With today, March 8, being International Women's Day, let's honor Margaret Hamilton, who led the team that developed the software for the Apollo Guidance Computer systems!
Posted by KrunchyPhrog@reddit | vintagecomputing | View on Reddit | 23 comments
Annual_Body_9931@reddit
https://www.reddit.com/r/badhistory/s/baKukFSfu9
AOClaus@reddit
Damn. That's some impressive work from all of you.
KrunchyPhrog@reddit (OP)
My father was an engineer who worked at IBM's huge R&D and PC manufacturing plant in north Austin during the 1980s. I also worked part-time at the Austin IBM PC plant on that same campus while attending UT-Austin full-time, doing final QA inspections of assembled IBM 5160/XT and 5170/AT PC motherboards (e.g. looking for tiny cracks in resistors and capacitors and bent pins on IC chips that were rapidly inserted into the through-hole motherboards by IBM's use of high-speed robotics at a time when the Detroit auto industry lagged far behind the Japanese auto industry for robotics use) before the motherboards went to wave soldering. After finishing grad school, I applied for positions at that same Austin IBM campus, but IBM was undergoing their first-ever hiring freeze and first-ever layoffs in their company's history. My dad survived those IBM layoffs in the early 1990s, but IBM did not hire me, and I ended up getting hired by one of the world's largest oil and gas companies, which led to my 4-year work assignment in London. So in many ways, along with having a British girlfriend while I lived in London, programming number-crunching code on Crays worked out really well, and I explored all of Western and Eastern Europe during that time, including the recently-reunified Germany.
Fun fact: The oil and gas companies of the world rank only behind the National Security Agency and government labs of the world for their acquisition and use of supercomputing power, and the world's oil and gas companies use more supercomputers than university research, medical research, meteorology, and other applications of supercomputing.
AOClaus@reddit
What do they use all that power for?
KrunchyPhrog@reddit (OP)
The bread-and-butter staple of the oil and gas industry is seismic data, both for exploring for new oil and gas reserves and for managing an underground oil reservoir to determine the best way to drill and extract the oil and gas. Seismic data is acquired by making sounds on the ground (using "thumper trucks" to thump the ground) to look for oil and gas underground or making sounds in the water (using airguns to make loud pops underwater) to look for oil and gas reserves under the water and then you record the sound waves that reflect back from all the layers of the earth. A good analogy is when you knock on a wall and listen for differences in the sound to locate a stud behind the wall if your house has drywall - a hollow sound means no stud and a muffled sound means a stud is behind the wall.
So all that seismic data, which is basically artificially generated and recorded sound waves, gets processed by supercomputers using various algorithms. You can google "reflection seismology" to see one common method for finding oil and gas. Most of the processing of the seismic data sound waves is very numerically-intensive floating point algorithms that must process lots of seismic data that is typically gathered over large areas of land or water, thus requiring all that supercomputer horsepower. You can google "Fast Fourier transform" and "deconvolution" (we physicists just call it "decon") for two common computations used to process seismic data.
Processing seismic data also tends to be very iterative; i.e. you rarely get the perfect output just from one single processing of a seismic dataset. You have to put the seismic data through a sequence of various algorithms, depending upon what you are trying to achieve. After each iteration of processing during the 1990s, I would view the output, often rendered as "seismic wiggle traces," on high-end Silicon Graphics workstations, and then adjust some of the processing parameters or changing which code to use, and put the data back into the wash-and-rinse cycle for another round of processing on the Crays, which could take up to several hours of runtime.
Fun fact: A fully-loaded fastest-model $25-million Cray Y-MP from 1988, equipped with 8 32-bit CPUs running at around 200 MHz, would get easily beat by a $500 i7 laptop. But the modern supercomputers are still used by energy companies so instead of waiting hours that can add up to several days of Cray runtimes before you are satisfied with your final processed data, you can get larger amounts of data processed all within one day!
Zakmackraken@reddit
I was browsing eBay a few years ago for SGI stuff and I came across a CD that contained very technical software for the mining industry, it may have been something like you describe. The seller still wanted about $13K for the disc! It was late 90’s or early 2000’s software. Sure it’s a shot in the dark but maybe there was some desperate buyer out there lol.
