0:00:17.160,0:00:18.060
JULIAN SIMIONI: All right. Everyone can hear[br]me well now?

0:00:18.060,0:00:19.279
They've turned it off. Everything's good.

0:00:19.279,0:00:21.750
All right. It says 3:50 on the little clock

0:00:21.750,0:00:23.240
here which means it's time for me to go.

0:00:23.240,0:00:26.610
So, hi everyone. I'm Julian. This is my first

0:00:26.610,0:00:29.990
RailsConf. It's great to be here. We are not

0:00:29.990,0:00:31.500
gonna talk about Rails at all. We're gonna[br]talk

0:00:31.500,0:00:34.720
about space. This is one of my favorite pictures

0:00:34.720,0:00:36.800
ever, right. It's the, the Earth rise picture[br]from

0:00:36.800,0:00:39.920
Apollo 8.

0:00:39.920,0:00:43.370
And my whole life, I've just loved space.[br]From

0:00:43.370,0:00:44.690
when I was a little kid, it just inspired

0:00:44.690,0:00:46.860
me so much. I know it inspires tons of

0:00:46.860,0:00:48.580
other people. I know, I got into a technical

0:00:48.580,0:00:51.180
field because, as a kid, I loved space. Raise

0:00:51.180,0:00:52.969
your hand if you think it's the same thing

0:00:52.969,0:00:55.250
for you. I know I've talked to so many

0:00:55.250,0:00:56.809
people where that's the case. Awesome.

0:00:56.809,0:00:59.719
And I mean, just, space is really cool, right.

0:00:59.719,0:01:03.430
There is, just, you know, everything about[br]astronomy or

0:01:03.430,0:01:07.280
manned space explorations is just really,[br]really awesome and

0:01:07.280,0:01:12.009
inspiring. And there's one thing, above all[br]else, that

0:01:12.009,0:01:14.759
inspires people just so much, and that, of[br]course,

0:01:14.759,0:01:17.869
is the Apollo program.

0:01:17.869,0:01:20.950
Landing on the moon is, maybe, the greatest[br]thing

0:01:20.950,0:01:24.970
that humans have ever done. And, and, you[br]know,

0:01:24.970,0:01:27.080
people will talk about awesome rockets, right.[br]Like, the

0:01:27.080,0:01:30.300
Saturn 5. It generates seven and a half million

0:01:30.300,0:01:33.530
pounds of thrust on takeoff. That's. That's[br]a lot.

0:01:33.530,0:01:36.830
Or they'll talk about, you know, astronauts[br]who, who

0:01:36.830,0:01:39.830
trained for years, and, you know, were brave[br]and

0:01:39.830,0:01:41.700
took insane risks that, you know, most of[br]us

0:01:41.700,0:01:44.929
can't even comprehend. Or they'll talk about[br]mathematicians and

0:01:44.929,0:01:47.530
awesome things they did for, like, the guidance[br]equations

0:01:47.530,0:01:49.020
and stuff to get people to the moon safely

0:01:49.020,0:01:49.950
and back.

0:01:49.950,0:01:53.270
And then they'll say things, like, oh, yeah,[br]we,

0:01:53.270,0:01:55.080
we landed on the moon with, you know, less

0:01:55.080,0:01:58.429
computing power in a wrist watch, or something.[br]And

0:01:58.429,0:02:01.330
as a software developer, you know, that makes[br]me

0:02:01.330,0:02:03.270
sad. You know, for a long time, I didn't

0:02:03.270,0:02:05.050
really think that there was anything in the[br]Apollo

0:02:05.050,0:02:09.610
program to inspire us software developers.[br]And, fortunately, it

0:02:09.610,0:02:11.880
turns out, that is not at all the case.

0:02:11.880,0:02:15.470
So this is the Apollo guidance computer. To[br]give

0:02:15.470,0:02:17.950
you a sense of the scale, the little display

0:02:17.950,0:02:19.040
is, like, a little bit bigger than your hand.

0:02:19.040,0:02:21.860
I think it's like, nine by nine inches. So

0:02:21.860,0:02:23.360
this thing was built to take people to the

0:02:23.360,0:02:29.590
moon. And it is awesome in ways you might

0:02:29.590,0:02:29.870
not expect.

0:02:29.870,0:02:33.320
So it only had two kilobytes of RAM, I

0:02:33.320,0:02:36.370
think. Not very much. It was really, really,[br]really

0:02:36.370,0:02:41.810
slow. I did some back-of-the-envelope calculations,[br]and it turns

0:02:41.810,0:02:45.340
out all of the computation done on all of

0:02:45.340,0:02:47.840
the Apollo missions, ever, could be done by[br]my

0:02:47.840,0:02:53.260
little laptop over there in under a second.

0:02:53.260,0:02:57.360
But, my laptop may or may not make it

0:02:57.360,0:03:02.760
through this talk. The Apollo guidance computer[br]never hit

0:03:02.760,0:03:06.230
an unknown bug, in thousands of hours of time

0:03:06.230,0:03:08.770
in space. It never failed, despite the fact[br]that

0:03:08.770,0:03:15.760
space is pretty hostile, right. There's, there's[br]like gamma

0:03:15.760,0:03:19.430
rays. Anything metallic that floats around[br]will, will find

0:03:19.430,0:03:21.060
the one place where it can short things out

0:03:21.060,0:03:24.840
and, and cause lots of problems.

0:03:24.840,0:03:26.540
And so this thing is pretty amazing. It worked

0:03:26.540,0:03:29.860
really, really well. But what's even more[br]amazing is

0:03:29.860,0:03:32.510
that it was built at a time when we

0:03:32.510,0:03:36.860
didn't know much about making software. They[br]started working

0:03:36.860,0:03:41.030
on it in 1961. And they finished in 1968.

0:03:41.030,0:03:46.980
So to put that into perspective, in 1965,[br]the

0:03:46.980,0:03:50.819
MIT instrumentation lab, that's the department[br]of MIT that

0:03:50.819,0:03:53.860
was building this thing, they got this new[br]computer.

0:03:53.860,0:03:56.410
It's called an IBM system 360. Anyone heard[br]of

0:03:56.410,0:03:59.709
that before? It's one of the more famous computers

0:03:59.709,0:04:01.170
in the history of computing.

