The Grand Narrative of the History of Computing – Professor Doron Swade



thank you very much Raymond and thank you of course to progression college and the British side of history of mathematics for the opportunity to speak here I must say that the British side of history mathematics say that they the friendliest and most productive group of folk in I've have come across when we talk about the grand narrative the history of computing and really what we're asking is what is the story we tell ourselves about how we got to where we are telling the story will allow us to locate the particular thread of automatic computation in the larger scheme of things and allow us to ask questions that is it sustainable is it sustainable what will he do how can we locate smartphones tablets in this long historical thread we can then look at some of the core ideas if there's time look at some of the core ideas in modern computing and actually see where they might have come from so the history of automatic computation is an emerging field it's a bit like a rebellious adolescent in the sense in which it defies containment what we're looking for is an aerial view but we're looking for an aerial view that will frame that will give us some basis for looking at the meaning and understanding the countless local narratives that populate this tale so what is that Springs them together all those stringers which we'll see is an inappropriate way of looking at things and despite the use of the field if we look at historical surveys canonical text books and popular histories a clear narrative has emerged despite the fact that the field is still changing and it's very difficult actually certainly to frame modern developments so what I propose to do is to give a very brief outline just the arc of the narrative and then look at how it's constructed is there some infrastructure that will give us some clues about how this narrative might evolve is it sustainable can it absorb these new developments or does it Frizzle outs like the Wicked Witch of the West in The Wizard of Oz wind doused with a bucket of water so what is the grand narrative the history of computing it looks to its root in early counting in number systems and typically an account will start with Roman calculate the pebbles that we used on counting boards it then moves on to keep ooh or knotted cords knotted cords we can actually here we go we can see those up there these are knotted cords that we used for counting it's a positional system the value of the position of the north the presence they're not in the position of the old actually gives the value the thing these go back very very early on sort of early antiquity and the dawn of time these expressions historians use when they haven't a clue so these go quite far back these were used by Inc American Indians Incas and we established quite prominently in the fifteenth and sixteenth centuries and we then find tally sticks these are knotted sticks and these go back the earliest one of those about 8,000 years old and these are notches which again record number that it usually amounts usually of money now these were used in the Exchequer for tax administration from the 13th century in responses the usage was stopped in 1826 so they're overly recent now the way it worked was you actually the nachos record the amount it's either a deposit or a debt and the nachos record the amount there's a code that allows you to go and then you split this long it's lengthwise to make two parts so they're not superior in both arms both section one is a thin bit and one is a thick bit the fin bit is called a foil after the French word for leaf and the thick of it is called the stock now the bank or the taxman keeps the thin bit and the the depositor or debtor keeps the the stock and the legacy of that is still here we now talk about counter foils and we talk about stock holders so this the legacy of that usage is still in our language some of these some of these sticks for one to two foot in length and if there may be an opportunity for the Financial Services Authority the idea was that when you then had a new settlement or a new payment you married the two halves and if the notches were intact then you were relative should that they hadn't been any forgery or fraud and there may be an opportunity over the Financial Services Authority is to actually introduce this again to prevent bank fraud well know if I had a two-foot stick and I was in a room with a banker there probably other things I could imagine doing other than cutting notches in it so so these are the systems kind of severed security so what's important is that computing looks back and sees its origins in counting and number systems and various paraphernalia the physical aids and artifacts to assist relieve or replace manual company little manual calculation so we have other bits and pieces the narrative then usually shifts from counting aids to devices with moving parts and of course a mechanical calculating adds in the abacus here has unchallenged priority there's a Japanese soroban and the Russians surety and has with the keeper they positional systems the value of the position of the bead on the in the wire and bead frame and so on during the rapini enlightenment there was a surge in calculating aides analog devices with graduated scales for calculation and measurement with the mainstay of calculation in the 17th century and onwards quadrants extras proportional compasses are some of these I won't go into any detail in those Napier's bones 1617 we have an example there there's nothing but car been calling the Napier's bones there's nothing the carving at at all they were called Napier's bones because the upmarket versions of these were actually carved in bone and not in wood as this example shows here and they were used for photosystem II multiplication division they were also specialized ones for doing things like square roots slide rules were the absolute mainstay of calculation for over 350 years the logarithms were introduced early 17th century and slide rules of which there are countless varieties were actually the main so they became a badge if you like of office for engineers and I was my generation of engineers always had slide rules this was until they were entirely wiped out and achieved a total obscurity