0:00:13.832,0:00:17.885
First of all, I just want to start off [br]by thanking the organizer.

0:00:17.885,0:00:20.074
They really put in a lot of effort

0:00:20.083,0:00:22.036
to make me feel really special as a speaker.

0:00:22.046,0:00:24.833
Also as an attendee,[br]watching these other talks

0:00:25.421,0:00:27.611
so I just want to give them a huge thanks

0:00:27.751,0:00:30.821
Can I get an extra round of [br]applaud for the organizers

0:00:36.175,0:00:38.879
And also, thanks everyone for being here.

0:00:39.991,0:00:40.930
So, I'm Blithe

0:00:40.939,0:00:43.374
You can find me @blithe[br]on Twitter and Github

0:00:45.103,0:00:47.517
I work for a company call Big Nerd Ranch

0:00:47.517,0:00:49.711
It's based in Atlanta and

0:00:49.711,0:00:51.905
some people know it. Alright!

0:00:51.905,0:00:54.916
It's based in Atlanta and [br]I work remotely full time in Oakland

0:00:54.916,0:00:57.216
We do a couple of things at[br]Big Nerd Ranch

0:00:57.220,0:01:01.295
We develop apps, we write books[br]and we also teach courses.

0:01:02.163,0:01:04.673
And of course, we're hiring[br]good people, always.

0:01:06.336,0:01:08.582
As you heard, I used to be a scientist

0:01:08.597,0:01:10.781
I have a PhD in Physical Chemistry

0:01:11.475,0:01:13.913
And a lot of people ask me [br]what is physical chemistry?

0:01:13.913,0:01:15.400
what does that mean?

0:01:15.416,0:01:17.956
Well, if you think about it[br]It's right on the border between

0:01:17.956,0:01:23.470
what you know as physics and[br]what you know as chemistry

0:01:23.491,0:01:26.566
That means I spent a [br]long time in grad school

0:01:27.156,0:01:30.122
It also means I've done a lot[br]of experiments

0:01:30.123,0:01:32.510
A lot of people asked me[br]"How did you do it?

0:01:32.510,0:01:34.397
How did you make it [br]through 6 years of grad school

0:01:34.622,0:01:36.160
make it to the end?"

0:01:36.194,0:01:39.026
And I did it by using the scientific method

0:01:41.118,0:01:44.992
and I'm still using it everyday

0:01:45.420,0:01:48.910
I'm a developer, but[br]I'm still using the scientific method everyday

0:01:49.470,0:01:51.247
How many of you have gotten an error

0:01:51.788,0:01:52.721
Google it

0:01:53.061,0:01:56.030
pull the first answer straight from StackOverflow

0:01:56.033,0:01:57.517
and put it in your code and run it

0:01:58.795,0:02:00.236
I think we all tried it

0:02:01.786,0:02:03.097
And did it solve your problem?

0:02:03.847,0:02:04.658
Maybe

0:02:04.678,0:02:06.878
Maybe it did.[br]Maybe if you were lucky

0:02:07.038,0:02:08.387
But probably not

0:02:08.532,0:02:09.788
Most of the time, probably not

0:02:09.788,0:02:11.400
and it was pretty inefficient

0:02:11.412,0:02:13.015
and the more important question is

0:02:13.015,0:02:15.564
Did you learn anything from that?

0:02:17.227,0:02:20.690
The most important part of troubleshooting is learning

0:02:20.921,0:02:23.991
If you solved a problem without learning[br]from it,

0:02:23.991,0:02:25.922
It's only a temporary fix

0:02:25.922,0:02:32.095
It's either going to come back later or[br]you're going to make the same mistake in the future

0:02:33.515,0:02:37.711
Learning is how you level up as a developer

0:02:37.711,0:02:40.381
by learning from your experiences

0:02:40.381,0:02:43.279
And that bring me back to the scientific method

0:02:43.282,0:02:48.225
How can using the scientific method[br]help you level up as a developer?

0:02:48.225,0:02:49.794
Well, What is it?

