TEDGlobal 2007
Ron Eglash: The fractals at the heart of African designs
Filmed:
Readability: 4.3
1,740,687 views
'I am a mathematician, and I would like to stand on your roof.' That is how Ron Eglash greeted many African families he met while researching the fractal patterns he'd noticed in villages across the continent.
Ron Eglash - Mathematician
Ron Eglash is an ethno-mathematician: he studies the way math and cultures intersect. He has shown that many aspects of African design -- in architecture, art, even hair braiding -- are based on perfect fractal patterns. Full bio
Ron Eglash is an ethno-mathematician: he studies the way math and cultures intersect. He has shown that many aspects of African design -- in architecture, art, even hair braiding -- are based on perfect fractal patterns. Full bio
Double-click the English transcript below to play the video.
00:13
I want to start my story in Germany, in 1877,
0
1000
3000
00:16
with a mathematician named Georg Cantor.
1
4000
2000
00:18
And Cantor decided he was going to take a line and erase the middle third of the line,
2
6000
5000
00:23
and then take those two resulting lines and bring them back into the same process, a recursive process.
3
11000
5000
00:28
So he starts out with one line, and then two,
4
16000
2000
00:30
and then four, and then 16, and so on.
5
18000
3000
00:33
And if he does this an infinite number of times, which you can do in mathematics,
6
21000
3000
00:36
he ends up with an infinite number of lines,
7
24000
2000
00:38
each of which has an infinite number of points in it.
8
26000
3000
00:41
So he realized he had a set whose number of elements was larger than infinity.
9
29000
4000
00:45
And this blew his mind. Literally. He checked into a sanitarium. (Laughter)
10
33000
3000
00:48
And when he came out of the sanitarium,
11
36000
2000
00:50
he was convinced that he had been put on earth to found transfinite set theory
12
38000
6000
00:56
because the largest set of infinity would be God Himself.
13
44000
3000
00:59
He was a very religious man.
14
47000
1000
01:00
He was a mathematician on a mission.
15
48000
2000
01:02
And other mathematicians did the same sort of thing.
16
50000
2000
01:04
A Swedish mathematician, von Koch,
17
52000
2000
01:06
decided that instead of subtracting lines, he would add them.
18
54000
4000
01:10
And so he came up with this beautiful curve.
19
58000
2000
01:12
And there's no particular reason why we have to start with this seed shape;
20
60000
3000
01:15
we can use any seed shape we like.
21
63000
4000
01:19
And I'll rearrange this and I'll stick this somewhere -- down there, OK --
22
67000
4000
01:23
and now upon iteration, that seed shape sort of unfolds into a very different looking structure.
23
71000
7000
01:30
So these all have the property of self-similarity:
24
78000
2000
01:32
the part looks like the whole.
25
80000
2000
01:34
It's the same pattern at many different scales.
26
82000
2000
01:37
Now, mathematicians thought this was very strange
27
85000
2000
01:39
because as you shrink a ruler down, you measure a longer and longer length.
28
87000
5000
01:44
And since they went through the iterations an infinite number of times,
29
92000
2000
01:46
as the ruler shrinks down to infinity, the length goes to infinity.
30
94000
6000
01:52
This made no sense at all,
31
100000
1000
01:53
so they consigned these curves to the back of the math books.
32
101000
3000
01:56
They said these are pathological curves, and we don't have to discuss them.
33
104000
4000
02:00
(Laughter)
34
108000
1000
02:01
And that worked for a hundred years.
35
109000
2000
02:04
And then in 1977, Benoit Mandelbrot, a French mathematician,
36
112000
5000
02:09
realized that if you do computer graphics and used these shapes he called fractals,
37
117000
5000
02:14
you get the shapes of nature.
38
122000
2000
02:16
You get the human lungs, you get acacia trees, you get ferns,
39
124000
4000
02:20
you get these beautiful natural forms.
40
128000
2000
02:22
If you take your thumb and your index finger and look right where they meet --
41
130000
4000
02:26
go ahead and do that now --
42
134000
2000
02:28
-- and relax your hand, you'll see a crinkle,
43
136000
3000
02:31
and then a wrinkle within the crinkle, and a crinkle within the wrinkle. Right?
44
139000
3000
02:34
Your body is covered with fractals.
45
142000
2000
02:36
The mathematicians who were saying these were pathologically useless shapes?