in the early 1970s when electronic pocket calculators if if you wiped them out and that last example he's actually the the slide where I carried around as an engineer so that tells you that I go back to biblical times so in the next chapter in the sequence is is mechanical calculators the beginnings of attempts to transfer human intelligence into mechanism that somehow the algorithm the rules of operation were now incorporated in some way in the machine in these cases only partial so it starts off with it's a 17th century early 17th century Shekhar's calculating clock which isn't here the Pascaline 1642 Blaise Pascal's calculator and Johan Mueller's Universal calculator so these were these were really objects on law they were they were paraded in front of the wealthy and the wealthy and the aristocracy they were if you like philosophical novelties which stimulate a lot of debate about the possibility of automating machines and calculation and if you like mechanizing thought but they weren't really suitable for practical use things change things get serious about in 1820 with introductions of the tacoma's era thermometer that's a later version but it was introduced in 1820 but it took nearly 60 years 60 70 years before the era of monitor actually matured as a product in the general register office by 1876 that's over half a century after the first introduction there were still only three Aerith monitors in use so took a huge amount of time that take up of this product was actually very very slow and the implications of acceptance of innovation is something that we'll be looking at and after that Babbage is usually we usually have a section on Babbage it's a manual calculator is what defines them as a class is that they require human intervention for the calculation to proceed you have to enter the into the numbers by hand you have to read them off you have to transcribe the results and so on the algorithm is only partially incorporated usually does addition and carriage to a limited extent but the actual complete augur they're not automatic they're manual in the sense that they require human intervention for the algorithm to proceed the algorithm is student human the big breakthrough of course is jobs Babbage and 1821 with his conception of automating this and then it's that's followed very quickly by Herman Hollerith with introduction 1890 of his tabulating machines his tabulators and sorters for the 1890 census so we then have a seamless segue from Hollerith into the great office automation industry that Martin touched on earlier of the early part of the 19th century so this is punch card machines and desktop calculators the celebrity devices Burroughs felt Baldwin brunsviga facet these were the great mechanical calculators of the age in part of the office automation era and then we have in the ensuing boom in the computer industry over the last eight years which brings us perhaps slightly breathlessly to the present so that's just a general sort of arc it goes from from from calculi to smartphones if you like via these mechanical calculator you started with counting and progressively trying to incorporate and replace more and more of human activity in this thing so what are the implications of telling a narrative of this kind well the serial serial nature of this tale suggests a single developmental thread a kind of comforting form of evolutionary gradualism the tale is essentially one of innovatory highlights resembling a you know welcome and cheerful progress there is a tendency that because something succeeds in time it is as a result of its immediate predecessor precedent that is succession in time may imply causal connection and do not mistake causal connection for the succession type of calls that was what my tutor in philosophy of science told me in Cambridge this is the post hoc ergo propter hoc fallacy this is a fallacy because it because later doesn't necessarily mean it occurred because of what immediately received it so telling the story in a linear way has perhaps an implication that there is a monic causal connection between events need is often posited as historical cause looking backwards through time knowing how it turned out it's quite easily to back project need often needs were fictional the idea is that there was a pre-existing need often a critical need even even when the outcomes are unanticipated and often these needs were created for rhetorical and political purposes Babbage Babbage is known to have created the primary motor for Babbage's machines was supposedly to eliminate errors in printed mathematical tables then the infallibility of machinery would eliminate errors in tables and this is supposedly history tells us is his primary motive his actual motives were and looks at his early writings was in the theoretical potential of these machines not in the utilitarian justification through tables the utilitarian justification the practical use of these machines to make error free table but it almost retrofitted by dinosaurs Lord Lardner a scientific publicist who needed to dramatize the story of Babbage's engines because the mathematics of the engine was actually too difficult for popular audiences and he wrote to Babbage and say how're we gonna sell this thing the mathematics is too difficult tell me about errors so if you like need so the whole story there's a problem problems or errors in tables and the solution is Babbage's engines was actually a piece of rhetoric if as it as it was fictionalized for political purposes that doesn't mean to say it isn't good history this history in fact is very intriguing but the idea that we should not mistake that the idea that all innovation was a satisfactory response to declared need now seeing each innovation as a response to a pre-existing need eliminates the need for further inquiry which means a lot of things we should be looking at don't get looked at because simply we already have an account so many techniques devices and processes were actually met with resistance and when acceptance was the outcome this was sometimes far from immediate and often very hard for the idea of the internal stored-program regarded as the defining feature of the modern computer was met with puzzlement in some quarters when it was first introduced was resisted in other quarters and it's