Coding for Crays and SGI is very cool, but can the Cray run DOOM? :-)
KrunchyPhrog@reddit (OP)
I think nearly *ALL* professional-grade software that is sold to the financial, engineering, and scientific industries have two things in common: (1) All their software licenses just for using it on one host has 4 to 5 dollar digits in the price, and that is for use on a single PC or Unix workstation, and (2) All of these really expensive pro-grade software packages verify license keys during installation; i.e. these companies are not stupid in knowing that even their own customers may cheat and try to install their software on multiple hosts. I worked for one geophysical company where we used a license manager called "FlexLM" and part of our routine software testing involved verifying that the FlexLM software worked properly to authorize installation and use of a new installation of a new software version on one host and that the same license key could not be used on installations of the same software on other non-authorized hosts. So you could not even fully install our software on an unauthorized PC or Unix workstation/server because you had to enter a license key just to begin software installation from the multi-CD package. And even if you found a way to copy the installed software from another host, every time that you started up our software or tried to directly run parts of the software without using the main GUI, the license key was verified each time that way too lol.
I am guessing that the $13K for the CD was possibly *STOLEN* by the eBay seller from a company that he had worked for. And if that CD was stolen from a mining company, or if it was an old CD from an old version of the software, the license key was likely already previously used by the mining company that originally purchased the software; i.e. you could not just pop the CD into your computer, install it, and immediately start using the mining software because all of this expensive technical software is protected by license keys and license managers that, just like PC software licenses and just like you cannot take a Windows 11 DVD and use it on 500 PCs for free, would make the mining CD useless. So a fool would have paid $13K to that eBay scammer and ended up with a useless shiny CD in his hand that could not be installed and used - $13K paid for a shiny plastic drink coaster lmao.
A Cray X-MP or Y-MP could probably run Doom if it was ported to the Cray. But being that a $25-million Cray Y-MP is less powerful than a current $500 i7 laptop, the Cray would utterly struggle with modern AAA games unless the games were set to lowest settings.
PlanItClaire@reddit
I think nearly *ALL* professional-grade software that is sold to the financial, engineering, and scientific industries have two things in common: (1) All their software licenses just for using it on one host has 4 to 5 dollar digits in the price, and that is for use on a single PC or Unix workstation, and (2) All of these really expensive pro-grade software packages verify license keys during installation; i.e. these companies are not stupid in knowing that even their customers may cheat and try to install their software on multiple hosts. I worked for one geophysical company where we used a license manager called "FlexLM" and part of our routine software testing involved verifying that the FlexLM software worked properly to authorize installation and use of a new installation of a new software version on one host and that the same license key could not be used on installations of the same software on other non-authorized hosts. So you could not even fully install our software on an unauthorized PC or Unix workstation/server because you had to enter a license key just to begin software installation from the multi-CD package. And even if you found a way to copy the installed software from another host, every time that you started up our software or tried to directly run parts of the software without using the main GUI, the license key was verified each time that way too lol.
I am guessing that the $13K for the CD was possibly *STOLEN* by the eBay seller from a company that he had worked for. And if that CD was stolen from a mining company, or if it was an old CD from an old version of the software, the license key was likely already previously used by the mining company that originally purchased the software; i.e. you could not just pop the CD into your computer, install it, and immediately start using the mining software because all of this expensive technical software is protected by license keys and license managers that, just like PC software licenses and just like you cannot take a Windows 11 DVD and use it on 500 PCs for free, would make the mining CD useless. So a fool would have paid $13K to that eBay scammer and ended up with a useless shiny CD in his hand that could not be installed and used - $13K paid for a shiny plastic drink coaster lmao.
A Cray X-MP or Y-MP could probably run Doom if it was ported to the Cray. But being that a $25-million Cray Y-MP is less powerful than a current $500 i7 laptop, the Cray would utterly struggle with modern AAA games unless the games were set to lowest settings.
AOClaus@reddit
This is similar to what is used in archeological endeavors then? I've heard of it being used to find or map ruins etc. I've heard about it being used to map the Earth's interior as well, though that requires natural quakes.
Do you think "AI" will improve this technology in the future? It seems to be good at recognizing patterns once it's fed enough data.
Thanks for answering my questions, I took some programming years ago and just didn't have mindset for it. Couldn't get myself to focus, so I didn't get the logic, got frustrated, and gave up. Regret that now.
KrunchyPhrog@reddit (OP)
Yes, all of the methods and sciences for locating underground oil and gas reserves, including other methods that do not involve making a loud noise on the ground such as ground-penetrating radar (GPR), are also used to find and map buried archeological ruins, detect underground aquifers, detect and track groundwater movements, etc. If there is something buried underground that changes the structure of the layers of the soil, then it can often be detected using methods similar to oil and gas exploration. GPR is also used to detect underground power lines, gas lines, water and sewer pipes, and GPR has often been successfully used to find disturbed soil that was due to either legally buried bodies in old forgotten gravesites or illegally murdered people buried out in the vast area of a remote forest.
