Redefining "big iron"

December 28, 2011 12:29 PM

The world's first general-purpose, programmable computer was Charles Babbage's mechanical Analytical Engine, which was a formidable accomplishment even if the cost of its construction was prohibitive. While Babbage focused on engineering challenges, mathematician Ada Lovelace wrote the first program for the Analytical Engine, and provided some important insights into the power of a programmable computer. Unfortunately, Babbage never completed an Analytical Engine. Mike James has written an interesting piece on his blog speculating about how our world would be different a working Analytical Engine had been constructed. This topic also was covered in William Gibson and Bruce Sterling's 1990 novel, "The Difference Engine", a seminal work in the steampunk genre. It's interesting to think about how the world would be different had engineers and scientists had access to fast, high-speed computers a hundred years before the birth of UNIVAC.

A previous discussion on MetaFilter focused on an ongoing effort to construct an Analytical Engine, but the impatient reader may prefer to investigate Fourmilab's software implementation.

posted by wintermind (33 comments total)

Obligatory link to Young Ada Lovelace.

(Also I'll be all snooty and say that The Difference Engine doesn't really fall under the area of modern Steampunk since it retains an awareness of actual history and engineering and isn't just people with goggles doing magic stuff with cogs. )
posted by Artw at 12:37 PM on December 28, 2011 [1 favorite]

Neuromancer also doesn't fit the cyberpunk genre, as Gibson would attest. It still produced the genre.
posted by mek at 12:43 PM on December 28, 2011

How would Moore's law apply to mechanical computers? (maybe this belongs in a different thread)
posted by mr vino at 12:44 PM on December 28, 2011

At the time it came out, I remember The Difference Engine being described as Steam Punk. It's not really the book's fault that the genre has changed since then.
posted by octothorpe at 12:48 PM on December 28, 2011

Unfortunately, Babbage never completed an Analytical Engine.

Hold up... are you telling me that Sydney Padua's 2D Goggles: The Adventures of Lovelace and Babbage is not historically accurate?
posted by PenDevil at 12:49 PM on December 28, 2011

Remember that "steampunk" originated as a portmanteau of cyberpunk (itself, of course, a portmanteau), probably driven by Gibson and Sterling being well-known cyberpunk authors already as much as anything else. The whole thing of sticking random gears on your goffik clothes was still years away.
posted by Halloween Jack at 12:53 PM on December 28, 2011 [2 favorites]

Don't forget The Thrilling Adventures of Lovelace and Babbage, by Metafilter's own Erasmouse (Sydney Padua).

(On preview, beaten again, but Erasmouse's comment that I linked is really interesting.)
posted by kmz at 12:53 PM on December 28, 2011

Damn, fast and high-speed? Even today, we have to settle for one or the other.
posted by LogicalDash at 1:03 PM on December 28, 2011

Sorry, LogicalDash, that should read "cheap, high-speed".
posted by wintermind at 1:05 PM on December 28, 2011

The 'fast' referred to how firmly attached to its mount Babbage's original design was, and also that it allowed a promiscuity of calculation.
posted by Abiezer at 1:12 PM on December 28, 2011 [4 favorites]

It still produced the genre.

It's the defining work of the genre (which had been around for nearly a decade by the time it came out). If Neuromancer wasn't cyberpunk, nothing was.
posted by empath at 1:23 PM on December 28, 2011

Plus also there aren't a bunch of cosplay idiots running around dressed as Neo claiming to be "cyberpunks".

Or are there?
posted by Artw at 1:33 PM on December 28, 2011

Were the inner workings of as computer to be as visible as a mechanical machine, I think it would cause many problems. I don't really understand how the computer beneath my fingers works, but it gives me no sense of its workings as it does so, and so is no distraction. But to see moving parts and spinning axles would annoy me endlessly as I tried to follow each input through the machine and figure out what it does. I doubt anything would get done if every keystroke were an entertaining mystery to be investigated, and not the blackboxed magic that it seems.
posted by Jehan at 2:00 PM on December 28, 2011

Are there any toys that work let you snap together logic gates the way that redstone works in minecraft? I think that would be a really fun thing, and I would buy it immediately.
posted by empath at 2:30 PM on December 28, 2011 [1 favorite]

Plus also there aren't a bunch of cosplay idiots running around dressed as Neo claiming to be "cyberpunks".

I'm fairly certain they were an early to mid-90s phenomenon.
posted by Apocryphon at 2:55 PM on December 28, 2011

The successful construction of the Difference engine is also a plot point in Michael Flynn's 'In The Country Of The Blind'

It has an excellent bibliography on the statistical modeling of history.
posted by the man of twists and turns at 3:20 PM on December 28, 2011 [2 favorites]

Neuromancer also doesn't fit the cyberpunk genre, as Gibson would attest. It still produced the genre.