only retrospectively when we actually say what is it that defines the modern computer well it's the internals program that this actually achieves the kind of historical prominence that it has so we should not assume that all new innovation was in fund of replacement technology ously received by these hard press suffering practitioners who were suddenly relieved of owners toil by this by being rescued and rescue narratives are actually quite prominent and they'll see some of those the other features of historical change that don't sit comfortably with this model of relentless determinism there's a tendency to omit long periods of dormancy period where nothing at all happens and I mean it's sort of said in a rather circular way and it's always the news that makes the headlines well when nothing happens you still need a headline and we'll see how this has influenced the construction of our others that the arc of this narrative that I've traced lateral klas cross influence where the explicit or implied and there's a there's a there's the the the the bitter dispute about the priority of ENIAC john mauchly was the co-designer of ENIAC wartime american first tend to use vacuum tubes for large-scale electronic calculation and there was a court case about the priority of these patents because Moakley hadn't had visited John Atanasoff at Iowa State University and see in the ABC computer and the question was was ENIAC based on atomic sauce work was more Klee influenced and the court case was told in 1973 and the evidence is voluminous and conflicting and the court ruled in favor of Atanasoff and invalidated the ENIAC patents so here we have a situation in which cross influence is very difficult to determine historically was he or was he not intimate that the court ruled that he had been influenced that there the seeds of it were actually at tennis orbs but the actual but history that the story that history told is quite unaffected by this court ruling and in fact a lot of issues of influence are actually nuanced in lo the level documentation and so court rulings are one thing and actually what was going on is probably unresolved but they say dispute that split the historical community but most bitterest and unseemly dispute sever that community was split taking one or another side and was a huge huge rift in the in the computing community finally in serial narratives as a tendency to deselect parallel development so when things are going on at the same time you're telling a serial tale there isn't place for everything so you know so at the same time XYZ was going on it actually interferes with the narrative so there's a tendency to channel this narrative in a in a linear and serial way now in music we can follow multiple voices in harmonic sounds visually we can actually apprehend we can register more than one particular area but it seems in language it seems to be a feature of our cognitive apparatus that when we talk about a language which is both written and spoken it needs to be serialized it needs to be broken up and serialized and it could be that the serious serial tale is actually an artifact of the medium chapter sequences need to follow one after the other you can't read two chapters at the same time so that we perhaps shouldn't be too harsh on the popularize of science in looking for a narrative made of headlines but there may be something deeper at play here the linear narrative may in fact be reflecting a presumption conscious or unconscious of technological determinism the idea that the engine of change is technology which then impacts Society as though somehow isolated from it and not influenced by it so the machines make history ours tale bran in a classic paper or does history make machines and we've got a way in which telling a serial narrative may in fact distort that picture so chronology at history make no mistake makes for very dramatic narratives but we can see how it does perhaps a disservice to a richly story ography now if we subtract shift from the stack eccentric or innovation driven model into a user based model then the landscape opens out and we can actually produce a map that can address a more dimension map a more nuanced method can address some of the difficulties that I've outlined so that's what I called the river diagram and that identifies three distinct categories or trajectories of human activities that we put aside for a moment the material culture of these things and just look at the activities going on then we have four threads calculation automatic computation information management and communication and in saw the seventies in this fusion chamber of solid-state physics or chip of integrated circuits they all reemerge and become rebound in a bundle again but if we look at these separate activities and then look at the material culture that's associated with them then actually we can begin to deal with things like cross influence we can begin to now okay so in this scheme this top calculation starts with calculating all the rest and as always through and ends perhaps with an epilogue of the electronic pocket calculator in which for function arithmetic becomes completely trivial to the point in which calculators or giveaways they software add-ons you hit a button on your keyboard you'll get pop-up a pop-up calculate this isn't the primary purpose of the computer so the point is the problem is solved the big struggles with desktop calculation point is that what class what what what distinguishes this category of artifact is that they're portable or semi portable they require human intervention to operate they deliver calculation at the point of need and they're not they're not automatic so that is a is a class of of physical objects and of human activity which is quite distinct which would get lost if we're actually talking about a serial narrative made of headlines of innovation there are a lot of artifacts that actually don't make the ælis to the innovations catalog and I don't mean by the innovations catalog that thing came as a supplement the Sunday newspaper in which you can buy a monogrammed coffee coaster that also acts as a compass I'm talking about if you look at the world first the a lists of innovation there are lot of things that affect it very dramatically the mechanical culture of the day that