0:00:16.910,0:00:18.180
SAM LIVINGSTON-GRAY: Hello. Welcome to the[br]very

0:00:18.180,0:00:20.400
last session of RailsConf.

0:00:20.400,0:00:21.080
When I leave this stage,

0:00:21.080,0:00:22.410
they are gonna burn it down.

0:00:22.410,0:00:24.210
AUDIENCE: Yeah!

0:00:24.900,0:00:26.240
S.L.: I have a couple of items of business

0:00:26.250,0:00:29.890
before I launch into my presentation proper.[br]The first

0:00:29.890,0:00:32.930
of which is turning on my clicker. I work

0:00:32.930,0:00:36.449
for LivingSocial. We are hiring. If this fact[br]is

0:00:36.449,0:00:38.230
intriguing to you, please feel free to come[br]and

0:00:38.230,0:00:41.400
talk to me afterwards. Also, our recruiters[br]brought a

0:00:41.400,0:00:43.790
ton of little squishy stress balls that are[br]shaped

0:00:43.790,0:00:46.650
like little brains. As far as I know, this

0:00:46.650,0:00:48.730
was a coincidence, but I love the tie-in so

0:00:48.730,0:00:50.110
I brought the whole bag. I had them leave

0:00:50.110,0:00:52.079
it for me. So if you would like an

0:00:52.079,0:00:54.960
extra brain, please come talk to me after[br]the

0:00:54.960,0:00:55.899
show.

0:00:55.899,0:00:59.050
A quick note about accessibility. If you have[br]any

0:00:59.050,0:01:02.079
trouble seeing my slides, hearing my voice,[br]or following

0:01:02.079,0:01:04.519
my weird trains of thought, or maybe you just

0:01:04.519,0:01:06.580
like spoilers, you can get a PDF with both

0:01:06.580,0:01:09.420
my slides and my script at this url. It's

0:01:09.420,0:01:14.270
tinyurl dot com, cog dash shorts dash railsconf.[br]I

0:01:14.270,0:01:15.750
also have it up here on a thumb drive,

0:01:15.750,0:01:18.220
so if the conference wi-fi does what it usually

0:01:18.220,0:01:19.970
does, please go see Evan Light up in the

0:01:19.970,0:01:21.020
second row.

0:01:21.020,0:01:23.830
I'm gonna leave this up for a couple more

0:01:23.830,0:01:25.780
minutes. And I also want to give a shoutout

0:01:25.780,0:01:30.110
to the opportunity scholarship program here.[br]To quote the

0:01:30.110,0:01:33.020
RailsConf site, this program is for people[br]who wouldn't

0:01:33.020,0:01:35.360
usually take part in our community or who[br]might

0:01:35.360,0:01:37.380
just want a friendly face during their first[br]time

0:01:37.380,0:01:40.750
at RailsConf. I'm a huge fan of this program.

0:01:40.750,0:01:42.940
I think it's a great way to welcome new

0:01:42.940,0:01:45.890
people and new voices into our community.[br]This is

0:01:45.890,0:01:47.630
the second year that I've volunteered as a[br]guide

0:01:47.630,0:01:48.770
and this is the second year that I've met

0:01:48.770,0:01:51.750
somebody with a fascinating story to tell.[br]If you're

0:01:51.750,0:01:54.780
a seasoned conference veteran, I strongly[br]encourage you to

0:01:54.780,0:01:56.869
apply next year.

0:01:56.869,0:02:03.869
OK. Programming is hard. It's not quantum[br]physics. But

0:02:04.410,0:02:06.700
neither is it falling off a log. And if

0:02:06.700,0:02:08.530
I had to pick just one word to explain

0:02:08.530,0:02:12.530
why programming is hard, that word would be[br]abstract.

0:02:12.530,0:02:15.709
I asked Google to define abstract, and here's[br]what

0:02:15.709,0:02:17.040
it said.

0:02:17.040,0:02:19.129
Existing in thought or as an idea, but not

0:02:19.129,0:02:21.629
having a physical or concrete existence.

0:02:21.629,0:02:24.459
I usually prefer defining things in terms[br]of what

0:02:24.459,0:02:26.500
they are, but in this case I find the

0:02:26.500,0:02:31.390
negative definition extremely telling. Abstract[br]things are hard for

0:02:31.390,0:02:34.150
us to think about precisely because they don't[br]have

0:02:34.150,0:02:35.950
a physical or a concrete existence, and that's[br]what

0:02:35.950,0:02:37.540
our brains are wired for.

0:02:37.540,0:02:40.670
Now, I normally prefer the kind of talk where

0:02:40.670,0:02:42.950
the speaker just launches right in and forces[br]me

0:02:42.950,0:02:45.590
to keep up, but this is a complex idea,

0:02:45.590,0:02:48.310
and it's the last talk of the last day,

0:02:48.310,0:02:49.480
and I'm sure you're all as fried as I

0:02:49.480,0:02:54.010
am. So, here's a little background. I got[br]the

0:02:54.010,0:02:56.570
idea for this talk when I was listening to

0:02:56.570,0:03:00.510
the Ruby Rogues podcast episode with Glenn[br]Vanderburg. This

0:03:00.510,0:03:03.360
is lightly edited for length, but in that[br]episode,

0:03:03.360,0:03:06.200
Glenn said, The best programmers I know all[br]have

0:03:06.200,0:03:09.170
some good techniques for conceptualizing or[br]modeling the programs

0:03:09.170,0:03:11.030
that they work with. And it tends to be

0:03:11.030,0:03:13.960
sort of a spatial/visual model, but not always.[br]And

0:03:13.960,0:03:16.069
he says, What's going on is our brains are

0:03:16.069,0:03:18.400
geared towards the physical world and dealing[br]with our

0:03:18.400,0:03:21.680
senses and integrating that sensory input.

0:03:21.680,0:03:23.100
But the work we do as programmers is all

0:03:23.100,0:03:25.210
abstract. And it makes perfect sense that[br]you would

0:03:25.210,0:03:27.140
want to find techniques to rope the physical[br]sensory

0:03:27.140,0:03:29.540
parts of your brain into this task of dealing

0:03:29.540,0:03:32.150
with abstractions. And this is the part that[br]really

0:03:32.150,0:03:34.170
got my attention. He says, But we don't ever

0:03:34.170,0:03:35.730
teach anybody how to do that or even that

0:03:35.730,0:03:37.950
they should do that.

0:03:37.950,0:03:40.380
When I heard this, I started thinking about[br]the

0:03:40.380,0:03:44.069
times that I've stumbled across some technique[br]for doing

0:03:44.069,0:03:46.200
something like this, and I've been really[br]excited to

0:03:46.200,0:03:48.830
find a way of translating a programming problem[br]into

0:03:48.830,0:03:50.630
some form that my brain could really get a

0:03:50.630,0:03:52.770
handle on. And I was like yeah, yeah, brains

0:03:52.770,0:03:54.970
are awesome. And we should be teaching people[br]that

0:03:54.970,0:03:57.880
this is a thing they can do.