What? First of all, the term Cyberpunk was around before Neuromancer was written. Second of all when did Gibson ever say his books weren't cyberpunk? What makes them not cyberpunk? For most people they are considered the canonical cyberpunk books. From wikipedia:

William Gibson with his novel Neuromancer (1984) is likely the most famous writer connected with the term cyberpunk. He emphasized style, a fascination with surfaces, and atmosphere over traditional science-fiction tropes. Regarded as ground-breaking and sometimes as "the archetypal cyberpunk work,"[10] Neuromancer was awarded the Hugo, Nebula, and Philip K. Dick Awards. After Gibson's popular debut novel, Count Zero (1986) and Mona Lisa Overdrive (1988) followed. According to the Jargon File, "Gibson's near-total ignorance of computers and the present-day hacker culture enabled him to speculate about the role of computers and hackers in the future in ways hackers have since found both irritatingly naïve and tremendously stimulating."

I think the steampunk aesthetic is kind of interesting, and so is the idea of powerful mechanical devices and so on, but the whole "steampunk thing" is pretty ridiculous these days.
posted by delmoi at 3:24 PM on December 28, 2011

Also, one cool thing about the Analytical engine: It was actually a stored-program computer, one that would actually have been Turing complete, a hundred years before Turing even came up with the concept.
posted by delmoi at 3:26 PM on December 28, 2011 [1 favorite]

This article is kind of ridiculous, though. I mean:

Had Babbage built the Analytical Engine then it might be that we would all be using IBM PCs - but in this case that would stand for International Babbage Machine and Powered Computers! You would also be reading “I-Analytical Engineer”!

And:

Of course it is also entirely possible that silicon lithography, the technique use to manufacture chips, would have been invented for a very different purpose.

If you know about transistors and electronic logic then presumably you can see how to use the doping of silicon to build integrated circuits but if what you know about is mechanical logic - presumably you start to think about ways of miniaturising mechanical devices.
...
With a start in mechanical computers micro-engineering would have been developed and the division between the machine that does work and the machine that thinks would have been less obvious.

This stuff is just nonsense. People were well aware of using mechanical calculators before the advent of electronic ones, and people were already making gears as small as they could in order to make watches. The first electrical computers were done using relays, and relays signaled by electrical signals rather then gears signaled pistons and cogs would have been much more efficient.
posted by delmoi at 3:35 PM on December 28, 2011 [2 favorites]

How would Moore's law apply to mechanical computers?. It wouldn't. Moore's law is not a physical law, it is merely a smart observation about the interactions of physics, consumer demand and capital.

The evolution of the internal combustion engine in the past 100 years is a good analogy for how mechanical computers would have evolved. An Alternate 21st Century Babbage Engine would likely be only 100 times better/faster than Babbage's original design, just like a Formula 1 engine is only a few hundred times more powerful than a 1910 automobile engine.
posted by monotreme at 3:43 PM on December 28, 2011 [1 favorite]

> I doubt anything would get done if every keystroke were an entertaining mystery to be
> investigated, and not the blackboxed magic that it seems.

It's hard to resist. In addition to my dad's big-enough-for-a-deadly-weapon e-school slide rule I also have his big heavy Marchant motorized mechanical calculator, which will still light up and play. I am sorry to say it is not steam powered but if you ask it to divide something by zero it will sit there all day and spin, and spin, and spin. Trying.

If Babbage's engine had a similar feature and a full-size one had been completed, it might--you know, maybe--might have been big enough.
posted by jfuller at 3:57 PM on December 28, 2011

Previously in MetaFilter:
http://www.metafilter.com/90380/Happy-Ada-Lovelace-Day
posted by Cranberry at 4:42 PM on December 28, 2011

Were the inner workings of as computer to be as visible as a mechanical machine, I think it would cause many problems. I don't really understand how the computer beneath my fingers works, but it gives me no sense of its workings as it does so, and so is no distraction. But to see moving parts and spinning axles would annoy me endlessly as I tried to follow each input through the machine and figure out what it does. I doubt anything would get done if every keystroke were an entertaining mystery to be investigated, and not the blackboxed magic that it seems.

First of all, there wouldn't be any 'keystrokes' You would put your programs down on punchcards, at home, then drop them off. If you're lucky, you could put them in yourself. Then, after whirring away for a few hours, you would get your result. Most likely, you'd just drop them off and come back a few hours later.

That's actually how most people used computers at first. After that, people used terminals hooked up to computers in other rooms. They worked like a typwriter, except you'd see the output of the compute on paper, rather then what you typed. (oh, and lots of people used mechanical typewriters just fine...)

How would Moore's law apply to mechanical computers?. It wouldn't. Moore's law is not a physical law, it is merely a smart observation about the interactions of physics, consumer demand and capital.

I think it's actually more of a self-fulfilling prophecy then anything else. Everyone else in the industry expected the '18 month' thing and didn't push any harder because they knew no one else would. Had Moore not made that observation (very, very early on by the way) we might have advanced faster. There's no physical reason why we couldn't have.
posted by delmoi at 4:45 PM on December 28, 2011

OK, so let's suppose that a working Analytical Engine had been constructed in the 1850s. Would it actually have changed the world? I doubt it.