the physical practice of the tyent which actually don't make the grade and we have a good example here which is the comp tater it's a chain adder it's a beautiful device hold in your hand and what it will do is keep a running total of numbers entered with a stylus on those things you can freeze it by operating the lever on the side which means you can meet the reed intermediate results now this was these were Hugh made in vast numbers huge very widely they don't feature at all in the narrative so their raft of minor devices which affect its practice daily practice which live if you like in the shadow of the celebrity devices which grab the headlines which of the Bruns videos the facets the boroughs machines and so on there many other devices of this kind now if we then look at the next thread which is automatic computation which is the thread I'm trying to isolate which will then give them more particular attention to and the starting point there okay all these are manual so there is a starting point there isn't in technological terms an automatic device prior to Babbage 1821 Babbage is sitting there checking mathematical tables with her show a lot of errors he closed his hand to his head and he says I wish to God these calculations had been executed by steam steam being a metaphor for the infallibility of machinery and also to solve the problem of supply through production so the starting point of what Mary computation is up there 1820s 1821 and it's it's a separate thread but there's a huge gap when Babbage died or really from 1840s 1850s onwards not a lot happens in many computation baby died in 1871 the movement to automate mechanized thought mechanized calculation lost its most vigorous advocate and its most active practitioner and then nothing happened till about the 1930s when Comrie revived some of those techniques and calm he talks about the hundred dark years in the history of computing where nothing happened now linear narratives don't deal well with big gaps so what fills the gap ok so we need something to fill the gap so so this this what time cereal narratives don't do well is deal with issues of that kind information management the practices go back to record-keeping the khipu and the tally sticks are actually not just they know that they've a medium of a record they ways they ways of managing information and have very little computational capability and so there's a fundamental distinction information major the automatic updating it's largely to do with emphasis information management systems are things that have lock vast data throughput relatively moderate and sometimes not exist in computational ability and automatic computation is computation intensive doesn't necessarily had a lot of data so is it justifiable so separate through both in terms of the human activity involved and also in terms of the machinery the material culture that's associated with these communication has its own rich traditions in a slightly out of our remit here but it goes back to Telegraph sir oops semaphore Telegraph's all that sort of stuff and is a very significant part of the merchants in the convergence over the age of electronic information age but we're going to leave that aside now if we look at the standard chapter sequence in the arc that I told you about it started with prehistory with keeping pebbles and all that stuff and then we have a mechanic class and then Babbage is the big thing then we have holler of the straight afterwards because that's 1890s he died days and one 1890s we then get the electronic era sanika bacterium broken down and so on now we can see immediately that this serial narrative has been spliced from segments taken from different threads which have neither traditional connections nor artefactual functional connection so they involve different activities involve different categories of artifacts associated with them so the way we've constructed the narrative is actually take pieces out of each of these which over us was was the innovatory highlight of the time and greatest serial narrative and say that is how if you like under the underlying construction of this now we can overlay time bands on this we can overlay time bands so you can use generation you can have various overlays on this you can look at mechanical you can look a generational technology the mechanical era electromechanical eerie electronic area and so on and so on you can overlay that onto those things and see how these narratives are constructed so we can deal with parallelism the fact that something was going on in automatic computation here doesn't mean to say that the quest for desktop calculators stopped although it may have dropped out of the story were telling and we should also shouldn't be to take these vertical bands to to literally if we look at yes here we have the whole earth tidal island sorta and that's the 1890s cover of Scientific American dealing with the use of the whole earth tablet in 99 so we'll mention that we will come back to that presently and what we want to look at is the curta calculator now that's a handheld thing it's like a little coffee grinder it fits in the palm of your hand and it's actually a circular with ma meter now this thing was induced in the 1940s and was use right through to the 1980s so you can see that it's in terms of its it's a like expecting gravity and utility and practical use it actually spanned several of these generation technologies so again we have an issue where but jointing you involving jump cuts from Babbage to Hollerith for example and there were a batter ending stories which actually are not part of the same tile the extreme example of bridging technologies of a single device which preach all these technologies is the Hibiscus of course this is a this is a version of brought back from Moscow in 1990 I went into a bread shop in Moscow next to the electronic tool there was an abacus and I asked the guide what's going on here and the guide said that country-folk mistrusted the electronic calculator unless the change was worked out on the abacus and there was an argument or or or a dispute and I thought this was quite charming and I took my camera out to shoot this cliched old a new image that is beloved of publicists and marketing people and I thought I would take a photograph of the the