0:03:57.880,0:03:59.650
And I thought about it, and some time later,

0:03:59.650,0:04:02.600
I was like, wait a minute. No. Brains are

0:04:02.600,0:04:06.540
horrible. And teaching people these tricks[br]would be totally

0:04:06.540,0:04:09.069
irresponsible, if we also didn't warn them[br]about cognitive

0:04:09.069,0:04:13.040
bias. I get to that in a little bit.

0:04:13.040,0:04:16.639
This is a talk in three parts. Part one,

0:04:16.639,0:04:19.000
brains are awesome. And as Glenn said, you[br]can

0:04:19.000,0:04:21.019
rope the physical and sensory parts of your[br]brain

0:04:21.019,0:04:23.020
as well as a few others I'll talk about

0:04:23.020,0:04:26.220
into helping you deal with abstractions. Part[br]two, brains

0:04:26.220,0:04:28.319
are horrible and they lie to us all the

0:04:28.319,0:04:31.470
time. But if you're on the look out for

0:04:31.470,0:04:33.210
the kinds of lies that your brain will tell

0:04:33.210,0:04:35.790
you, in part three I have an example of

0:04:35.790,0:04:37.470
the kind of amazing hack that you just might

0:04:37.470,0:04:42.290
be able to come up with.

0:04:42.290,0:04:45.360
Our brains are extremely well-adapted for[br]dealing with the

0:04:45.360,0:04:50.250
physical world. Our hindbrains, which regular[br]respiration, temperature, and

0:04:50.250,0:04:52.310
balance, have been around for half a billion[br]years

0:04:52.310,0:04:57.139
or so. But when I write software, I am

0:04:57.139,0:04:58.750
leaning hard on parts of the brain that are

0:04:58.750,0:05:01.800
relatively new in evolutionary terms, and[br]I'm using some

0:05:01.800,0:05:04.370
relatively expensive resources.

0:05:04.370,0:05:05.770
So over the years I have built up a

0:05:05.770,0:05:09.400
small collection of techniques and shortcuts[br]that engage specialized

0:05:09.400,0:05:10.960
structures of my brain to help me reason about

0:05:10.960,0:05:13.870
programming problems. Here's the list.

0:05:13.870,0:05:17.880
I'm gonna start with a category of visual[br]tools

0:05:17.880,0:05:20.410
that let us leverage our spatial understanding[br]of the

0:05:20.410,0:05:23.500
world and our spatial reasoning skills to[br]discover relationships

0:05:23.500,0:05:25.850
between different parts of a model. Or just[br]to

0:05:25.850,0:05:27.760
stay oriented when we're trying to reason[br]through a

0:05:27.760,0:05:29.090
complex problem.

0:05:29.090,0:05:31.570
I'm just gonna list out a few examples of

0:05:31.570,0:05:34.350
this category quickly, because I think most[br]developers are

0:05:34.350,0:05:37.320
likely to encounter these, either in school[br]or on

0:05:37.320,0:05:38.740
the job. And they all have the same basic

0:05:38.740,0:05:41.870
shape. They're boxes and arrows.

0:05:41.870,0:05:45.620
There's Entity-Relationship Diagrams, which[br]help us understand how our

0:05:45.620,0:05:48.790
data is modeled. We use diagrams to describe[br]data

0:05:48.790,0:05:53.210
structures like binary trees, linked lists,[br]and so on.

0:05:53.210,0:05:56.090
And for state machines of any complexity,[br]diagrams are

0:05:56.090,0:05:57.580
often the only way to make any sense of

0:05:57.580,0:06:00.330
them. I could go on, but like I said,

0:06:00.330,0:06:01.949
most of us are probably used to using these,

0:06:01.949,0:06:02.919
at least occasionally.

0:06:02.919,0:06:05.470
There are three things that I like about these

0:06:05.470,0:06:09.720
tools. First, they lend themselves really[br]well to standing

0:06:09.720,0:06:12.120
up in front of a white board, possibly with

0:06:12.120,0:06:15.260
a co-worker, and just standing up and moving[br]around

0:06:15.260,0:06:17.030
a little bit will help get the blood flowing

0:06:17.030,0:06:19.510
and, and get your brain perked up.

0:06:19.510,0:06:22.570
Second, diagrams help us offload some of the[br]work

0:06:22.570,0:06:26.169
of keeping track of things, of different concepts,[br]by

0:06:26.169,0:06:29.250
attaching those concepts to objects in a two-dimensional[br]space.

0:06:29.250,0:06:30.919
And our brains have a lot of hardware support

0:06:30.919,0:06:34.419
for keeping track of where things are in space.

0:06:34.419,0:06:36.510
And third, our brains are really good at pattern

0:06:36.510,0:06:39.690
recognition, so visualizing our designs can[br]give us a

0:06:39.690,0:06:42.560
chance to spot certain kinds of problems just[br]by

0:06:42.560,0:06:45.590
looking at their shapes before we ever start[br]typing

0:06:45.590,0:06:49.449
code in an editor, and I think that's pretty

0:06:49.449,0:06:50.010
cool.

0:06:50.010,0:06:52.560
Here's another technique that makes use of[br]our spatial

0:06:52.560,0:06:54.889
perception skills, and if you saw Sandi's[br]talk yesterday,

0:06:54.889,0:06:57.680
you'll know this one. It's the squint test.[br]It's

0:06:57.680,0:07:00.370
very straight forward. You open up some code[br]and

0:07:00.370,0:07:02.370
you either squint your eyes at it or you

0:07:02.370,0:07:04.729
decrease the font size. The point is to look

0:07:04.729,0:07:06.850
past the words and check out the shape of

0:07:06.850,0:07:08.280
the code.

0:07:08.280,0:07:10.940
This is a pathological example that I used[br]in

0:07:10.940,0:07:14.520
a refactoring talk last year. You can use[br]this

0:07:14.520,0:07:17.610
technique as an aid to navigation, as a way

0:07:17.610,0:07:19.800
of zeroing in on high-risk areas of code,[br]or

0:07:19.800,0:07:22.210
just plain to get oriented in a new code

0:07:22.210,0:07:25.479
base. There are a few specific patterns that[br]you

0:07:25.479,0:07:27.199
can look for, and you'll find others as you,

0:07:27.199,0:07:28.800
as you do more of it.

0:07:28.800,0:07:31.960
Is the left margin ragged, as it is here?

0:07:31.960,0:07:34.139
Are there any ridiculously long lines? There's[br]one towards

0:07:34.139,0:07:37.639
the bottom. What does your syntax highlighting[br]tell you?

0:07:37.639,0:07:39.330
Are there groups of colors or are colors sort

0:07:39.330,0:07:43.820
of spread out? And there's a lot of information

0:07:43.820,0:07:47.759
you can glean from this. Incidentally, I have[br]only

0:07:47.759,0:07:50.070
ever met one blind programmer, and we didn't[br]really

0:07:50.070,0:07:52.060
talk about this stuff. If any of you have

0:07:52.060,0:07:55.610
found that a physical or a cognitive disability[br]gives

0:07:55.610,0:07:58.960
you a, an interesting way of looking at code,

0:07:58.960,0:08:01.840
or understanding code I suppose, please come[br]talk to

0:08:01.840,0:08:05.710
me, because I'd love to hear your story.

