Skip to content
Hamilton Archive
0 / 17
Pre-launch Review

The Rope Mothers

Software You Could Hold in Your Hands

Section titled “Software You Could Hold in Your Hands”

The Apollo Guidance Computer’s software didn’t exist as files on a disk. It existed as copper wire threaded through and around tiny magnetic ferrite cores — core rope memory. A wire passing through a core stored a 1. A wire routed around it stored a 0. Every bit of the flight software was a physical, permanent decision.

Macro close-up of core rope memory showing copper wires woven through and around ferrite cores — each threading path encodes a single bit of the Apollo flight software

Core rope memory at the bit level. Each ferrite core (copper-colored ring) has wires threaded through it (a 1) or routed around it (a 0). This is what software looked like before there were files. Photo via Ken Shirriff, righto.com.

The numbers are staggering:

  • 36,864 words of read-only core rope memory per AGC
  • 6 rope modules per computer, each storing ~6,144 words
  • 512 cores per module, arranged as two physical layers of 256
  • 192 sense wires per core (storing 12 words of 16 bits each)
  • Half a mile of wire threaded through each module
  • 589,824 individual wire-through-or-around-core decisions per complete AGC

There was no undo. There was no patch. If the software changed after the ropes were woven, the ropes had to be rewoven. Code had to be frozen months in advance. This is part of why Hamilton’s insistence on getting the software right before the fact wasn’t just methodology — it was a manufacturing constraint.

Diagram showing how core rope memory works — sense lines (blue) pass through or around cores, while set/reset (green) and inhibit (red) lines select which core to read

How core rope selection works. Blue sense lines carry the data (through = 1, around = 0). The green set/reset line and red inhibit lines select which core to read. Each core holds 192 bits — 12 words of 16 bits each. Diagram via Ken Shirriff, righto.com.

The Women at Raytheon

Section titled “The Women at Raytheon”

The core rope modules were manufactured at Raytheon’s plant in Waltham, Massachusetts. The assembly work was done primarily by women — many hired from the local textile industry for their sewing skills, others from the Waltham Watch Company, which had a long tradition of precision handwork. (The same watch company also helped manufacture the high-precision gyroscopes used in the Apollo navigation system.)

A woman at Raytheon's Waltham plant threading wire through a core rope memory module for the Apollo Guidance Computer

A Raytheon worker threads wire through a core rope memory module at the Waltham plant. The matrix of ferrite cores is visible in the frame she’s holding. Photo via Ken Shirriff, righto.com / Raytheon.

Sitting across from one another at long desks, they passed a hollow needle containing wire back and forth through a matrix of eyelet holes, each holding a magnetic core bead. An automated system — driven by punched tape generated from the AGC’s YUL assembler — read the program and positioned an aperture over the correct core. The weaver threaded the needle through the aperture, installing each wire in the right location. Then the system jogged down to pull the wire around a nylon pin before the next threading.

Two women at Raytheon's Waltham plant sitting across from each other, passing a needle through a core rope memory module frame between them — one wearing a Raytheon badge

Two Raytheon workers weaving core rope memory at the Waltham plant, sitting across from each other and passing wire back and forth through the module frame. The rows of ferrite cores are visible between them. The woman on the right wears a Raytheon badge. Photo via Ken Shirriff, righto.com / Raytheon.

A woman at Raytheon weaving core rope memory — she threads a needle through a rack of ferrite cores while a wire spool feeds from the right

Weaving core rope memory at Raytheon’s Waltham plant. The worker threads a hollow needle through the eyelet matrix while a wire spool (right) feeds the sense line. The metal housing behind the core rack is part of the automated positioning system driven by punched tape. Smithsonian National Air and Space Museum, image 13006h.

Manufacturing a single rope module took approximately 8 weeks and cost about $15,000 — roughly $130,000 in today’s dollars. 100% accuracy was required. Each component was inspected by three or four people before being stamped off.

