How History Informs Today: Babbage, Turing, Lovelace, & the Discovery of Computation
Best of: Internet Summit 2018
- Sydney Padua - Cartoonist & Visual Effects Artist
- Doron Swade - Historian of Computing
- Moderator: John Graham-Cumming - CTO, Cloudflare
Transcript (Beta)
🎵Outro music plays🎵 Alright, let's sit down.
Thank you very much. Alright, good afternoon. A slight change of theme, we're going to go backwards in time, into the history of computing, and see if we can figure out the relevance of the history of computing to today.
And the two people to help me out with that are Sydney Padua, who is, yeah exactly, whoo!
Who is a fan here, so he's a cartoonist and visual effects artist who wrote a very well-received, and in fact award-winning, graphic novel called The Thrilling Adventures of Lovelace and Babbage.
I'm going to ask her why thrilling in a minute.
But it was so thrilling that she won a prize, the Newman Prize for writing about the history of mathematics, which I'm sure everybody congratulates you on.
And Dorrance Wade is an engineer who wanted to answer the question about why certain things hadn't been done in computing, in particular things that Babbage had done, and so turned into a historian, ended up at the Science Museum, and if you've seen The Difference Engine No.
2, Babbage's machine, this is the guy who masterminded that project and caused it to happen, and has turned that into a career of history of computing, everything from how we started counting on our fingers until where we are today with the silicon world in which we live.
Why did you bring an iPad on stage? Apparently I'm live drawing.
Okay, you're allowed to live draw, alright. Here we go, get ready. So let me ask you a question.
Why do you say that Lovelace and Babbage was thrilling? Why was it a thrilling adventure?
It started as a joke because it wasn't that thrilling because they never built this machine.
They just sat in a room and talked about it.
So it's somewhat ironic as a title, but in the book they actually build the machine and run around and have adventures.
Adventures that we imagine they would have had.
Doran, why did you have to build the machine and why didn't Babbage do it?
Is it a question of why I built it or why Babbage didn't? Well, why Babbage didn't build it and then why was the right time for you to build it?
Almost any history book will tell you that Babbage did not succeed in building this mechanical machine because of limitations of 19th century mechanical engineering, the idea being that you couldn't make parts with sufficient precision.
Anyone who knows about the 19th century knows that achievable precision is not a limitation.
So I became Curator of Computing at the Science Museum, took the books off the shelf, started reading about Babbage and it said limitations of technology, that's why the machine was not built.
So I kept turning the pages to get then, well, when these limitations were removed, like in our modern times, why wasn't the machine built?
And there was nothing and the book ended. That's where the story ended.
So the question is, was the fact that it did not build it masking some logical or technical impossibility in the prospects for having a workable computer?
So that's what started the whole venture. And why did you end up building it?
So you got fascinated by this, but then you decided to actually build the artifact.
Now, why bother building a machine that is, I mean, it's very, very slow, right?
And it only is very limited in use. Why bother with that? The purpose wasn't to produce a machine that produced anything useful in contemporary terms.
The reason for building the machine was actually to explore an historical thesis and to, in a sense, vindicate Babbage.
We did not know whether the machine would or would not work.
The idea was to build it in order to understand why. Why it did not work if it did not work.
As it happened, it worked perfectly. There were practically no deficiencies in the logic or the mechanical design of the thing.
So we built it to actually, as a piece of experimental history, if the received perception is its limitations of technology and we believe that not to be the case, then let's prove it not to be the case.
So we built the machine and no more precisely than we know from measurement, Babbage could himself have done.
So we didn't use original 19th century manufacturing techniques, but we did measurements on parts that Babbage himself left and measured the tolerance.
How precisely could you make...
If you gave an engineer a part, how precisely could make another part to match it, near identical parts?
And we worked to know more precisely than we know from those measurements, Babbage could himself have done.
And more sloppily whenever we could. So we could say that, yes, he could have built the machine.
And therefore, if history is the human account of the past, then we rewrote history because we now have a different narrative about why Babbage didn't succeed.
