TEDxImperialCollege
John Graham-Cumming: The greatest machine that never was
Filmed:
Readability: 3.9
1,242,570 views
Computer science began in the '30s ... the 1830s. John Graham-Cumming tells the story of Charles Babbage's mechanical, steam-powered "analytical engine" and how Ada Lovelace, mathematician and daughter of Lord Byron, saw beyond its simple computational abilities to imagine the future of computers.
John Graham-Cumming - Computer programmer
Computer programmer John Graham-Cumming keeps geek history alive by raising awareness for it's heroes and their inventions. Full bio
Computer programmer John Graham-Cumming keeps geek history alive by raising awareness for it's heroes and their inventions. Full bio
Double-click the English transcript below to play the video.
00:16
So the machine I'm going to talk you about
0
106
1424
00:17
is what I call the greatest machine that never was.
1
1530
1674
00:19
It was a machine that was never built,
2
3204
1999
00:21
and yet, it will be built.
3
5203
2380
00:23
It was a machine that was designed
4
7583
1992
00:25
long before anyone thought about computers.
5
9575
2351
00:27
If you know anything about the history of computers,
6
11926
2149
00:29
you will know that in the '30s and the '40s,
7
14075
2520
00:32
simple computers were created
8
16595
2555
00:35
that started the computer revolution we have today,
9
19150
2793
00:37
and you would be correct,
10
21943
1286
00:39
except for you'd have the wrong century.
11
23229
2482
00:41
The first computer was really designed
12
25711
1640
00:43
in the 1830s and 1840s, not the 1930s and 1940s.
13
27351
3672
00:46
It was designed, and parts of it were prototyped,
14
31023
2328
00:49
and the bits of it that were built are here
15
33351
2136
00:51
in South Kensington.
16
35487
1993
00:53
That machine was built by this guy, Charles Babbage.
17
37480
3319
00:56
Now, I have a great affinity for Charles Babbage
18
40799
1756
00:58
because his hair is always completely unkempt like this
19
42555
2609
01:01
in every single picture. (Laughter)
20
45164
2341
01:03
He was a very wealthy man, and a sort of,
21
47505
1847
01:05
part of the aristocracy of Britain,
22
49352
2013
01:07
and on a Saturday night in Marylebone,
23
51365
2339
01:09
were you part of the intelligentsia of that period,
24
53704
2431
01:12
you would have been invited round to his house
25
56135
1496
01:13
for a soiree — and he invited everybody:
26
57631
2959
01:16
kings, the Duke of Wellington, many, many famous people —
27
60590
3432
01:19
and he would have shown you one of his mechanical machines.
28
64022
2521
01:22
I really miss that era, you know, where you could
29
66543
3040
01:25
go around for a soiree and see a mechanical computer
30
69583
2348
01:27
get demonstrated to you. (Laughter)
31
71931
1113
01:28
But Babbage, Babbage himself was born
32
73044
3266
01:32
at the end of the 18th century,
33
76310
1734
01:33
and was a fairly famous mathematician.
34
78044
2044
01:35
He held the post that Newton held at Cambridge,
35
80088
3143
01:39
and that was recently held by Stephen Hawking.
36
83231
2512
01:41
He's less well known than either of them because
37
85743
2855
01:44
he got this idea to make mechanical computing devices
38
88598
3150
01:47
and never made any of them.
39
91748
2285
01:49
The reason he never made any of them, he's a classic nerd.
40
94033
3285
01:53
Every time he had a good idea, he'd think,
41
97318
2011
01:55
"That's brilliant, I'm going to start building that one.
42
99329
1448
01:56
I'll spend a fortune on it. I've got a better idea.
43
100777
2595
01:59
I'm going to work on this one. (Laughter) And I'm going to do this one."
44
103372
2360
02:01
He did this until Sir Robert Peel, then Prime Minister,
45
105732
2880
02:04
basically kicked him out of Number 10 Downing Street,
46
108612
2464
02:06
and kicking him out, in those days, that meant saying,
47
111076
2441
02:09
"I bid you good day, sir." (Laughter)
48
113517
3152
02:12
The thing he designed was this monstrosity here,
49
116669
1980
02:14
the analytical engine. Now, just to give you an idea of this,
50
118649
3439
02:17
this is a view from above.