0:04:01.170,0:04:03.810
So there's one particularly famous person[br]that worked on

0:04:03.810,0:04:05.810
it. And anyone know who I'm talking about?[br]Anyone

0:04:05.810,0:04:08.750
have an idea? Come on, you all know. Fred

0:04:08.750,0:04:11.140
Brooks, right. And he, he wrote this book.[br]What

0:04:11.140,0:04:14.660
was the book called? Anyone? The Mythical[br]Man Month.

0:04:14.660,0:04:17.720
Yeah. So when did he write that book? Not

0:04:17.720,0:04:20.630
until 1975. And by then, you know, we had

0:04:20.630,0:04:24.470
been gone from the moon for many, many years.

0:04:24.470,0:04:27.240
So think about what that means, right. The,[br]the

0:04:27.240,0:04:31.289
software developers writing code for the Apollo[br]Guidance Computer

0:04:31.289,0:04:34.600
were learning the same things as Fred Brooks[br]at

0:04:34.600,0:04:37.150
the same time, right. So if you're a manager

0:04:37.150,0:04:39.000
at a, you know, you have some software developers

0:04:39.000,0:04:40.210
working under you, and your boss comes to[br]you

0:04:40.210,0:04:42.160
and says hey, you know, your projects a little

0:04:42.160,0:04:45.020
bit late. I'm gonna, you know, give you some

0:04:45.020,0:04:48.990
more developers to hopefully speed it up,[br]you say,

0:04:48.990,0:04:52.160
OK. Whereas today you might say, no. Read[br]The

0:04:52.160,0:04:55.000
Mythical Man Month, and you know that that[br]won't

0:04:55.000,0:04:55.700
work.

0:04:55.700,0:05:00.690
So, another thing. The word software was first[br]used

0:05:00.690,0:05:03.850
in 1950. It was used in, like, a research

0:05:03.850,0:05:09.220
paper. And there's some back story here. It's[br]pretty

0:05:09.220,0:05:11.180
important. We're gonna do a little aside.

0:05:11.180,0:05:15.460
So, the first computers in, in the, you know,

0:05:15.460,0:05:19.790
20s, 30s, 40s. They were built by men. Like,

0:05:19.790,0:05:25.060
predominantly, overwhelmingly men. And, and[br]they would build these

0:05:25.060,0:05:28.470
computers, and the computers did one thing,[br]right. You

0:05:28.470,0:05:31.710
would build a computer to, you know, calculate[br]the

0:05:31.710,0:05:33.490
Fibonacci sequence or something, and if you[br]wanted a

0:05:33.490,0:05:37.210
computer to, to calculate prime numbers, like,[br]you built

0:05:37.210,0:05:39.240
another computer, or you took the computer[br]you had

0:05:39.240,0:05:42.560
and, and modified it really heavily, right.[br]And over

0:05:42.560,0:05:44.900
time, you know, computers got a little bit[br]better

0:05:44.900,0:05:46.650
and, and then they could sort of be configured,

0:05:46.650,0:05:49.270
right. They could, you could like, flip some[br]switches

0:05:49.270,0:05:50.930
and set some bits and then the computers would

0:05:50.930,0:05:56.120
do different things and, and, and the men[br]building

0:05:56.120,0:05:58.130
these computers that, they hated this task,[br]right. They

0:05:58.130,0:06:01.840
thought it was boring and that it was easy

0:06:01.840,0:06:04.060
and that it wasn't very interesting at all.[br]So

0:06:04.060,0:06:07.169
they hired a bunch of women to do all

0:06:07.169,0:06:10.430
these things, right. And only later, like,[br]as time

0:06:10.430,0:06:13.210
goes on, did they figure out that, that this

0:06:13.210,0:06:17.050
task of configuring the computer that, that[br]the men

0:06:17.050,0:06:19.389
had built was actually the entire field of[br]software

0:06:19.389,0:06:23.410
development, and that it was actually really,[br]really hard

0:06:23.410,0:06:26.949
and challenging and it was a field onto itself.

0:06:26.949,0:06:29.790
And only then did, did men start coming in

0:06:29.790,0:06:33.710
and start saying, hey, I would like to do

0:06:33.710,0:06:36.500
this. During the Apollo program, one of the[br]managers

0:06:36.500,0:06:40.790
of, of one of the software teams, when he

0:06:40.790,0:06:44.020
was assigned to that team, his wife was told

0:06:44.020,0:06:45.960
by him, hey, honey, don't tell any of our

0:06:45.960,0:06:48.070
friends that I manage a software team. Cause[br]it

0:06:48.070,0:06:51.220
wasn't, it wasn't cool for a guy to write

0:06:51.220,0:06:54.080
software, yet, in the, in the 60s. It was,

0:06:54.080,0:06:55.030
it was getting there.

0:06:55.030,0:06:59.100
So, it's kind of interesting, right. So here's[br]my

0:06:59.100,0:07:02.180
takeaway from that. If you're a woman and[br]you're

0:07:02.180,0:07:05.590
writing some software, and you're maybe thinking,[br]hey, you

0:07:05.590,0:07:07.259
know, maybe I'm not cut out for this or,

0:07:07.259,0:07:08.900
or some other people are saying things that[br]say

0:07:08.900,0:07:11.500
that I shouldn't be here. Don't listen to[br]them.

0:07:11.500,0:07:15.570
That's bull shit. You're grandma was writing[br]code. And,

0:07:15.570,0:07:17.400
and she did a damn fine job. So don't

0:07:17.400,0:07:17.840
worry about that.

0:07:17.840,0:07:19.930
And if you're a guy, a man, or woman

0:07:19.930,0:07:22.710
I guess, and you're thinking or saying or[br]writing

0:07:22.710,0:07:25.900
things like, you know, women shouldn't be[br]here. They're

0:07:25.900,0:07:29.030
not as good at writing code as men. Stop.

0:07:29.030,0:07:31.600
Stop saying those things. That is all bull[br]shit.

0:07:31.600,0:07:34.840
It, that does not stand up to the history

0:07:34.840,0:07:38.150
of computing. So. Just a little aside there.

