35: The Analytical Engine
Anyone who has been exposed to early computer history will no doubt have heard of Charles Babbage (1791–1871), a British mathematician and overall supernerd who rebelled as a teenager against the status quo by teaching himself calculus at four in the morning (I’m not making this up). Eventually, he went on to study mathematics at the university, and even there, he was dissatisfied with the teaching, given his high-level ability when first entering was beyond that of his class-mates. Instead of rebelling again, he went through the motions and eventually graduated. Good for him.
In time, like most academics, he got a teaching position at his university and, from there, did what other mathematicians did: research. In particular, he focused some of his curiosity on electromagnetism, although didn’t achieve much in this field. But while working on electromagnetism, he came to glance over logarithmic calculation tables, which gave him the initial inspiration for a computational machine. It takes people immense amount of effort to compute these numbers, he thought. And cecause each number was computed by a human being*, and human beings are fallible, it meant that there were slight errors in these tables, which could be avoided if they had been computed by machines.
Being smack in the middle of the industrial revolution in 1812, Babbage imagined that there must be an easier way to generate these numbers. If he could build a machine could compute logarithms, the tables could be done cheaper, faster, with perfect accuracy, and also without human input.
Because no such machine existed, he decided to build one himself. He was a man with a plan, and so, designed all the physical and mathematical machinery himself. MOSFETs (1960) or any types of transistors/transistor-like components (1907), at this point, were still nearly a century away. Even digital gates were still around 80 years away from first being formally studied, so Babbage had to invent computing mechanisms on his own if he was going to make it a reality.
Given the previous articles, we already know how to digitally perform long addition. However, we don’t know how to do it mechanically: The striking aspect of such a machine was that the technology of the time only allowed him to work with steam-powered gears.
For the overall design, this isn’t a large problem: As long as we can perform the instructions somehow, then moving data from different registers to another still works. If additionally, the instructions do what is expected of them, his mechanical computer will logically work exactly like an electronic one but much slower. Thus, instead of using electricity to compute and store memory, Babbage used the presence or absence of gear teeth, which would mechanically lock further processing, increment a value, or oave a certain offset of rotational degrees to represent a number.
Many of the physical designs for digital adders, subtractors, etc., turns out, were invented by Babbage himself, just in a different form than what we are used to. It was only when George Boole (1815–1864) came along that the mathematics of binary numbers was formalized, signaling the future of discrete mathematics, that adders & subtractors were built using logic gates.
Either way, from the initial inspiration of the computing machine, Babbage began work on what eventually became know as the Difference Engine, a mechanical computer that would use the method of divided differences to find interpolated (read: estimated) values of logarithms, and a few other functions, which had recently been discovered by mainly French mathematicians. The mathematical developments of interpolation of a mathematical function, a special case of which is polynomial interpolation, had a profound impact most fields of engineering, including in the form of smooth curves of glyphs (letters) on a computer screen. But more on that later.
For now, the Difference Machine was supposed only to compute numbers. The design was ready in 1822, and with it, he secured funding to start building this one-of-a-kind computing machines for the Royal Astronomical Academy, which is how the building of the first “computer” began.
It took ten years for Babbage and his colleague Joseph Clement to physically build a prototype of the Difference Engine in 1832, all the while being funded by the government, who felt that they might save some money by letting a machine handle computations instead of a person. Sadly, because of a payment disagreement with his colleague*, Clement exited the project and took with it any realistic way forward to build the machine. From then on, the project was suspended that same year and abandoned altogether another ten years later. In other words, the whole process of building the machine took 20 years, and even after all this time, Babbage had nothing to show to his financiers. Like always, the government just wasted a lot of money.
Get 365 Days of Computer’s stories in your inbox
Join Medium for free to get updates from this writer.
Or did it? Babbage didn’t keep the focus on this one project for twenty years, every single day. That would be like doing an endless Ph.D. Instead, like software product owners, he thought of too many features to add to his Difference Engine. He was, in effect, designing not just functional interpolation within a new machine, but a universal computer/calculator that would compute a couple of functions but could, in theory, be used as a tool to compute anything that people needed computing, such as iterative calculations, which required previous calculations’ results to be remembered.
At this time, in 1840, at a conference in Italy, the (now historically famous) mathematician Ada Lovelace entered the picture. After Babbage held a lecture in Turin on the Analytical Engine (the universal computer he was designing), arguing how much more powerful it was than the Difference Engine and how this universal “computer” was going to simplify all computation, Babbage hired Lovelace for her still underground talent in mathematics to operate it. Simply by playing around with the ideas of the engine, Lovelace decided to try and mechanically spell out instructions of how to compute a special sequence of numbers known as the Bernoulli numbers, each element of which could be computed by using previous elements in the sequence (like the Fibonacci sequence).
Little did she know, historians would look at this event as marking one of the first, if not the first, computer programs to be written in the literature. It seems unusual that a newcomer would have created the first algorithm with the machine, as it was Babbage who created the machine in the first place. But no matter: if Lovelace wasn’t the first programmer, she was unarguably the first pure programmer, who didn’t necessarily understand every detail about the physical mechanisms of computation, just like today’s programmers often don’t know much about the hardware their programs are running on.
Where Lovelace did contribute, which set her apart from Babbage, was in the form of her thoughts on the Analytical Engine. She speculated, in poetic verse, on what computers might be able to do someday. Calculations are a given, but they might also one day compose music or write poetry, an outlandish claim at the time that only in recent decades has seen genuine fruition. It’s unlikely she predicted the mechanisms behind artificial intelligence, as the development of these techniques emerged 100–150 years later. But she was right about machines becoming more and more intelligent, to the point where it is sometimes hard to distinguish what is done by a computer and what is done by a human.
While the Analytical Engine ended up never being built, just like the Difference Engine, the paperwork marked the first time someone designed a “general purpose” computer/calculator and the first time someone wrote an algorithm to program it. The idea was so far ahead of its time that nothing became of the design, per se.
Several new machines based on the Difference Engine were created by such people as George Scheutz, a Swedish lawyer. His machine, invented in 1837, whose design was later on improved on by adding a “printing” mechanism. Other designs had similar purposes before the Difference Engine, along with the Analytical Engine, eventually fell into obscurity.
Babbage, too, eventually left his computational engines behind, most likely because he was also interested in things other than just machinery. After that, nothing became of his designs, except for the idea of them, which survived the test of time.
One could argue that before Babbage, there were several other “computers,” e.g., calculators, that could do simple computations. Are these computers? Or are they simply tools to compute things? Hegel had just entered the public sphere at the time of the invention of these engines, so perhaps it’s best to understand this problem with Hegelian dialectics and to state that computers were emerging evolutionarily: What we perceived as the “computer”, back then, was the idea of the computer still developing. What exactly a computer is, is a problem we’ll face at a later time, which came with the developments of the 1930s, when the definiton of what a “computational machine” was in the first place, was formally (mathematically) defined for the first time.
There was still a long way to go before computers would become a physical reality. But the fact that such a step had been taken in the first place, that someone could compute things automatically in the first place, marks a big step in the direction of ubiquitous, universal computing.
Footnotes
*This was, in fact, the origin of the word “computer”: a person who computes, just like a cook who cooks or a driver who drives.
†Like with many “great inventors”, Babbage was apparently a dick
