Next Post Analysis

Post coverage audit as of 2026-04-10


1. All 14 Posts and Their Main Characters

# Post Date Main Characters Machines
1 The Man Who Forecasted Weather with a Pencil 2026-03-24 Richardson (manual calculation)
2 The Number That Connects Turbulence to War 2026-03-25 Richardson (manual)
3 The Line That Models Cannot Draw, Part 1 2026-03-26 V. Bjerknes, J. Bjerknes
4 The Line That Models Cannot Draw, Part 2 2026-03-27 Browning (conveyor belt)
5 The Line That Models Cannot Draw, Part 3 2026-03-28 (ECMWF/WRF framing) modern NWP
6 The Man Who Tamed the Equations 2026-03-29 Charney ENIAC
7 The First Climate Model Had 5 KB of RAM 2026-03-30 Phillips IAS machine
8 The Butterfly That Broke the Forecast 2026-03-31 Lorenz LGP-30
9 From Cables to Chaos: Four Computers 2026-04-02 ENIAC/IAS/IBM704/LGP-30 ensemble; Frankel, Backus, Amdahl mentioned ENIAC, IAS, IBM 704, LGP-30
10 The Swedes Got There First 2026-04-03 Rossby, Bolin BESK
11 The Magician Who Told No Secrets 2026-04-04 Beurling (FRA context)
12 The Blueprint Von Neumann Gave Away 2026-04-08 von Neumann; Metropolis, Ulam, Barricelli (supporting) IAS, MANIAC, JOHNNIAC, ILLIAC, ORDVAC, WEIZAC, SILLIAC, BESM
13 The Machine That Learned Too Early 2026-04-09 Rosenblatt, Minsky IBM 704 (Mark I Perceptron)
14 The Machine That Built IBM 2026-04-10 Watson Jr., Hurd, Rochester, Cressman, Samuel, Dostert, Shuman IBM 701

The following people have dedicated research files but have not yet been the focus of any post. “Supporting mention” is noted where relevant.

Never or barely appeared

Person Research File Notes on Research Depth
Jule Charney Charney.md Already main character (Post 6). DONE.
Tor Bergeron Bergeron.md Bergen School - no dedicated post yet; appeared in Posts 3-5 peripherally
Sverre Petterssen Petterssen.md D-Day forecast angle; no post yet
Halvor Solberg Solberg.md Polar front co-author; minor figure
Jorgen Holmboe Holmboe.md Recruited Charney to UCLA; no post yet
Eric Eady Eady.md Baroclinic instability; no post yet
Klara Dan von Neumann Klara_von_Neumann.md Mentioned in passing; full story not told
Adele Goldstine Adele_Goldstine.md Not yet featured; significant women-of-computing file
Six ENIAC Women ENIAC_programmers.md Not yet featured; substantial 6-person file
Herman Goldstine Herman_Goldstine.md Supporting in Post 12; not main character
Julian Bigelow Bigelow.md IAS chief engineer; no post yet
Jay Forrester Forrester.md Whirlwind/core memory; no post yet
Stan Frankel Frankel.md Mentioned in Post 9 (cables to chaos); full story not told
Nicholas Metropolis Metropolis.md Supporting in Post 12; not main character
Stanislaw Ulam Ulam.md Supporting in Post 12; not main character
Joseph Smagorinsky Smagorinsky.md ENIAC team; GFDL founder; no post yet
Syukuro Manabe Manabe.md Nobel laureate; no post yet
George Platzman Platzman.md ENIAC team; NWP historian; no post yet
Ragnar Fjortoft Fjortoft.md ENIAC team; 2D turbulence theorem; no post yet
Barry Saltzman Saltzman.md Convection model Lorenz simplified; mentioned in Post 8 but not main character
Akio Arakawa Arakawa.md Arakawa grids; no post yet
Kerry Emanuel Emanuel.md Hurricane physics/chaos; supporting
Peter Lynch Peter_Lynch.md Richardson vindication; NWP historian; no post yet
Bert Bolin Bolin.md Swedish NWP; appeared in Post 10 as supporting character
Erik Stemme Stemme.md BESK builder; appeared peripherally in Post 10/11
Gosta Neovius Neovius.md BESK architect; appeared peripherally in Post 10
Carl-Erik Froberg Froberg.md IAS visitor 1947-48; built SMIL at Lund
Conny Palm Conny_Palm.md Queueing/statistics at Ericsson; died 1951 before BESK completed
Polish codebreakers Polish_codebreakers.md Rejewski/Rozycki/Zygalski; mentioned briefly in Post 11
Gene Amdahl Amdahl.md IBM 704 architect; mentioned in Post 9
Nathaniel Rochester Rochester.md IBM 701 designer; featured supporting in Post 14