Baker in his white outfit and this thing and the shop cleared into takes the shop evaporated I mean there were no customs lived I I save as a bread shot out of kind of social convention there was no bread in the shop they were just empty shelves and there are lots of queues and this was perestroika which which is during the dismantling of the state controls of supply of food and I began to wonder actually whether it was a bread shop at all so where does this leave us in relation to material culture what sustains the threads if you like since I mean you've heard perhaps from the introduction that you know material culture being a curator is actually looking at what can object stell us about history I would say their threads are actually sustained but the categories of artifacts that support them and so and the point about is that an arithmetic doesn't look like a speaking tube and a telegraph doesn't look like a tabulator and a telephone doesn't look like a fax machine so the idea is that something that is functionally distinct is actually physically distinct and that both sustains the the the threads and also allows us to categorize artifacts so I believe that prior to the convergence where all those things converge if you were organizing a collection you were sitting in a store knew how to label the shelves since this communication is to sort of at a computation this is what you you would not have too many anomalies you could resolve it by looking at so the phones don't look are there clearly communication devices and a clearly part of that class of telephones they don't look out Telegraph or we just on telephone so there are differences and the differences are actually that have been respected now I'm going to show you a very alarming diagram whoops which seems to vanish I don't know where it's gone no matter it was a diagram about convergence is very startling diagram and what it shows you shows that threads is a kind of sausage and it goes into kind of fusion chamber and the fusion chamber solid-state physics this is where tens of thousands of transistors be placed on a small little thing with submicron ticket sub micron geometry now to give you some idea of sub micron Beza transistors which are smaller than a micron to give you some idea what a micron is a micron is the length your fingernail grows in about 10 minutes that's how big these little devices were so suddenly you have the opportunity of massive density of functionality functionality as Americans call and in a single device and you could assimilate into a compound device all these multiple functions so you can have an automated computing device a burger all that stuff all in one device so you get this fusion chamber in the 70s which if you like destroys the set of distinct categories that supported our threads and I when I was developing these ideas that was the Computer History Museum in Silicon Valley and I spoke about that the collapse of categories but when she hit this fusion chamber where all these things converge you get the collapse of artefactual categories material culture actually suffers a kind of breakdown at the same time Gordon Bell who is famous for his contribution to the Digital Equipment Corporation and the dominance of the deck mini computers he was there and he was working at the same time he's a trustee of the and very active trustee of the Museum he was working the same time and his work on the death of classes I was talking about the collapse of categories as a way of trying to say to develop a concept diagram for 2,000 years of history of computing we were doing a computer gallery I was the guest curator there how do you tell the tale what do you say about two beside the idea was that you had this fusion chamber after the silicon revolution 70s 80s you had this massive profusion of devices of smartphones work which defied conventional categories up to that point the conventions had been respected all the way back and he was thought about the death of classes on the basis that the many computer ones proliferates devices computing devices anywhere and they ceased to exist and so he was generalizing from many computers and mainframes and saying what is it that defines a class and can we say anything more general in a kind of robustly Universal way about the longevity of classes so something seems to be happening happened there in the 70s convergence death of classes and the collapse of categories so is this the death of the grand narrative if progress is what underpins the ground narrative would say no the smartphone is just another thing in this succession of artifacts if it's a question of material culture then I would say yes it is the death of the grand narrative so what are the core ideas we've now isolated the thread of automatic computation and we can now look at its to think of history uncluttered if you liked by the kind of interference the noise interference from other threads in the way we tell the tale so some of the core ideas in modern computing firstly mechanical process which we're going to talk about a bit more digital logic algorithms systems architecture the internal organization of a machine software and universality and the internal stored program and say those ideas that are sort of picked out of a basket of many which we would say are uniquely or particularly modern they say that you know Turing is associated with many of them so let's look at these in sequence we going to stop the Charles Babbage because that was the start of the automatic computation thread we've said he had his epiphany in 1821 he was going to devote the rest of his life he'd like to mechanizing arithmetic and when he said I wish to God these calculations be next not when he said it he said I wish to God these calculations been executed by Stephen Peter Ackroyd an historian a novelist wrote in in his book the lion lion a scholar man avec novel he said that was one of the most wonderful sentences of the nineteenth century I wish to god these calculations had been executed by steam because what it's capturing is the extension of the industrial metaphor to mental activity from physical production to mental activity and that's what it actually symbolized so the first of his engines was different engine number one this is an assembler demonstration piece assembled