Interior of a completed AGC core rope memory module showing 512 ferrite cores arranged in a matrix with hundreds of sense wires threaded through them

The finished product: interior of a core rope memory module. 512 ferrite cores arranged in a matrix, with sense wires woven through them encoding ~6,144 words of flight software. The connector pins along the edges carry signals to the AGC’s sense amplifiers. Photo via Ken Shirriff, righto.com / Computer History Museum.

The Names We Know

Section titled “The Names We Know”

Most of the women who wove Apollo’s software into hardware remain anonymous. But we know some names, thanks in part to a Raytheon news photograph and a 2019 search by Caroline Kennedy.

The Five Women in the Raytheon Photograph

Section titled “The Five Women in the Raytheon Photograph”

A photograph from Raytheon’s records — captioned with their names — shows five women who worked on AGC core rope memory:

  • Vernell Norman
  • Caroline Butler
  • Helen Lennon
  • Edna Walcott
  • Mary Julian

In June 2019, on the 50th anniversary of Apollo 11, Caroline Kennedy appeared on WCVB-TV Boston asking for the public’s help in finding these five women or their descendants.

Mary Julian’s family responded. Born around 1920, Julian had worked as a seamstress before joining Raytheon. She died in 2004. Her granddaughter Lisa Julian told reporters: “I knew she worked for Raytheon, but I had absolutely no idea that she was part of this.” Her son Jim Julian said: “I knew my mother worked on the Apollo project but never knew what she did. She wouldn’t talk about it. Now here we are, 50 years later… I’m so proud of my mom and what she did.”

Other Named Workers

Section titled “Other Named Workers”

Mary Lou Rogers of Waltham worked on a different part of the Apollo production line at Raytheon. In a BBC interview, she recalled the quality control process: “Each piece had to be looked at by three or four people before it was stamped off.”

Mrs. Christina Palcos, a Raytheon worker, suggested the use of laminar workflow benches for the assembly of AGC microcircuits — a practical innovation that prevented contaminants from settling on the work. Her name appears in Raytheon production records.

Two Meanings of “Rope Mother”

Section titled “Two Meanings of “Rope Mother””

The term “rope mother” is used in Apollo histories to mean two different things, and most accounts conflate them.

At MIT’s Instrumentation Laboratory, a “rope mother” was a programmer — an engineer responsible for parenting the development of the software for a specific mission. They made sure the code was correct before it was sent to Raytheon for weaving into rope. Rope mothers at MIT were generally male engineers.

The notable exception: Margaret Hamilton was the rope mother for LUMINARY — the Lunar Module flight software that landed on the Moon. As the Smithsonian puts it: “Getting the programs right was the responsibility of Ms. Hamilton, the ‘Rope Mother.’”

Margaret Hamilton working at her desk at the Charles Stark Draper Laboratory during the Apollo program

Margaret Hamilton at the Charles Stark Draper Laboratory (formerly MIT Instrumentation Lab) during the Apollo program. The software she wrote here would be translated into wire lists and sent to Raytheon for weaving into rope. Smithsonian National Air and Space Museum, image 13007h.

At Raytheon’s factory in Waltham, a different group of people — the women on the manufacturing floor — physically wove the ropes. They were not called “rope mothers” in Raytheon’s own documentation. They were called “LOLs” or “space age needleworkers.”

This archive uses the term “rope mothers” for the page title because it has become the popular shorthand for the entire story — the programmers who wrote the software, the women who wove it, and the manufacturing process that turned code into copper. But precision matters: the women at Raytheon and the engineers at MIT did different work, and both deserve recognition on their own terms.

The Craft Evolves, the Discipline Endures

Section titled “The Craft Evolves, the Discipline Endures”

The women at Raytheon represent one end of a sixty-year arc in how humans turn ideas into working systems. Each era brought new tools that changed what “building software” meant — and each time, people worried the craft was dying.