Already covered as main characters (DO NOT repeat without new angle)

Richardson (Posts 1-2), V. Bjerknes + J. Bjerknes (Post 3), Browning (Post 4), Charney (Post 6), Phillips (Post 7), Lorenz (Post 8), Beurling (Post 11), von Neumann (Post 12), Rosenblatt + Minsky (Post 13), Watson Jr. + Hurd (Post 14).


3. Computers with Research Files Not Covered in Any Post

Computer Research File Status
EDVAC EDVAC.md + EDVAC_additional.md No dedicated post; mentioned only as stored-program context
ORDVAC ORDVAC.md Covered briefly in Post 12 (IAS clone diaspora)
JOHNNIAC JOHNNIAC.md Covered briefly in Post 12
ILLIAC ILLIAC.md Covered briefly in Post 12
WEIZAC WEIZAC.md Covered briefly in Post 12
SILLIAC SILLIAC.md Covered briefly in Post 12
BESM BESM.md Covered briefly in Post 12
MANIAC MANIAC.md Covered briefly in Post 12
LGP-21 LGP-21.md Mentioned in Post 9 only
RPC-4000 RPC-4000.md No dedicated post; Lorenz’s later machine
IBM 709/7090 IBM_709_7090.md No dedicated post; only listed in Post 9 overview
Whirlwind Whirlwind.md No dedicated post; mentioned in context of SAGE
Cray-1 Cray-1.md No dedicated post yet
Cray-2 Cray-2.md No dedicated post yet

Already covered with dedicated posts

ENIAC (Post 6, 9), IAS machine (Post 7, 12), IBM 701 (Post 14), IBM 704 (Post 9, 13), LGP-30 (Post 8, 9), BESK (Post 10, 11).


4. The 7 Open Threads: Research Assessment

Thread 1: Polish Codebreakers

Research file: Polish_codebreakers.md

What exists:

  • Deep biographies of all three: Rejewski (1905-1980), Rozycki (1909-1942), Zygalski (1908-1978)
  • Full technical story of the Enigma crack (1932): permutation group theory, cyclometer, bomba, Zygalski sheets
  • Pre-war intelligence sharing: the July 1939 handoff to Britain and France at Pyry
  • Post-war suppressions: Polish contribution hidden until 1974 (Polish law) then 1980s (Hinsley official history)
  • Individual post-war fates: Rejewski lived to see recognition; Rozycki drowned 1942; Zygalski died in UK 1978 months before the full story broke
  • Connection to Bletchley Park clearly established
  • No dedicated Geheimschreiber_crypto.md angle (that’s Beurling’s story, already covered in Post 11)

What’s missing for a post:

  • No research on Marian Rejewski’s post-war accounting job (a Nobel-calibre cryptanalyst working as a bookkeeper)
  • The story of how Denniston and Dilly Knox initially dismissed Polish input
  • The specific mathematical elegance of the Enigma solution could be more narrative-friendly

Narrative strength: HIGH. The injustice angle (suppressed for 40+ years), the three distinct personalities, the pre-war tension, the drowning of Rozycki - this is dramatic. Strong link to Post 11 (Beurling/Geheimschreiber) as a parallel crypto story. Polish blog author personal connection is significant.


Thread 2: Women of Computing

Research files: Klara_von_Neumann.md, Adele_Goldstine.md, ENIAC_programmers.md

What exists:

  • Klara von Neumann: substantial file (~5,000 words estimated from earlier reading). Full biography: Budapest ice skating champion, four marriages, coded the ENIAC weather forecast and H-bomb simulations, suicide in 1963 at La Jolla. The tragedy arc is complete.
  • Adele Goldstine: trained the ENIAC programmers, wrote the technical manual, died 1964 aged 44. File appears to be of moderate depth.
  • Six ENIAC programmers: comprehensive individual biographies for all six (McNulty/Antonelli, Jennings/Bartik, Snyder/Holberton, Wescoff/Meltzer, Bilas/Spence, Lichterman/Teitelbaum). The 1946 unnamed-in-photos story is documented.
  • Thread note in INDEX.md also mentions Mary Tsingou, Arianna Rosenbluth, Thelma Estrin, core memory threaders - but no dedicated files for these.