51
122088
1872
02:19
Every one of these circles is a cog, a stack of cogs,
52
123960
3113
02:22
and this thing is as big as a steam locomotive.
53
127073
3236
02:26
So as I go through this talk, I want you to imagine
54
130309
1991
02:28
this gigantic machine. We heard those wonderful sounds
55
132300
2644
02:30
of what this thing would have sounded like.
56
134944
1760
02:32
And I'm going to take you through the architecture of the machine
57
136704
1777
02:34
— that's why it's computer architecture —
58
138481
1479
02:35
and tell you about this machine, which is a computer.
59
139960
3327
02:39
So let's talk about the memory. The memory
60
143287
3403
02:42
is very like the memory of a computer today,
61
146690
2247
02:44
except it was all made out of metal,
62
148937
2702
02:47
stacks and stacks of cogs, 30 cogs high.
63
151639
3544
02:51
Imagine a thing this high of cogs,
64
155183
2070
02:53
hundreds and hundreds of them,
65
157253
1755
02:54
and they've got numbers on them.
66
159008
1890
02:56
It's a decimal machine. Everything's done in decimal.
67
160898
2419
02:59
And he thought about using binary. The problem
68
163317
1585
03:00
with using binary is that the machine would have been so
69
164902
1718
03:02
tall, it would have been ridiculous. As it is, it's enormous.
70
166620
3317
03:05
So he's got memory.
71
169937
2059
03:07
The memory is this bit over here.
72
171996
2407
03:10
You see it all like this.
73
174403
2330
03:12
This monstrosity over here is the CPU, the chip, if you like.
74
176733
4535
03:17
Of course, it's this big.
75
181268
2250
03:19
Completely mechanical. This whole machine is mechanical.
76
183518
2913
03:22
This is a picture of a prototype for part of the CPU
77
186431
4141
03:26
which is in the Science Museum.
78
190572
2139
03:28
The CPU could do the four fundamental functions of arithmetic --
79
192711
3632
03:32
so addition, multiplication, subtraction, division --
80
196343
2453
03:34
which already is a bit of a feat in metal,
81
198796
3008
03:37
but it could also do something that a computer does
82
201804
2629
03:40
and a calculator doesn't:
83
204433
1699
03:42
this machine could look at its own internal memory and make a decision.
84
206132
3938
03:45
It could do the "if then" for basic programmers,
85
210070
2866
03:48
and that fundamentally made it into a computer.
86
212936
2140
03:50
It could compute. It couldn't just calculate. It could do more.
87
215076
4598
03:55
Now, if we look at this, and we stop for a minute,
88
219674
2681
03:58
and we think about chips today, we can't
89
222355
1871
04:00
look inside a silicon chip. It's just so tiny.
90
224226
3815
04:03
Yet if you did, you would see something
91
228041
1801
04:05
very, very similar to this.
92
229842
1822
04:07
There's this incredible complexity in the CPU,
93
231664
2947
04:10
and this incredible regularity in the memory.
94
234611
2692
04:13
If you've ever seen an electron microscope picture,
95
237303
1662
04:14
you'll see this. This all looks the same,
96
238965
1969
04:16
then there's this bit over here which is incredibly complicated.
97
240934
2566
04:19
All this cog wheel mechanism here is doing is what a computer does,
98
243500
3983
04:23
but of course you need to program this thing, and of course,
99
247483
2093
04:25
Babbage used the technology of the day
100
249576
3025
04:28
and the technology that would reappear in the '50s, '60s and '70s,
101
252601
3646
04:32
which is punch cards. This thing over here
102
256247
2869
04:35
is one of three punch card readers in here,
103
259116
2824
04:37
and this is a program in the Science Museum, just
104
261940
3680
04:41
not far from here, created by Charles Babbage,
105
265620
4393
04:45
that is sitting there — you can go see it —
106
270013
1868
04:47
waiting for the machine to be built.
107
271881
2441
04:50
And there's not just one of these, there's many of them.
108
274322
3420
04:53
He prepared programs anticipating this would happen.