EraMediumThe Craft
1960sCopper wire through ferrite coresHand-woven at Raytheon from assembler wire lists
1970sPunch cards and paper tapeKeypunched by operators, batch-submitted to mainframes
1980sFloppy disks and terminalsTyped into editors, compiled locally
1990sHard drives and networksWritten in IDEs, version-controlled, deployed over networks
2000sCloud infrastructureCommitted to repositories, built by CI pipelines, deployed globally
2020sAI-assisted developmentDesigned by humans, generated and reviewed collaboratively with models

The medium changed completely. The discipline didn’t. Hamilton’s principles — prevent errors by construction, design the interfaces first, make entire categories of failure impossible — apply whether you’re weaving wire or prompting a model. The question was never how the software gets built. It was always whether the people building it understood what could go wrong and designed against it.

The women at Raytheon knew that every wire they threaded was permanent. That knowledge shaped how carefully they worked. Today, software is infinitely mutable — you can deploy, revert, patch, and redeploy in minutes. The ease of change makes it tempting to be less careful. But the systems we depend on — medical devices, avionics, financial infrastructure — still demand their mindset: do it right, because the consequences of getting it wrong are real.

Margaret Hamilton standing next to a stack of Apollo Guidance Computer program listings that is taller than she is

Margaret Hamilton with the Apollo Guidance Computer program listings. All of this code — every instruction, every constant, every address — was translated into wire lists and woven by hand into core rope memory at Raytheon’s Waltham plant. Smithsonian National Air and Space Museum, image 12982.

What We’re Still Looking For

Section titled “What We’re Still Looking For”
  • Names and oral histories of additional Raytheon workers
  • Raytheon internal manufacturing documentation for AGC core rope production
  • Production timelines — how the schedule for rope manufacturing interacted with software freeze dates
  • Error rates and rejection data from the quality control process
  • Photographs from the Raytheon factory floor (beyond the known press kit images)
  • Any connection to the Navajo women who assembled AGC integrated circuits at Fairchild Semiconductor’s plant in Shiprock, New Mexico — a related story that deserves its own telling

Sources and Further Reading

Section titled “Sources and Further Reading”

Primary and Archival Sources

Section titled “Primary and Archival Sources”
  • Raytheon Apollo 11 Press Kit photographs — Collection of David Meerman Scott, author of Marketing the Moon (MIT Press, 2014). Includes the photograph of Norman, Butler, Lennon, Walcott, and Julian.
  • Jack Poundstone files — Raytheon Space Division, Waltham, MA. Archived via Caltech/MIT digital collections.
  • Smithsonian National Air and Space Museum — Apollo Lunar Equipment Collection. Hamilton’s papers and program listings. Image references: 13006h (core rope weaving), 13007h (Hamilton at Draper), 13005h (program listings).
  • Computer History Museum — Shustek Research Archives. Restored AGC hardware. Photography by Mark Richards.
  • MIT Museum Collections — “Computer for Apollo” (1965 MIT Science Reporter segment, WGBH-TV Boston). Features Eldon Hall, Ramon Alonso, Albert Hopkins (MIT) and Jack Poundstone (Raytheon).

Secondary Sources

Section titled “Secondary Sources”

Watch: Core Rope Weaving

Section titled “Watch: Core Rope Weaving”

Weaving Software into Core Memory by Hand

Watch on YouTube — footage of the core rope weaving process, showing how wires were threaded through and around ferrite cores to physically encode the AGC’s software.

Interactive Visualizer

Section titled “Interactive Visualizer”

See how core selection actually works — activate inhibit, set, and reset wires and watch 512 cores respond in real time:

Section titled “Related Documents”
Apollo Flight Software (2019) Hamilton's retrospective on the software these women wove into hardware
The Software Effort (1971) MIT's account of the software development process that produced the wire lists for Raytheon
GNC Hardware Overview (2009) The guidance, navigation, and control hardware that housed the core rope modules
Preventative Software (1994) Hamilton's methodology — shaped in part by the reality that woven rope couldn't be patched