Judging by the history of punched cards in the later nineteenth century, it would have made data processing more efficient, but the difference would have been quantitative rather than qualitative: i.e. people would have used the Analytical Engine to speed up their work, rather than to change the way they worked. The obvious technological applications would have been in textile manufacturing, but since British cotton manufacturers already dominated the world market, it's not clear that the existence of an Analytical Engine would have given them any commercial edge they didn't already have. Perhaps it might have encouraged the development of scientific management, but that happened anyway in the 1880s and 1890s and it's not clear how much influence it actually had on industrial productivity. All in all, I suspect a mid-Victorian Analytical Engine would have been a solution in search of a problem.

On the other hand, I'd still like to know what difference the Analytical Engine might have made to the history of late nineteenth-century mathematics, and what practical applications this might have had. Any mathematicians care to comment?
posted by verstegan at 4:48 PM on December 28, 2011 [1 favorite]

OK, so let's suppose that a working Analytical Engine had been constructed in the 1850s. Would it actually have changed the world? I doubt it.

I very much doubt this. What's important here isn't so much the implementation details but the ideas promoted by the underlying technology. The reality is that had the engine been completed it would've been only a matter of time before somebody started networking them together and it's likely that we'd have something like the net maybe 50 years earlier. Think about that.
posted by nixerman at 5:39 PM on December 28, 2011

(oh, and lots of people used mechanical typewriters just fine...)

A mechanical typewriter is not quite the same. They're so simple you can figure them out in a few minutes once the cover is removed. I know this because I did it several times as a child, which kind feeds my suspicion that the curiosity of watching a hugely complex mechanical computer working would be very distracting.
posted by Jehan at 5:49 PM on December 28, 2011

The obvious technological applications would have been in textile manufacturing, but since British cotton manufacturers already dominated the world market, it's not clear that the existence of an Analytical Engine would have given them any commercial edge they didn't already have. Perhaps it might have encouraged the development of scientific management, but that happened anyway in the 1880s and 1890s and it's not clear how much influence it actually had on industrial productivity. All in all, I suspect a mid-Victorian Analytical Engine would have been a solution in search of a problem.

Huh? you're assuming that everything would have been the same, except with mechanical computers. Obviously, it would have changed things. The biggest differences wouldn't be obvious right away, but businesses could start using computers to do their books. For scientists and engineers they could use computers to do their mathematical work rather then spending most of their time doing arithmetic.

People today hardly ever do pure arithmetic on their jobs, it's all done with spreadsheets and software and so we don't even see it. So we don't even think about how things used to be done.

But basically it would have freed up lots of smart people, otherwise occupied doing arithmetic to do other things.
posted by delmoi at 6:04 PM on December 28, 2011 [1 favorite]

Though the ingenuity required to get around the lack of computational power wasn't all wasted, I am sure.
posted by Zalzidrax at 7:13 PM on December 28, 2011

I derived sheer pleasure from seeing this FPP here.

I'd also like to throw out the question whether mechanical calculating tools such as the abacus count as early computers, the original sense of the world. I am reminded of Sean McMullen's Souls in the Great Machine
posted by infini at 10:22 PM on December 28, 2011

The reality is that had the engine been completed it would've been only a matter of time before somebody started networking them together and it's likely that we'd have something like the net maybe 50 years earlier.

Sure, after super accurate rockets and the atomic bomb.
posted by empath at 10:24 PM on December 28, 2011

For scientists and engineers they could use computers to do their mathematical work rather then spending most of their time doing arithmetic.

People today hardly ever do pure arithmetic on their jobs, it's all done with spreadsheets and software and so we don't even see it. So we don't even think about how things used to be done.

But basically it would have freed up lots of smart people, otherwise occupied doing arithmetic to do other things

Enigma
posted by infini at 10:24 PM on December 28, 2011

There were some great dystopian concepts in The Difference Engine. One was the notion that scientific progress as we knew it in the 19th century might have been stifled in some ways (because of the advent of the AE?). The characters in The Difference Engine live in a modern police state, where the computers are used to keep copious information on the population. However they have no germ theory of disease. Possibly the authors were speculating that people like Pasteur and Koch would have gone into programming instead of what they did in real life - analogous to their description of Disraeli as a journalist and ghostwriter rather than a politician. I found that to be the most thought-provoking element of the novel.
posted by zomg at 9:20 AM on December 29, 2011 [1 favorite]

One of the points someone made in the past (I heard this on the radio, so no source...) was that Babbage's machine was not built not because it was too expensive, but because the machining tolerances of the age were too large to allow its practical construction. Kind of a proto-Moore's law, if you will. (I know this could be translated into "too expensive", but I like the idea that in an era when British machinists were capable of amazing feats of invention, they ran up against this specific technical limitation, rather than a purely economic one.) The person on the radio also said that Babbage's plans were the most complex set of draughts (drafts?) devised to that point, and remained so until the early 20th C.

One thing I still appreciate about the Gibson/Sterling book was their conception of computing as an ivory tower that required special status to enter, and only special channels to access the services of the machine itself. First I thought that was parody, but now I'm not so sure.
posted by sneebler at 9:29 AM on December 29, 2011