109
277742
3090
04:56
Now, the reason they used punch cards was that Jacquard,
110
280832
1973
04:58
in France, had created the Jacquard loom,
111
282805
2172
05:00
which was weaving these incredible patterns controlled by punch cards,
112
284977
2678
05:03
so he was just repurposing the technology of the day,
113
287655
2632
05:06
and like everything else he did, he's using the technology
114
290287
2105
05:08
of his era, so 1830s, 1840s, 1850s, cogs, steam,
115
292392
4747
05:13
mechanical devices. Ironically, born the same year
116
297139
3938
05:16
as Charles Babbage was Michael Faraday,
117
301077
2172
05:19
who would completely revolutionize everything
118
303249
2677
05:21
with the dynamo, transformers, all these sorts of things.
119
305926
2513
05:24
Babbage, of course, wanted to use proven technology,
120
308439
3158
05:27
so steam and things.
121
311597
1553
05:29
Now, he needed accessories.
122
313150
1673
05:30
Obviously, you've got a computer now.
123
314823
1672
05:32
You've got punch cards, a CPU and memory.
124
316495
2389
05:34
You need accessories you're going to come with.
125
318884
1935
05:36
You're not just going to have that,
126
320819
1628
05:38
So, first of all, you had sound. You had a bell,
127
322447
2828
05:41
so if anything went wrong — (Laughter) —
128
325275
2154
05:43
or the machine needed the attendant to come to it,
129
327429
2345
05:45
there was a bell it could ring. (Laughter)
130
329774
1970
05:47
And there's actually an instruction on the punch card
131
331744
1536
05:49
which says "Ring the bell." So you can imagine this "Ting!"
132
333280
2902
05:52
You know, just stop for a moment, imagine all those noises,
133
336182
2200
05:54
this thing, "Click, clack click click click,"
134
338382
1081
05:55
steam engine, "Ding," right? (Laughter)
135
339463
2937
05:58
You also need a printer, obviously, and everyone needs a printer.
136
342400
2435
06:00
This is actually a picture of the printing mechanism for
137
344835
3008
06:03
another machine of his, called the Difference Engine No. 2,
138
347843
2483
06:06
which he never built, but which the Science Museum
139
350326
1935
06:08
did build in the '80s and '90s.
140
352261
2171
06:10
It's completely mechanical, again, a printer.
141
354432
2275
06:12
It prints just numbers, because he was obsessed with numbers,
142
356707
2698
06:15
but it does print onto paper, and it even does word wrapping,
143
359405
3517
06:18
so if you get to the end of the line, it goes around like that.
144
362922
2772
06:21
You also need graphics, right?
145
365694
1650
06:23
I mean, if you're going to do anything with graphics,
146
367344
1556
06:24
so he said, "Well, I need a plotter. I've got a big piece of paper
147
368900
2596
06:27
and an ink pen and I'll make it plot."
148
371496
2108
06:29
So he designed a plotter as well,
149
373604
1830
06:31
and, you know, at that point, I think he got pretty much
150
375434
3925
06:35
a pretty good machine.
151
379359
1531
06:36
Along comes this woman, Ada Lovelace.
152
380890
2690
06:39
Now, imagine these soirees, all these great and good comes along.
153
383580
2721
06:42
This lady is the daughter of the mad, bad
154
386301
3092
06:45
and dangerous-to-know Lord Byron,
155
389393
2422
06:47
and her mother, being a bit worried that she might have
156
391815
2520
06:50
inherited some of Lord Byron's madness and badness,
157
394335
2857
06:53
thought, "I know the solution: Mathematics is the solution.
158
397192
3238
06:56
We'll teach her mathematics. That'll calm her down."
159
400430
2949
06:59
(Laughter) Because of course,
160
403379
3756
07:03
there's never been a mathematician that's gone crazy,
161
407135
3915
07:06
so, you know, that'll be fine. (Laughter)
162
411050
2401
07:09
Everything'll be fine. So she's got this mathematical training,
163
413451
3338
07:12
and she goes to one of these soirees with her mother,
164
416789
2738
07:15
and Charles Babbage, you know, gets out his machine.