0:08:05.710,0:08:07.620
Next up, I have a couple of techniques that

0:08:07.620,0:08:10.560
involve a clever use of language. The first[br]one

0:08:10.560,0:08:13.990
is deceptively simple, but it does require[br]a prop.

0:08:13.990,0:08:18.470
Doesn't have to be that big. You can totally

0:08:18.470,0:08:20.419
get away with using the souvenir edition.[br]This is

0:08:20.419,0:08:24.720
my daughter's duck cow bath toy. What you[br]do

0:08:24.720,0:08:26.710
is you keep a rubber duck on your desk.

0:08:26.710,0:08:28.210
When you get stuck, you put the rubber deck

0:08:28.210,0:08:30.840
on top- rubber duck, excuse me, on top of

0:08:30.840,0:08:33.349
your keyboard, and you explain your problem[br]out loud

0:08:33.349,0:08:34.429
to the duck.

0:08:34.429,0:08:35.328
[laughter]

0:08:35.328,0:08:39.578
Really. I mean, it sounds absurd, right. But[br]there's

0:08:39.578,0:08:41.360
a good chance that in the process of putting

0:08:41.360,0:08:44.269
your problem into words, you'll discover that[br]there's an

0:08:44.269,0:08:46.579
incorrect assumption that you've been making[br]or you'll think

0:08:46.579,0:08:48.490
of some other possible solution.

0:08:48.490,0:08:50.709
I've also heard of people using teddy bears[br]or

0:08:50.709,0:08:53.689
other stuffed animals. And one of my co-workers[br]told

0:08:53.689,0:08:56.649
me that she learned this as the pet-rock technique.

0:08:56.649,0:08:58.929
This was a thing in the seventies. And also

0:08:58.929,0:09:00.869
that she finds it useful to compose an email

0:09:00.869,0:09:03.740
describing the problem. So for those of you[br]who,

0:09:03.740,0:09:07.040
like me, think better when you're typing or[br]writing

0:09:07.040,0:09:08.429
than when you're speaking, that can be a nice,

0:09:08.429,0:09:13.259
a nice alternative.

0:09:13.259,0:09:15.069
The other linguistic hack that I got, I got

0:09:15.069,0:09:18.639
from Sandi Metz, and in this book, Practical[br]Oriented

0:09:18.639,0:09:22.050
Design in Ruby, PODR for short, she describes[br]a

0:09:22.050,0:09:25.100
technique that she uses to figure out which[br]object

0:09:25.100,0:09:28.730
a method should belong. I tried paraphrasing[br]this, but

0:09:28.730,0:09:30.730
honestly Sandi did a much better job than[br]I

0:09:30.730,0:09:32.779
would, describing it, so I'm just gonna read[br]it

0:09:32.779,0:09:33.459
verbatim.

0:09:33.459,0:09:35.819
She says, How can you determine if a Gear

0:09:35.819,0:09:39.040
class contains behavior that belongs somewhere[br]else? One way

0:09:39.040,0:09:41.170
is to pretend that it's sentient and to interrogate

0:09:41.170,0:09:43.769
it. If you rephrase every one of its methods

0:09:43.769,0:09:45.600
as a question, asking the question ought to[br]make

0:09:45.600,0:09:47.100
sense.

0:09:47.100,0:09:49.309
For example, "Please Mr. Gear, what is your[br]ratio?"

0:09:49.309,0:09:52.079
seems perfectly reasonable, while "Please[br]Mr. Gear, what are

0:09:52.079,0:09:54.449
your gear_inches?" is on shaky ground, and[br]"Please Mr.

0:09:54.449,0:09:59.889
Gear, what is your tire(size)?" is just downright[br]ridiculous.

0:09:59.889,0:10:01.809
This is a great way to evaluate objects in

0:10:01.809,0:10:04.350
light of the single responsibility principle.[br]Now I'll come

0:10:04.350,0:10:07.240
back to that thought in just a minute, but

0:10:07.240,0:10:10.240
first, I described the rubber duck and Please,[br]Mr.

0:10:10.240,0:10:13.579
Gear? as techniques to engage linguistic reasoning,[br]but that

0:10:13.579,0:10:17.269
doesn't quite feel right. Both of these techniques[br]force

0:10:17.269,0:10:20.439
us to put our questions into words, but words

0:10:20.439,0:10:23.189
themselves are tools. We use words to communicate[br]our

0:10:23.189,0:10:25.749
ideas to other people.

0:10:25.749,0:10:28.619
As primates, we've evolved a set of social[br]skills

0:10:28.619,0:10:30.970
and behaviors for getting our needs met as[br]part

0:10:30.970,0:10:34.619
of a community. So, while these techniques[br]do involve

0:10:34.619,0:10:36.869
using language centers of your brain, I think[br]they

0:10:36.869,0:10:39.989
reach beyond those centers to tap into our[br]social

0:10:39.989,0:10:41.709
reasoning.

0:10:41.709,0:10:44.499
The rubber duck technique works because, putting[br]your problem

0:10:44.499,0:10:47.309
into words forces you to organize your understanding[br]of

0:10:47.309,0:10:49.230
a problem in such a way that you can

0:10:49.230,0:10:53.529
verbally lead another mind through it. And[br]Please, Mr.

0:10:53.529,0:10:56.420
Gear? let's us anthropomorphize an object[br]and talk to

0:10:56.420,0:10:58.720
it to discover whether it conforms to the[br]single

0:10:58.720,0:11:00.910
responsibility principle.

0:11:00.910,0:11:03.550
To me, the tell-tale phrase in Sandi's description[br]of

0:11:03.550,0:11:06.459
this technique is, asking the question ought[br]to make

0:11:06.459,0:11:13.459
sense. Most of us have an intuitive understanding[br]that

0:11:13.809,0:11:16.410
it might not be appropriate to ask Alice about

0:11:16.410,0:11:20.350
something that is Bob's responsibility. Interrogating[br]an object as

0:11:20.350,0:11:22.420
though it were a person helps us use that

0:11:22.420,0:11:25.239
social knowledge, and it gives us an opportunity[br]to

0:11:25.239,0:11:28.119
notice that a particular question doesn't[br]make sense to

0:11:28.119,0:11:31.009
ask any of our existing objects, which might[br]prompt

0:11:31.009,0:11:32.300
us to ask if we should create a new

0:11:32.300,0:11:35.019
object to fill that role instead.

0:11:35.019,0:11:39.610
Now, personally, I would have considered PODR[br]to have

0:11:39.610,0:11:41.679
been a worthwhile purchase if Please, Mr.[br]Gear was

0:11:41.679,0:11:44.089
the only thing I got from it. But, in

0:11:44.089,0:11:45.649
this book, Sandi also made what I thought[br]was

0:11:45.649,0:11:49.209
a very compelling case for UML Sequence Diagrams.

