19th century inventor Charles Babbage never lived to finish his analytical engine, a punchcard computer that was decades ahead of its time, but now Babbage aficionados are finally building this device. Jennifer Oullette of Cocktail Party Physics has the details.
In 1990, cyberpunk authors William Gibson and Bruce Sterling collaborated on a historical thriller, The Difference Engine, in which the 19th century inventor Charles Babbage perfected his steam-powered Analytical Engine. It's an awesome, visionary read, detailing an alternate history in which the computer age arrives a century earlier, Lord Byron becomes prime minister instead of dying of syphilis at 32, and a rebellious group of subversive, anti-technology Luddites — the Insane Clown Posse of their day perhaps? — conspires to overthrow the intellectual elite in power. We can only imagine how our daily lives, science, and the course of history may have been altered if Babbage's ingenious engines had given us comparable computing power in the 19th century.
Except there is exciting news! Plans are afoot to raise funds to finally build Babbage's Analytical Engine, the last and most complex of his many designs for "thinking machines." This is not entirely unprecedented. In the 1990s, a team of scientists at London's Science Museum, led by Doron Swade, built a working model of Babbage's second Difference Engine (the precursor to the Analytical Engine), using only the materials and tools that would have been available in Babbage's day. And it worked. The machine is now prominently displayed in the museum, and Swade wrote a marvelous book about his experiences: The Difference Engine: Charles Babbage and the Quest to Build the First Computer. Here's Swade talking with Wired about his model Difference Engine:
As impressive a feat as building the Difference Engine was, the Analytical Engine is even more impressive. John Graham-Cumming, a programmer and science blogger, now hopes to realise Babbage's vision. Various folks have assembled different elements of the engine over the last 173 years, but this would be the first complete working model. Graham-Cummings told the London Telegraph, "The big difference between it and machines which came 100 years later was that the programme was stored externally, in punch cards. It is basically a giant number-crunching machine–which is effectively what modern computers are today, it's just that those numbers appear to us as words or images on a screen."
Just who was this visionary whose designs were a century before their time? Well, Babbage's personal motto may have been "Born To TInker." The son of a wealthy London banker, he was one of those kids who loved to take apart his toys to see how they worked — probably with the usual mixed results when it came time to put the toys back together again. He had a knack for math, too, teaching himself algebra before attending Cambridge University's Trinity College. He loved numbers and found minute details endlessly fascinating, even compiling a collection of "jest books" to scientifically analyze "the causes of wit." I guess nobody had the heart to tell him that explaining a joke never works. He even cracked the supposedly unbreakable "Vigenere" cipher around 1854, considered by many historians to be the most significant breakthrough in cryptoanalysis since the 9th century.
Babbage could be tiresome, even pompous, but he could also be quite charming. Once, as a "diversion," he drew up a set of mortality tables, now a basic tool of the modern insurance industry. "A man with such a head for numbers and flair for flattery was bound to end up in life insurance," historian Benjamin Woolley quipped in his biography of Ada Lovelace (a Babbage fan and collaborator), The Bride of Science.
He wasn't handsome, either: the poet Thomas Carlyle, who once described Babbage as "a cross between a frog and a viper." But despite his physical shortcomings and pedantic tendences, Babbage had a strong romantic streak, marrying the woman he loved without his father's permission, and suffering the consequences: being cut out of the will (giving up his chunk of the substantial family fortune in the process).
So, he wasn't wealthy, but he loved his wife, and he was sufficiently well off financially to pursue his love of invention. Among other things, he invented a speedometer and a "cowcatcher," a device that could be affixed to steam locomotives to clear cattle from the tracks.
Babbage first conceived of the idea of a calculating engine in 1821, when he was examining a set of mathematical tables with astronomer John Herschel (the son of William Herschel). Such tables were used to make calculations for astronomy, engineering and nautical navigation, but they were calculated by hand and were riddled with errors. So the answers were often wrong, no doubt causing any number of shipwrecks and engineering travesties. "I wish to God these calculations had been executed by steam!" Babbage exclaimed in exasperation after finding more than a thousand errors in one table. And so began his lifelong quest to mechanize the process.