in 1832 it's one seventh of the whole engine and so initial values are set by hand on the oops on these figure wheels you crank the handle on the top to crank the handle on the top and the Machine calculates automatically and the result appears on the logs to call them so now the the significance of this machine cannot be overstated this is the first that's probably the most celebrated icon in the prehistory of computing it's the first successful automatic computational device they used to say it's the first device that successfully embodies mathematical ruling mechanism you crank the handle you did not need to understand the principle on which it's based or the mechanism to get useful results you get results which up to that point in time you could only get by thinking you exert physical energy you get results by thinking now the implications of this were not lost on Babbage and his contemporaries lady Byron saw the machine in 1933 and wrote in a diary last week we saw the thinking machine for such it seems heroine what Buxton a junior colleague of Babbage's said that and that Babbage had replaced the marvelous pulp and fiber of the brain he with with brass and steel he Babbage had he Babbage had thought will work to think and he adds or at least to do the office of thought at least in the officer thought so already at that stage they were aware that brains weren't cogwheels that what it is a machine did mechanically may produce the same outcome as a thinking process but it wasn't doing what brains do and we have of course Turing's great thing the idea of mechanical process that what we mean by mechanical in the Turing Newman census its unthinking and we use that now we do something mechanically we do it without thinking and repetitively so mechanical in baby systems was literal it was cogs levers and ratchets and things but there's a much deeper significance which is actually articulated by Turing and a new Newman and that is mechanical that is it removes what what Turing did we know is he watched somebody do a calculation and all he saw was marks on paper and he said if you put a machine to make the same marks on paper by rules then you have mechanical calculation which the point being it does not presuppose higher orders of consciousness of human in order to the calculation which is what he meant by mechanical mechanical process so a mechanical process from so that's the first of how ideas now in broad cultural terms we can see this as the extension of the industrial metaphor and and this was actually evident in language at the time that the industrialization of thought is actually reflected in contemporary language they spoke about a real manufactory of numbers mechanical fabrication of tables and emphatically a machine for manufacturing tables so they were aware both of industrial implications what was happening which was the culture of the times and this reflected in the way they referred to these things now next one is digital what do you mean by digital so yeah we have mechanical process what I mean by digital digital and value design there's no decimal digital machines they're decimal in the sense they used the familiar number system naught to 9 they they're digital in two ways firstly if you want to represent a multi digit number you put you whoops you put units at the bottom tens hundreds thousands of em you've got it so each digit in the number has its own figure wheel so if you want to represent 2.5 you have a figure wheel with five at the bottom and a figure will two above a wheel between 2 & 3 is indeterminate is logically indeterminate it's not 2.5 it's actually a mistake now these figure wheels are inherently analog devices all the transitional states between 2 & 3 are stable they can sit there quite indefinitely so the question is how can they be digital machines in the answer is the security measures that Babbage incorporating they are digital only because Babbage's control mechanisms make them so and I'll show you some examples of the techniques he used this is that's from inside he's first to print the 1832 version and you can see there there is a roller on a sprung lever and you can see these hemispheres which are the kind of teeth we can look at a bit closer these are the hemispheres of kind of teeth now that's a sprung roller as this tries to turn you can see that it's biased so that that's sprung it's trying to press in between the two cheeks all the time and that biases the wheel in favor of one or another now the dwelling angle is actually too shallow for it not to actually purchase on top but if that was a triangle you'd have the equivalent of a modern flip-flop something which is ice table one side and stable the other side but the transitional values or not and so this is this is 1820s as you've mid-1820s design you can see that the attempt digitised to discretize the emotion to make it incremental in fixed jumps is absolutely completely evident today later use more sophisticated devices he used a wedge if you imagine that as looking down on a column of figure wheels there's a wedge which is like a sword blade which runs the entire length of the columns and that goes between the teeth of the thing which stops it now if a tooth D ranges that the lock which is what he called them will hit the end of the tooth and the machine will jam so the jamming is not a calamity the calamity it would appear to be the jamming actually is error detection it says a wheel is an indeterminate indeterminate position that the integrity of the calculation has been compromised and those locks performs three things as they enter that collects more they make small error Corrections if they're locked in place all the time they prevent the wheels D ranging during periods which they're not needed and finally their forward error detection that it is a there's a half a tooth derangement then the machine will jam and that you know and Babbage boasts he says he says that the machine will jam or break but never deceive that was his confidence in his mechanisms that is actually a video of the lock working but I don't think this time for it you can see the lock there is that vertical saw blade that's the wedge-shaped edge and it's going into a series of racks which are triangular which act more like a flip-flop the error correction as it goes