0:02:49.794,0:02:54.665
Well, if you Google it and go straight to[br]Wikipedia you'll see this definition

0:02:54.665,0:02:58.295
It's a set of techniques for acquiring knowledge

0:02:58.295,0:03:01.005
And I love this definition because it can apply to anything

0:03:01.005,0:03:04.699
any career, including being a developer

0:03:04.699,0:03:07.915
The scientific method, it's methodical

0:03:07.915,0:03:13.322
It's an established procedure and[br]scientists have been using it since the 17th century

0:03:13.322,0:03:15.452
It's also systematic

0:03:15.452,0:03:19.198
It's a step-by-step process

0:03:19.198,0:03:23.881
and it's for gaining knowledge[br]and incorporating that into your process

0:03:23.881,0:03:28.660
and the time for reflection is built right into the[br]scientific method

0:03:30.130,0:03:35.196
The other day, a coworker of mine popped up[br]in the company chat room and said

0:03:35.196,0:03:39.790
"Has anyone seen Ruby string interpolation fail?"

0:03:39.790,0:03:41.617
and I thought that's a pretty interesting question

0:03:41.617,0:03:45.147
Why do you ask? Why are you asking that?

0:03:45.147,0:03:50.115
So we dug a little deeper and they were [br]using factory girl sequencing method

0:03:50.115,0:03:54.155
to create a series of users [br]with unique email addresses.

0:03:54.155,0:03:57.154
So email 1, email 2, email 3...

0:03:57.154,0:03:58.872
and it wasn't working. It was failing.

0:03:58.872,0:04:01.867
It was trying to create users [br]with the same email address

0:04:01.867,0:04:04.133
and it was failing and [br]they couldn't figure out why

0:04:04.133,0:04:06.399
and their solution,

0:04:06.399,0:04:09.115
they ended up just rolling [br]their own sequencing method

0:04:09.115,0:04:11.831
Well, that may have solved [br]their problem temporarily,

0:04:11.831,0:04:14.547
that wasn't really the [br]most elegant solution

0:04:14.547,0:04:16.374
and it turned out,

0:04:16.374,0:04:20.061
the real problem was that they [br]were missing the hash symbol

0:04:20.061,0:04:22.819
that was required for interpolation

0:04:22.819,0:04:26.202
so simple mistakes, but we all [br]make simple mistakes like that

0:04:26.202,0:04:28.151
We all make them all the times

0:04:28.151,0:04:32.320
But maybe using the scientific[br]method would have helped in this case

0:04:33.120,0:04:37.452
The first step of the scientific [br]method is defining the problem

0:04:37.452,0:04:43.296
As a scientist, this means saying what is [br]the information that you're trying to gain?

0:04:43.296,0:04:49.007
What knowledge are you[br]trying to gain or learn?

0:04:49.007,0:04:52.315
What's important to keep in mind[br]when you're defining the problem is

0:04:52.315,0:04:56.803
forget what you think you know [br]and question everything

0:04:56.803,0:05:01.622
Just think about open up your mind, question [br]everything and just define the problem

0:05:01.622,0:05:06.050
It's also important to get rid of [br]your emotions at this point

0:05:06.050,0:05:10.299
You want to, maybe you're on a tight deadline[br]maybe you think you know what the answer is

0:05:10.299,0:05:14.906
I'm just going to go do this. I already [br]know what's wrong blah blah blah

0:05:14.906,0:05:18.457
Get rid of all that and [br]just define the problem

0:05:18.457,0:05:20.904
You can do that by [br]answering a few questions

0:05:20.904,0:05:23.533
What is behavior that [br]you're expecting to happen?

0:05:23.533,0:05:26.162
What's actually happening?

0:05:26.162,0:05:28.792
and how are those two things different?

0:05:28.792,0:05:31.950
and how will you know [br]when you solved your problem?

0:05:31.950,0:05:35.550
What is your criteria for success?

0:05:35.550,0:05:38.616
This is particularly important when

0:05:38.616,0:05:42.492
when other people come and [br]ask you to help them troubleshoot

0:05:42.492,0:05:44.375
Don't listen to their assumptions.

0:05:44.375,0:05:46.238
Get rid of that out of the way.

0:05:46.238,0:05:49.192
and just start from scratch and [br]answer these simple questions

0:05:49.192,0:05:52.142
What do you want to happen? [br]What's actually happening?