0:07:38.150,0:07:42.110
Anyways, OK. We're talking about the Apollo[br]Guidance Computer,

0:07:42.110,0:07:45.979
and things that people learned from that.[br]So, what,

0:07:45.979,0:07:48.690
let me ask you guys a question. How many

0:07:48.690,0:07:51.750
of you have built some sort of web app,

0:07:51.750,0:07:54.900
maybe a Rails app or something, and you depend

0:07:54.900,0:07:58.199
on someone else to service. Right, an API[br]or

0:07:58.199,0:07:59.780
something like that. Someone, something that[br]another team at

0:07:59.780,0:08:01.400
your company built or something external,[br]right. And what

0:08:01.400,0:08:03.620
happened to your site the first time that[br]thing

0:08:03.620,0:08:08.949
went down? Yeah, I heard a very sad noise

0:08:08.949,0:08:10.590
of something breaking.

0:08:10.590,0:08:13.130
Yeah. Exactly. Your pager goes off. Although[br]maybe in

0:08:13.130,0:08:16.509
the 60s, I don't think they even had pagers.

0:08:16.509,0:08:19.389
So what did you do after awhile, right? You,

0:08:19.389,0:08:22.009
you, the first time you fixed that one thing,

0:08:22.009,0:08:23.949
right, you fixed it so that that particular[br]thing

0:08:23.949,0:08:28.280
happens again, you know, it'll be OK. And[br]then

0:08:28.280,0:08:30.740
something else happens and something else[br]happens, and hopefully

0:08:30.740,0:08:36.589
after awhile you, your, your system is smart[br]enough

0:08:36.589,0:08:40.429
that anything that this, this other system[br]throws at

0:08:40.429,0:08:44.430
you, something you've never even seen before,[br]you know,

0:08:44.430,0:08:48.860
it doesn't matter. You'll keep running just[br]fine.

0:08:48.860,0:08:51.379
And so the same thing happened during the[br]Apollo

0:08:51.379,0:08:54.869
11 landing. So this is the Lunar module. So,

0:08:54.869,0:08:57.769
so picture, this is the most critical moment[br]of

0:08:57.769,0:08:59.970
the entire Apollo space program, right. Neil[br]Armstrong and

0:08:59.970,0:09:02.449
Buzz Aldrin are, are in this little thing.[br]They're

0:09:02.449,0:09:05.589
heading down towards the surface of the moon.[br]And

0:09:05.589,0:09:07.059
this, the last ten minutes of this are by

0:09:07.059,0:09:11.829
far the most, absolutely the most critical[br]and, and

0:09:11.829,0:09:14.759
high intensity part of the entire, of the[br]entire

0:09:14.759,0:09:15.649
mission.

0:09:15.649,0:09:18.769
Neil Armstrong said that those last couple[br]of minutes,

0:09:18.769,0:09:20.879
on a difficulty scale of one to ten were

0:09:20.879,0:09:22.740
a thirteen, right. And he's, he's a pretty[br]good

0:09:22.740,0:09:24.970
pilot. So that means a lot.

0:09:24.970,0:09:28.230
So, so, so they're cruising down towards,[br]towards the

0:09:28.230,0:09:31.279
surface. They're, they're busy. They're in[br]the zone. All

0:09:31.279,0:09:33.740
of the sudden, on their dashboard, you know,[br]a

0:09:33.740,0:09:37.079
big light comes on. The master alarm. Everything[br]is

0:09:37.079,0:09:40.670
wrong. And their computer is, is giving them[br]this

0:09:40.670,0:09:42.579
error, this, like, twelve-oh-one error.

0:09:42.579,0:09:45.459
And, you know, the astronauts are distracted[br]and mission

0:09:45.459,0:09:47.339
control, they're freaking out. They're trying[br]to figure out

0:09:47.339,0:09:50.679
what this is. You know, what does this mean?

0:09:50.679,0:09:53.410
One of the MIT engineers in, in mission control

0:09:53.410,0:09:56.879
had, very wisely, made a little cheat sheet[br]of

0:09:56.879,0:10:00.350
every error code that the computer could throw[br]up.

0:10:00.350,0:10:02.999
Very, very smart.

0:10:02.999,0:10:05.790
And he, you know, he, he looks up, OK,

0:10:05.790,0:10:08.009
twelve-oh-one. What does it say? It says,[br]if, if

0:10:08.009,0:10:09.389
nothing else seems wrong, you know, you're[br]good to

0:10:09.389,0:10:13.610
go for the landing, and he actually was too

0:10:13.610,0:10:16.199
scared to form words, so when he was asked,

0:10:16.199,0:10:17.110
you know, are we good to go? He just

0:10:17.110,0:10:19.970
sort of like, made a thumbs up. And then

0:10:19.970,0:10:21.709
he was awarded a medal from the President[br]of

0:10:21.709,0:10:24.129
the United States for his, for his decision[br]to,

0:10:24.129,0:10:26.119
to not abort the moon landing.

0:10:26.119,0:10:28.300
So what happened? What was this master alarm[br]that

0:10:28.300,0:10:31.179
scared the crap out of everyone for actually[br]no

0:10:31.179,0:10:36.540
reason? So, today we're pretty lucky with[br]our computers,

0:10:36.540,0:10:39.569
right. If you, if you write some code that

0:10:39.569,0:10:41.249
does an infinite loop or something, your computer[br]will

0:10:41.249,0:10:43.629
probably handle it just fine, right. You have[br]to,

0:10:43.629,0:10:45.779
you have to do something pretty crazy to crash

0:10:45.779,0:10:48.619
your computer.

0:10:48.619,0:10:52.739
The very idea that multiple programs could[br]even be

0:10:52.739,0:10:54.339
running on a computer was new in the 90s,

0:10:54.339,0:10:57.480
1960s. They, they didn't have words like scheduling[br]or

0:10:57.480,0:11:01.759
tasks or processes like we have today. They[br]called

0:11:01.759,0:11:07.480
it, they called it, like, like, timesharing,[br]right. And,

0:11:07.480,0:11:10.879
and up until the Apollo, Apollo guidance computer[br]had

0:11:10.879,0:11:14.839
been built, they didn't do anything sophisticated[br]like, like

0:11:14.839,0:11:17.579
have priorities assigned to different tasks.[br]They just said,

0:11:17.579,0:11:18.119
anything that wants to run, we'll let it run

0:11:18.119,0:11:19.999
for a little while, and then we'll run something

0:11:19.999,0:11:21.459
else.