Some underground imaging techniques are also used in medical imaging, including the science of tomography - google "tomography" if you like reading about science. When you get a CT scan of your body at a hospital, CT is "computed tomography." Medical imaging has the advantage of being able to collect data from different angles of a human body, while searching for underground oil, ruins, groundwater only collects data from above the target area (one-dimensional data collection vs. 3D data collection).
AI can certainly help with *some* pattern recognitions, such as automatically detecting underground water pipes or bodies buried underground, but AI has a lonnnnggg way to go before AI can actually find oil reserves by itself without humans. After we physicists design algorithms and process seismic data, the geologists ultimately make educated guesses on where oil and gas may be trapped inside a reservoir based on numerous factors including stratigraphy and geologic history of the area being explored.
Even some of the pioneers of the current AI movement will tell you that we still need a quantum leap in the programming methods of AI before it really starts to replace more human thought. *ALL* AI applications are still a brute force approach - collect as much data as you can and use large amounts of existing data to look for a pattern or to generate a result. But real human thought collects and organizes data in very non-linear patterns and abstractions that current AI only roughly mimics.
When the LISP programming language was invented in 1958, followed by Prolog language in 1972, everyone proclaimed, "Oooooo aaaaaahhh, here comes the AI revolution!!! Here comes the robots!!!!!!" Does that sound familiar to today's "AI is coming" chant? The huge difference is that CPUs are much faster now than the 1960s so the current brute force approaches to AI work better and faster than before. The concept of neural networks was invented during the 1940s and the oil industry already uses neural networks for pattern recognition and sifting through what is signal and what is noise. Neural networks were supposed to mimic how the human brain works, but it also has not taken over the world since the 1940s, despite lots of (exaggerated) promises.
If you look at much of the AI-generated imagery that gets posted online right now, AI-generated imaging is basically Artificial Compositing and Aggregation, or Artificial Photoshopping, because it totally relies on saved and stored images that were previously created. And despite using its huge database of existing images, including hundreds of thousands of stored images of humans, you often see funny AI-generated images of humans with 4 fingers or 6 or 8 fingers on their hands. Anyone who has non-expert intermediate Photoshop skills can create far better images than a AI-generated image.
Also don't drink the glossy marketing brochure Kool-Aid of Tesla's Autopilot feature. Since 2019, there have been more than 700 Tesla crashes due to Autopilot guidance, with about 17 fatalities. AI in cars can be very useful for slowing a car down to be fully stopped if it determines that someone has fallen asleep at the wheel, or if it detects the car rapidly approaching a stationary object, or assisting the erratic driving of a senile 80-year old. But the hope that a driver can just fall asleep or go hands-free and watch a movie is still a long way off, again because all of the mental thoughts and neuromuscular actions that a good experienced driver makes are still roughly duplicated by current self-driving cars. The Waymo driverless cars have a better safety record than Tesla's Autopilot, but Waymo also makes some incredibly stupid decisions that look like a senile 90-year old is driving the car, e.g. totally not seeing and hitting a cyclist, or not seeing a big city bus and hitting it.
DorkyMcDorky@reddit
It's cool that they honor her - but it's better to do this:
List 10 women writers you would read? Can't come up with 10, go to the bookstore.
List 10 women who influenced tech in the last decade. Can't name 10? Google it.
I can't do this easily, but I'm making it a priority.
SuperConductiveRabbi@reddit
As a big fan of the AGC, I hate this myth.
Margaret Hamilton didn't write "all of that code." She was part of a team that wrote it. She wasn't a lead developer, she was just a developer at a lower level that joined in 1965, years after coding started. She was promoted to assistant director of the Command Module team by the man she was dating, whom she married a year later. When she was awarded the Presidential Medal of Freedom by Obama her teammates were disappointed that the proper people who actually did groundbreaking work were once again overlooked.
In that famous photograph of the printouts, it's actually just copies of multiple versions of the code made for marketing purposes.
https://www.reddit.com/r/badhistory/comments/18yum8s/no_margaret_hamilton_at_nasa_is_not_standing_next/
KrunchyPhrog@reddit (OP)
Yes , the "she wrote by hand" in that first college image is misleading as it implies that she single-handedly wrote everything, which should be clearly obvious that is not true for ANY kind of technical software of that complex and massive scope, especially since the codebase directly affects the lives of astronauts way out on the Moon. I just had that image saved from a 2016 posting that someone else made, but the wording really should say "that she and her team wrote by hand" (and yes software programming during the 1950s and 1960s was all initially done by manually writing it out, whether you are physically wiring core-rope memory or programming Fortran code on keypunch cards - it all started by writing out code on paper).