What’s missing:

  • No files for Mary Tsingou (first Monte Carlo simulation, attribution dispute with Fermi), Arianna Rosenbluth (first published Monte Carlo code), Margaret Hamilton (Apollo software)
  • Core memory threaders (the “housewives with knitting experience” who assembled BESK’s ferrite cores) mentioned in BESK.md but no individual identities known

Narrative strength: VERY HIGH. Three self-contained dramatic arcs (six women erased from the 1946 photo; Adele Goldstine who taught them and died young; Klara who coded the weather forecast and the H-bomb and drowned in the Pacific). Could be done as a single thematic post or split into 2. The erasure from history structure is proven powerful in popular science writing.


Thread 3: IBM 7090 and 1960s NWP

Research file: IBM_709_7090.md

What exists:

  • Full technical specs for IBM 709 (1957), 7090 (1959 transistorized), 7094 (1962)
  • Data channels / DMA innovation
  • NASA/space program use (early satellite trajectory calculations)
  • “Daisy Bell” (HAL 9000’s song origin story) - IBM 7094 at Bell Labs, 1961
  • Song: “A Computer Sang to Me” by John Kelly and Carol Lochbaum

What’s missing (for NWP angle):

  • The 7090 replaced the IBM 704 at JNWPU (this transition is important for NWP history but not richly documented in the research file)
  • No file on Frederick Shuman’s work on primitive equations at JNWPU in the early 1960s (Shuman.md is a planned but not yet created file)
  • The JNWPU’s transition from barotropic to baroclinic models happened on the 7090 - this is a crucial NWP milestone not yet written up

Narrative strength: MODERATE. Technically interesting (transistors, 6x speed, NASA use) but the NWP angle is thin without more research on the 1960s JNWPU operations. The “Daisy Bell” / HAL 9000 connection is a delightful hook but pulls the story away from weather. Would need either new research on 1960s NWP operations or reframing as a pure computing post.


Thread 4: Cray-1 / NCAR

Research file: Cray-1.md, Cray-2.md

What exists in Cray-1.md:

  • Full technical specs: 160 MFLOPS peak, 80 MHz, vector registers, chaining
  • NCAR as first paying customer (serial #3, arrived July 1977, $8.86 million, accepted December 1977)
  • The famous C-shaped cabinet with ring of padded benches (“world’s most expensive love seat”)
  • Cooling system engineering (Freon, 6-month delay from lubricant leakage)
  • ECMWF installation (96-hour MTBF noted)
  • Seymour Cray’s design philosophy: CDC STAR-100 failure analysis, scalar/vector balance
  • NSA as early customer
  • All Cray-1 variants (1S, 1M) documented
  • Notable achievements section (not yet read in full)

What’s missing for NWP angle:

  • No research on what specifically NCAR ran on the Cray-1 (which atmospheric models, which breakthroughs)
  • No file on Seymour Cray himself as a biographical subject
  • The CDC 7600 predecessor at NCAR and why the 4.5x throughput improvement mattered for weather is not detailed
  • No file on ECMWF’s Cray-1 purchase and the European Centre’s founding story

Narrative strength: HIGH. Seymour Cray is an extraordinary character (hermit engineer who built tiny rooms to dig tunnels and “talk to the elves” for inspiration - documented in legend). The NCAR angle directly connects to weather. The 1977 transition from CDC 7600 to Cray-1 at NCAR is a genuine NWP milestone. Would need a Seymour_Cray.md file written to do full justice to the human story.


Thread 5: Smagorinsky → Manabe → Nobel

Research files: Smagorinsky.md, Manabe.md

What exists:

  • Smagorinsky: Full biography. Son of Belarusian immigrants, Stuyvesant HS, WWII bomber weather observer, ENIAC team 1950, GFDL founded 1955 at von Neumann’s instigation (age 29). 1963 primitive-equation GCM paper. Smagorinsky-Lilly turbulence model. 28 years as GFDL director.
  • Manabe: Full biography. Japanese village doctor’s son, D.Sc. University of Tokyo 1958, recruited to GFDL by Smagorinsky 1959. 1967 CO2 paper with Wetherald (quantifying greenhouse effect for first time), 1969 coupled ocean-atmosphere model, 1975 CO2-doubling study, 1979 Charney Report, 2021 Nobel Prize. Rich quotes: “great fun,” “curiosity is the thing.”

