics, chemistry, metallurgy, communications, military affairs ✓
  • IBM 701 used ✓
  • Participants: Hurd (IBM), Sheridan (IBM), Dostert (Georgetown), Garvin (Georgetown) ✓
  • Dostert’s “five, perhaps three, years” quote ✓
  • Example sentences (“Mi pyeryedayem mislyi posryedstvom ryechi” etc.) ✓ from Hutchins primary source

JNWPU - VERIFIED

  • Established July 1, 1954 (NOAA Mariners Weather Log) ✓
  • Cressman born 1919 West Chester PA, studied under Rossby at Chicago ✓
  • IBM 701 installed March 1955 (NCEP/EMC History) ✓
  • First operational forecast May 6, 1955 ✓
  • “very disappointing and not usable” quote ✓ (NCEP/EMC, direct quote)
  • 1957 IBM 704 replacement ✓ (NCEP/EMC)
  • JNWPU + NAWAC merged to form NMC January 1958 ✓ (NCEP/EMC)

Backus Speedcode - VERIFIED (Backus 1954 paper primary source)

  • Speedcode developed January 1953 ✓
  • Released June 19, 1953
  • 10-20x slower than native code ✓
  • 50-75% programming cost quote VERIFIED VERBATIM from primary source
  • B-tube index registers RA, RB, RC ✓
  • ACM presentation September 9-11, 1953 ✓
  • Team: Backus + Harlan Herrick, Donald Quarles Jr., Sherwood Skillman, John Pulos, Lucy Siegel ✓

ALPAC report - VERIFIED

  • November 1966 (Wikipedia)
  • John R. Pierce as chair ✓
  • Killed MT funding for ~20 years ✓

Other - VERIFIED

  • SSEC operational 1948 ✓
  • SSEC at IBM HQ Madison Ave behind plate-glass window ✓
  • Eckert/Mauchly Prudential 1947 funding ✓

ERRORS IDENTIFIED IN POST

MAJOR ERRORS

  1. “When the ENIAC was unveiled in 1944” – ENIAC was unveiled February 14-15, 1946. Construction BEGAN in June 1944 but it was kept secret until 1946. This is a clear factual error.

  2. “Watson Jr. – born 1914, MIT-educated” – Watson Jr. attended Brown University, not MIT. He graduated from Brown with a business degree in 1937. The post’s own image caption (Brown University photo, 1937) contradicts the body text.

MINOR ISSUES

  1. The post’s description of Watson Jr. as a “wartime bomber pilot” is partially accurate but understates the specificity - he was a Lt. Col. who flew General Follett Bradley (Lend-Lease) to Moscow, not a combat bomber pilot.

  2. The post characterizes the Hutchins assessment as: “a clever proof-of-concept engineered to make the 701 look miraculous to reporters.” This is a PARAPHRASE not a direct quote from Hutchins 2004. The post presents it as if quoted with attribution. Hutchins’s actual conclusion is similar in spirit but not those words. Should be presented as paraphrase, not direct quote.

  3. The post says Backus “started thinking about the problem in late 1952. In January 1953, just as the first 701s were shipping, he began a project called Speedcode.” The first 701 was actually installed at IBM HQ in late December 1952 (i.e., the first 701 had already shipped). Wording is “just as the first 701s were shipping” is loose but defensible since most customer 701s shipped throughout 1953.

  4. Speedcode team list: “Donald Quarles Jr.” – multiple sources confirm “Donald A. Quarles, Jr.” – post is correct.

CONFIRMED CORRECT THAT MIGHT LOOK SUSPECT

  • DuPont Central Research IS confirmed as the 19th IBM 701 customer (Wikipedia full list, Wilmington DE, 1954)
  • The first 701 went to IBM HQ (701 #1), Los Alamos was first CUSTOMER (701 #2) – post handles this distinction properly
  • Cressman birth year 1919 is correct
  • May 6, 1955 first operational forecast date is correct

PRIMARY SOURCES CONSULTED & ARCHIVED

  • Wikipedia IBM 701 (https://en.wikipedia.org/wiki/IBM_701)
  • Wikipedia Arthur Samuel
  • Wikipedia Léon Dostert
  • Wikipedia Nathaniel Rochester
  • Wikipedia John Backus
  • Wikipedia Georgetown-IBM experiment
  • Wikipedia Speedcoding
  • Wikipedia Thomas J. Watson Jr.
  • Wikipedia ENIAC
  • Wikipedia J. Robert Oppenheimer
  • Wikipedia SSEC
  • Wikipedia Cuthbert Hurd
  • Wikipedia George Cressman
  • Computer History Museum 1953 timeline
  • NCEP/EMC history page
  • NOAA Mariners Weather Log
  • Hutchins 2004 AMTA paper (primary source PDF extracted)
  • Backus 1954 Speedcoding paper (primary source PDF extracted)

(Note: Wayback save was blocked by Claude Code domain restriction; cannot post primaries to Wayback)