The Reddit post that you link to is not entirely accurate. Those stacked binders were *NOT* a "marketing" gimmick that NASA or Margaret staged lmao. On one hand, Margaret was not a programming superhero who did all the coding for the AGC software. But on the other hand, Margaret was indeed the first developer for the Apollo flight software project at MIT, the first female programmer on that project (which was also very unique at that time), and she was the director and team lead for both the Command and Lunar Module flight software and also later for NASA's Skylab space station's flight software. To be totally dismissive by saying that Margaret was an inexperienced bimbo who was just promoted by the man she was dating has misogynist overtones for anyone who quickly attributes her promotion entirely to favoritism. If she was not dating the guy, she may still have received the promotion due to either job performance or seniority, although we will never know.
Many years ago, another member of the Apollo Guidance Computer team was interviewed who said that those 17 binders stacked up were indeed all the code listings of the on-board Guidance Computer during the Apollo missions. Those binders did not contain printed outputs from simulations or code testing. Those 17 binders were not the code for just one Apollo mission or just one Apollo computer because that much code would not fit into the tiny memory space. And that stacked binder photo was absolutely NOT "just copies of multiple versions of the code made for marketing purposes" lol. It just happens to be a photo that was publicized more over the years.
All of the Apollo Guidance Computer software was written using NASA's own semi-interpreted assembly code - assembly code with lots of pre-assembled advanced trigonometry and matrix operations. The flight guidance code was assembled on a Honeywell 1800 mainframe and data center printers printed out all the code listings that you see in those 17 binders in the photo. The assembly code was then physically woven into the 36K of core-rope ROM memory - a wire goes through a core for a 1 bit and a wire goes around for a 0 bit.
The Apollo missions to the Moon used two Apollo Guidance Computers. The AGC in the Command Module ran a program called "Colossus". The AGC in the Lunar Module ran a program called "Luminary".
During the 1960s, Margaret Hamilton stood right at about 5'0" in height. If you estimate the stack of 17 binders in that photo to be around 5'4", people have estimated that stack of code listings contains around 15,000 code pages with almost 700,000 lines of code. There were different versions of Colossus and Luminary software with the Apollo missions, kept in separate binders, and that is what you are looking at in that stacked photo - the entirety of the printed code listings during the Apollo missions, not just from one flight.
The 14,000 to 15,000 pages of Apollo Guidance Computer code listings from those 17 binders in that photo have been partially scanned/photographed and you can see what the code looks like here:
https://archive.org/details/luminary21000miti/page/453/mode/2up
And there is a Virtual AGC archive that preserves other code listings and engineering drawings here:
SuperConductiveRabbi@reddit
I said "for marketing purposes," and yes it was:
It has to contain duplicates, probably older versions, because the AGC listing is only 2000 pages. As the post says:
That's not at all what I said. I don't know how you got there, but you seem to buy into the historical-hero narrative that it's not suspicious that a spouse promoted his direct report whom he was dating at the time, and later married. If there was nepotism, we certainly wouldn't learn about it now. Also don't ignore the words of her teammates expressing frustration over decades that the innovators on their team were overlooked for PR reasons.
KrunchyPhrog@reddit (OP)
YAWNNNNN... I already wrote early on: "There were different versions of Colossus and Luminary software with the Apollo missions, kept in separate binders, and that is what you are looking at in that stacked photo - the entirety of the printed code listings during the Apollo missions, not just from one flight." You obviously have never worked in a software department because different versions of code revisions are *NOT* "duplicates" that you and that other Reddit link call it. Sure, updated versions of code are not entirely unique, but "duplicates" implies redundant copies of code listings in those binders, which they are not. And there were *2* sets of AGC code, one for Colossus and one for Luminary.