0:11:21.459,0:11:24.009
And a lot of engineers thought that this approach

0:11:24.009,0:11:25.929
was safe, because it was more predictable,[br]and it

0:11:25.929,0:11:28.369
was easier to reason about. But it has a

0:11:28.369,0:11:34.670
downside, in that you, with something like[br]priority-based scheduling,

0:11:34.670,0:11:39.249
you can, you can figure out what are the

0:11:39.249,0:11:41.119
most important things that need to be running,[br]right.

0:11:41.119,0:11:43.249
The program to fire the thrusters should probably[br]run

0:11:43.249,0:11:45.540
exactly when it wants. The program to update[br]the

0:11:45.540,0:11:47.670
display for the astronauts, if it's a couple[br]milliseconds

0:11:47.670,0:11:49.339
late, I think they'll be OK.

0:11:49.339,0:11:51.230
And so this was a brand new approach for

0:11:51.230,0:11:54.139
the Apollo program, or for computing in general.[br]So,

0:11:54.139,0:11:58.360
the takeaway here is handle failure gracefully.[br]Just do

0:11:58.360,0:12:00.470
a little bit of actual work so that something

0:12:00.470,0:12:03.019
unexpected can come up and you'll handle it[br]just

0:12:03.019,0:12:03.899
fine.

0:12:03.899,0:12:05.559
Step two of that, by the way, is if

0:12:05.559,0:12:07.569
you get to the point where you can handle

0:12:07.569,0:12:11.259
an unknown failure gracefully, don't throw,[br]like a master

0:12:11.259,0:12:15.939
caution alarm and raise a ruckus when, you[br]know,

0:12:15.939,0:12:17.860
you're actually handling everything great.

0:12:17.860,0:12:21.519
OK. So, who would have ever thought that in

0:12:21.519,0:12:26.049
a RailsConf, talking about testing would be,[br]would be

0:12:26.049,0:12:28.350
weird? But I have an entire segment about[br]testing.

0:12:28.350,0:12:30.970
Thanks DHH for that.

0:12:30.970,0:12:36.389
So, there were lives on the line. They're[br]traveling

0:12:36.389,0:12:37.899
into space, which is pretty difficult. There[br]was a

0:12:37.899,0:12:39.889
ton of testing for the software, for the software

0:12:39.889,0:12:44.720
for the Apollo Guidance Computer. They, they[br]had unit

0:12:44.720,0:12:46.290
tests. They actually even called them unit[br]tests but

0:12:46.290,0:12:48.329
with, with quotes, cause it was like a brand

0:12:48.329,0:12:50.189
new word for them.

0:12:50.189,0:12:53.670
But the, the engineers didn't really talk[br]much about

0:12:53.670,0:12:55.660
their unit tests. They said they were OK,[br]but

0:12:55.660,0:12:59.670
they didn't, like, rant and rave about them.[br]And

0:12:59.670,0:13:00.869
I have a theory as to why that is,

0:13:00.869,0:13:03.720
and that's because most of the code for the

0:13:03.720,0:13:08.179
Apollo Guidance Computer was, was, was like[br]functional code.

0:13:08.179,0:13:09.929
Not in like the programming sense but al,[br]also

0:13:09.929,0:13:12.480
in the mathematical sense on which it's based.[br]Right?

0:13:12.480,0:13:16.519
You have, you have code for guidance equations.[br]And

0:13:16.519,0:13:18.839
furthermore you have mathematicians writing[br]them.

0:13:18.839,0:13:21.069
The idea of a computer science degree did[br]not

0:13:21.069,0:13:23.149
exist in the 1960s. I think the last, the

0:13:23.149,0:13:25.689
first computer science degree was actually[br]awarded, like, in

0:13:25.689,0:13:28.959
the late 60s. So no one writing these computers

0:13:28.959,0:13:31.850
was trained as a software programmer, right,[br]or as

0:13:31.850,0:13:34.929
a computer scientist. They were trained as[br]mathematicians. And

0:13:34.929,0:13:37.449
they were writing code to implement mathematical[br]equations. And,

0:13:37.449,0:13:40.420
as you can imagine, eventually they'll get[br]it right.

0:13:40.420,0:13:42.689
So, what they do talk really, really highly[br]about

0:13:42.689,0:13:45.079
is their, their integration tests. And now,[br]I have

0:13:45.079,0:13:46.749
no idea. Should we call these system tests[br]or

0:13:46.749,0:13:50.970
some other kind of tests? But anyways.

0:13:50.970,0:13:55.459
In 1972, they published a retrospective on[br]basically the

0:13:55.459,0:13:59.589
entire software development effort, and they[br]said things. They,

0:13:59.589,0:14:03.470
they talked specifically about things that[br]plagued the software

0:14:03.470,0:14:05.639
development effort throughout the entire program.[br]And they listed

0:14:05.639,0:14:07.739
two things, and they're both really interesting.[br]I have

0:14:07.739,0:14:09.109
a link to it at the end of my

0:14:09.109,0:14:11.670
slide, so definitely check it out when you[br]get

0:14:11.670,0:14:11.970
a chance.

0:14:11.970,0:14:13.850
The first thing that they just could not get

0:14:13.850,0:14:18.480
right was estimating schedules. And we don't[br]get that

0:14:18.480,0:14:19.980
right today so, great. We've made a lot of

0:14:19.980,0:14:22.379
progress there. One cool thing they did actually[br]was

0:14:22.379,0:14:24.189
they, they cataloged a bunch of ways in which

0:14:24.189,0:14:25.509
they can go wrong, right. Like if you have,

0:14:25.509,0:14:30.329
if you have two projects competing, or two[br]projects

0:14:30.329,0:14:33.910
that are running in parallel, and one of them

0:14:33.910,0:14:37.769
is gonna take longer, and you make the one

0:14:37.769,0:14:40.569
that is not gonna take as long happen even

0:14:40.569,0:14:43.980
faster, you haven't actually made your whole[br]project progress

0:14:43.980,0:14:47.730
any faster, right. Just to give an example.

0:14:47.730,0:14:50.809
The other thing they talked about, that plagued[br]their

0:14:50.809,0:14:55.420
whole project, was getting up to date specs[br]and

0:14:55.420,0:15:01.149
requirements and information about different[br]systems. The Apollo program,

0:15:01.149,0:15:02.999
you know, was in a rush throughout its entire

0:15:02.999,0:15:05.970
development. Everything, every part of every[br]system in the

0:15:05.970,0:15:09.989
space craft was constantly changing. And getting[br]up to

0:15:09.989,0:15:12.629
date information about these other things[br]is hard, right.