46
144000
3000
02:39
They were breathing those words with fractal lungs.
47
147000
2000
02:41
It's very ironic. And I'll show you a little natural recursion here.
48
149000
4000
02:45
Again, we just take these lines and recursively replace them with the whole shape.
49
153000
5000
02:50
So here's the second iteration, and the third, fourth and so on.
50
158000
5000
02:55
So nature has this self-similar structure.
51
163000
2000
02:57
Nature uses self-organizing systems.
52
165000
2000
02:59
Now in the 1980s, I happened to notice
53
167000
3000
03:02
that if you look at an aerial photograph of an African village, you see fractals.
54
170000
4000
03:06
And I thought, "This is fabulous! I wonder why?"
55
174000
4000
03:10
And of course I had to go to Africa and ask folks why.
56
178000
2000
03:12
So I got a Fulbright scholarship to just travel around Africa for a year
57
180000
6000
03:18
asking people why they were building fractals,
58
186000
2000
03:20
which is a great job if you can get it.
59
188000
2000
03:22
(Laughter)
60
190000
1000
03:23
And so I finally got to this city, and I'd done a little fractal model for the city
61
191000
7000
03:30
just to see how it would sort of unfold --
62
198000
3000
03:33
but when I got there, I got to the palace of the chief,
63
201000
3000
03:36
and my French is not very good; I said something like,
64
204000
3000
03:39
"I am a mathematician and I would like to stand on your roof."
65
207000
3000
03:42
But he was really cool about it, and he took me up there,
66
210000
3000
03:45
and we talked about fractals.
67
213000
1000
03:46
And he said, "Oh yeah, yeah! We knew about a rectangle within a rectangle,
68
214000
3000
03:49
we know all about that."
69
217000
2000
03:51
And it turns out the royal insignia has a rectangle within a rectangle within a rectangle,
70
219000
4000
03:55
and the path through that palace is actually this spiral here.
71
223000
4000
03:59
And as you go through the path, you have to get more and more polite.
72
227000
4000
04:03
So they're mapping the social scaling onto the geometric scaling;
73
231000
3000
04:06
it's a conscious pattern. It is not unconscious like a termite mound fractal.
74
234000
5000
04:11
This is a village in southern Zambia.
75
239000
2000
04:13
The Ba-ila built this village about 400 meters in diameter.
76
241000
4000
04:17
You have a huge ring.
77
245000
2000
04:19
The rings that represent the family enclosures get larger and larger as you go towards the back,
78
247000
6000
04:26
and then you have the chief's ring here towards the back
79
254000
4000
04:30
and then the chief's immediate family in that ring.
80
258000
3000
04:33
So here's a little fractal model for it.
81
261000
1000
04:34
Here's one house with the sacred altar,
82
262000
3000
04:37
here's the house of houses, the family enclosure,
83
265000
3000
04:40
with the humans here where the sacred altar would be,
84
268000
3000
04:43
and then here's the village as a whole --
85
271000
2000
04:45
a ring of ring of rings with the chief's extended family here, the chief's immediate family here,
86
273000
5000
04:50
and here there's a tiny village only this big.
87
278000
3000
04:53
Now you might wonder, how can people fit in a tiny village only this big?
88
281000
4000
04:57
That's because they're spirit people. It's the ancestors.
89
285000
3000
05:00
And of course the spirit people have a little miniature village in their village, right?
90
288000
5000
05:05
So it's just like Georg Cantor said, the recursion continues forever.
91
293000
3000
05:08
This is in the Mandara mountains, near the Nigerian border in Cameroon, Mokoulek.
92
296000
4000
05:12
I saw this diagram drawn by a French architect,
93
300000
3000
05:15
and I thought, "Wow! What a beautiful fractal!"
94
303000
2000
05:17
So I tried to come up with a seed shape, which, upon iteration, would unfold into this thing.
95
305000
6000
05:23
I came up with this structure here.
96
311000
2000
05:25
Let's see, first iteration, second, third, fourth.
97
313000
4000
05:29
Now, after I did the simulation,
98
317000
2000
05:31
I realized the whole village kind of spirals around, just like this,
99
319000
3000
05:34
and here's that replicating line -- a self-replicating line that unfolds into the fractal.
100
322000
6000
05:40
Well, I noticed that line is about where the only square building in the village is at.