0:05:52.142,0:05:55.012
How will you know when [br]you fixed the problem?

0:05:55.012,0:05:57.048
Once you defined your problem,[br]

0:05:57.048,0:05:59.474
you can start by doing [br]some background research

0:05:59.474,0:06:01.866
and in reality, you guys are already doing this.

0:06:01.866,0:06:04.258
Everyone's doing it.

0:06:04.258,0:06:06.652
In grad school, we would just be constantly reading

0:06:06.652,0:06:08.669
Before we would ever start an experiment,[br]

0:06:08.669,0:06:11.716
we would just be reading for days and days and days

0:06:11.716,0:06:14.190
Reading recent articles

0:06:14.190,0:06:18.034
or sometimes really old article that we would have to go to the library and dig up

0:06:18.034,0:06:20.780
We were just constantly reading

0:06:20.780,0:06:25.167
And the reason is because you want [br]to have enough background information

0:06:25.167,0:06:30.506
So that you know you're starting [br]with the right experiment

0:06:30.506,0:06:33.556
As a scientist, it's really expensive both[br]

0:06:33.556,0:06:36.606
It cost a lot of time and money to do an experiment

0:06:36.606,0:06:39.656
So you want to make sure you're starting with the right experiment

0:06:39.656,0:06:44.627
and you do that by making sure you have enough background information

0:06:44.627,0:06:47.878
As a developer, you can do that by knowing your environment

0:06:47.878,0:06:51.072
So you need to know the language that you're working in

0:06:51.072,0:06:54.285
and you can always become a little more familiar with it

0:06:54.285,0:06:58.227
go dig in that dark area that you're unfamiliar with

0:06:58.227,0:07:00.884
or how does this work? What's going on here?

0:07:00.884,0:07:04.063
When you're working on a specific application,

0:07:04.063,0:07:07.622
you can look at what are the dependencies that you're using

0:07:07.622,0:07:13.183
Get to know those and even which version of those dependency are you using

0:07:13.183,0:07:16.859
For the case of the factory girl sequencing,

0:07:16.859,0:07:20.165
probably just taking a look at the documentation might have helped

0:07:20.165,0:07:23.802
Oh, I'm missing the hash symbol. That's it.

0:07:23.802,0:07:26.147
You also want to be reading the literature

0:07:26.147,0:07:29.294
You want to stay up to date with blog post

0:07:29.294,0:07:32.380
Always be reading what's coming out, [br]what's new

0:07:32.380,0:07:36.571
There's always a lot of content out there,[br]so you want to be reading it

0:07:36.571,0:07:38.635
For your specific dependency,

0:07:38.635,0:07:40.401
you can look at the README

0:07:40.401,0:07:43.307
or the changelog for the most recent version

0:07:43.307,0:07:45.645
What changed, what's different

0:07:45.645,0:07:49.754
You can also look at the specific source code[br]of the version you're using

0:07:49.754,0:07:53.901
Either on GitHub or using bundle open and the gem name

0:07:53.901,0:07:56.796
You can pull it right up, [br]right on your computer

0:07:56.796,0:07:59.734
And dig around in there.[br]Look at what's going on.

0:07:59.734,0:08:03.122
And also a helpful thing is to look at the test suite

0:08:03.122,0:08:05.612
of this dependency that you're looking at

0:08:05.612,0:08:08.280
Sometimes that helps you tell [br]the expect behavior of the gem[br]

0:08:08.280,0:08:10.528
What's suppose to be happening

0:08:11.348,0:08:14.996
Another helpful source of background information is[br]just discussion

0:08:14.996,0:08:17.138
with your friends and colleagues

0:08:17.138,0:08:19.840
Maybe they have already encounter your problem

0:08:19.840,0:08:22.263
and they know the perfect solution

0:08:22.263,0:08:25.531
The best first experiment for you to start with[br]

0:08:25.531,0:08:27.244
That's really hellpful

0:08:27.244,0:08:30.949
Particularly, when you're struggling on a problem

0:08:30.949,0:08:33.274
maybe you been working on it for a long time

0:08:33.274,0:08:34.881
and you're not getting anywhere

0:08:34.881,0:08:37.842
it's time to get up and go ask somebody for help

0:08:37.842,0:08:39.743
Have you seen anything like this?