0:11:49.209,0:11:53.110
Where Please, Mr. Gear is a good tool for

0:11:53.110,0:11:56.329
discovering which objects should be responsible[br]for a particular

0:11:56.329,0:11:59.100
method, a Sequence Diagram can help you analyze[br]the

0:11:59.100,0:12:04.019
runtime interaction between several different[br]objects. At first glance,

0:12:04.019,0:12:05.929
this looks kind of like something in the boxes

0:12:05.929,0:12:08.499
and arrows category of visual and spatial[br]tools, but

0:12:08.499,0:12:10.429
again, this feels more like it's tapping into[br]that

0:12:10.429,0:12:12.959
social understanding that we have. This can[br]be a

0:12:12.959,0:12:14.339
good way to get a sense for when an

0:12:14.339,0:12:18.089
object is bossy or when performing a task[br]involves

0:12:18.089,0:12:22.309
a complex sequence of several, several interactions.[br]Or if

0:12:22.309,0:12:25.009
there are just plain too many different things[br]to

0:12:25.009,0:12:27.319
keep track of.

0:12:27.319,0:12:29.100
Rather than turn this into a lecture on UML,

0:12:29.100,0:12:30.910
I'm just gonna tell you to go buy Sandi's

0:12:30.910,0:12:33.779
book, and if for whatever reason, you cannot[br]afford

0:12:33.779,0:12:35.220
it, come talk to me later and we'll work

0:12:35.220,0:12:36.329
something out.

0:12:36.329,0:12:41.389
Now for the really hand-wavy stuff. Metaphors[br]can be

0:12:41.389,0:12:45.069
a really useful tool in software. The turtle[br]graphic

0:12:45.069,0:12:47.399
system in Logo is a great metaphor. Has anybody

0:12:47.399,0:12:50.509
used Logo at any point in their life? About

0:12:50.509,0:12:52.850
half the people. That's really cool.

0:12:52.850,0:12:55.269
We've probably all played with drawing something[br]on the

0:12:55.269,0:12:57.679
screen at some point, but most of the rendering

0:12:57.679,0:13:00.069
systems that I've used are based on a Cartesian

0:13:00.069,0:13:04.910
coordinate system, a grid. And this metaphor[br]encourages the

0:13:04.910,0:13:07.569
programmer to imagine themselves as the turtle,[br]and to

0:13:07.569,0:13:09.449
use that understanding to figure out, when[br]they get

0:13:09.449,0:13:11.670
stuck, what they should be doing next.

0:13:11.670,0:13:14.369
One of the original creators of Logo called[br]this

0:13:14.369,0:13:17.920
Body Syntonic Reasoning, and specifically[br]developed it to help

0:13:17.920,0:13:21.529
children solve problems. But the turtle metaphor,[br]the turtle

0:13:21.529,0:13:24.480
metaphor works for everybody, not just for[br]kids.

0:13:24.480,0:13:31.480
Cartesian grids are great for drawing boxes.[br]Mostly great.

0:13:31.769,0:13:33.489
But it can take some very careful thinking[br]to

0:13:33.489,0:13:36.529
figure out how to, how to use x, y

0:13:36.529,0:13:41.239
coordinate pairs to draw a spiral or a star

0:13:41.239,0:13:44.860
or a snowflake or a tree. Choosing a different

0:13:44.860,0:13:47.739
metaphor can make different kinds of solutions[br]easy, where

0:13:47.739,0:13:49.319
before they seemed like too much trouble to[br]be

0:13:49.319,0:13:51.139
worth bothering with.

0:13:51.139,0:13:58.139
James Ladd, in 2008, wrote a couple of interesting

0:13:58.819,0:14:03.609
blog posts about what he called East-oriented[br]code. Imagine

0:14:03.609,0:14:07.420
a compass overlaid on top of your screen.[br]In

0:14:07.420,0:14:10.069
this, in this model, messages that an object[br]sends

0:14:10.069,0:14:13.319
to itself go South, and any data returned[br]from

0:14:13.319,0:14:17.420
those calls goes North. Communications between[br]objects is the

0:14:17.420,0:14:20.359
same thing, rotated ninety degrees. Messages[br]sent to other

0:14:20.359,0:14:22.879
objects go East, and the return values from[br]those

0:14:22.879,0:14:26.920
messages flow West.

0:14:26.920,0:14:29.759
What James Ladd suggests is that, in general,[br]code

0:14:29.759,0:14:32.970
that sends messages to other objects, code[br]where information

0:14:32.970,0:14:36.290
mostly flows East, is easier to extend and[br]maintain

0:14:36.290,0:14:38.249
than code that looks at data and then decides

0:14:38.249,0:14:40.269
what to do with it, which is code where

0:14:40.269,0:14:42.100
information flows West.

0:14:42.100,0:14:45.420
Really, this is just the design principle,[br]tell, don't

0:14:45.420,0:14:49.299
ask. But, the metaphor of the compass, compass[br]recasts

0:14:49.299,0:14:51.839
this in a way that helps us use our

0:14:51.839,0:14:55.170
background spatial awareness to keep this[br]principle in mind

0:14:55.170,0:14:58.910
at all times. In fact, there are plenty of

0:14:58.910,0:15:01.629
ways we can use our background level awareness[br]to

0:15:01.629,0:15:03.139
analyze our code.

0:15:03.139,0:15:08.639
Isn't this adorable? I love this picture.

0:15:08.639,0:15:11.220
Code smells are an entire category of metaphors[br]that

0:15:11.220,0:15:13.779
we use to talk about our work. In fact,

0:15:13.779,0:15:16.959
the name code smell itself is a metaphor for

0:15:16.959,0:15:19.639
anything about your code that hints at a design

0:15:19.639,0:15:25.389
problem, which I suppose makes it a meta-metaphor.

0:15:25.389,0:15:29.230
Some code smells have names are extremely[br]literal. Duplicated

0:15:29.230,0:15:30.559
code, long method and so on. But some of

0:15:30.559,0:15:36.049
these are delightfully suggestive. Feature[br]envy. Refused bequest. Primitive

0:15:36.049,0:15:39.149
obsession. To me, the names on the right have

0:15:39.149,0:15:41.769
a lot in common with Please, Mr. Gear. They're

0:15:41.769,0:15:44.730
chosen to hook into something in our social[br]awareness

0:15:44.730,0:15:47.739
to give a name to a pattern of dysfunction,

0:15:47.739,0:15:50.319
and by naming the problems it suggests a possible

0:15:50.319,0:15:51.720
solution.

0:15:51.720,0:15:54.730
So, these are most of the shortcuts that I've

0:15:54.730,0:15:57.399
accumulated over the years, and I hope that[br]this

0:15:57.399,0:15:59.199
can be the start of a similar collection for

0:15:59.199,0:16:04.480
some of you.

0:16:04.480,0:16:05.359
Now the part where I try to put the

0:16:05.359,0:16:09.629
fear into you. Evolution has designed our[br]brains to

0:16:09.629,0:16:13.529
lie to us. Brains are expensive. The human[br]brain

0:16:13.529,0:16:16.129
accounts for just two percent of body mass,[br]but

0:16:16.129,0:16:19.329
twenty percent of our caloric intake. That's[br]a huge

0:16:19.329,0:16:22.879
energy requirement that has to be justified.