101
328000
5000
05:45
So, when I got to the village,
102
333000
2000
05:47
I said, "Can you take me to the square building?
103
335000
2000
05:49
I think something's going on there."
104
337000
2000
05:51
And they said, "Well, we can take you there, but you can't go inside
105
339000
3000
05:54
because that's the sacred altar, where we do sacrifices every year
106
342000
3000
05:57
to keep up those annual cycles of fertility for the fields."
107
345000
3000
06:00
And I started to realize that the cycles of fertility
108
348000
2000
06:02
were just like the recursive cycles in the geometric algorithm that builds this.
109
350000
4000
06:06
And the recursion in some of these villages continues down into very tiny scales.
110
354000
4000
06:10
So here's a Nankani village in Mali.
111
358000
2000
06:12
And you can see, you go inside the family enclosure --
112
360000
3000
06:15
you go inside and here's pots in the fireplace, stacked recursively.
113
363000
4000
06:19
Here's calabashes that Issa was just showing us,
114
367000
4000
06:23
and they're stacked recursively.
115
371000
2000
06:25
Now, the tiniest calabash in here keeps the woman's soul.
116
373000
2000
06:27
And when she dies, they have a ceremony
117
375000
2000
06:29
where they break this stack called the zalanga and her soul goes off to eternity.
118
377000
5000
06:34
Once again, infinity is important.
119
382000
3000
06:38
Now, you might ask yourself three questions at this point.
120
386000
4000
06:42
Aren't these scaling patterns just universal to all indigenous architecture?
121
390000
4000
06:46
And that was actually my original hypothesis.
122
394000
2000
06:48
When I first saw those African fractals,
123
396000
2000
06:50
I thought, "Wow, so any indigenous group that doesn't have a state society,
124
398000
4000
06:54
that sort of hierarchy, must have a kind of bottom-up architecture."
125
402000
3000
06:57
But that turns out not to be true.
126
405000
2000
06:59
I started collecting aerial photographs of Native American and South Pacific architecture;
127
407000
4000
07:03
only the African ones were fractal.
128
411000
2000
07:05
And if you think about it, all these different societies have different geometric design themes that they use.
129
413000
6000
07:11
So Native Americans use a combination of circular symmetry and fourfold symmetry.
130
419000
6000
07:17
You can see on the pottery and the baskets.
131
425000
2000
07:19
Here's an aerial photograph of one of the Anasazi ruins;
132
427000
3000
07:22
you can see it's circular at the largest scale, but it's rectangular at the smaller scale, right?
133
430000
5000
07:27
It is not the same pattern at two different scales.
134
435000
4000
07:31
Second, you might ask,
135
439000
1000
07:32
"Well, Dr. Eglash, aren't you ignoring the diversity of African cultures?"
136
440000
3000
07:36
And three times, the answer is no.
137
444000
2000
07:38
First of all, I agree with Mudimbe's wonderful book, "The Invention of Africa,"
138
446000
4000
07:42
that Africa is an artificial invention of first colonialism,
139
450000
3000
07:45
and then oppositional movements.
140
453000
2000
07:47
No, because a widely shared design practice doesn't necessarily give you a unity of culture --
141
455000
5000
07:52
and it definitely is not "in the DNA."
142
460000
3000
07:55
And finally, the fractals have self-similarity --
143
463000
2000
07:57
so they're similar to themselves, but they're not necessarily similar to each other --
144
465000
4000
08:01
you see very different uses for fractals.
145
469000
2000
08:03
It's a shared technology in Africa.
146
471000
2000
08:06
And finally, well, isn't this just intuition?
147
474000
3000
08:09
It's not really mathematical knowledge.
148
477000
2000
08:11
Africans can't possibly really be using fractal geometry, right?
149
479000
3000
08:14
It wasn't invented until the 1970s.
150
482000
2000
08:17
Well, it's true that some African fractals are, as far as I'm concerned, just pure intuition.
151
485000
5000
08:22
So some of these things, I'd wander around the streets of Dakar
152
490000
3000
08:25
asking people, "What's the algorithm? What's the rule for making this?"
153
493000
3000
08:28
and they'd say,
154
496000
1000
08:29
"Well, we just make it that way because it looks pretty, stupid." (Laughter)
155
497000
3000
08:32
But sometimes, that's not the case.