0:08:39.743,0:08:42.176
And that's exactly what my colleague did

0:08:42.176,0:08:43.755
They popped in the company chatroom

0:08:43.755,0:08:46.102
"Hey, I need some help on this"

0:08:46.102,0:08:47.754
We have a rule at Big Nerd Ranch

0:08:47.754,0:08:50.059
If you're working on something for more than an hour

0:08:50.059,0:08:51.424
and you're not making progress

0:08:51.424,0:08:52.450
Don't!

0:08:52.450,0:08:54.800
It's time to stop and go ask somebody for help[br]

0:08:54.800,0:08:57.600
It's better to get help from other people

0:08:57.600,0:09:02.769
And of course, sometime the good [br]old rubber duck is a good helpful solution

0:09:02.769,0:09:04.798
Maybe the rubber duck doesn't talk back,

0:09:04.798,0:09:06.934
but just by explaining your problem

0:09:06.934,0:09:12.140
and how you got there is a good way to get ideas

0:09:12.140,0:09:17.698
You also want to make sure you can [br]replicate the problem that you're having

0:09:17.698,0:09:19.185
Make it fail.

0:09:19.185,0:09:23.132
I mean can you even reproduce it? [br]Is it an error you can reproduce?

0:09:23.132,0:09:26.255
And what are the steps that it takes to do that

0:09:26.255,0:09:27.927
Once you have that information,

0:09:27.927,0:09:30.099
you can go ahead and write an automated test

0:09:30.099,0:09:32.223
to help you solve that problem

0:09:32.803,0:09:35.507
After you gather a lot of background information,

0:09:35.507,0:09:38.213
you can then establish a hypothesis

0:09:38.213,0:09:43.768
A hypothesis needs to be educated and based in evidence

0:09:43.768,0:09:46.388
So this goes back to all the research [br]that you've been doing

0:09:46.388,0:09:49.696
And it also need to be something that you can test

0:09:49.696,0:09:52.436
So this goes back to defining your problem

0:09:52.436,0:09:55.236
and your criteria for success

0:09:55.236,0:09:57.200
So once you have a hypothesis,

0:09:57.200,0:09:59.724
then you can start to design an experiment

0:09:59.724,0:10:02.141
And the most important part of an experiment design

0:10:02.141,0:10:04.781
is that it test your hypothesis, of course.

0:10:04.781,0:10:08.687
Don't do an experiment that doesn't test your hypothesis

0:10:08.687,0:10:11.715
Couple of tips for designing an experiment

0:10:11.715,0:10:16.294
One good thing to start with is just dividing and conquering your code base

0:10:16.294,0:10:19.387
Well first of all, hopefully you've [br]written an automated test

0:10:19.387,0:10:21.918
So you can just run that test over and over again

0:10:21.918,0:10:24.939
to focus on the condition that you're working on

0:10:24.939,0:10:30.240
So that isolate a lot other behavior that you don't [br]need to test immediately right now

0:10:30.240,0:10:33.821
Of course, you can always comment [br]out big chunks of codes

0:10:33.821,0:10:35.597
If you're like "I know this stuff is working.

0:10:35.597,0:10:37.201
Let me kind of comment it out

0:10:37.201,0:10:40.855
and just run the part of the code [br]that I need to look at right now."

0:10:40.855,0:10:45.359
Breakpoints are excellent helpful way to do it

0:10:45.359,0:10:47.515
Say your code has to go through 10 steps

0:10:47.515,0:10:49.751
and only the last one isn't working

0:10:49.751,0:10:51.462
Go ahead and enter a breakpoint

0:10:51.462,0:10:53.343
or a binding .pry right in there

0:10:53.343,0:10:58.276
You can stop and just look at the code is broken,[br]that is not working

0:11:00.446,0:11:02.230
You can also

0:11:02.230,0:11:06.285
Say you suspect there's a problem [br]with the dependency or a gem

0:11:06.285,0:11:08.590
What you can do is maybe create a new app

0:11:08.590,0:11:12.155
and only use that one dependency or API

0:11:12.155,0:11:13.872
Are you still having the problem?