What’s missing:

  • No file on Richard Wetherald (Manabe’s key collaborator on the 1967 and 1975 papers)
  • No file on Kirk Bryan (first coupled ocean-atmosphere model)
  • The GFDL institutional story (move from Washington to Princeton 1968, relationship with Princeton University) not fully developed as narrative
  • The chain from ENIAC to Nobel (Smagorinsky was on the 1950 ENIAC team that started the whole thing) is not yet written as a single narrative arc

Narrative strength: VERY HIGH. The chain is perfect: immigrant’s son from NYC runs the first computer weather forecast in 1950, founds GFDL, recruits a Japanese village doctor’s son in 1959, that recruit quantifies the greenhouse effect in 1967, wins the Nobel Prize 54 years later. All primary files exist. The CO2 paper is historically significant for current readers in a way no other thread can match. Manabe’s quoted personality (“great fun”) makes him very writable.


Thread 6: Stan Frankel’s Arc

Research file: Frankel.md

What exists:

  • Full biography: Berkeley Ph.D. under Oppenheimer, Los Alamos T-Division 1943, organized Group T-5 (human computers assembly line)
  • The Feynman story: Frankel caught “the computer disease” computing arctangents while critical bomb calculations waited; Feynman took over
  • ENIAC H-bomb calculations November 1945 (first nuclear physics calculation on electronic computer)
  • Red Scare clearance revocation; cut off from national labs
  • Reinvention: Caltech, designed MINAC (113 vacuum tubes), licensed to Librascope as LGP-30
  • LGP-30 sold 500+ units; Lorenz bought his chaos-discovery machine because it was the only affordable option
  • Frankel died May 1978 aged 58; never knew his machine produced chaos theory

What’s missing:

  • The circumstances of his clearance revocation are not well documented
  • His later years after the LGP-30 (1956-1978) are thin
  • No quotes from Frankel himself; most documentation comes from Feynman’s account

Narrative strength: HIGH. The arc from Los Alamos to butterfly wings is extraordinary. However, the documentation is thinner than the Smagorinsky/Manabe thread, particularly for the post-LGP-30 years, and Frankel left few direct personal records. Also: Frankel was mentioned in Post 9 (cables to chaos) as one of the key figures, so there’s setup already. Could be the missing deeper dive that Post 9 promised.


Thread 7: Whirlwind / Forrester / Core Memory

Research files: Whirlwind.md (~370 lines, the largest computer file), Forrester.md

What exists in Whirlwind.md:

  • Complete technical specs (16-bit, 1 MHz, real-time, 5000 vacuum tubes, 2000 sq ft)
  • Design philosophy: bit-parallel for real-time; deliberate departure from serial (EDVAC) approach
  • Origins: Navy flight simulator that became a general computer
  • Core memory installation 1953: doubled speed to 40,000 operations/second
  • SAGE connection: Whirlwind as prototype for $8 billion air defense system; largest computer project in history
  • First computer graphics (bouncing ball on oscilloscope, 1948-49)
  • First real-time interactive computer: operator could intervene mid-calculation

What exists in Forrester.md:

  • Nebraska cattle ranch upbringing; wind-powered electrical system from car parts at age 14
  • MIT graduate student 1939; servomechanisms under Gordon Brown
  • Core memory invention 1949: 3D ferrite ring matrix, patent 1956
  • William Papian fabricated first 2x2 test array October 1950
  • SAGE development; Lincoln Laboratory
  • DEC founders trained at Whirlwind (not in file - well-known context)
  • Later career: system dynamics, “Limits to Growth” connection

What’s missing:

  • No weather/NWP connection for Whirlwind or Forrester
  • SAGE was air defense, not meteorology
  • Core memory was used in IBM 704 (which did run NWP) - that’s the NWP connection, but it’s indirect

Narrative strength: HIGH for a computing history post, but LOW for NWP series specifically. The story of core memory threading by women workers (the “housewives with knitting experience” at BESK - same technology) could provide the women-of-computing connection. But the narrative hook for connecting this to weather is weaker than the other threads. Would need a strong framing device.