156
500000
3000
08:35
In some cases, there would actually be algorithms, and very sophisticated algorithms.
157
503000
5000
08:40
So in Manghetu sculpture, you'd see this recursive geometry.
158
508000
3000
08:43
In Ethiopian crosses, you see this wonderful unfolding of the shape.
159
511000
5000
08:48
In Angola, the Chokwe people draw lines in the sand,
160
516000
4000
08:52
and it's what the German mathematician Euler called a graph;
161
520000
3000
08:55
we now call it an Eulerian path --
162
523000
2000
08:57
you can never lift your stylus from the surface
163
525000
2000
08:59
and you can never go over the same line twice.
164
527000
3000
09:02
But they do it recursively, and they do it with an age-grade system,
165
530000
3000
09:05
so the little kids learn this one, and then the older kids learn this one,
166
533000
3000
09:08
then the next age-grade initiation, you learn this one.
167
536000
3000
09:11
And with each iteration of that algorithm,
168
539000
3000
09:14
you learn the iterations of the myth.
169
542000
2000
09:16
You learn the next level of knowledge.
170
544000
2000
09:19
And finally, all over Africa, you see this board game.
171
547000
2000
09:21
It's called Owari in Ghana, where I studied it;
172
549000
3000
09:24
it's called Mancala here on the East Coast, Bao in Kenya, Sogo elsewhere.
173
552000
5000
09:29
Well, you see self-organizing patterns that spontaneously occur in this board game.
174
557000
5000
09:34
And the folks in Ghana knew about these self-organizing patterns
175
562000
3000
09:37
and would use them strategically.
176
565000
2000
09:39
So this is very conscious knowledge.
177
567000
2000
09:41
Here's a wonderful fractal.
178
569000
2000
09:43
Anywhere you go in the Sahel, you'll see this windscreen.
179
571000
4000
09:47
And of course fences around the world are all Cartesian, all strictly linear.
180
575000
4000
09:51
But here in Africa, you've got these nonlinear scaling fences.
181
579000
4000
09:55
So I tracked down one of the folks who makes these things,
182
583000
2000
09:57
this guy in Mali just outside of Bamako, and I asked him,
183
585000
4000
10:01
"How come you're making fractal fences? Because nobody else is."
184
589000
2000
10:03
And his answer was very interesting.
185
591000
2000
10:05
He said, "Well, if I lived in the jungle, I would only use the long rows of straw
186
593000
5000
10:10
because they're very quick and they're very cheap.
187
598000
2000
10:12
It doesn't take much time, doesn't take much straw."
188
600000
3000
10:15
He said, "but wind and dust goes through pretty easily.
189
603000
2000
10:17
Now, the tight rows up at the very top, they really hold out the wind and dust.
190
605000
4000
10:21
But it takes a lot of time, and it takes a lot of straw because they're really tight."
191
609000
5000
10:26
"Now," he said, "we know from experience
192
614000
2000
10:28
that the farther up from the ground you go, the stronger the wind blows."
193
616000
5000
10:33
Right? It's just like a cost-benefit analysis.
194
621000
3000
10:36
And I measured out the lengths of straw,
195
624000
2000
10:38
put it on a log-log plot, got the scaling exponent,
196
626000
2000
10:40
and it almost exactly matches the scaling exponent for the relationship between wind speed and height
197
628000
5000
10:45
in the wind engineering handbook.
198
633000
1000
10:46
So these guys are right on target for a practical use of scaling technology.
199
634000
5000
10:51
The most complex example of an algorithmic approach to fractals that I found
200
639000
5000
10:56
was actually not in geometry, it was in a symbolic code,
201
644000
2000
10:58
and this was Bamana sand divination.
202
646000
3000
11:01
And the same divination system is found all over Africa.
203
649000
3000
11:04
You can find it on the East Coast as well as the West Coast,
204
652000
5000
11:09
and often the symbols are very well preserved,
205
657000
2000
11:11
so each of these symbols has four bits -- it's a four-bit binary word --
206
659000
6000
11:17
you draw these lines in the sand randomly, and then you count off,
207
665000
5000
11:22
and if it's an odd number, you put down one stroke,
208
670000
2000
11:24
and if it's an even number, you put down two strokes.