0:11:13.872,0:11:17.469
OK, well maybe you just got rid of all [br]of the other code that you have written

0:11:17.469,0:11:18.886
and you know it just with that

0:11:18.886,0:11:21.070
you can replicate it in a fresh app

0:11:21.070,0:11:25.946
That's another way you can isolate the code

0:11:27.246,0:11:30.125
Another important thing is to limit the variables

0:11:30.125,0:11:31.484
that you're testing

0:11:31.484,0:11:34.355
When you look at beginner troubleshoot sometimes

0:11:34.355,0:11:37.374
They try to change many variables at once

0:11:37.374,0:11:39.337
Maybe it's a problem with the controller over here

0:11:39.337,0:11:42.970
and the view or maybe it's in the model. I don't know.

0:11:42.970,0:11:46.953
and sometimes they get down a road[br]that's really hard to come back from

0:11:46.953,0:11:50.885
and they can't even answer the question [br]"what has changed?"

0:11:50.885,0:11:52.543
So you really don't want to do that

0:11:52.543,0:11:56.125
You want to limit your experiment [br]to one variable at a time.

0:11:56.125,0:11:58.218
Let's test one variable

0:11:58.218,0:12:01.043
and finish our experiment before we start another one

0:12:01.043,0:12:02.698
so that we know

0:12:02.698,0:12:07.644
and we need to make sure we do [br]the analysis part of the experiment

0:12:07.644,0:12:11.734
Don't be afraid to try something weird

0:12:11.734,0:12:14.050
Sometimes, you know you think, you're thinking

0:12:14.050,0:12:16.816
"Ah this never gonna work[br]but I'll try it anyway."

0:12:16.816,0:12:20.282
And you try it and you find your solution

0:12:20.282,0:12:23.107
Don't be afraid to do that.[br]Try something strange.

0:12:23.107,0:12:24.857
You can also try maybe

0:12:24.857,0:12:28.171
try something that you know is [br]going to break your program

0:12:28.171,0:12:30.488
Go throw some gibberish in there. [br]Type it in there.

0:12:30.488,0:12:32.618
And you know it's going to break it.

0:12:32.618,0:12:34.748
You know it's going to fail.

0:12:34.748,0:12:38.285
What happen when you run your program[br]and it doesn't fail?

0:12:38.285,0:12:40.308
Maybe that code was never even getting run

0:12:40.308,0:12:43.039
Maybe you're not including that file or something

0:12:43.039,0:12:45.592
So try something strange

0:12:45.592,0:12:48.108
A friend of mine told me a story the other day

0:12:48.108,0:12:50.724
about how they've created a table

0:12:50.724,0:12:53.495
anytime they were trying to do a join with that table,

0:12:53.495,0:12:56.266
it was throwing some obscure error [br]and they couldn't figure it out

0:12:56.266,0:12:59.037
What's wrong with this one table?[br]What's wrong with it?

0:12:59.037,0:13:01.462
So they ended up, they just

0:13:01.462,0:13:03.757
they got rid of the table

0:13:03.757,0:13:06.192
They created a new one and they named it Bananas. [br]

0:13:06.192,0:13:07.626
Just something weird.

0:13:07.626,0:13:09.410
Like just do it weird.

0:13:09.410,0:13:11.387
And that actually fixed their problem.[br]

0:13:11.387,0:13:12.924
So Bananas worked.

0:13:12.924,0:13:14.407
And they're just like "What? What is going on?

0:13:14.407,0:13:16.240
Why? Why did that work"

0:13:16.240,0:13:21.195
and it turned out they had named[br]their original table attributes

0:13:21.945,0:13:25.321
So of course that's a protected name [br]so it didn't work

0:13:25.321,0:13:26.983
So trying something weird,

0:13:26.983,0:13:30.385
naming the table Bananas, [br]that worked for them.

0:13:31.555,0:13:34.847
How many of you have ever [br]done something like this?

0:13:34.847,0:13:37.710
Where you dig, you super deep dive. Yeah.