0:16:22.879,0:16:25.949
Evolution, as a designer, does one thing and[br]one

0:16:25.949,0:16:29.559
thing only. It selects for traits that allow[br]an

0:16:29.559,0:16:32.449
organism to stay alive long enough to reproduce.[br]It

0:16:32.449,0:16:34.899
doesn't care about getting the best solution.[br]Only one

0:16:34.899,0:16:38.029
that's good enough to compete in the current[br]landscape.

0:16:38.029,0:16:40.259
Evolution will tolerate any hack as long as[br]it

0:16:40.259,0:16:42.649
meets that one goal.

0:16:42.649,0:16:44.249
As an example, I want to take a minute

0:16:44.249,0:16:46.269
to talk about how we see the world around

0:16:46.269,0:16:50.059
us. The human eye has two different kinds[br]of

0:16:50.059,0:16:52.389
photo receptors. There are about a hundred[br]and twenty

0:16:52.389,0:16:55.279
million rod cells in each eye. These play[br]little

0:16:55.279,0:16:57.730
or no role in color vision, and they're mostly

0:16:57.730,0:16:59.809
used for night time and peripheral vision.

0:16:59.809,0:17:02.019
There are also about six or seven million[br]cone

0:17:02.019,0:17:04.500
cells in each eye, and these give us color

0:17:04.500,0:17:05.939
vision, but they require a lot more light[br]to

0:17:05.939,0:17:09.869
work. And the vast majority of cone cells[br]are

0:17:09.869,0:17:11.890
packed together in a tight little cluster[br]near the

0:17:11.890,0:17:13.900
center of the retina. This area is what we

0:17:13.900,0:17:15.970
use to focus on individual details, and it's[br]smaller

0:17:15.970,0:17:20.640
than you might think. It's only fifteen degrees[br]wide.

0:17:20.640,0:17:23.660
As a result, our vision is extremely directional.[br]We

0:17:23.660,0:17:25.520
have a very small area of high detail and

0:17:25.520,0:17:27.430
high color, and the rest of our field of

0:17:27.430,0:17:31.140
vision is more or less monochrome. So when[br]we

0:17:31.140,0:17:38.140
look at this, our eyes see something like[br]this.

0:17:40.390,0:17:42.580
In order to turn the image on the left

0:17:42.580,0:17:44.920
into the image on the right, our brains are

0:17:44.920,0:17:46.960
doing a lot of work that we're mostly unaware

0:17:46.960,0:17:49.280
of.

0:17:49.280,0:17:51.400
We compensate for having such highly directional[br]vision by

0:17:51.400,0:17:54.500
moving our eyes around a lot. Our brains combine

0:17:54.500,0:17:57.050
the details from these individual points of[br]interest to

0:17:57.050,0:17:59.710
construct a persistent mental model of whatever[br]we're looking

0:17:59.710,0:18:03.400
at. These fast point to point movements are[br]called

0:18:03.400,0:18:06.230
saccades. And they're actually the fastest[br]movements that the

0:18:06.230,0:18:09.230
human body can make. The shorter saccades[br]that you

0:18:09.230,0:18:11.020
make, might make when you're reading, last[br]for twenty

0:18:11.020,0:18:13.960
to forty milliseconds. Longer ones that travel[br]through a

0:18:13.960,0:18:16.610
wider arc might take two hundred milliseconds,[br]or about

0:18:16.610,0:18:18.520
a fifth of a second.

0:18:18.520,0:18:20.990
What I find so fascinating about this is that

0:18:20.990,0:18:25.030
we don't perceive saccades. During a saccade,[br]the eye

0:18:25.030,0:18:27.460
is still sending data to the brain, but what

0:18:27.460,0:18:29.010
it's sending is a smeary blur. So the brain

0:18:29.010,0:18:32.680
just edits that part out. This process is[br]called

0:18:32.680,0:18:35.970
saccadic masking. You can see this effect[br]for yourself.

0:18:35.970,0:18:37.880
Next time you're in front of a mirror, lean

0:18:37.880,0:18:40.470
in close and look back and forth from the

0:18:40.470,0:18:43.070
reflection of one eye to the other. You won't

0:18:43.070,0:18:45.520
see your eyes move. As far as we can

0:18:45.520,0:18:48.110
tell, our gaze just jumps instantaneously[br]from one reference

0:18:48.110,0:18:50.220
point to the next. And here's where I have

0:18:50.220,0:18:51.990
to wait for a moment while everybody stops[br]doing

0:18:51.990,0:18:56.900
this.

0:18:56.900,0:18:58.940
When I was preparing for this talk, I found

0:18:58.940,0:19:01.940
an absolutely wonderful sentence in the Wikipedia[br]entry on

0:19:01.940,0:19:06.100
saccades. It said, Due to saccadic masking,[br]the eye/brain

0:19:06.100,0:19:07.810
system not only hides the eye movements from[br]the

0:19:07.810,0:19:10.130
individual, but also hides the evidence that[br]anything has

0:19:10.130,0:19:12.160
been hidden.

0:19:12.160,0:19:17.890
Hides. The evidence. That anything has been[br]hidden. Our

0:19:17.890,0:19:20.100
brains lie to us. And they lie to us

0:19:20.100,0:19:22.610
about having lied to us. And this happens[br]to

0:19:22.610,0:19:26.390
you multiple times a second, every waking[br]hour, every

0:19:26.390,0:19:29.060
day of your life. Of course, there's a reason

0:19:29.060,0:19:30.390
for this.

0:19:30.390,0:19:32.240
Imagine if, every time you shifted your gaze[br]around,

0:19:32.240,0:19:34.200
you got distracted by all the pretty colors.[br]You

0:19:34.200,0:19:36.330
would be eaten by lions.

0:19:36.330,0:19:40.320
But, in selecting for this design, evolution[br]made a

0:19:40.320,0:19:42.680
trade off. The trade off is that we are

0:19:42.680,0:19:45.570
effectively blind every time we move our eyes[br]around.

0:19:45.570,0:19:48.330
Sometimes for up to a fifth of a second.

0:19:48.330,0:19:51.050
And I wanted to talk about this, partly because

0:19:51.050,0:19:53.430
it's a really fun subject, but also to show

0:19:53.430,0:19:55.760
that just one of the ways that our brains

0:19:55.760,0:19:58.500
are doing a massive amount of work to process

0:19:58.500,0:20:01.250
information from our environment and present[br]us with an

0:20:01.250,0:20:03.430
abstraction.

0:20:03.430,0:20:05.980
And as programmers, if we know anything about[br]abstractions,

0:20:05.980,0:20:09.240
it's that they're hard to get right. Which[br]leads

0:20:09.240,0:20:11.860
me to an interesting question. Does it make[br]sense

0:20:11.860,0:20:13.440
to use any of the techniques that I talked

0:20:13.440,0:20:15.900
about in part one, to try to coral different

0:20:15.900,0:20:17.690
parts of our brains into doing our work for

0:20:17.690,0:20:19.690
us, if we don't know what kinds of shortcuts

0:20:19.690,0:20:26.650
they're gonna take?