0:15:12.629,0:15:13.939
You have to go out and talk to every

0:15:13.939,0:15:16.509
team, and if they don't happen to tell you

0:15:16.509,0:15:19.009
that something's changing, you might not know.

0:15:19.009,0:15:21.949
Again, they didn't come up with a great way

0:15:21.949,0:15:24.720
to solve this, and again we still struggle[br]with

0:15:24.720,0:15:26.359
this today. I think communication is the hardest[br]part

0:15:26.359,0:15:29.279
of software development. But they did find[br]out a

0:15:29.279,0:15:32.129
great way to figure out if you have it

0:15:32.129,0:15:34.249
wrong. And that's to write integration tests,[br]where you

0:15:34.249,0:15:37.299
take everything together, right. Every part[br]of your system,

0:15:37.299,0:15:38.720
and you test the hell out of it and

0:15:38.720,0:15:40.709
see what breaks.

0:15:40.709,0:15:47.709
And, and, and, so. What happens is, there's[br]a

0:15:49.629,0:15:52.899
really interesting thing where your unit tests[br]with test

0:15:52.899,0:15:54.949
your own code, and then you have integration[br]tests,

0:15:54.949,0:15:57.799
which are also code. But they're not there[br]so

0:15:57.799,0:16:01.109
much, they're not there so much to test your

0:16:01.109,0:16:04.919
code as they are to test your communication.[br]So

0:16:04.919,0:16:11.470
to put it another way, if you. How do

0:16:11.470,0:16:14.470
I put this? Let's see.

0:16:14.470,0:16:17.350
If you, if your integration tests are failing,[br]what

0:16:17.350,0:16:20.489
it means, not necessarily that you wrote bad[br]code,

0:16:20.489,0:16:23.259
but you wrote great code that does the wrong

0:16:23.259,0:16:29.410
thing, right. So, so that's the takeaway there.

0:16:29.410,0:16:32.279
Another great thing they did that we don't[br]do

0:16:32.279,0:16:34.199
today, is they tested what they call the off

0:16:34.199,0:16:37.209
nominal cases, which is just an awesome word,[br]right.

0:16:37.209,0:16:40.449
We don't use nominal nearly enough anymore.

0:16:40.449,0:16:45.819
So, so. They said for every single, for every

0:16:45.819,0:16:47.549
one test they had that tested like the normal

0:16:47.549,0:16:49.209
case of things going right during a mission,[br]they

0:16:49.209,0:16:53.509
had a hundred that tested things going wrong.

0:16:53.509,0:16:54.649
So how many of you guys have a test

0:16:54.649,0:16:56.970
in your Rails app, something like, if a user

0:16:56.970,0:17:00.579
logs in with the right username and the right

0:17:00.579,0:17:03.019
password, they get logged in, right. I, I[br]do.

0:17:03.019,0:17:04.929
Right. A lot of hands go up. How many

0:17:04.929,0:17:07.510
of you guys have a test where if someone

0:17:07.510,0:17:09.770
puts in the right username and the wrong password,

0:17:09.770,0:17:12.859
they don't get logged in, right. Fewer hands[br]go

0:17:12.859,0:17:13.339
up.

0:17:13.339,0:17:16.040
I didn't have that test until I was making

0:17:16.040,0:17:20.429
this talk. So I mean, there's so many different

0:17:20.429,0:17:22.250
tests for things going wrong that, that we[br]don't

0:17:22.250,0:17:24.819
write today, I think. Anything from, you know,[br]security

0:17:24.819,0:17:27.940
related or, or like, checking of all sorts[br]of,

0:17:27.940,0:17:33.240
you know, inputs and stuff like that. But[br]the

0:17:33.240,0:17:36.450
Apollo Guidance Computer team got it right,[br]and there

0:17:36.450,0:17:38.930
were all sorts of cases during the Apollo[br]program

0:17:38.930,0:17:42.310
where, where something went a little bit wrong[br]and,

0:17:42.310,0:17:44.570
and something unexpected happened, and, and[br]like Guidance Computer,

0:17:44.570,0:17:47.280
because they had done all these off nominal[br]tests,

0:17:47.280,0:17:49.660
came out just fine.

0:17:49.660,0:17:52.120
So let's talk about their teams.

0:17:52.120,0:17:58.740
In 1961 when, when the instrument lab signed[br]the

0:17:58.740,0:18:01.400
contract to deliver the Apollo Guidance Computer,[br]that contract

0:18:01.400,0:18:04.440
didn't mention software at all. It said, you[br]just

0:18:04.440,0:18:06.810
have to give us this, this box, with a

0:18:06.810,0:18:09.590
computer in it that can take men to the

0:18:09.590,0:18:10.230
moon.

0:18:10.230,0:18:12.140
And, so they didn't know if it was just

0:18:12.140,0:18:14.260
gonna be, you know, just a bunch of hardware

0:18:14.260,0:18:15.700
and it did this one thing and that's all

0:18:15.700,0:18:17.400
it needed or if it was gonna be like

0:18:17.400,0:18:20.200
a, you know, a fully programmable computer[br]like it

0:18:20.200,0:18:22.480
turned out to be.

0:18:22.480,0:18:26.570
They, they didn't know, in 1961, what a space

0:18:26.570,0:18:27.790
craft that goes to the moon would look like

0:18:27.790,0:18:30.790
at all. They didn't know, you know, would[br]it

0:18:30.790,0:18:32.320
be one giant ship that goes all the way

0:18:32.320,0:18:33.360
to the moon and then comes all the way

0:18:33.360,0:18:35.580
back? Would it be several little ships that[br]are

0:18:35.580,0:18:37.630
combined together? Would they build two rockets[br]and have

0:18:37.630,0:18:39.410
them meet up in space and then go from

0:18:39.410,0:18:40.830
there? Would they build one giant rocket?[br]No one

0:18:40.830,0:18:42.300
knew.

0:18:42.300,0:18:45.350
No one knew if, even in principle, it would

0:18:45.350,0:18:48.510
be possible to navigate to the moon, right.[br]They

0:18:48.510,0:18:51.550
didn't have the math to, to prove that you

0:18:51.550,0:18:53.920
could go to the moon. And no one knew,

0:18:53.920,0:18:56.180
even given math that can help you navigate[br]to

0:18:56.180,0:18:58.070
the moon, if a computer could be built to

0:18:58.070,0:18:59.510
do it.