It wasn't because of limitations of technology directly. If by that is construed, achievable precision.
There were other reasons. He needed...
If you look at a machine, of any of those machines of that period, there's a lot of repetition.
Every digit wheel, every number has its own wheel. So you have these stacks of near identical parts.
They did not have the production technology to produce repeat parts, stamping, pressure die casting, that would inherently have repeatability.
So everything was made... So the pinch point was manufacture.
So technology, social organisation... If you have a nut made in Manchester and a bolt made in London, you could not be sure they would fit because the lead screws and all the lathes were different.
There was no standardisation. So the reason we built it was to show that had he not had these limitations of credibility, time and cost, would it have worked?
So it was an historical thesis we were testing.
Sidney, also some limitations of personality as well, right? So you've got obsessed by Lovelace and Babbage's actual personages.
I think you're going to help draw us as well.
I mean, he had some personality limitations as well, right?
Yeah, I think... I mean, at the time when people were writing about what was happening with this machine, because it was very famous, the analytical engine shows up in the popular press of the time as Babbage's incredible calculating machine which was getting more and more elaborate and ever actually being built and running through all this government funding.
And I think the perception at the time was the problem was Babbage.
He was a very difficult person. He had a really hard time working with other people and he was also a perfectionist.
So he would keep redesigning and improving the design on paper without ever actually building anything physical.
So I got interested in, I guess, how much the personalities and the failings of the people involved were driving the history of the time.
And you end up drawing them. Yes. So it's hard to draw and talk at the same time.
So Doran came at it from an engineering background. You came from an animation background and started drawing these people.
But talking to you earlier, they ended up obsessing you a bit.
Yeah, well, they're very... I'd actually started the comic as someone who hated computers.
So it was a satire on computing.
How did that work out? I still hate computers. Sorry, I'm the wrong crowd for that sort of thing.
There's a back entrance. We'll go out in a minute. We'll be safe.
I mean, I'm an animator. I started in hand-drawn and then I moved into computer animation, which is very abstract and mathematical and a bit alienating, I think, for someone like me.
So I was fascinated by the analytical engine because it's so tactile.
I mean, it's a very physical machine that I could understand.
I grew to love the analytical engine, I think.
The analytical engine is such a beautiful thing, even as a concept, even though that one doesn't exist.
It's still very, very beautiful.
So there's an attraction, I think, to that for someone like me.
It's just the beauty of it. Did you come to love Babbage and Lovelace? I'm very fond of them.
I think, you know, when you're dealing with personalities... I mean, the ones in the comic are characters, of course, of the people that they actually were, but they were certainly very difficult people, which makes them great for storytelling because they were always causing problems in their own lives and other people's lives, and that makes for a good story.
Give us an example.
What sort of trouble did they cause in other people's lives? Oh, I mean, they caused problems in their own lives with each other, I think.
I mean, there's the very famous battle that they had.
If you're not familiar with the history, Lovelace was an aristocrat who wrote the only publication on the analytical engine in that period, and she wrote it as a set of footnotes to a paper that somebody else had written.
But Babbage objected... Babbage was very happy with her notes, but he wanted to introduce the paper with this giant screed about the government funding and how terrible it was, so they had a god -awful fight about it.
I mean, it's one of the reasons the engine wasn't built, aside from the tolerances of threads in Manchester.
I think a lot of it was simply, to solve that sort of problem, you need people who can work together, I think, more effectively than they did.
Well, I'm going to let you draw from it and go back to Doran.
So they don't build it for these sort of manufacturing reasons, but there's another thing that's going on at the same time, which is Michael Faraday is over at the Royal Institution.
He's got dynamos and motors and solar noise and all these things he's inventing, and Babbage persists with cogs all through his life, well into his 80s.
In parallel, electronics is about to appear.
Why does he not suddenly switch gears in the middle?
It's a very intriguing question, and it's one that is sort of incomprehensible to us.
Why didn't we see the potential of electromagnetics to begin with?
