---
id: "20260709-1808-hop-purdue-first-cs-department"
title: "Computer science's founding department wasn't Stanford's — and both origin stories trace back to numerical analysts"
type: "capture"
status: "promoted"
origin: "hop-batch"
promoted_to: ["30-notes/claim-purdue-established-first-us-cs-department-1962.md","30-notes/claim-forsythe-coined-computer-science-and-founded-stanford-cs.md","30-notes/observation-numerical-analysts-founded-first-us-cs-departments.md"]
questions_routed: ["50-questions/question-verify-purdue-first-us-cs-department-priority.md","50-questions/question-verify-forsythe-coined-computer-science-cs26-quote.md"]
not_promoted: ["Cleve Moler / MATLAB / LINPACK-EISPACK biography — the hop's entry path, not one of the capture's three core claims and not independently grounded for promotion; folded as supporting detail into the Forsythe and observation notes. Left in inbox.","Forsythe's WWII meteorology → numerical weather prediction lead — explicitly deferred by the capture as a future chain; folded only as a one-clause descriptor.","Saved hooks not followed (Peter Neumann/RISKS, Tony Hoare, Richard Tapia, Conte's textbook lineage, LINPACK→TOP500) — the capture itself set these aside; nothing to promote."]
promoted_by_model: "claude-opus-4-8"
promoted_date: "2026-07-11T00:00:00.000Z"
model: "claude-sonnet-5"
date_created: "2026-07-09T00:00:00.000Z"
hop_chain: ["seed: ACM In Memoriam (awards.acm.org) -> Cleve Moler obituary (max_cosine 0.726)","Cleve Moler (MATLAB creator) -> George Forsythe, Moler's Stanford PhD advisor (max_cosine 0.683)","George Forsythe / Stanford CS dept (1965) -> Purdue's 1962 department, the actual first CS department in the US (max_cosine 0.715)","Purdue priority claim -> Samuel Conte, Purdue's founding department head (max_cosine 0.536)"]
novelty_max_cosine: 0.776
tags: ["hop","computer-science-history","numerical-analysis","institutional-history","myth-ledger-candidate"]
source_1_url: "https://www.cs.purdue.edu/history/index.html"
source_1_tier: 3
source_2_url: "https://mathshistory.st-andrews.ac.uk/Biographies/Forsythe/"
source_2_tier: 3
source_3_url: "https://computerhistory.org/blog/in-memoriam-cleve-moler-1939-2026/"
source_3_tier: 2
---


## Core claims

**1. Purdue, not Stanford, founded the first computer science department in the US.** Purdue split its Division of Mathematical Sciences into math, statistics, and computer science in October 1962 — over two years before Stanford's department existed. Grounding quote (Purdue CS history page): "The first Department of Computer Sciences in the United States was established at Purdue University in October 1962." Source tier: 3 (institutional self-history page, historical/biographical claim, uncontested — acceptable per sources.md).

**2. Stanford's founder, George Forsythe, is credited with coining "computer science" — and defined it as a numerical-analysis-adjacent discipline.** Forsythe (a WWII meteorologist turned numerical analyst) built a Computer Science Division inside Stanford's Math department in 1961, spun it independent in January 1965, and in a 1965 technical report wrote: "I consider computer science to be the art and science of exploiting automatic digital computers" (Technical Report CS26, 1965, per MacTutor). Source tier: 3.

**3. Purdue's actually-first department was built and initially staffed around numerical analysis.** Founding head Samuel Conte came from TRW in 1962; Purdue's first two PhDs, awarded in 1966, were both in Numerical Analysis. Source tier: 3 (Purdue CS department history/bio pages).

## Why this was hop-worthy

Both the department that actually came first (Purdue, via Conte) and the department that gets popular credit for coming first (Stanford, via Forsythe) were built by numerical analysts — the same discipline this vault has already traced as the quiet ancestor of backpropagation (Linnainmaa 1970). "Computer science" the institution, not just "the gradient" the technique, has numerical analysis as a hidden root.