0:13:37.710,0:13:39.820
You're like oh it's way down here.

0:13:39.820,0:13:41.690
There's some weird problem.

0:13:41.690,0:13:43.930
And it ends up being [br]just right on the surface.

0:13:43.930,0:13:45.660
Something simple.

0:13:46.140,0:13:47.978
When you're designing your experiment,

0:13:47.978,0:13:49.546
you want to focus on

0:13:49.546,0:13:52.156
Think about the hierarchy of blame.

0:13:52.156,0:13:53.441
Think about OK,

0:13:53.441,0:13:57.246
what is the most likely [br]candidate for this problem?

0:13:57.246,0:14:00.474
Once you eliminated [br]the most likely candidates,

0:14:00.474,0:14:03.952
then you can go on[br]to the harder problems

0:14:03.952,0:14:05.052
For example,

0:14:05.052,0:14:07.246
more recently changed code

0:14:07.246,0:14:10.250
is probably more likely to be at fault

0:14:10.250,0:14:13.894
than codes that's been in [br]production for a long time

0:14:13.894,0:14:16.178
and something that you wrote

0:14:16.178,0:14:19.332
is probably more likely to[br]be the problem than say,

0:14:19.332,0:14:21.106
something in Ruby

0:14:22.156,0:14:26.413
But through your experiments go ahead [br]and eliminate the most likely candidate

0:14:26.413,0:14:28.230
and then work your way down

0:14:28.230,0:14:32.419
Don't start by trying to figure out [br]what's wrong with Ruby

0:14:32.979,0:14:35.138
After you designed your experiment,

0:14:35.138,0:14:36.997
go ahead and carry it out

0:14:36.997,0:14:38.867
and you're going to start [br]gathering your data

0:14:38.867,0:14:41.787
that you're going to use for your analysis

0:14:42.687,0:14:44.441
So you need to think about OK,

0:14:44.441,0:14:46.825
what's the current status of your program now?

0:14:46.825,0:14:48.267
What's happening now?

0:14:49.167,0:14:50.795
Were there any changes?

0:14:50.795,0:14:52.487
Are there any new messages?

0:14:52.487,0:14:55.539
Are there any new results from your experiment?

0:14:55.539,0:14:57.390
What else changed recently?

0:14:57.390,0:14:59.597
Were there any other things that you changed?

0:14:59.597,0:15:01.644
What does the log file say?

0:15:01.644,0:15:04.142
Gather all that information

0:15:04.522,0:15:06.250
And this one may seem obvious

0:15:06.250,0:15:09.498
but it need to be said [br]even for senior developers

0:15:09.498,0:15:11.737
Read the error message!

0:15:11.737,0:15:12.781
Just read it.

0:15:12.781,0:15:13.927
Read it out loud.

0:15:13.927,0:15:15.447
Even that alone will help it.

0:15:15.447,0:15:16.357
Just read it.

0:15:16.917,0:15:18.385
Take a look at the stack trace.

0:15:18.385,0:15:19.827
Look at the line number.

0:15:19.827,0:15:21.122
What's going on?

0:15:21.122,0:15:24.113
That's very useful information for a reason

0:15:24.113,0:15:26.027
So gather that information

0:15:26.027,0:15:28.091
as part of your experiment.

0:15:29.011,0:15:30.440
Once you have all your data,

0:15:30.440,0:15:32.519
you can start doing your analysis

0:15:32.519,0:15:35.140
and you really don't [br]want to skip this part

0:15:35.140,0:15:36.912
This is the most important part

0:15:36.912,0:15:39.424
because this where the learning happens

0:15:39.424,0:15:42.058
This is where you level up as a developer

0:15:43.048,0:15:46.441
and it's built right into the scientific method

0:15:47.801,0:15:49.218
So now that you've done your experiment

0:15:49.218,0:15:51.138
Did it solve your problem?

0:15:51.138,0:15:53.828
That test that you wrote, is it passing now?

0:15:53.828,0:15:56.030
and what about all the rest of your test suite?