209
672000
2000
11:26
And they did this very rapidly,
210
674000
3000
11:29
and I couldn't understand where they were getting --
211
677000
2000
11:31
they only did the randomness four times --
212
679000
2000
11:33
I couldn't understand where they were getting the other 12 symbols.
213
681000
2000
11:35
And they wouldn't tell me.
214
683000
2000
11:37
They said, "No, no, I can't tell you about this."
215
685000
2000
11:39
And I said, "Well look, I'll pay you, you can be my teacher,
216
687000
2000
11:41
and I'll come each day and pay you."
217
689000
2000
11:43
They said, "It's not a matter of money. This is a religious matter."
218
691000
3000
11:46
And finally, out of desperation, I said,
219
694000
1000
11:47
"Well, let me explain Georg Cantor in 1877."
220
695000
3000
11:50
And I started explaining why I was there in Africa,
221
698000
4000
11:54
and they got very excited when they saw the Cantor set.
222
702000
2000
11:56
And one of them said, "Come here. I think I can help you out here."
223
704000
4000
12:00
And so he took me through the initiation ritual for a Bamana priest.
224
708000
5000
12:05
And of course, I was only interested in the math,
225
713000
2000
12:07
so the whole time, he kept shaking his head going,
226
715000
2000
12:09
"You know, I didn't learn it this way."
227
717000
1000
12:10
But I had to sleep with a kola nut next to my bed, buried in sand,
228
718000
4000
12:14
and give seven coins to seven lepers and so on.
229
722000
3000
12:17
And finally, he revealed the truth of the matter.
230
725000
4000
12:22
And it turns out it's a pseudo-random number generator using deterministic chaos.
231
730000
4000
12:26
When you have a four-bit symbol, you then put it together with another one sideways.
232
734000
6000
12:32
So even plus odd gives you odd.
233
740000
2000
12:34
Odd plus even gives you odd.
234
742000
2000
12:36
Even plus even gives you even. Odd plus odd gives you even.
235
744000
3000
12:39
It's addition modulo 2, just like in the parity bit check on your computer.
236
747000
4000
12:43
And then you take this symbol, and you put it back in
237
751000
4000
12:47
so it's a self-generating diversity of symbols.
238
755000
2000
12:49
They're truly using a kind of deterministic chaos in doing this.
239
757000
4000
12:53
Now, because it's a binary code,
240
761000
2000
12:55
you can actually implement this in hardware --
241
763000
2000
12:57
what a fantastic teaching tool that should be in African engineering schools.
242
765000
5000
13:02
And the most interesting thing I found out about it was historical.
243
770000
3000
13:05
In the 12th century, Hugo of Santalla brought it from Islamic mystics into Spain.
244
773000
6000
13:11
And there it entered into the alchemy community as geomancy:
245
779000
6000
13:17
divination through the earth.
246
785000
2000
13:19
This is a geomantic chart drawn for King Richard II in 1390.
247
787000
5000
13:24
Leibniz, the German mathematician,
248
792000
3000
13:27
talked about geomancy in his dissertation called "De Combinatoria."
249
795000
4000
13:31
And he said, "Well, instead of using one stroke and two strokes,
250
799000
4000
13:35
let's use a one and a zero, and we can count by powers of two."
251
803000
4000
13:39
Right? Ones and zeros, the binary code.
252
807000
2000
13:41
George Boole took Leibniz's binary code and created Boolean algebra,
253
809000
3000
13:44
and John von Neumann took Boolean algebra and created the digital computer.
254
812000
3000
13:47
So all these little PDAs and laptops --
255
815000
3000
13:50
every digital circuit in the world -- started in Africa.
256
818000
3000
13:53
And I know Brian Eno says there's not enough Africa in computers,
257
821000
5000
13:58
but you know, I don't think there's enough African history in Brian Eno.
258
826000
5000
14:03
(Laughter) (Applause)
259
831000
3000
14:06
So let me end with just a few words about applications that we've found for this.
260
834000
4000
14:10
And you can go to our website,
261
838000
2000
14:12
the applets are all free; they just run in the browser.
262
840000
2000
14:14
Anybody in the world can use them.
263
842000
2000
14:16
The National Science Foundation's Broadening Participation in Computing program
264
844000
5000
14:21
recently awarded us a grant to make a programmable version of these design tools,
265
849000
7000
14:28
so hopefully in three years, anybody'll be able to go on the Web
266
856000
2000
14:30
and create their own simulations and their own artifacts.