165
419527
2790
07:18
The Duke of Wellington is there, you know,
166
422317
1834
07:20
get out the machine, obviously demonstrates it,
167
424151
1572
07:21
and she gets it. She's the only person in his lifetime, really,
168
425723
3751
07:25
who said, "I understand what this does,
169
429474
1292
07:26
and I understand the future of this machine."
170
430766
2207
07:28
And we owe to her an enormous amount because we know
171
432973
3087
07:31
a lot about the machine that Babbage was intending to build
172
436060
2977
07:34
because of her.
173
439037
1603
07:36
Now, some people call her the first programmer.
174
440640
2757
07:39
This is actually from one of -- the paper that she translated.
175
443397
3386
07:42
This is a program written in a particular style.
176
446783
2911
07:45
It's not, historically, totally accurate that she's the first programmer,
177
449694
3569
07:49
and actually, she did something more amazing.
178
453263
2053
07:51
Rather than just being a programmer,
179
455316
1570
07:52
she saw something that Babbage didn't.
180
456886
2172
07:54
Babbage was totally obsessed with mathematics.
181
459058
3184
07:58
He was building a machine to do mathematics,
182
462242
3949
08:02
and Lovelace said, "You could do more than mathematics
183
466191
3259
08:05
on this machine." And just as you do,
184
469450
2835
08:08
everyone in this room already's got a computer on them
185
472285
1625
08:09
right now, because they've got a phone.
186
473910
2054
08:11
If you go into that phone, every single thing in that phone
187
475964
2228
08:14
or computer or any other computing device
188
478192
2015
08:16
is mathematics. It's all numbers at the bottom.
189
480207
2081
08:18
Whether it's video or text or music or voice, it's all numbers,
190
482288
4693
08:22
it's all, underlying it, mathematical functions happening,
191
486981
3980
08:26
and Lovelace said, "Just because you're doing
192
490961
2105
08:28
mathematical functions and symbols
193
493066
3322
08:32
doesn't mean these things can't represent
194
496388
2247
08:34
other things in the real world, such as music."
195
498635
3353
08:37
This was a huge leap, because Babbage is there saying,
196
501988
2734
08:40
"We could compute these amazing functions and print out
197
504722
2222
08:42
tables of numbers and draw graphs," — (Laughter) —
198
506944
3668
08:46
and Lovelace is there and she says, "Look,
199
510612
1896
08:48
this thing could even compose music if you
200
512508
2476
08:50
told it a representation of music numerically."
201
514984
3548
08:54
So this is what I call Lovelace's Leap.
202
518532
1569
08:56
When you say she's a programmer, she did do some,
203
520101
3737
08:59
but the real thing is to have said the future is going to be
204
523838
3137
09:02
much, much more than this.
205
526975
2196
09:05
Now, a hundred years later, this guy comes along,
206
529171
2179
09:07
Alan Turing, and in 1936, and invents the computer all over again.
207
531350
5453
09:12
Now, of course, Babbage's machine was entirely mechanical.
208
536803
2577
09:15
Turing's machine was entirely theoretical.
209
539380
2531
09:17
Both of these guys were coming from a mathematical perspective,
210
541911
2791
09:20
but Turing told us something very important.
211
544702
2553
09:23
He laid down the mathematical foundations
212
547255
2935
09:26
for computer science, and said,
213
550190
2113
09:28
"It doesn't matter how you make a computer."
214
552303
3187
09:31
It doesn't matter if your computer's mechanical,
215
555490
1878
09:33
like Babbage's was, or electronic, like computers are today,
216
557368
4410
09:37
or perhaps in the future, cells, or, again,
217
561778
2804
09:40
mechanical again, once we get into nanotechnology.
218
564582
3146
09:43
We could go back to Babbage's machine
219
567728
2037
09:45
and just make it tiny. All those things are computers.
220
569765
2576
09:48
There is in a sense a computing essence.
221
572341
1632
09:49
This is called the Church–Turing thesis.
222
573973
2005
09:51
And so suddenly, you get this link where you say
223
575978
2667
09:54
this thing Babbage had built really was a computer.