0:20:26.650,0:20:30.510
According to the Oxford English Dictionary,[br]the word bias

0:20:30.510,0:20:32.580
seems to have entered the English language[br]around the

0:20:32.580,0:20:35.770
1520s. It was used as a, a technical term

0:20:35.770,0:20:38.600
in the game of lawn bowling, and it referred

0:20:38.600,0:20:40.230
to a ball that was constructed in such a

0:20:40.230,0:20:42.650
way that it would curve, it would roll in

0:20:42.650,0:20:44.640
a curved path instead of in a straight line.

0:20:44.640,0:20:47.150
And since then, it's picked up a few additional

0:20:47.150,0:20:49.580
meanings, but they all have that same different[br]connotation

0:20:49.580,0:20:54.440
of something that's skewed or off a little[br]bit.

0:20:54.440,0:20:57.260
Cognitive bias is a term for systematic errors[br]in

0:20:57.260,0:20:59.490
thinking. These are patterns of thought that[br]diverge in

0:20:59.490,0:21:02.860
measurable and predictable ways from what[br]the answers that

0:21:02.860,0:21:07.070
pure rationality might give are. We have some[br]free

0:21:07.070,0:21:09.100
time. I suggest that you go have a look

0:21:09.100,0:21:11.810
at the Wikipedia page called List of cognitive[br]biases.

0:21:11.810,0:21:13.890
There are over a hundred and fifty of them

0:21:13.890,0:21:16.610
and they are fascinating reading.

0:21:16.610,0:21:18.950
And this list of cognitive biases has a lot

0:21:18.950,0:21:20.670
in common with the list of code smells that

0:21:20.670,0:21:22.740
I showed earlier. A lot of these names are

0:21:22.740,0:21:25.160
very literal. But there are a few that stand

0:21:25.160,0:21:28.640
out, like cursive knowledge, or the Google[br]effect or,

0:21:28.640,0:21:32.680
and I kid you not, the Ikea effect. But

0:21:32.680,0:21:34.490
the parallel goes deeper than that.

0:21:34.490,0:21:37.850
This lif - excuse me - this list gives

0:21:37.850,0:21:40.640
names to patterns of dysfunction, and once[br]you have

0:21:40.640,0:21:41.950
a name for a thing, it's a lot easier

0:21:41.950,0:21:43.510
to recognize it and figure out what to do

0:21:43.510,0:21:46.330
about it. I do want to call your attention

0:21:46.330,0:21:48.280
to one particular item on this list. It's[br]called

0:21:48.280,0:21:51.920
the bias blind spot. This is the tendency[br]to

0:21:51.920,0:21:54.990
see oneself as less biased than other people,[br]or

0:21:54.990,0:21:57.150
to be able to identify more cognitive biases[br]in

0:21:57.150,0:22:04.150
others than in oneself. Sound like anybody[br]you know?

0:22:10.820,0:22:17.820
Just let that sink in for a minute. Seriously,

0:22:24.350,0:22:25.730
though.

0:22:25.730,0:22:27.630
In our field, we like to think of ourselves

0:22:27.630,0:22:30.240
as more rational than the average person,[br]and it

0:22:30.240,0:22:33.940
just isn't true. Yes, as programmers, we have[br]a

0:22:33.940,0:22:37.250
valuable, marketable skill that depends on[br]our ability to

0:22:37.250,0:22:40.130
reason mathematically. But we do ourselves[br]and others a

0:22:40.130,0:22:42.280
disservice if we allow ourselves to believe[br]that being

0:22:42.280,0:22:45.260
good at programming means anything other than,[br]we're good

0:22:45.260,0:22:48.570
at programming. Because as humans we are all[br]biased.

0:22:48.570,0:22:51.220
It's built into us, in our DNA. And pretending

0:22:51.220,0:22:53.630
that we aren't biased only allows our biases[br]to

0:22:53.630,0:22:54.780
run free.

0:22:54.780,0:22:57.500
I don't have a lot of general advice for

0:22:57.500,0:22:59.640
how to look for bias, but I think an

0:22:59.640,0:23:02.830
obvious and necessary first step is just to[br]ask

0:23:02.830,0:23:07.510
the question, how is this biased? Beyond that,[br]I

0:23:07.510,0:23:09.520
suggest that you learn about as many specific[br]cognitive

0:23:09.520,0:23:11.230
biases as you can so that your brain can

0:23:11.230,0:23:13.010
do what it does, which is to look for

0:23:13.010,0:23:17.090
patterns and make associations and classify[br]things.

0:23:17.090,0:23:19.190
I think everybody should understand their[br]own biases, because

0:23:19.190,0:23:21.610
only by knowing how you're biased can you[br]then

0:23:21.610,0:23:23.910
decide how to con- how to correct for that

0:23:23.910,0:23:26.610
bias in the decisions that you make. If you're

0:23:26.610,0:23:28.630
not checking your work for bias, you can look

0:23:28.630,0:23:30.630
right past a great solution and you'll never[br]know

0:23:30.630,0:23:32.200
it was there.

0:23:32.200,0:23:35.340
So for part three of my talk, I have

0:23:35.340,0:23:37.860
an example of a solution that is simple, elegant,

0:23:37.860,0:23:41.460
just about the last thing I ever would have

0:23:41.460,0:23:44.690
thought of.

0:23:44.690,0:23:45.850
For the benefit of those of you who have

0:23:45.850,0:23:48.810
yet to find your first gray hair, Pac-Man[br]was

0:23:48.810,0:23:52.420
a video game released in 1980 that let people

0:23:52.420,0:23:54.990
maneuver around a maze eating dots while trying[br]to

0:23:54.990,0:23:58.300
avoid four ghosts. Now, playing games in fun,[br]but

0:23:58.300,0:23:59.960
we're programmers. We want to know how things[br]work.

0:23:59.960,0:24:02.720
So let's talk about programming Pac-Man.

0:24:02.720,0:24:04.450
For the purposes of this discussion, we'll[br]focus on

0:24:04.450,0:24:10.640
just three things. The Pac-Man, the ghosts,[br]and the

0:24:10.640,0:24:11.930
maze. The Pac-Man is controlled by the player.[br]So

0:24:11.930,0:24:14.720
that code is basically just responding to[br]hardware events.

0:24:14.720,0:24:17.550
Boring. The maze is there so that the player

0:24:17.550,0:24:19.780
has some chance at avoiding the ghosts. But[br]the

0:24:19.780,0:24:22.650
ghost AI, that's what's gonna make the game[br]interesting

0:24:22.650,0:24:25.880
enough that people keep dropping quarters[br]into a slot,

0:24:25.880,0:24:27.220
and by the way, video games used to cost

0:24:27.220,0:24:28.400
a quarter.

0:24:28.400,0:24:31.800
When I was your age.