You are also now deeply committed to a kneejerk hair-splitting pissing contest that bores me. I do not "buy into the historical-hero narrative" because I already wrote "Margaret was not a programming superhero who did all the coding for the AGC software" smdh
For many decades since the 1960s, and maybe earlier, *MANY* companies have tried to encourage women to go to college for engineering or computer science degrees. NASA, IBM, oil companies, defense contractors, telecom companies have all featured women working in technical roles and women in leadership in their companies. So most of these women in management positions at tech companies also sweet-talked or slept their way to the top? I see lots of misogynists sneering at women in high-ranking positions and discrediting everything about them just because they look pretty or wear a skirt. To downplay featuring female tech workers as mere "marketing" or "public relations" is being totally ignorant and clueless about the need that all U.S. companies have for tech talent, as women are still, even to this day, an underutilized source of tech talent. And, yes, women can do as well as men in tech, even though they are often paid less than men for doing the same job. I am speaking as a guy who has worked with many amazing female physicists and programmers. Many of the biggest American tech companies like Google and Apple sponsor H-1B visas not because they prefer foreign workers over Americans, just as NASA did not only prefer Margaret over her male co-workers, but because of their need to employ more tech talent, both foreign talent and American women, than what the U.S. tech talent can supply to U.S. companies. Get a clue, please! smdh
If you or a friend was actually a member of the AGC software team, just say so and I would fully understand your animosity. Some of these rants about Margaret sound like the incels, misogynists, and racists who spout hate on 4Chan. Mention any achievement by a non-Caucasian or woman on 4Chan and it immediately triggers the boiling hatred from all those 4Chan idiots. I purposely use hundreds of user accounts on social media, and sometimes I masquerade as a woman on Facebook to discuss world events, and even when I post highly-informed comments, I immediately draw fire from all the misogynists and I get scathing insults when posting from one of my accounts listing me as a female, whereas I get no insults when posting as a male.... quite an interesting (and expected) social science experiment lol
SuperConductiveRabbi@reddit
If you simply wrote accurate information then no criticism of your points would be warranted. Your image in your OP is extremely wrong, and now you're ranting about unrelated things.
We're actually in agreement on some points, but you're willfully spreading a myth, and seem to be trying to bait for things that are totally unrelated to my argument.
Like I said, I'm a fan of the AGC and its history and that's why I want accurate information about it, and get annoyed when people like you do what you're doing here. Also I'm pissed that any criticism of what you're saying and you put words in my mouth like "Margaret Hamilton is a bimbo." Your style of argument seems to be throwing chum in the water and going for low-effort replies.
KrunchyPhrog@reddit (OP)
YAWNN.. my very first reply to you was:
Yes , the "she wrote by hand" in that first college image is misleading as it implies that she single-handedly wrote everything, which should be clearly obvious that is not true for ANY kind of technical software of that complex and massive scope, especially since the codebase directly affects the lives of astronauts way out on the Moon. I just had that image saved from a 2016 posting that someone else made, but the wording really should say "that she and her team wrote by hand"
You are getting totally triggered by that one "she wrote by hand" connotation. BTW, I posted this EXACT thing on a Facebook group yesterday and actually corrected the wording of the image that triggers you so much to say "she and her team wrote by hand", but Reddit does not allow me to just replace that one photo whereas I replaced that image on my Facebook post. I never said that you used the word "bimbo", but the your previous statements imply that she was totally not worthy of that promotion. As I mentioned, even if she was not dating him, she may have been promoted just on the basis of her seniority. Maybe, maybe not, neither you nor I can adamantly state that we are right.
In regards to "people like you do what you're doing here", WHOOSH!!!! , you totally missed the concept and spirt of International Women's Day. Or are you subconsciously wondering why there is no matching International Men's Day???? Unless you are a total misogynist (and most racists and misogynists deny it by saying "I'm not racist/misogynist, but..."), there is no denying that (1) excluding the "she wrote by hand" semantics which I mentioned was not originally worded by me in that collage, and (2) excluding the debate about her being favored for promotion, there is absolutely no denying that Margaret broke a lot of class ceilings BY HER OWN EFFORTS at a time when scientists and engineers in high-tech were an exclusive male fraternity. In that regard, not only did you miss the bullseye for why I mentioned Margaret yesterday (and also because my sister worked on Space Shuttle flight software for NASA because I could have mentioned many hundreds of other women that I admire), but you missed the entire dartboard and you are now throwing darts out through the window of the pub in deviating from the single simple reason for mentioning her. Do I think that Margaret was a godlike superhero? Hell no!!! I have worked with many brilliant female physicists who were likely smarter than her. But AT THE TIME, in that era of computers with 2K of memory, Margaret broke many glass ceilings in the world of tech.
2raysdiver@reddit
Man did not land on the moon. People put another human on the moon, he just happened to be a man (not to take ANYTHING away from Neil Armstrong or Buzz Aldrin - still heroes of mine). NASA employed people of all sexes, races and creeds to put a man on the moon, yet when you look at pictures from that era, all you see are white men. We would not have been able to put a man on the moon without women of all races.