224
578645
2223
09:56
In fact, it was capable of doing everything we do today
225
580868
2825
09:59
with computers, only really slowly. (Laughter)
226
583693
4832
10:04
To give you an idea of how slowly,
227
588525
2106
10:06
it had about 1k of memory.
228
590631
3839
10:10
It used punch cards, which were being fed in,
229
594470
2918
10:13
and it ran about 10,000 times slower the first ZX81.
230
597388
5760
10:19
It did have a RAM pack.
231
603148
1603
10:20
You could add on a lot of extra memory if you wanted to.
232
604751
3179
10:23
(Laughter) So, where does that bring us today?
233
607930
2326
10:26
So there are plans.
234
610256
1608
10:27
Over in Swindon, the Science Museum archives,
235
611864
2933
10:30
there are hundreds of plans and thousands of pages
236
614797
1694
10:32
of notes written by Charles Babbage about this analytical engine.
237
616491
3469
10:35
One of those is a set of plans that we call Plan 28,
238
619960
3961
10:39
and that is also the name of a charity that I started
239
623921
2154
10:41
with Doron Swade, who was the curator of computing
240
626075
2734
10:44
at the Science Museum, and also the person who drove
241
628809
2239
10:46
the project to build a difference engine,
242
631048
1430
10:48
and our plan is to build it.
243
632478
2558
10:50
Here in South Kensington, we will build the analytical engine.
244
635036
3866
10:54
The project has a number of parts to it.
245
638902
2002
10:56
One was the scanning of Babbage's archive.
246
640904
2520
10:59
That's been done. The second is now the study
247
643424
1922
11:01
of all of those plans to determine what to build.
248
645346
3110
11:04
The third part is a computer simulation of that machine,
249
648456
4444
11:08
and the last part is to physically build it at the Science Museum.
250
652900
2923
11:11
When it's built, you'll finally be able to understand how a computer works,
251
655823
2576
11:14
because rather than having a tiny chip in front of you,
252
658399
1834
11:16
you've got to look at this humongous thing and say, "Ah,
253
660233
2610
11:18
I see the memory operating, I see the CPU operating,
254
662843
3335
11:22
I hear it operating. I probably smell it operating." (Laughter)
255
666178
3867
11:25
But in between that we're going to do a simulation.
256
670045
2599
11:28
Babbage himself wrote, he said,
257
672644
1757
11:30
as soon as the analytical engine exists,
258
674401
1618
11:31
it will surely guide the future course of science.
259
676019
3704
11:35
Of course, he never built it, because he was always fiddling
260
679723
1845
11:37
with new plans, but when it did get built, of course,
261
681568
2082
11:39
in the 1940s, everything changed.
262
683650
3510
11:43
Now, I'll just give you a little taste of what it looks like
263
687160
1823
11:44
in motion with a video which shows
264
688983
2633
11:47
just one part of the CPU mechanism working.
265
691616
4594
11:55
So this is just three sets of cogs,
266
699210
2999
11:58
and it's going to add. This is the adding mechanism
267
702209
2822
12:00
in action, so you imagine this gigantic machine.
268
705031
2657
12:03
So, give me five years.
269
707688
1159
12:04
Before the 2030s happen, we'll have it.
270
708847
2344
12:07
Thank you very much. (Applause)
271
711191
2970
ABOUT THE SPEAKER
John Graham-Cumming - Computer programmerComputer programmer John Graham-Cumming keeps geek history alive by raising awareness for it's heroes and their inventions.
Why you should listen
John Graham-Cumming received his Ph.D. in computer security from Oxford University. A peripatetic programmer, he has worked in Silicon Valley, New York, the UK, Germany and France. In 2004, his open source POPFile program won a Jolt Productivity Award. He has authored two books: The Geek Atlas, a travel companion for math and science history enthusiasts; and GNU Make Unleashed, a guide for programmers. In 2009, he spearheaded a successful petition to the British Government demanding an apology for its persecution of Alan Turing for his homosexuality. His personal blog was rated among the "top 30 science blogs" by The Times in February 2010. And in October, 2010, he founded an organization dedicated to build Charles Babbage's famous analytical engine.
John Graham-Cumming | Speaker | TED.com