5. Top 3 Candidates for Next Post

Rank 1: Smagorinsky → Manabe → Nobel

Why: This is the thread that closes the NWP-to-climate-change arc the entire series has been building toward. The research is fully sufficient - Smagorinsky.md and Manabe.md together contain everything needed. The chronology is already established in the series (ENIAC 1950 forecast → GFDL 1955 → 1967 CO2 paper → 2021 Nobel). The contemporary relevance is unmatched: climate change is the reason NWP history matters to most readers today. Manabe’s quoted personality (“great fun”) and Smagorinsky’s immigrant-son-to-GFDL-founder arc are both very writable. No new research files needed.

Risks: Post must avoid being a climate-change polemic. The hook should be biographical/computational (how did a 1963 primitive-equation model lead to a Nobel 58 years later?) rather than political.

Suggested title angle: “The Man Who Modeled the Future” or “The Quiet Man Who Quantified the Greenhouse” – focusing on Manabe as the protagonist, with Smagorinsky as the enabling mentor.


Rank 2: Polish Codebreakers

Why: The research file is exceptionally rich. The three distinct personalities (the quiet mathematician Rejewski, the young geographer Rozycki who drowned at 32, the meticulous Zygalski who died in UK just before the story was finally told) support a multi-character narrative. The suppression arc (40+ years) is a genuine historical injustice with a satisfying resolution. Strong connection to Post 11 (Beurling/Geheimschreiber) as a parallel European crypto story – could be explicitly framed as the companion piece. The author’s Polish identity is a legitimate personal hook.

What would make it stronger: A note about how Rejewski’s permutation group theory was the first application of abstract algebra to cryptanalysis – the mathematical sophistication is comparable to Charney’s quasi-geostrophic filtering. This comparison makes it relevant to the NWP series.

Risk: The NWP connection is tenuous (the post is really a crypto/WWII post, not a weather post). Works best if framed as part of the broader “mathematics saves the world” theme of the series, or if it follows the Beurling post explicitly as “Part 2.”


Rank 3: Women of Computing (ENIAC Programmers + Klara von Neumann)

Why: Three distinct, research-backed tragic arcs in one post. The erasure-from-history structure mirrors the Polish codebreakers’ suppression. The ENIAC programmers file is the most thorough multi-person file in the research collection (six detailed biographies). Klara von Neumann’s life - Budapest ice skating champion to ENIAC programmer to suicide in La Jolla - is extraordinary. Adele Goldstine (who taught them all and died at 44) is the quietly devastating connective tissue.

What would make it stronger: The core memory threaders at BESK (“housewives with knitting experience” who assembled the ferrite rings) would extend the pattern to Sweden. This ties Posts 10-11 backward to the Women thread and shows the pattern was not American exceptionalism.

Risk: The post requires careful structuring so it doesn’t fragment across too many characters. The Klara von Neumann arc (suicide by drowning, preceded by profound depression after JvN’s death) needs sensitive handling.


6. Threads 4, 6, 7 Assessment for Later Posts

Cray-1/NCAR (Thread 4): Strong candidate for a post about the 1977-1985 supercomputing era in NWP. Needs a Seymour_Cray.md file. Best placed after the Smagorinsky/Manabe post to complete the GFDL story.

Stan Frankel’s Arc (Thread 6): Best placed as a standalone “deeper dive” post that explicitly revisits Post 9 (cables to chaos). The post exists and promised more – this delivers it. Research is adequate but thin on post-1956 Frankel.

Whirlwind/Forrester (Thread 7): Lowest NWP relevance. Best saved for a post explicitly about real-time computing and SAGE that can justify departure from the weather theme. The core memory / women threaders angle provides the connecting hook to the Women thread.


7. Research Gaps to Fill Before Writing

Thread Missing Files Needed
Smagorinsky → Manabe → Nobel Wetherald.md (optional but useful); GFDL_history.md (useful)
Polish Codebreakers None – Polish_codebreakers.md is sufficient
Women of Computing Mary_Tsingou.md, Arianna_Rosenbluth.md (for completeness; core six already documented)
Cray-1/NCAR Seymour_Cray.md (essential); NCAR_history.md (useful)
Stan Frankel’s Arc No gaps; existing Frankel.md is sufficient for a post
Whirlwind/Forrester No gaps in existing files; weak NWP angle is the structural problem, not a research problem
IBM 7090/1960s NWP Shuman.md (already anticipated, file not yet created); 1960s_JNWPU.md