Then Babbage learned of a novel scheme employed by the French mathematician Gaspard Riche de Prony. France had recently switched to the metric system of measurement. This gave scientists a much-needed standardized system to measure and compare results, but it also required a whole new set of calculating tables with which to carry out increasingly complex scientific calculations. So de Prony established calculating "factories" to manufacture logarithms the same way workers manufactured mercantile goods.
These human "computers" were mostly out-of-work hairdressers who had found their skills at constructing elaborate pompadours for aristrocrats much less in demand after so many former clients lost their heads (literally) at the height of the French Revolution. De Prony devised a rote system of compiling results based on a set of given values and formulae, and the workers just cranked out the answers in what must have been the world's first mathematical assembly line. Babbage figured that if an army of untrained hairdressers could make the calculations, so could a computing "engine." In fact, a mechanical calculator for adding and subtracting numbers, called the arithmometer, had only just been invented.
Suitably inspired, Babbage designed his first "Difference Engine," which created tables of values by finding the common difference between terms in a sequence. Powered by steam, it was limited only by the number of digits the machine had available. His demonstration model was the size of a steamer trunk, with two brass columns supporting two thick plates sandwiching a complicated array of gears and levers. Small wheels with numbers engraved in their rims displayed the terms and results of the calculations. Babbage proudly displayed his prototype to the many visitors he entertained in his Dorset Street home, although most did not share his love of numbers.
In an effort to pique their interest, he put on a dog-and-pony show: he would announce that the machine would add numbers by two, then begin turning the crank. The wheels with the carved figures would begin displaying the predictable sequence: 0, 2, 4, 6, 8, and so on. The machine would do this for roughly 50 iterations. Just when the audience was becoming bored and restless, the number would suddenly leap to a new, seemingly random value and then continue the adding-by-two sequence from there. It seemed miraculous, but really, all Babbage had done was program the engine to perform one routine for a given number of turns of the crank, and then to jump to a "subroutine" for another few turns before returning to the original routine.
Why was a full-scale version of this amazing machine never built? Well, Babbage was a perfectionist who was never satisfied with his work, and continually revised his blueprints. He was more of a tinkerer than, say, a "finisher." He spent thousands of pounds of government funding to rebuild the same parts over and over. Eventually the British government lost patience with his lack of progress and decided to suspend funding for his work in 1832, terminating the project altogether in the 1840s.
Nor was the government inclined to look kindly on the successor to the Difference Engine: the Analytical Engine. This new, improved machine wouldn't just calculate a specific set of tables; it would solve a variety of math problems. Babbage based his design on the cotton mill. There was a memory function called a "store", and processing function called the "mill."
But the Analytical Engine would use punched cards to control the cogs, based on the weaving cards developed to "program" looms to weave particular patterns — essentially pre-Victorian floppy disks or CD-ROMs. When strung together, these cards would enable the machine to perform "loops," whereby a sequence of instructions would be repeated over and over, or "conditional branching," where one series of cards would be skipped and another read if certain conditions were met.
Babbage's design provoked a heated reaction from those who just couldn't grasp what he was after. Robert Peel, the head of England's Tory administration at the time, denounced the Analytical Engine as a "worthless device, its only conceivable use being to work out exactly how little benefit it would be to science." It would have required tens of thousands of parts, intricately assembled into a frame the size of a small locomotive. Some historians have speculated that Babbage's machines were never built because they demanded a level of engineering sophistication that simply didn't exist in pre-Victorian England. But it turns out it was mostly a question of money.
Frankly, it's still a question of money, since Graham-Cumming estimates it will cost about $640,000 to complete the project — assuming there aren't any over-runs. He's sufficiently motivated that he's been soliciting donations from various sources. The idea is to raise the money by this January so he can start building his Analytical Engine. And he's off to a great start, with around 1,600 supporters so far who have pledged funds to the project. He will base his project on the original Babbage blueprints (and we don't envy him the task of sifting through Babbage's many revisions), which will be digitized the better to decipher the inventor's many annotations. Then Graham-Cummings and hs team will build a 3D simulation of the gadget on a computer before attempting to build the real thing. Once he's done, he'll donate the machine to a museum — hopefully the same one that houses Swade's Difference Engine. Keep Babbage's rebuilt machines together, I say. That's how he would have wanted it.
This post originally appeared at Cocktail Party Physics.