> [!note] Seek's commentary:
> The vault already knows numerical analysis fathered reverse-mode AD in obscurity for a decade. Tonight it turns out numerical analysts also fathered the department that studies AI in the first place — twice, independently, at two universities that don't agree on who was first.

## Further leads

- Samuel Conte's own numerical-analysis textbook lineage (*Elementary Numerical Analysis*) — unexplored, orphan-territory (max_cosine 0.536).
- Forsythe's WWII meteorology background as a cross-domain bridge into early numerical weather prediction (von Neumann's ENIAC group) — not followed tonight, flagged for a future chain.
- LINPACK's benchmark descendant, the TOP500 supercomputer list — mechanism hook into modern compute-ranking culture; deliberately not chased tonight per the redirect away from AI-compute territory.

## Hop chain

Hop 1: ACM In Memoriam — https://awards.acm.org/in-memoriam
- Hook type: The unfamiliar name / the person behind the thing
- Hook: Cleve Moler's name on the memoriam list, next to "creator of MATLAB" — a tool everyone in quant fields knows, a founder almost nobody names.
- Why followed: Moler's obituary was the most information-dense entry on a list otherwise dominated by figures already flagged off-limits (AI history) by tonight's redirect.
- Key findings: Moler did his PhD under George Forsythe at Stanford, co-authored the LINPACK/EISPACK Fortran libraries, and built MATLAB in the 1970s as a classroom shortcut around Fortran before commercializing it in 1984.

Hop 2: Cleve Moler obituary (Computer History Museum) — https://computerhistory.org/blog/in-memoriam-cleve-moler-1939-2026/
- Hook type: The person behind the thing (zoom out to mentor)
- Hook: "PhD from Stanford under pioneering computer scientist George Forsythe."
- Why followed: Forsythe was unnamed in the vault and his role sounded foundational, not incidental.
- Key findings: Forsythe founded Stanford's Computer Science Division in 1961 (independent department, 1965) and is credited with coining the term "computer science."

Hop 3: George Forsythe biography (MacTutor) — https://mathshistory.st-andrews.ac.uk/Biographies/Forsythe/
- Hook type: The surprising claim
- Hook: MacTutor calls Stanford's 1965 department "the first of its kind" — a claim that contradicts commonly cited computing history.
- Why followed: A specific, checkable, contested priority claim — exactly the shape of the vault's existing myth-ledger notes (Amari/SGD, Perceptrons/connectionism, Werbos/backprop).
- Key findings: Purdue's Department of Computer Sciences was established October 1962, more than two years before Stanford's — Purdue, not Stanford, was actually first.

Hop 4: Purdue Computer Science department history page — https://www.cs.purdue.edu/history/index.html
- Hook type: The unfamiliar name
- Hook: Samuel Conte, named as founding department head, entirely absent from the Forsythe-centric version of this story.
- Why followed: If the priority myth was going to resolve, it needed a name attached to the actual first department.
- Key findings: Conte came from TRW Systems in 1962, led the department 17 years, and directed Purdue's Computing Center 1962-68; Purdue's first two PhDs (1966) were both in Numerical Analysis.

Saved hooks not followed:
- Peter G. Neumann and the 41-year ACM RISKS Forum on computer failures — from ACM In Memoriam — reason saved: strong "institution built by one person's persistence" story, but the seed list had a richer numerical-analysis thread to pull.
- Antony Hoare (Quicksort, Hoare logic) — from ACM In Memoriam — reason saved: formal verification is a live vault-adjacent theme (mechanism/gap-detection), but tonight's thread went institutional-history instead.
- Richard Tapia and diversity-in-computing conferences — from ACM In Memoriam — reason saved: orphan-territory (0.585), no clear vault connection yet.
- Forsythe's WWII meteorology work as precursor to numerical weather prediction — from MacTutor bio — reason saved: promising cross-domain bridge (meteorology <-> computing) for a dedicated future chain.

post-worthy: maybe — a clean myth-ledger-style capture in the vault's established genre, but it would benefit from one more corroborating primary source (e.g. the 1962 Purdue trustees minutes or a CACM retrospective) before promotion.