0:15:56.030,0:15:58.622
Make sure it's also passing

0:15:58.622,0:16:02.106
Make sure you solved your problem [br]without breaking something else

0:16:02.106,0:16:06.297
And make sure you met all [br]your criteria for success

0:16:07.227,0:16:09.079
And you also want to think about

0:16:09.079,0:16:12.191
what did you learn from that experiment?

0:16:12.191,0:16:14.924
What information can you get from that?

0:16:14.924,0:16:17.366
Don't just move on to the next thing and say

0:16:17.366,0:16:19.628
"OK, it's fixed now. Let's go"

0:16:20.068,0:16:21.632
And take a moment,

0:16:21.632,0:16:24.251
Take the time to reflect on your experiment.

0:16:24.251,0:16:27.250
What went well? [br]What didn't go well?

0:16:27.250,0:16:28.931
Think about that

0:16:29.431,0:16:30.799
And most importantly,

0:16:30.799,0:16:32.840
you need to understand

0:16:32.840,0:16:35.892
why you solved your problem

0:16:35.892,0:16:37.674
Have you ever said,

0:16:37.674,0:16:41.187
"Well, I don't why it works, [br]but it works."

0:16:41.997,0:16:44.185
I have news for you. [br]You didn't fix it.

0:16:44.185,0:16:46.553
You didn't solve your problem.

0:16:47.223,0:16:49.140
It's the understanding

0:16:49.140,0:16:52.077
that cement the knowledge in your brain.

0:16:52.077,0:16:54.574
That's what make you a better developer.

0:16:54.574,0:16:57.076
And for those of you in leadership positions,

0:16:57.076,0:16:59.908
maybe you're teaching a newer developer

0:16:59.908,0:17:03.741
It's important that you make sure[br]that they understand the why

0:17:03.741,0:17:06.126
So make them explain it to you,

0:17:06.126,0:17:08.761
Why did this change solved your problem?

0:17:10.441,0:17:12.564
Your results from your experiment

0:17:12.564,0:17:14.379
will also help you determine

0:17:14.379,0:17:17.704
what's going to happen in your next experiment.

0:17:17.704,0:17:19.809
Maybe you solved your problem already

0:17:19.809,0:17:22.534
and you don't have to do another experiment

0:17:22.534,0:17:25.881
but it will influence your experiment in the future

0:17:25.881,0:17:29.259
Maybe now you know [br]a better first experiment to start with

0:17:29.259,0:17:30.040
"I know.

0:17:30.040,0:17:32.691
I know the fix. [br]I'll start with this next time."

0:17:32.691,0:17:34.883
Or if you didn't solve your problem,

0:17:34.883,0:17:37.255
maybe you have a new error message

0:17:37.255,0:17:38.797
"Ok, you can work with that.

0:17:38.797,0:17:40.649
Now what's the new experiment?"

0:17:40.649,0:17:42.997
Or "Oh maybe it's a problem with this gem.

0:17:42.997,0:17:46.605
I should do more [br]background research on that"

0:17:47.925,0:17:49.899
Another important part of the analysis

0:17:49.899,0:17:52.423
is embracing your successes.

0:17:52.723,0:17:54.206
This is really important.

0:17:54.206,0:17:55.368
A lot of people they just,

0:17:55.368,0:17:57.133
they sort of skip over this part

0:17:57.133,0:17:58.998
and go back to work

0:17:59.278,0:18:00.974
But it's really important.

0:18:01.164,0:18:03.498
If you did solve your problem,

0:18:03.498,0:18:05.302
let that sink in

0:18:05.302,0:18:06.884
because that's going to drive you

0:18:06.884,0:18:09.286
for future troubleshooting problems

0:18:09.736,0:18:11.610
You need to remember all those problems

0:18:11.610,0:18:13.139
that you solved before

0:18:13.139,0:18:14.758
and know that you can do it

0:18:14.758,0:18:16.907
and that you can keep going

0:18:17.027,0:18:18.956
And even if you didn't solved [br]your problem this time

0:18:18.956,0:18:20.645
and you're still working on it

0:18:20.645,0:18:23.000
maybe you got a new error message.