in it correct small derangement is like pulse shaping in modern electronics it cleans up a jagged edge of a degraded pulse so we've got mechanical process we've got digital now we need to go to the next one which is from calculation to computation the difference engine is what we'd call a calculator it only had a specific set of functions and it crunched numbers the only way of knew how called you the method of finite there whatever initial values are put in your crank down it'll do the same thing to them you couldn't alter what it did to the numbers to get general-purpose computation we need to be able to alter the rules by which the the information is manipulated and the next leap actually that's when Babbage said it would Jam break whenever you see he was right but that's different engine number two built never booked in Babbage's died completely 2002 at the Science Museum that's a thousand parts five tons and it works that's the back of it's not because I have time but because I couldn't resist it okay so it's the analytical engine conceived in 1833 that actually marks the transition from calculation to computation to general-purpose computation and that's the next step in automation digital automation is the one step the next step is to actually do to generalize this and here we have and those are right one of the point about the analytical engine one things it was programmable using punch cards and there were various kinds of punch cards one set of pankot was an operations card which holds the instructions and other kind of card was variable chord which tells you where the information is to be placed into memory and those are two examples we have an example of operation card that is a set of just under 30 cards which is solves some tiniest differential equations what signal does simultaneous equations and those are variable cards some of these cards are 26 inches in length but they're quite big things so here we have the algorithm has now been transferred to the operations cards the machine operates without human intervention it's automatic and it is general-purpose in that it's programmable that is a 1840s plan playing twenty five of Babbage's of the analytical engine and just to give you some idea of physical scale so Babbage bursts out of nowhere like a historical jack-in-the-box and confounds us with with with schemes of a logical conception and physical scale that were unprecedented the diameter of those central wheels so if you imagine each of these circles is a column of wheels 3250 deep so we're looking at it from the top and those go into the page these the diameter of that is intended five to six feet 15 feet high the diamond across there is eight feet that lots 15 foot high that is the memory over there this is the central processor over here the memories alone a memory can stand indefinitely there are 17 registers these things labeled V which he called variables which is not what we mean by variable he meant that the content to the register could change its number the variables there they 17 there and this machine is about 17 foot long he spoke about machines a minimum machine an entry-level machine would have 100 variables and he spoke with machines with a thousand variables he talked about something the length of this room so the conception in physical scale was massive the logical conception as we can see was quite extraordinary so information groups information gets taken from the stall under program control sent to the central processor via buffer registers over there that gets processed and sent back to the memory so there's a fetch execute cycle there's a separation of memory and central processor and those are just three of the of the of the features of it it has an internal repet of instruction so it can do four function arithmetic automatically multiplication division or dependent so long so it's embodies a set of logical and architectural features that are we would now recognize as part of modern computing just to give some idea I'll run through it they may not mean a whole lot to to some they may mean more to others it was programmable using punch cards it was it had microprogramming automatic execution of a large an instruction by the automatic execution of small instructions separation of store and low will be spoken about conditional control if then it could take a one or a number of roots you know in a calculation depending on the result of another calculations and so on and so on a parallel processing at a parallel data bus so all these features are explicit these aren't wishful things these on backwards projection for own time this isn't the coded vague vagueness of Nostradamus these are in there and these are very specific design features and we can talk about jockeyed you got the idea of punch cars from jacquard from from the jacquard loom we see a jacquard loom yeah and that is the famous sort of jacquard woven in silk it took 24,000 cards to make it is so finds often mistaken as an etching and this has both cultural significance as well as technical significance it was used by advocates for industrial for the industrial arts to show that that subtlety aesthetic subtlety was not the exclusive province of the human agent the translucency of the of the drapes up there and that shading the subtle shading was shown to be was particular was used as an attempt to win acceptance for objects of industrial manufacture mm-hmm and so one of the points about jacquard it wasn't just okay what the choco glue allowed you to do was to weave textiles with a pattern that is determined by the whole pattern on a bunch of punch cards so the machines read the punch cards now the point about this is this is a general machine for weaving it doesn't weave anything specific what is product specific about it is actually in the software what is product specific so you have the idea of a generalized machine and what is functionally specific is not in the hardware but actually in the software in the cards so the room is just much more than just a staging post for Babbage getting the idea about cards it actually has kind of generic universality the idea of a universal machine where the functional specificity is provided from outside the machine which is why this implications of the internal stored-program are so profound so here