0:18:59.510,0:19:02.830
So there were, in these early days, there[br]were,

0:19:02.830,0:19:06.850
you know, no deadlines. No managers. No requirements.[br]Very

0:19:06.850,0:19:12.000
little communication with, with other teams.[br]And what that

0:19:12.000,0:19:13.980
means is they were free to experiment and,[br]and

0:19:13.980,0:19:15.980
figure things out, right.

0:19:15.980,0:19:19.220
A couple years later, OK. Now they know, we're

0:19:19.220,0:19:21.700
gonna go to the moon. One giant rocket. Two

0:19:21.700,0:19:23.540
little space crafts. One is just for landing[br]on

0:19:23.540,0:19:27.080
the moon. We have figured out that, just like

0:19:27.080,0:19:29.020
at sea, you can use a sextant and measure

0:19:29.020,0:19:30.890
the angles between, like, stars and part of[br]the

0:19:30.890,0:19:31.990
moon or part of the Earth and figure out

0:19:31.990,0:19:34.750
exactly where you are. And now we know what

0:19:34.750,0:19:37.660
the math looks like to take those, those navigation

0:19:37.660,0:19:39.690
measurements and plug them into a computer,[br]which we

0:19:39.690,0:19:41.260
now know how to build, and we know it

0:19:41.260,0:19:43.490
can fit in a space craft and it can

0:19:43.490,0:19:44.760
take you to the moon.

0:19:44.760,0:19:47.890
So now there's four hundred software developers[br]in the

0:19:47.890,0:19:53.160
Apollo Guidance Computer project. And now[br]there's, there's lots

0:19:53.160,0:19:57.190
of managers. There's lots of deadlines. There's[br]lots of

0:19:57.190,0:20:03.740
requirements. There's lots of documentation.[br]So what happened is

0:20:03.740,0:20:06.480
they built, they had teams that were the right

0:20:06.480,0:20:08.360
shape and size for what they were trying to

0:20:08.360,0:20:10.410
accomplish, right.

0:20:10.410,0:20:13.560
If, if you were talking to someone who is

0:20:13.560,0:20:16.170
founding a startup, it's brand new, and they[br]said

0:20:16.170,0:20:19.940
they had four hundred employees, you might[br]be a

0:20:19.940,0:20:23.820
little worried. Likewise, if, if you were[br]talking to

0:20:23.820,0:20:27.120
a programmer that happened to work for your[br]bank,

0:20:27.120,0:20:28.600
and he said, we have a team of, of

0:20:28.600,0:20:30.660
twenty and we don't really have any requirements[br]for,

0:20:30.660,0:20:32.850
you know, security or anything like that,[br]you'd probably

0:20:32.850,0:20:35.250
find another bank.

0:20:35.250,0:20:38.780
So, so teams have to be the right shape

0:20:38.780,0:20:41.680
and size for what they want to accomplish.[br]Another,

0:20:41.680,0:20:44.630
a corollary to that is switching between those[br]team

0:20:44.630,0:20:49.680
sizes will be difficult and, at any given[br]time,

0:20:49.680,0:20:52.480
there will be people unhappy with how your[br]current

0:20:52.480,0:20:55.970
team is shaped. So NASA and the other contractors

0:20:55.970,0:20:58.610
building the Apollo space craft did not like[br]that

0:20:58.610,0:21:02.030
there were these, these rogue engineers over[br]at MIT

0:21:02.030,0:21:05.580
just fooling around doing whatever they wanted,[br]you know,

0:21:05.580,0:21:06.830
and then that they were gonna have to somehow

0:21:06.830,0:21:09.300
deliver a guidance computer.

0:21:09.300,0:21:12.530
Likewise, as the team grew, a lot of the

0:21:12.530,0:21:19.530
smartest engineers, they left the, the guidance[br]project once,

0:21:19.800,0:21:22.090
especially once Apollo 8 had gone to the moon,

0:21:22.090,0:21:24.010
right. A lot of the engineers said, OK. We've

0:21:24.010,0:21:26.060
done it. You know, a couple years ago we

0:21:26.060,0:21:27.850
had this cool challenge, and now we've figured[br]it

0:21:27.850,0:21:29.520
out. Onto the next thing.

0:21:29.520,0:21:32.310
But the, the mission to the moon was successful,

0:21:32.310,0:21:34.200
right. The goal was accomplished. We had landed[br]on

0:21:34.200,0:21:38.980
the moon before 1970. So the team was always

0:21:38.980,0:21:40.010
the right size.

0:21:40.010,0:21:42.660
All right, what's next. Let's talk a little[br]bit

0:21:42.660,0:21:45.810
about working with users. Except, in this[br]case, it's

0:21:45.810,0:21:50.890
actually working with astronauts. So, who's[br]seen The Right

0:21:50.890,0:21:53.150
Stuff or read the book? Right, they're both[br]great.

0:21:53.150,0:21:54.310
OK, lots of you.

0:21:54.310,0:22:00.660
How did they describe astronauts? Anyone?[br]What's that? Reckless,

0:22:00.660,0:22:04.670
OK. Cowboys. Yes. There's a lot of scenes[br]of

0:22:04.670,0:22:10.060
them riding horses around and stuff. They[br]were arrogant.

0:22:10.060,0:22:12.690
They were extremely talented. They were hard[br]to work

0:22:12.690,0:22:16.760
with. They're just like software developers.

0:22:16.760,0:22:22.410
So, just like any other user, astronauts don't[br]know

0:22:22.410,0:22:26.160
what they want, right. As one of the most

0:22:26.160,0:22:28.340
important things you can, you can learn about[br]working

0:22:28.340,0:22:30.710
with, with your users is they won't tell you

0:22:30.710,0:22:33.730
what they actually want. There's a great example[br]of

0:22:33.730,0:22:34.080
this.

0:22:34.080,0:22:35.870
So the Apollo Guidance Computer was not just[br]a

0:22:35.870,0:22:38.250
guidance computer. It was also an autopilot.[br]It could

0:22:38.250,0:22:41.060
take you from Earth orbit to, you know, a

0:22:41.060,0:22:42.460
couple feet above the lunar surface, and it[br]could

0:22:42.460,0:22:44.740
take you from the lunar surface all the way

0:22:44.740,0:22:46.140
back to Earth.

0:22:46.140,0:22:49.280
So, so the, the engineers at MIT, you know,

0:22:49.280,0:22:51.590
would talk to the astronauts, and they said,[br]hey,

0:22:51.590,0:22:53.920
hey astronauts. You know, you know how, at[br]the

0:22:53.920,0:22:55.450
end of your mission you have this, this re-entry

0:22:55.450,0:22:58.880
process and, and it's, it's pretty dangerous,[br]right. If

0:22:58.880,0:23:01.350
you, if you're coming in too steep you can

0:23:01.350,0:23:03.370
burn up. If you don't come in steep enough

0:23:03.370,0:23:05.200
you can ricochet out into space and, and you'll

0:23:05.200,0:23:07.300
never come back. But, but we know we can

0:23:07.300,0:23:10.260
write an autopilot that will do this for you

0:23:10.260,0:23:12.560
automatically. You just have to press a couple[br]buttons

0:23:12.560,0:23:15.340
and we'll get you right through every time.

0:23:15.340,0:23:17.150
And the astronauts said no. Do not do that.

0:23:17.150,0:23:20.000
You will be wasting your time. We are astronauts

0:23:20.000,0:23:22.020
and we are trained to do these hard things,

0:23:22.020,0:23:25.430
like, like do a manual re-entry. And astronauts,[br]by

0:23:25.430,0:23:29.000
the way, were very distrustful of all automation.[br]Partially

0:23:29.000,0:23:29.840
for good reason.

0:23:29.840,0:23:32.310
In the, in the 50s and 60s, a lot

0:23:32.310,0:23:36.160
of early flight tests with, especially analogue[br]computers flying

0:23:36.160,0:23:38.910
aircraft had gone, had gone wrong. And a lot

0:23:38.910,0:23:42.040
of their friends had died. But on the other

0:23:42.040,0:23:46.080
hand, also, astronauts, you know, are super[br]macho, right.

0:23:46.080,0:23:49.550
They don't like the idea of a computer taking

0:23:49.550,0:23:51.780
away something that, you know, they use to[br]show

0:23:51.780,0:23:52.810
their skills.

0:23:52.810,0:23:54.680
So the astronauts said no. Don't even bother[br]building

0:23:54.680,0:23:57.740
this, building this re-entry program, cause[br]we'll never use

0:23:57.740,0:24:02.000
it. So the engineers built it anyways. And[br]then

0:24:02.000,0:24:06.050
it was used on every single mission.

0:24:06.050,0:24:08.770
It turns out that when you've been in a

0:24:08.770,0:24:13.470
box in space with no showers and you're eating,

0:24:13.470,0:24:16.610
like, frozen, like, dehydrated food for like,[br]up to

0:24:16.610,0:24:20.310
a week, and then there's a button you can

0:24:20.310,0:24:22.430
press so that you just have to sit back

0:24:22.430,0:24:25.450
and relax, and like four minutes later you're[br]in,

0:24:25.450,0:24:27.230
you know, the warm air of the Pacific ocean

0:24:27.230,0:24:28.810
or something like that, you're gonna press[br]those buttons

0:24:28.810,0:24:31.000
every single time, instead of doing one more[br]hard

0:24:31.000,0:24:32.480
thing just to get home, when you've already[br]done

0:24:32.480,0:24:35.470
everything you needed to do. So your users[br]don't

0:24:35.470,0:24:37.160
know what they want.

0:24:37.160,0:24:40.720
Now let's talk about interfaces. Let's imagine[br]you've built

0:24:40.720,0:24:43.370
this computer that can go to the moon, and

0:24:43.370,0:24:45.200
you've got your astronauts ready to go to[br]the

0:24:45.200,0:24:48.710
moon, how do they talk to each other? Keep

0:24:48.710,0:24:50.970
in mind, this is the 60s. The field of

0:24:50.970,0:24:54.830
human computer action, human computer interaction[br]isn't invented until

0:24:54.830,0:24:57.150
1980.

0:24:57.150,0:25:00.100
You can't use a CRT screen, because they're[br]too

0:25:00.100,0:25:04.270
heavy. You can't use an LCD screen, because[br]they

0:25:04.270,0:25:07.810
haven't been invented yet, either. And you[br]can't actually

0:25:07.810,0:25:09.600
use any type of screen at all. It turns

0:25:09.600,0:25:11.170
out the Apollo Guidance Computer doesn't have[br]the power

0:25:11.170,0:25:12.850
for anything like that. So you can use, you

0:25:12.850,0:25:15.660
know, little seven segment LEDs and you can[br]use,

0:25:15.660,0:25:17.000
like, dials and stuff like that.

0:25:17.000,0:25:20.460
So what do you build that astronauts can use?

0:25:20.460,0:25:22.000
It has to be powerful, right. It can't just

0:25:22.000,0:25:23.320
be one button that says take me to the

0:25:23.320,0:25:24.980
moon and maybe another button that says take[br]me

0:25:24.980,0:25:28.750
back, right. There's a little bit more going[br]on.

0:25:28.750,0:25:31.550
And it has to be something that astronauts[br]in

0:25:31.550,0:25:34.140
space suits can, can deal with, right. So[br]I

0:25:34.140,0:25:35.450
don't think an iPhone would have worked even[br]if

0:25:35.450,0:25:39.080
they had it. It has to be something that

0:25:39.080,0:25:41.200
you can operate very quickly, right. There's[br]a little

0:25:41.200,0:25:42.860
bit of time pressure when you're moving at,[br]you

0:25:42.860,0:25:45.490
know, five miles per second.

0:25:45.490,0:25:47.050
So I want to show you what they came

0:25:47.050,0:25:50.580
up with, and it's, it's this thing. And this

0:25:50.580,0:25:53.160
is not a very pretty picture, but it's here

0:25:53.160,0:25:54.600
just as an assurance, because I have a live

0:25:54.600,0:25:58.430
demo of it actually working. So this is the

0:25:58.430,0:26:00.360
display and keyboard unit. So we're gonna[br]switch over

0:26:00.360,0:26:01.090
here.

0:26:01.090,0:26:03.770
So, some very smart people who are not me

0:26:03.770,0:26:07.120
have used ASM and taken an emulator running[br]actually

0:26:07.120,0:26:09.340
Apollo Guidance Computer code. We have all[br]the code,

0:26:09.340,0:26:10.850
by the way, and you can take a look

0:26:10.850,0:26:14.200
at it online. And they build an actual, functional

0:26:14.200,0:26:16.780
interface in your web browser to the guidance[br]computer,

0:26:16.780,0:26:18.290
and I'm gonna show you guys how it works.