And the answer, I think, is that Babbage's machines were digital. That is to say, the motions of each of the parts were discrete.
They were finitely incremental.
The idea for that was, he realised very early on, his very earliest conception was it had to be digital, because that's the only way you can quantify the inaccuracy, quantify the uncertainty.
So a number between two and three, you knew it wasn't less than two, you knew it wasn't more than three, but where it was in between didn't matter.
And that was the power of the digital stuff. It was also decimal.
And the idea then was, this was the age of mechanism, that mechanics and mechanical things would exhaust explanation.
Any quality could be reduced to quantity, and quantity could be calculated.
So the idea is that you didn't need anything outside that.
This was the terminus. This was the explanatory terminus of all things, was mechanism.
So they weren't open to the idea there was something else.
They weren't open to the idea that there was another way of implementing digital logic, mechanical logic.
But there's another aspect to that, and that is that they didn't see digital as necessarily to do with switching.
They saw it as incrementalized motion, of discretized motion.
So I think that was another.
Also, it was very, very early days. They were doing the earliest experiments with magnetism.
They didn't have Faraday's formula yet. Electromagnetism came.
So the idea they would have switching relays and then leap, as it were, jump horses from mechanics to electromagnetics didn't arise.
But he saw Wheatstone's thing.
He wrote a note in his diary saying, yesterday we saw Wheatstone's thing, and he says my machine can calculate faster than his.
The one that didn't exist.
Yes. He was also blinkered about the machine being purely for mathematics, whereas Lovelace saw something else much more fundamental, much more important.
Yes. Babbage was a mathematician at heart. That was his starting point for the whole thing.
The idea was to build a machine that could find the value of any mathematical function.
And he never saw, even after 50 years of mature reflection, the machine outside mathematics.
Lovelace was the first to articulate something profoundly important, which is why she should be celebrated, but not for the reasons that she's most commonly celebrated for.
She saw that number could represent entity other than quantity.
So because machines operate on a number, you assign meaning to the number.
You assign meaning from the world to the number.
The machine operates on the number. You then map those meanings back to the world, and you've got the machine talking about the world.
Babbage always saw this as mathematical, as a machine for mathematics.
So Lovelace saw that the power of a computer was in its ability to represent according to rules, to represent aspects of the world in symbols and manipulate according to rules.
So the power laid in the representation of the world contained in symbols.
So the machine would manipulate symbols represented by numbers, and you would map this back.
So she speaks about numbers representing letters of the alphabet, notes of music.
Okay, now that was unprecedented, because the idea there was that mathematics was the grammar of the world.
She's got a wonderful thing she said.
She said, mathematics is the language of the hidden connection between things.
So once you had mathematics, you didn't need symbolism. You didn't need abstract symbolism.
Say this number represents something else. But she articulated the idea that numbers could represent things outside mathematics, and that actually, it's not just she's saying it.
She's actually banging the table and saying, this is what is important about the analytical engine.
It's not just some whimsical retrospective attribution to her.
She actually is thumping the table saying, this is what is important about this machine.
It can do things outside any precedent of a machine.
And then she dies, age 36, and Babbage carries on for the rest of his life, and he must have read this, and he didn't grok it.
He carries on with the mathematics.
He read it, and he complimented her, and didn't take a blind bit of notice.
Right, exactly. So he completely missed the fact that there's gradients.
Or he believed that mathematics was the grammar of the world, and therefore you didn't need abstraction.
Yes, he didn't need to explain it. And to be clear, all the machines we use today are doing what Lovelace talked about, which is they're all doing numerical stuff underneath, and that's representing music, and what you've just been drawing, et cetera.
I mean, I got the sense that Babbage, he thought it was an interesting idea, but it wasn't what he was obsessed with.
I mean, he was obsessed with this very, very specific machine, and just improving that machine, rather than kind of re-approaching his original conception.
He was still looking to build this huge stack of cogs for dealing with very large numbers and printing the results.
I mean, that was his thing. And I think this idea of this music-composing, symbol -manipulating thing was a completely different machine than the one he was designing, because so much of what Babbage was working on, like the carry mechanism, which took up so much of his time, that was about speeding up carrying ones when you're multiplying two giant numbers.
I mean, that was his kind of obsession.
He wasn't kind of looking at a computer and trying to figure out how to get there.
He was kind of going off. It does make you wonder if she had not died so young, if she might have persuaded him later that maybe there's a natural use for this, and someone will have to be more interested in it.
He wasn't a very persuadable person. He wasn't a persuadable person, fair enough.
So with that, I mean, with Lovelace and Babbage eventually dying much later on, that sort of bit of history kind of came to an end.
And then if we jump forward to 1936, Turing comes along and looks at some theoretical problem, and he designs a machine, again on paper.
This one even more impractical than Babbage's machine because it requires an indefinitely long tape.
So we had two goes at computing in Britain.
We blew it both times, right? So we had two chances. But is there a link between these two?
Do they know about, does Turing know about Babbage?
One is they were both mathematicians, which I think is no accident.
So they both saw that the computational challenge was the central one.
Babbage, Turing knew of Lovelace, but it's questionable whether he actually read her paper, her 1843 paper.
It seems evident that he knows of Lovelace through a paper written through Hartree, one of the British pioneers of computing.
And he talks about the Lovelace objection in the context of his artificial intelligence work.
And Lovelace says, ventriloquizing Babbage as it happens, that a machine cannot originate anything.
It can only do what you tell it to do. It's a familiar trope in artificial intelligence, machine intelligence.
And Turing seizes this and argues against it, saying that we should not prescribe the limits of what machines could do.
Lovelace was just repeating, not repeating. I mean, she understood it and was convinced.
But there was a question of whether they were trying to, in a sense, appease the religious constituency people, whereby if you said the machine was autonomous and capable of autonomous activity, thought was sort of God-given, and you then said a machine could actually appropriate human intelligence, then you may be a great kind of offense.
So it could be a kind of faux defense of saying, don't worry, the machine is a dumb thing.
It only does what you tell it to do. It will never originate and compete with the divine qualities of humanity.
And so it could be that it actually wasn't meant that this is actually placing constraints of what, in principle, a machine could do, but actually trying to appease a political faction.
Sidney, what have you been drawing? I've not been watching. Is this Babbage?
I was doodling Babbage. This is Babbage. Now, just to make sure we understand their personality, how do you determine that Babbage stood like that or looked like that?
I mean, we have a few pictures of him, I guess. Yeah, I mean, there's photographs of Babbage.
There we go. Hang on. There's photographs of him, but most of his sort of energy, I guess, the way I draw it, comes from his autobiography.
And there's a lot of anecdotes about him as well. I mean, his personality is extremely defined.
He's a very, very, gosh, how would you put it, Dorian?
I can't. It's easier to articulate in a drawing than it is in words, actually.
He's completely predictable, highly principled, very robust, will take anybody on.
His friends could do no wrong. His enemies could do no right. Pugnacious. Pugnacious.
Dogmatic, pugnacious. Cumbative. Cumbative. He would take anyone on.
He would denounce people publicly. Oh, God, yes. People like the Prime Minister or anybody.
Absolutely. Sir Humphrey Davy, the whole of the British scientific community.
The whole of the British scientific community. It's extremely easy to imagine his Twitter feed.
He had very, very strong opinions on literally everything.
But Lovelace did too.
She had quite a strong opinion about herself. Yeah. Lovelace was definitely convinced that she was a genius and had very important things to say, which she did.
Which she did, right? I mean, she recognized that the computer was much more than this mathematical object.
Yeah, I guess she was more variable than Babbage.
I mean, Babbage presents one very strong personality and Lovelace presents about six or seven different very strong personalities.
But they're both very eccentric in a very Victorian way.
You know, if you think of Dickens' characters that seem over the top, I think, you know, it was a great flourishing of the art of personality, the Victorian period.