it's all outside the machine the functional specific facilities outside it so we've got almost all of our core ideas except one and that is we haven't there's a final step to go from where Babbage left things into modern computing what is music but number closed in sound so the question now is Babbage saw his machines as a new technology of mathematics he never saw his machines as anything other than bounded by a number this is evident from both all his writings and a smoking gun quote which I found in his notebooks recently which shows even late in life two years before he died when he sets out to describe to the world his analytical engine what he says its objects are or all mathematical realism and they are all mathematical so how do we get from a mathematical machine to a general-purpose computer and the step there is the realization that number can represent entity other than quantity that number can represent a note of the note the other a note of music or letter of the alphabet and that the power of computing comes from the representational power of number so if we've got a machine a generalized machine that can manipulate symbols according to rules and the representational power of those symbols is to take aspects of the world then we can manipulate those symbols and representations and produce knowledge of the world which was possibly not accessible by other means and that is where the power of the machine comes the generality of the general-purpose computer which is that the machine from impolitely symbols according to rules comes from that representation of power of number and the ability for the machine to manipulate a symbol now for Babbage the machine was always manipulating a number and the number didn't represent a symbol it always represented a quantity and that leap was made not by Babbage but by Ada Lovelace who was a met Babbage in 18 when she was 17 in 1833 and then subsequently in 1843 published a sketch of the analytical engine her only publication in which is the most penetrating account of the analytical engine in contemporary times of the time and it's still studied to actually figure out what it is they did or didn't understand but this is but this but the idea that number can represent that is the alphabet notes of music is actually exist is explicitly articulated in Lovelace's work and this is not Lovelace and making a suggestion this is Lovelace banging the table saying it is thus very significant about computers and for everything else that has been said about her she actually deserves recognition for this she saw in a sense the implications she may be an entirely mistaken I believe actually that it may mean she actually misunderstood what was going on in the analytical engine that she believed it was capable of symbolic manipulation that it could do algebra maybe she didn't think if you do algebra algebra mean can manipulate symbols regardless of its value and produce results in algebraic notation now Lovelace clearly believed that the analytical engine was capable of doing this it's not evident that it was so it may be that you spoke at the point is she made this he managed to bleep and said actually what's significant about a computer is its ability to manipulate symbols according to rule and the power of the computer to our lives is the representational power of number that allows the machine to say things about the world which then get mapped back so whether or not the grammar of the rules is the same as the grammar of the phenomena being modeled doesn't really matter in many cases they are they are mathematical models of physical processes and they've match there is exact but you have the opportunity of simulation and modeling which is the exploration of things in the world through the representation of power remember right the elephant in the room of course is assuming you would accept that these core ideas became evident in the 19th century which is 100 years earlier than is commonly credited so we associate those ideas that six core ideas with modern computation and what I've tried to show is that actually they were all made explicit in the 19th century almost without exception internal stored-program was not and so the elephant in the room is to what extent did this influence modern computing the implications are that had very little influence that the principles of computing these core ideas amongst them were reinvented by the pioneers of electronic computing largely in ignorance is what gone before so that seems to be the conclusion they are sort of weak threads but the the early designs cannot be held the 19th century mechanics could not be held to be the DNA of modern computing that level of detail wasn't known so what we end up with is a very startling congruence if it is the case that there was no influence which seems very likely what we end up is at the rod extraordinary congruence between the ideas of the 19th century and the ideas of the modern computing age and we are left with a conclusion both reassuring and confining that these ideas articulate a century apart embody something fundamental about the nature of computing thank you you

5 Comments

  1. hyphen point said:

    Astonishing with how little our ancients were able to work.

    June 26, 2019
    Reply
  2. Vin D said:

    In general, there are lots of exceptional people with the aptitude to do what other brilliant people have done, but without acknowledgment. Wthout their input, we would neither be here to view, nor comment, on this video about the history of computing. I'm grateful for their innovations.

    June 26, 2019
    Reply
  3. jiten patel said:

    Hay  you  Chiman  and  girish  patel  watch this video..The  great  person  in  the  world  and  I  Thank to  him…jzpatelut..28th.mar.2015 SATURDAY 17:35 pm..jzpatelut..

    June 26, 2019
    Reply
  4. Liz McLean-Knight said:

    38:43 "Jamming [of a machine] actually is error detection. It says the wheel is in a logically indeterminate position–the integrity of the calculation has been compromised." That's a rather clever perspective.

    June 26, 2019
    Reply
  5. Love & Division said:

    Interesting how the history of computing more closely follows intelligent design than evolution.

    June 26, 2019
    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *