Jul 29, 2021
The amount published in scientific journals has exploded over the past few hundred years. This helps in putting together a history of how various sciences evolved. And sometimes helps us revisit areas for improvement - or predict what’s on the horizon. The rise of computers often begins with stories of Babbage. As we’ve covered a lot came before him and those of the era were often looking to automate calculating increasingly complex mathematic tables.
Charles Babbage was a true Victorian era polymath. A lot was happening as the world awoke to a more scientific era and scientific publications grew in number and size. Born in London, Babbage loved math from an early age and went away to Trinity College in Cambridge in 1810. There he helped form the Analytical Society with John Herschel - a pioneer of early photography and a chemist and invented of the blueprint. And George Peacock, who established the British arm of algebraic logic, which when picked up by George Boole would go on to form part of Boolean algebra, ushering in the idea that everything can be reduced to a zero or a one.
Babbage graduated from Cambridge and went on to become a Fellow of the Royal Society and helped found the Royal Astronomical Society. He published works with Herschel on electrodynamics that went on to be used by Michael Faraday later and even dabbled in actuarial tables - possibly to create a data driven insurance company. His father passed away in 1827, leaving him a sizable estate. And after applying multiple times he finally became a professor at Cambridge in 1828.
He and the others from the Analytical Society were tinkering with things like generalized polynomials and what we think of today as a formal power series, all of which an be incredibly tedious and time consuming. Because it’s iterative. Pascal and Leibnitz had pushed math forward and had worked on the engineering to automate various tasks, applying some of their science. This gave us Pascal’s calculator and Leibnitz’s work on information theory and his calculus ratiocinator added a stepped reckoner, now called the Leibniz wheel where he was able to perform all four basic arithmetic operations.
Meanwhile, Babbage continued to bounce around between society, politics, science, mathematics, and even coining a book on manufacturing where he looked at rational design and profit sharing. He also looked at how tasks were handled and made observations about the skill level of each task and the human capital involved in carrying them out. Marx even picked up where Babbage left off and looked further into profitability as a motivator. He also invented the pilot for trains and was involved with lots of learned people of the day.
Yet Babbage is best known for being the old, crusty gramps of the computer. Or more specifically the difference engine, which is different from a differential analyzer. A difference engine was a mechanical calculator that could perform polynomial functions. A differential analyzer on the other hand solves differential equations using wheels and disks.
Babbage expanded on the ideas of Pascal and Leibniz and added to mechanical computing, making the difference engine, the inspiration of many a steampunk work of fiction. Babbage started work on the difference engine in 1819. Multiple engineers built different components for the engine and it was powered by a crank that spun a series of wheels, not unlike various clockworks available at the time. The project was paid for by the British Government who hoped it could save time calculating complex tables. Imagine doing all the work in spreadsheets manually. Each cell could take a fair amount of time and any mistake could be disastrous.
But it was just a little before its time. The plans have been built and worked and while he did produce a prototype capable of raising numbers to the third power and perform some quadratic equations the project was abandoned in 1833. We’ll talk about precision in a future episode.
Again, the math involved in solving differential equations at the time was considerable and the time-intensive nature was holding back progress. So Babbage wasn’t the only one working on such ideas. Gaspard-Gustave de Coriolis, known for the Coriolis effect, was studying the collisions of spheres and became a professor of mechanics in Paris. To aid in his works, he designed the first mechanical device to integrate differential equations in 1836.
After Babbage scrapped his first, he moved on to the analytical engine, adding conditional branching, loops, and memory - and further complicating the machine. The engine borrowed the punchcard tech from the Jacquard loom and applied that same logic, along with the work of Leibniz, to math. The inputs would be formulas, much as Turing later described when concocting some of what we now call Artificial Intelligence. Essentially all problems could be solved given a formula and the output would be a printer. The analytical machine had 1,000 numbers worth of memory and a logic processor or arithmetic unit that he called a mill, which we’d call a CPU today. He even planned on a programming language which we might think of as assembly today. All of this brings us to the fact that while never built, it would have been a Turing-complete in that the simulation of those formulas was a Turing machine.
Ada Lovelace contributed the concept of Bernoulli numbers in algorithms giving us a glimpse into what an open source collaboration might some day look like. And she was in many ways the first programmer - and daughter of Lord Byron and Anne Millbanke, a math whiz. She became fascinated with the engine and ended up becoming an expert at creating a set of instructions to punch on cards, thus the first programmer of the analytical engine and far before her time. In fact, there would be no programmer for 100 years with her depth of understanding. Not to make you feel inadequate, but she was 27 in 1843. Luigi Menabrea took the idea to France. And yet by the time Babbage died in 1871 without a working model.
During those years, Per Georg Scheutz built a number of difference engines based on Babbage’s published works - also funded by the government and would evolve to become the first calculator that could print. Martin Wiberg picked up from there and was able to move to 20 digit processing. George Grant at Harvard developed calculating machines and published his designs by 1876, starting a number of companies to fabricate gears along the way.
James Thomson built a differential analyzer in 1876 to predict tides. And that’s when his work on fluid dynamics and other technology seemed to be the connection between these machines and the military. Thomson’s work would Joe added to work done by Arthur Pollen and we got our first automated fire-control systems.
Percy Ludgate and Leonardo Torres wrote about Babbages work in the early years the 1900s and other branches of math needed other types of mechanical computing. Burroughs built a difference engine in 1912 and another in 1929.
The differential analyzer was picked up by a number of scientists in those early years. But Vaneevar Bush was perhaps one of the most important. He, with Harold Locke Hazen built one at MIT and published an article on it in 1931. Here’s where everything changes. The information was out there in academic journals. Bush published another in 1936 connecting his work to Babbage’s.
Bush’s designs get used by a number of universities and picked up by the the Balistic Research Lab in the US. One of those installations was in the same basement ENIAC would be built in. Bush did more than inspire other mathematicians. Sometimes he paid them. His research assistant was Claude Shannon, who built the General Purpose Analog Computer in 1941 and went on to become founder of the whole concept of information theory, down to the bits to bytes. Shannon’s computer was important as it came shortly after Alan Turing’s work on Turing machines and so has been seen as a means to get to this concept of general, programmable computing - basically revisiting the Babbage concept of a thinking, or analytical machine.
And Howard Aiken went a step further than mechanical computing and into electromechanical computing with he Mark I, where he referenced Babbage’s work as well. Then we got the Atanasoff-Berry Computer in 1942. By then, our friend Bush had gone on to chair the National Defense Research Committee where he would serve under Roosevelt and Truman and help develop radar and the Manhattan Project as an administrator where he helped coordinate over 5,000 research scientists. Some helped with ENIAC, which was completed in 1945, thus beginning the era of programmable, digital, general purpose computers.
Seeing how computers helped break Enigma machine encryption and solve the equations, blow up targets better, and solve problems that held science back was one thing - but unleashing such massive and instantaneous violence as the nuclear bomb caused Bush to write an article for The Atlantic called As We May Think, that inspired generations of computer scientists. Here he laid out the concept of a Memex, or a general purpose computer that every knowledge worker could have. And thus began the era of computing.
What we wanted to look at in this episode is how Babbage wasn’t an anomaly. Just as Konrad Zuse wasn’t. People published works, added to the works they read about, cited works, pulled in concepts from other fields, and we have unbroken chains in our understanding of how science evolves. Some, like Konrad Zuse, might have been operating outside of this peer reviewing process - but he eventually got around to publishing as well.