267
858000
3000
14:33
We've focused in the U.S. on African-American students as well as Native American and Latino.
268
861000
5000
14:38
We've found statistically significant improvement with children using this software in a mathematics class
269
866000
6000
14:44
in comparison with a control group that did not have the software.
270
872000
3000
14:47
So it's really very successful teaching children that they have a heritage that's about mathematics,
271
875000
6000
14:53
that it's not just about singing and dancing.
272
881000
4000
14:57
We've started a pilot program in Ghana.
273
885000
3000
15:00
We got a small seed grant, just to see if folks would be willing to work with us on this;
274
888000
5000
15:05
we're very excited about the future possibilities for that.
275
893000
3000
15:08
We've also been working in design.
276
896000
2000
15:10
I didn't put his name up here -- my colleague, Kerry, in Kenya, has come up with this great idea
277
898000
5000
15:15
for using fractal structure for postal address in villages that have fractal structure,
278
903000
5000
15:20
because if you try to impose a grid structure postal system on a fractal village,
279
908000
4000
15:24
it doesn't quite fit.
280
912000
2000
15:26
Bernard Tschumi at Columbia University has finished using this in a design for a museum of African art.
281
914000
5000
15:31
David Hughes at Ohio State University has written a primer on Afrocentric architecture
282
919000
8000
15:39
in which he's used some of these fractal structures.
283
927000
2000
15:41
And finally, I just wanted to point out that this idea of self-organization,
284
929000
5000
15:46
as we heard earlier, it's in the brain.
285
934000
2000
15:48
It's in the -- it's in Google's search engine.
286
936000
5000
15:53
Actually, the reason that Google was such a success
287
941000
2000
15:55
is because they were the first ones to take advantage of the self-organizing properties of the web.
288
943000
4000
15:59
It's in ecological sustainability.
289
947000
2000
16:01
It's in the developmental power of entrepreneurship,
290
949000
2000
16:03
the ethical power of democracy.
291
951000
2000
16:06
It's also in some bad things.
292
954000
2000
16:08
Self-organization is why the AIDS virus is spreading so fast.
293
956000
3000
16:11
And if you don't think that capitalism, which is self-organizing, can have destructive effects,
294
959000
4000
16:15
you haven't opened your eyes enough.
295
963000
2000
16:17
So we need to think about, as was spoken earlier,
296
965000
4000
16:21
the traditional African methods for doing self-organization.
297
969000
2000
16:23
These are robust algorithms.
298
971000
2000
16:26
These are ways of doing self-organization -- of doing entrepreneurship --
299
974000
3000
16:29
that are gentle, that are egalitarian.
300
977000
2000
16:31
So if we want to find a better way of doing that kind of work,
301
979000
4000
16:35
we need look only no farther than Africa to find these robust self-organizing algorithms.
302
983000
5000
16:40
Thank you.
303
988000
1000
ABOUT THE SPEAKER
Ron Eglash - MathematicianRon Eglash is an ethno-mathematician: he studies the way math and cultures intersect. He has shown that many aspects of African design -- in architecture, art, even hair braiding -- are based on perfect fractal patterns.
Why you should listen
"Ethno-mathematician" Ron Eglash is the author of African Fractals, a book that examines the fractal patterns underpinning architecture, art and design in many parts of Africa. By looking at aerial-view photos -- and then following up with detailed research on the ground -- Eglash discovered that many African villages are purposely laid out to form perfect fractals, with self-similar shapes repeated in the rooms of the house, and the house itself, and the clusters of houses in the village, in mathematically predictable patterns.
As he puts it: "When Europeans first came to Africa, they considered the architecture very disorganized and thus primitive. It never occurred to them that the Africans might have been using a form of mathematics that they hadn't even discovered yet."
His other areas of study are equally fascinating, including research into African and Native American cybernetics, teaching kids math through culturally specific design tools (such as the Virtual Breakdancer applet, which explores rotation and sine functions), and race and ethnicity issues in science and technology. Eglash teaches in the Department of Science and Technology Studies at Rensselaer Polytechnic Institute in New York, and he recently co-edited the book Appropriating Technology, about how we reinvent consumer tech for our own uses.
Ron Eglash | Speaker | TED.com