0:26:18.290,0:26:19.870
And I'm gonna see what you guys think.

0:26:19.870,0:26:24.330
So, so, you can see here, this thing on

0:26:24.330,0:26:27.970
the, on the left pretty much is the actual

0:26:27.970,0:26:29.250
interface. And then this stuff on the right[br]is

0:26:29.250,0:26:32.130
just for running the, like the emulation.[br]This sort

0:26:32.130,0:26:35.140
of represents like the, the simulated environment.[br]They didn't

0:26:35.140,0:26:35.860
get this thing.

0:26:35.860,0:26:37.510
So, so what do we have here? We've got

0:26:37.510,0:26:40.920
a bunch of lights for, you know, warnings[br]and,

0:26:40.920,0:26:44.260
and status and stuff like that. We've got[br]just

0:26:44.260,0:26:46.710
a couple little bits of display down here.[br]And

0:26:46.710,0:26:48.640
then we've got just a couple buttons. So,[br]so

0:26:48.640,0:26:52.500
let me give you an example of how this

0:26:52.500,0:26:53.500
works.

0:26:53.500,0:26:57.710
So, what you can do, is you hit, you

0:26:57.710,0:27:00.380
pick a verb and then you pick a noun.

0:27:00.380,0:27:04.230
So a verb is something like, show me some

0:27:04.230,0:27:07.230
data. So that's verb zero-six. And then a[br]noun

0:27:07.230,0:27:10.030
is something like, basically the current system[br]up time

0:27:10.030,0:27:12.930
is noun sixty-five. And you hit enter. And[br]it

0:27:12.930,0:27:15.320
shows you. You don't get any decimal points.[br]So

0:27:15.320,0:27:18.430
you have. And you don't get any labels.

0:27:18.430,0:27:19.670
So you have to know what these mean. So

0:27:19.670,0:27:21.980
I'll, I'll tell you. So this means, this computer's

0:27:21.980,0:27:25.480
been running for one hour, thirty-one minutes[br]and forty-nine

0:27:25.480,0:27:29.040
point eight eight seconds. OK, let's try something[br]else.

0:27:29.040,0:27:32.940
Let's hit another verb. That's verb one-six.[br]So that's

0:27:32.940,0:27:35.720
similar to verb six, except what it means[br]is

0:27:35.720,0:27:38.050
show me some data and then keep updating it.

0:27:38.050,0:27:41.280
And then we'll do noun sixty-five again.

0:27:41.280,0:27:43.550
So now, again, one hour thirty-two minutes,[br]and the

0:27:43.550,0:27:45.380
number of seconds are going up, and you can

0:27:45.380,0:27:46.300
see there's some activity.

0:27:46.300,0:27:49.700
So can anyone think of anything that maybe[br]we

0:27:49.700,0:27:52.490
use every single day, some of us, at least,

0:27:52.490,0:27:54.440
that lets you take verbs and nouns and combine

0:27:54.440,0:27:56.430
them in all sorts of interesting and powerful[br]ways?

0:27:56.430,0:27:57.090
Anyone?

0:27:57.090,0:27:58.420
AUDIENCE: Language.

0:27:58.420,0:28:02.130
J.S.: L- well. All right. Someone over there[br]is

0:28:02.130,0:28:05.490
a smart ass. A computer program that we use

0:28:05.490,0:28:07.240
every single day with verbs and nouns.

0:28:07.240,0:28:08.320
AUDIENCE: HTTP.

0:28:08.320,0:28:12.559
J.S.: HTTP is a good one, but no. I

0:28:12.559,0:28:15.220
was thinking of VIM. VIM. Let's hear it for

0:28:15.220,0:28:20.640
VIM. Yes. Exactly.

0:28:20.640,0:28:22.590
I would love to, I would love to find

0:28:22.590,0:28:25.940
out that, that the interface in VI was influenced

0:28:25.940,0:28:27.920
by the Apollo Guidance Computer. I don't think[br]it

0:28:27.920,0:28:29.790
was. If someone happens to know, please let[br]me

0:28:29.790,0:28:32.000
know. That would be awesome.

0:28:32.000,0:28:33.610
But a funny thing, by the way, is that

0:28:33.610,0:28:39.150
this interface was, was intended to be temporary.[br]All

0:28:39.150,0:28:40.960
the, all the engineers said, well, we don't[br]really

0:28:40.960,0:28:43.120
know what to build. But we'll build this thing

0:28:43.120,0:28:44.370
with these verbs and nouns and it's kind of

0:28:44.370,0:28:47.220
clever and, and we'll use that until we build

0:28:47.220,0:28:49.790
something better. And, of course, we now know,[br]like,

0:28:49.790,0:28:52.780
if you want something to stick around forever,[br]just

0:28:52.780,0:28:54.760
say it's temporary.

0:28:54.760,0:28:57.970
But it's cool that the astronauts were, actually[br]became

0:28:57.970,0:29:00.000
very adept at this, and it actually functioned[br]really

0:29:00.000,0:29:01.960
well. And the fact that we still use, you

0:29:01.960,0:29:04.290
know, something similar in VIM today, like.[br]Raise your

0:29:04.290,0:29:06.140
hand if you use VIM as your primary editor

0:29:06.140,0:29:08.180
for all the code you write? Exactly. It's[br]maybe

0:29:08.180,0:29:09.890
twenty-five percent of the people in the room.

0:29:09.890,0:29:13.730
So I just think that's really, really cool.[br]Let's

0:29:13.730,0:29:15.580
see. What else have I got for you? We're

0:29:15.580,0:29:17.750
gonna go back to the slides here, but just

0:29:17.750,0:29:23.250
to say I'm Julian Simioni. I work at 42Floors.

0:29:23.250,0:29:25.360
You can Tweet at me there. I am a

0:29:25.360,0:29:28.809
pilot in real life, and it's really fun. Again,

0:29:28.809,0:29:29.990
another thing I was inspired to do by the

0:29:29.990,0:29:32.730
space program. And I would love to take some

0:29:32.730,0:29:33.090
questions.

0:29:33.090,0:29:34.950
OK guys. Thank you so much.