0:24:31.800,0:24:34.470
So to keep things simple, we'll start with[br]one

0:24:34.470,0:24:37.940
ghost. How do we program its movement? We[br]could

0:24:37.940,0:24:40.410
choose a random direction and move that way[br]until

0:24:40.410,0:24:42.090
we hit a wall and then choose another random

0:24:42.090,0:24:44.550
direction. This is very easy to implement,[br]but not

0:24:44.550,0:24:46.220
much of a challenge for the player.

0:24:46.220,0:24:49.580
OK, so, we could compute the distance to the

0:24:49.580,0:24:52.180
Pac-Man in x and y and pick a direction

0:24:52.180,0:24:54.920
that makes one of those smaller. But then[br]the

0:24:54.920,0:24:56.890
ghost is gonna get stuck in corners or behind

0:24:56.890,0:24:58.630
walls cause it won't go around to catch the

0:24:58.630,0:25:00.500
Pac-Man. And, again, it's gonna be too easy[br]for

0:25:00.500,0:25:01.210
the player.

0:25:01.210,0:25:04.370
So how about instead of minimizing linear[br]distance, we

0:25:04.370,0:25:09.110
focus on topological distance? We can compute[br]all possible

0:25:09.110,0:25:12.760
paths through the maze, pick the shortest[br]one that

0:25:12.760,0:25:14.600
gets us to the Pac-Man and then step down

0:25:14.600,0:25:16.520
it. And when we get to the next place,

0:25:16.520,0:25:18.990
we'll do it all again.

0:25:18.990,0:25:21.170
This works find for one ghost. But if all

0:25:21.170,0:25:23.640
four ghosts use this algorithm, then they're[br]gonna wind

0:25:23.640,0:25:25.210
up chasing after the player in a tight little

0:25:25.210,0:25:29.610
bunch instead of fanning out. OK. So each[br]ghost

0:25:29.610,0:25:31.990
computes all possible paths to the Pac-Man[br]and rejects

0:25:31.990,0:25:34.670
any path that goes through another ghost.[br]That shouldn't

0:25:34.670,0:25:37.760
be too hard, right?

0:25:37.760,0:25:41.530
I don't have a statistically valid sample,[br]but my

0:25:41.530,0:25:43.820
guess is that when asked to design an AI

0:25:43.820,0:25:45.730
for the ghosts, most programmers would go[br]through a

0:25:45.730,0:25:47.310
thought process more or less like what I just

0:25:47.310,0:25:51.590
walked through. So, how is this solution biased?

0:25:51.590,0:25:54.420
I don't have a name, a good name for

0:25:54.420,0:25:56.390
how this is biased, so the best way I

0:25:56.390,0:25:58.670
have to communicate this idea is to walk you

0:25:58.670,0:26:00.840
through a very different solution.

0:26:00.840,0:26:04.560
In 2006, I attended Oopsla, this is a conference

0:26:04.560,0:26:07.840
put on by the ACM, as a student volunteer,

0:26:07.840,0:26:09.620
and I happened to sit in on a presentation

0:26:09.620,0:26:14.270
by Alexander Repenning from the University[br]of Colorado. And

0:26:14.270,0:26:18.060
in his presentation, he walked through the[br]Pac-Man problem,

0:26:18.060,0:26:19.980
more or less the way I just did, and

0:26:19.980,0:26:21.140
then he presented this idea.

0:26:21.140,0:26:23.870
What you do is you give the Pac-Man a

0:26:23.870,0:26:25.970
smell, and then you model the diffusion of[br]that

0:26:25.970,0:26:31.340
smell throughout the environment. In the real[br]world, smells

0:26:31.340,0:26:33.640
travel through the air. We certainly don't[br]need to

0:26:33.640,0:26:36.440
model each individual air molecule. What we[br]can do,

0:26:36.440,0:26:39.180
instead, is just divide the environment up[br]into reasonably

0:26:39.180,0:26:42.260
sized logical chunks, and we model those.

0:26:42.260,0:26:44.760
Coincidentally, we already do have an object[br]that does

0:26:44.760,0:26:47.120
exactly that for us. It's the tiles of the

0:26:47.120,0:26:49.470
maze itself. They're not really doing anything[br]else, so

0:26:49.470,0:26:51.620
we can borrow those as a convenient container[br]for

0:26:51.620,0:26:56.360
this computation. We program the game as follows.

0:26:56.360,0:26:58.770
We say that the Pac-Man gives whatever floor[br]tile

0:26:58.770,0:27:02.030
it's standing on a Pac-Man smell value, say[br]a

0:27:02.030,0:27:05.620
thousand. The number doesn't really matter.[br]And that, that

0:27:05.620,0:27:08.030
tile then passes a smaller value off to each

0:27:08.030,0:27:10.270
of its neighbors, and they pass a smaller[br]value

0:27:10.270,0:27:12.520
off to each of their neighbors and so on.

0:27:12.520,0:27:13.750
Iterate this a few times and you get a

0:27:13.750,0:27:16.830
diffusion contour that we can visualize as[br]a hill

0:27:16.830,0:27:18.600
with its peak centered on the Pac-Man.

0:27:18.600,0:27:21.380
It's a little hard to see here. The Pac-Man

0:27:21.380,0:27:23.110
is at the bottom of that big yellow bar

0:27:23.110,0:27:30.110
on the left. So we've got the Pac-Man. We've

0:27:34.140,0:27:37.080
got the floor tiles. But in order to make

0:27:37.080,0:27:38.550
it a maze, we also have to have some

0:27:38.550,0:27:41.110
walls. What we do is we give the walls

0:27:41.110,0:27:43.770
a Pac-Man smell value of zero. That chops[br]the

0:27:43.770,0:27:50.770
hill up a bit.

0:27:56.050,0:27:58.210
And now all our ghost has to do is

0:27:58.210,0:28:03.559
climb the hill. We program the first ghost[br]to

0:28:03.559,0:28:06.650
sample each of the floor tiles next to it,

0:28:06.650,0:28:08.800
pick the one with the biggest number, go that

0:28:08.800,0:28:12.580
way. It barely seems worthy of being called[br]an

0:28:12.580,0:28:16.590
AI, does it? But check this out. When we

0:28:16.590,0:28:18.280
add more ghosts to the maze, we only have

0:28:18.280,0:28:20.750
to make one change to get them to cooperate.

0:28:20.750,0:28:23.360
And interestingly, we don't change the ghosts'[br]movement behaviors

0:28:23.360,0:28:26.380
at all. Instead, we have the ghosts tell the

0:28:26.380,0:28:29.390
floor tile that they're, I guess, floating[br]above, that

0:28:29.390,0:28:33.130
its Pac-Men smell value is zero. This changes[br]the

0:28:33.130,0:28:35.680
shape of that diffusion contour. Instead of[br]a smooth

0:28:35.680,0:28:38.320
hill that always slopes down away from the[br]Pac-Man,

0:28:38.320,0:28:40.030
there are now cliffs where the hill drops[br]immediately

0:28:40.030,0:28:40.780
to zero.

0:28:40.780,0:28:45.030
In effect, we turn the ghosts into movable[br]walls,

0:28:45.030,0:28:46.650
so that when one ghost cuts off another one,

0:28:46.650,0:28:49.559
the second one will automatically choose a[br]different route.