I mean, people were very idiosyncratic.
And they were certainly even more, I think, idiosyncratic than others.
We're also lucky in that there's so many descriptions of them from people who knew them.
So after she died, we don't know a lot about their relationship. But you uncovered this description of Babbage in the United States, I think, where he gets asked about her.
Oh, it's an American visitor. He tried to persuade Babbage to go to the States.
I think he's meeting him in London. But, I mean, so much of the book came from browsing around Google Books for kind of anecdotes of them.
And that gave me this sort of... I guess I always want to give a shout-out to the Google Books Project because so much of what people are like, wow, this original research that she did was so amazing.
It's just because I was the first person to type Babbage, Lovelace, time frame, you know, search on Google Books.
And it just pours out this tremendous amount of stuff. And this letter that this American writes about Babbage expressing his admiration for Lovelace and how fond of her he was.
And, I mean, I think more information about their relationship than in any other historical document.
And that just turned up in some church circular, you know, from the 1860s.
And no one else would have found that, I think, without search.
And, in fact, the visitors were slightly embarrassed because he was so emotional about the fact that she was no longer there as the one person who understood what the hell he was trying to do.
I mean, the reason why this was an exceptional find is that none of Babbage's personal letters survive.
His technical archive is well-documented. But, you know, we don't actually know what this relationship was.
We can speculate. We have Lovelace's as one of Babbage's.
Well, she speculated enormously. Absolutely. That's the beauty of having someone outside the silo of academia.
Right, right. To summarize Babbage, we were saying, well, what was he like?
Babbage thought being right entitled him to be rude.
You know, that was actually his public... A lot of people in the computer industry I've met actually think that, too.
Honestly, yeah. So let's just talk a little bit about Turing because we're in the UK.
So Turing does this thing, and then he goes, obviously we know about Bletchley and all the co -breaking thing, and then there's this period at NPL building a machine.
And that machine also doesn't really get anywhere.
But there are other contemporary machines, based on what von Neumann was doing, which do get built.
Why is that? Why does Turing not carry on as to be the great builder of machines?
The short answer is that the von Neumann architecture and the developments in America effectively overwhelmed what was going on in England and Europe.
So there were two independent strands that were running parallel right until probably the late 60s, early 70s.
This is Turing developing very idiosyncratic logic, ingenious stuff.
But his idea was that because processing power was so precious, you minimised what the machine did and maximised what the human did.
And the whole trend of the von Neumann architecture, the whole American developments, which include in England EDSAC, one built by Wilkes in Cambridge, followed the von Neumann architecture, which has actually dominated computer design ever since.
And so Turing was, in a sense, advocating something which turned out to be a dead end.
He went to Manchester.
There's no doubt the expertise from the war disseminated to other people. But his main influence was actually through dissemination through people rather than every particular hardware.
There are two schools of thought. One says that Turing was a genius.
He's the father of the computer age. And the other one says modern computing would be absolutely no different had he never lived.
And so he's a controversial character.
I don't think history has resolved Turing's contribution.
There's no question the 1936 paper, which formalises, in an abstract way, gave intellectual respectability to computer science.
That is, without question, a remarkable thing.
And without question, the work of Bletchley, Balkan, Kobrak is remarkable.
Whether his contribution to computing actually is significant is questionable.
And also there's this remarkable thing, which is Alonzo Church, who is his PhD supervisor.
We have this idea of the Church-Turing thesis, which is that all these models of computing are the same.
And that does give us a link back to Babbage, which was the thing he was trying to build would have been a computer, as we know it, even if it was mechanical.
Because of this, there's a fundamental essence of computing in there.
Yes, I think that's absolutely correct. All right, listen, we are going to get out of time.
And I could sit here and talk to you for another hour about this and watch Sydney draw things.
Is this Babbage and Lovelace right here?
What are they up to? Yep. All right, so why don't we open this up to questions, because I'm sure people have got questions for both of you about the history of computing, how you draw, how you got obsessed with Babbage and Lovelace.