One of my favorite accounts... During the Eagle descent on the Apollo 11 mission, the computer signaled a fault. NASA knew they could continue with this fault because a female test engineer encountered it and documented what it was and if it could be corrected, bypassed, or ignored long before the Apollo 11 space craft was even rolled out onto the launch pad.
These women deserve more than just a day of recognition.
KrunchyPhrog@reddit (OP)
Yes, white Caucasian men did not just sprout wings and fly to the moon by themselves lol. Humans landed on the moon, of species Homo sapiens. Especially during the past two millennia, but also all throughout numerous previous millennia, any feat of human ingenuity involved a huge beehive of humans of multiple ethnic backgrounds and genders contributing knowledge that kept growing with each generation and allowed for the latest round of inventions and creations to be possible, like an onion that keeps adding layers. And even if just one person had an "Aha!" epiphany, even if that person was Albert Einstein, that person's technical brilliance is still piggybacked onto many generations of knowledge contributions by men and women from around the world, a hive mind formed from many centuries of collective intelligence gathered from around the world.
Since the world always seems to be involved in some kind of war, the Apollo astronauts were also like the soldiers sent to war - predominantly men who, while exhibiting courage, also likely felt fear as their rocket ignited or as they reached a war zone. But for better and for worse, the knowledge and know-how that let the astronauts reach the moon using what amounts to a tiny-memory TRS-80 computer guiding their spacecraft was achieved by centuries of previous knowledge that had to be discovered first, just as humans are capable of inventing better and faster ways to kill each other during war, again based on the collective wisdom of previous experiences of war.
Fun fact: About 7 years ago, researchers at the University of Chicago studied warfare during the 14th through the 19th centuries, and discovered Europe's queens were more likely to wage war than kings in Europe, and when they did wage war, the queens acquired more territory, from Spain to Russia, during that period than kings who initiated war. And married queens were especially more likely to be warmongers who added new territories to their control because the queens were better at creating alliances and waging war with allies instead of men who usually waged war without allies joining them.
VK6FUN@reddit
The finest baby frogs, dew picked and flown from Iraq, cleansed in finest quality spring water, lightly killed, and then sealed in a succulent Swiss quintuple smooth treble cream milk chocolate envelope and lovingly frosted with glucose.
BobT21@reddit
... and in 1970 the practice of commenting code was initiated.
miniscant@reddit
From back in a time when programs could be certified as error-free.
KrunchyPhrog@reddit (OP)
In honor of International Women's Day today, March 8, I honor computer scientist and mathematician Margaret Hamilton, who led the team that programmed the on-board flight software for NASA's Apollo command modules and lunar modules. Back then, the Apollo Guidance Computer that was on the Apollo spacecraft only had 2K to 4K of erasable magnetic-core memory and 36K of read-only core-rope memory. It was the first computer to use silicon integrated circuits and had the approximate computing power of a TRS-80 or Commodore PET.
I also honor my sister who previously worked on NASA's team who wrote the on-board flight software for NASA's Space Shuttle program, although she had far more computing power and memory to work with on the Space Shuttle. NASA's Space Shuttles had 5 on-board computers, each with 106K of 32-bit ferrite magnetic-core memory. All of the Space Shuttle's flight and on-board operational software was too large to fit into primary memory so it had to be divided into 8 memory overlays. The 5 on-board Space Shuttle computers each used a top-of-the-line IBM System/4 Pi avionics computer that was also used in NASA's Skylab space station, B-52 and B-1 bombers, F-15 and other fighter jets, and some anti-ship and anti-submarine missiles and rockets.
I am a physicist and mathematician myself, with 20+ years of experience writing numerically-intensive signal processing and signal analysis algorithms in Fortran and C++ that would run for hours on supercomputers. My first job after graduate school was getting transferred from my U.S. office to London for four years where I developed signal processing code that ran on Cray X-MP and Y-MP supercomputers, but coding for the 16 megabytes of 64-bit words on a Cray X-MP was far easier than what Margaret had to work with on the Apollo missions.
Here are photos of Margaret Hamilton being awarded the Presidential Medal of Freedom by President Obama in 2016. Other people who received the medal in that ceremony included Bill Gates, Tom Hanks, Robert Redford, Robert De Niro, Kareem Abdul-Jabbar, Michael Jordan, Diana Ross, and Bruce Springsteen.