0:28:49.559,0:28:52.990
This lets the ghosts cooperate without needing[br]to be

0:28:52.990,0:28:56.750
aware of each other. And halfway through this[br]conference

0:28:56.750,0:28:58.270
session that I was sitting in where I saw

0:28:58.270,0:29:00.820
this, I was like.

0:29:00.820,0:29:04.650
What just happened?

0:29:04.650,0:29:07.220
At first, like, my first level of surprise[br]was

0:29:07.220,0:29:09.970
just, what an interesting approach. But then[br]I got

0:29:09.970,0:29:13.330
really, completely stunned when I thought[br]about how surprising

0:29:13.330,0:29:16.760
that solution was. And I hope that looking[br]at

0:29:16.760,0:29:19.520
the second solution helps you understand the[br]bias in

0:29:19.520,0:29:21.809
the first solution.

0:29:21.809,0:29:24.940
In his paper, Professor Repenning wrote, The[br]challenge to

0:29:24.940,0:29:27.520
find this solution is a psychological, not[br]a technical

0:29:27.520,0:29:30.930
one. Our first instinct, when we're presented[br]with this

0:29:30.930,0:29:33.860
problem, is to imagine ourselves as the ghost.[br]This

0:29:33.860,0:29:36.920
is the body syntonic reasoning that's built[br]into Logo,

0:29:36.920,0:29:39.080
and in this case it's a trap.

0:29:39.080,0:29:40.780
Because it leads us to solve the pursuit problem

0:29:40.780,0:29:44.730
by making the pursuer smarter. Once we started[br]down

0:29:44.730,0:29:48.210
that road, it's very unlikely that we're going[br]to

0:29:48.210,0:29:51.600
consider a radically different approach, even,[br]or, perhaps, especially

0:29:51.600,0:29:55.740
if it's a very much simpler one. In other

0:29:55.740,0:29:59.020
words, body syntonicity biases us towards[br]modeling objects in

0:29:59.020,0:30:03.830
the foreground, rather than objects in the[br]background.

0:30:03.830,0:30:06.970
Oops, sorry.

0:30:06.970,0:30:10.600
OK. Does this mean that you shouldn't use[br]body

0:30:10.600,0:30:12.780
syntonic reasoning? Of course not. It's a[br]tool. It's

0:30:12.780,0:30:15.430
right for some jobs. It's not right for others.

0:30:15.430,0:30:16.910
I want to take a look at one more

0:30:16.910,0:30:21.110
technique from part one. What's the bias in[br]Please

0:30:21.110,0:30:23.590
Mr. Gear, what is your ratio? Aside from the

0:30:23.590,0:30:27.770
gendered language, which is trivially easy[br]to address. This

0:30:27.770,0:30:31.110
technique is explicitly designed to give you[br]an opportunity

0:30:31.110,0:30:34.120
to discover new objects in your model. But[br]it

0:30:34.120,0:30:35.650
only works after you've given at least one[br]of

0:30:35.650,0:30:38.050
those objects a name.

0:30:38.050,0:30:42.180
Names have gravity. Metaphors can be tar pits.[br]It's

0:30:42.180,0:30:44.620
very likely that the new objects that you[br]discover

0:30:44.620,0:30:46.860
are going to be fairly closely related to[br]the

0:30:46.860,0:30:51.390
ones that you already have. Another way to[br]help

0:30:51.390,0:30:53.420
see this is to think about how many steps

0:30:53.420,0:30:56.610
it takes to get from Please, Ms. Pac-Man,[br]what

0:30:56.610,0:30:59.730
is your current position in the maze? To,[br]please

0:30:59.730,0:31:01.610
Ms. Floor Tile, how much do you smell like

0:31:01.610,0:31:04.540
Pac-Man?

0:31:04.540,0:31:05.880
For a lot of people, the answer to that

0:31:05.880,0:31:09.660
question is probably infinity. It certainly[br]was for me.

0:31:09.660,0:31:11.390
My guess is that you don't come up with

0:31:11.390,0:31:14.400
this technique unless you've already done[br]some work modeling

0:31:14.400,0:31:18.950
diffusion in another context. Which, incidentally,[br]is why I

0:31:18.950,0:31:21.840
like to work on diverse teams. The more different

0:31:21.840,0:31:24.730
backgrounds and perspectives we have access[br]to, the more

0:31:24.730,0:31:27.520
chances we have to find a novel application[br]of

0:31:27.520,0:31:31.720
some seemingly unrelated technique, because[br]somebody's worked with it

0:31:31.720,0:31:33.370
before.

0:31:33.370,0:31:37.600
It can be exhilarating and very empowering[br]to find

0:31:37.600,0:31:39.900
these techniques that let us take shortcuts[br]in our

0:31:39.900,0:31:43.929
work by leveraging these specialized structures[br]in our brains.

0:31:43.929,0:31:46.890
But those structures themselves take shortcuts,[br]and if you're

0:31:46.890,0:31:48.890
not careful, they can lead you down a primrose

0:31:48.890,0:31:49.290
path.

0:31:49.290,0:31:51.550
I want to go back to that quote that

0:31:51.550,0:31:53.309
got me thinking about all this in the first

0:31:53.309,0:31:56.080
place. About how we don't ever teach anybody[br]how

0:31:56.080,0:31:58.510
to do that or even that they should.

0:31:58.510,0:32:00.920
Ultimately, I think we should use techniques[br]like this,

0:32:00.920,0:32:03.090
despite the, the biases in them. I think we

0:32:03.090,0:32:06.650
should share them. And I think, to paraphrase[br]Glenn,

0:32:06.650,0:32:08.059
we should teach people that this is a thing

0:32:08.059,0:32:11.980
that you can and should do. And, I think

0:32:11.980,0:32:14.240
that we should teach people that looking critically[br]at

0:32:14.240,0:32:16.960
the answers that these techniques give you[br]is also

0:32:16.960,0:32:19.490
a thing that you can and should do.

0:32:19.490,0:32:20.640
We might not always be able to come up

0:32:20.640,0:32:23.960
with a radically simpler or different approach,[br]but the

0:32:23.960,0:32:27.059
least we can do is give ourselves the opportunity

0:32:27.059,0:32:33.050
to do so, by asking how is this biased?

0:32:33.050,0:32:34.690
I want to say thank you, real quickly, to

0:32:34.690,0:32:36.490
everybody who helped me with this talk, or[br]the

0:32:36.490,0:32:39.120
ideas in it. And also thank you to LivingSocial

0:32:39.120,0:32:41.820
for paying for my trip. And also for bringing

0:32:41.820,0:32:45.790
these wonderful brains. So, they're gonna[br]start tearing this

0:32:45.790,0:32:48.630
stage down in a few minutes. Rather than take

0:32:48.630,0:32:50.050
Q and A up here, I'm gonna pack up

0:32:50.050,0:32:52.380
all my stuff and then I'm gonna migrate over

0:32:52.380,0:32:54.570
there, and you can come and bug me. pick

0:32:54.570,0:32:55.929
up a brain. Whatever.

0:32:55.929,0:32:56.790
Thank you.