Why don't we take some questions? So if Babbage's machine had been built, do you think it would have sat in a dusty corner as a kind of oddity, or do you think it would have kick-started an almost information revolution in the Victorian era?
I appreciate that speculation, but what do you think?
What if history? What if history, yes, counterfactual history. It's very intriguing, because Babbage failed to build his difference engine.
The purpose of the difference engine was to have a factory for producing mathematical tables, because that's what everyone relied on for calculation.
So he failed. It was a false dawn.
But other people in the 19th century, inspired by Babbage, and that's the Schweitzers, Weiberg in Sweden, others, built successful difference engines, small ones.
And the ones that were technically successful failed commercially.
So one thing, yes, Babbage failed, but if the object of this was to produce mathematical tables, then there was no market for it.
There was no economic case for spending huge capital investment in a machine when the experts of the day, this was unverifiable at the time, the experts they argued, tables were sufficiently accurate, and they could be improved by progressively finding fault.
They could be purified, as it were, with time.
So one question is, it would have made no difference had he succeeded, because other people who did succeed didn't.
The other thing was whether it would have given rise to an information age.
The problem was communications. And if you look at the four strands, if you like, which converged in the 70s in the great fusion of solid state physics, you've got calculation, you've got automatic computation, you've got communication, and you've got information management.
The communication element wasn't there to create anything else.
Babbage himself did not foresee that there would be more than one major computer in the major capitals of the world.
That was it.
So that's one side. It wouldn't have been any different. The other side is the world would have been completely different, and the reason for that is a phrase that Francis Spufford, a writer, said.
When he looked at the difference engine, he said, what is this thing, the thing we built?
Is it a fake antique? Is it a new primary source?
And he said, when you stand in front of it and watch it work with these magnificent double helices turning, he said, it allows you to appreciate the wonder of a future already past.
So you put yourself in the position of a Victorian, and even though electronics technology has so far superseded anything Babbage did, it is still a wondrous object.
So I think it's difficult to imagine that any Victorian standing in front of this extraordinary machine would not have been inspired to say, what is this?
How do we carry this on?
So two answers. One, nothing. The second one is everything. Do we have a question for Sydney?
Let's see if there's a question. If there's not, we'll take a question.
I would like to. I have a question. Why don't you ask her, right? She's right there.
Why, when you portray Lovelace, does she have a pipe? Why is she smoking a pipe?
Yeah, she's not smoking a pipe here, but she's generally smoking a pipe because people who smoke pipes get taken seriously.
Ah, wait. I'll remember that.
You always look very thoughtful. We have a couple more minutes. Is there another question in general from the audience?
I've got a question for Doran. Go ahead.
So given the dawn of the quantum age, does the von Neumann architecture have a place in the future, or is it going to belong in a museum?
I would have thought that there are any number of people present before me that are better positioned to answer that question than I would.
I would say this, that nobody predicted, nobody predicted, including right from the 1960s, alone people in the 19th century, predicted what the computer was to become.
It was Lovelace standing alone as this kind of light in the darkness that foresaw that the application of computer was outside number, outside the boundaries of mathematics.
So I would find it completely impossible to speculate on what the implication is.
We simply don't know, and I would say that as a kind of plug for blue sky research money, is that there needs to be a source of funding in addition to and other than markets, because we really don't know where the stuff is going, and people should be free to pursue it in ways that aren't necessarily constrained by markets, as they inevitably are, as useful.
You have to prescriptively say what the utility of a product is before you have the freedom to develop it.
And here there were things that had such far-reaching outcomes, which were developed because these were gentlemen of science, independently wealthy, they could indulge themselves.
So I say there need to be institutions and sources of funding that allow the kind of things that go on without actually specific knowledge of what its utility and benefit might be.
Right, sort of blue sky type things.
All right, I think we are just about out of time. Thank you, Sydney, for drawing and telling us about Babbage and Lovelace personalities and all that.
Darren, thank you for the historical perspective.
Thanks very much. Thank you.