0:18:23.000,0:18:27.655
That's progress and that's [br]something to be celebrated

0:18:29.535,0:18:32.101
Another important part of being a scientist

0:18:32.101,0:18:34.137
is keeping a good lab notebook

0:18:34.137,0:18:35.981
and the reason you do this is because

0:18:35.981,0:18:39.545
it tells the story of your experiences

0:18:39.545,0:18:43.258
either to you in the future when [br]you're going back to look at it

0:18:43.258,0:18:45.231
or to somebody else

0:18:45.231,0:18:47.844
maybe another developer in the future

0:18:47.924,0:18:51.054
So this is the picture of one of my [br]lab notebook from grad school.

0:18:51.054,0:18:55.402
I probably filled about 10 of these [br]and they're like 300 pages each

0:18:55.402,0:18:58.716
So I did a lot of documenting in grad school.

0:18:58.716,0:19:01.026
The reason I included this is just to say

0:19:01.026,0:19:05.336
" it's OK to have a paper lab notebook [br]and write the things that you're doing

0:19:05.336,0:19:08.606
That's OK. Digital work also."

0:19:08.826,0:19:11.076
And the reason you keep [br]a lab notebook is because

0:19:11.076,0:19:13.796
you're just not going to remember it

0:19:13.796,0:19:15.898
you're not going to remember

0:19:15.898,0:19:16.899
As a grad student,

0:19:16.899,0:19:20.900
we probably did about 20 to [br]30 experiments a day

0:19:21.340,0:19:24.273
each just changing one variable at a time.

0:19:24.273,0:19:26.621
And there's just no way we would've remember

0:19:26.621,0:19:28.879
all of the things that we tried

0:19:28.879,0:19:32.551
And think about how many experiments [br]you do as a developer each day

0:19:32.551,0:19:34.783
when you're troubleshooting

0:19:35.363,0:19:36.961
And maybe you think

0:19:36.961,0:19:40.089
"OK, well I'm already just [br]recording all this stuff in the log.

0:19:40.089,0:19:41.993
It's all saved over there.

0:19:41.993,0:19:44.257
I can always go back and look

0:19:44.257,0:19:48.572
But that's not an sufficient lab notebook

0:19:48.572,0:19:53.096
Too much information can be just[br]as bad as too little information

0:19:53.096,0:19:54.379
So it's up to you

0:19:54.379,0:19:58.512
to glean the meaning from [br]these experiment that you're doing

0:19:58.512,0:20:00.736
and record that

0:20:00.736,0:20:04.416
One of the best first step for[br]keeping a good lab notebook

0:20:04.416,0:20:06.796
is writing a good commit message

0:20:06.806,0:20:08.917
It's all fresh in your brain

0:20:08.917,0:20:11.685
and so you can maybe you solve the problem

0:20:11.685,0:20:14.073
so you can commit that change and push it

0:20:14.073,0:20:16.242
So write some explanation in that

0:20:16.242,0:20:19.634
that commit message and let [br]people know what's going on

0:20:19.634,0:20:21.771
and people will see this in the future

0:20:21.771,0:20:23.985
Either you'll come back and look at it

0:20:23.985,0:20:28.529
or maybe another developer who's working [br]in the codebase will come back and look

0:20:28.889,0:20:33.068
Additionally , if you added test for [br]the scenario that you're working on,

0:20:33.068,0:20:36.397
that's also a good documentation for what's going on.

0:20:36.397,0:20:38.340
"OK, this was the the problem.

0:20:38.340,0:20:40.193
We added a test for it now."

0:20:42.133,0:20:44.443
Also, right after you solve a problem,

0:20:44.443,0:20:46.393
you want to update the documentation.

0:20:46.393,0:20:48.285
Is it missing anything?

0:20:48.285,0:20:50.039
Did you have that error because

0:20:50.039,0:20:52.693
the documentation wasn't good enough?

0:20:52.693,0:20:54.698
So go ahead and go update the README.

0:20:54.698,0:20:58.153
Or maybe you were missing [br]some environment variable

0:20:58.153,0:21:01.627
Go ahead and set up an example[br]environment variable file

0:21:01.627,0:21:05.141
so people know "Hey, you need this thing."

0:21:05.141,0:21:06.935
to save them that time

0:21:06.935,0:21:09.135
So that's all part of good documentation.

0:21:09.135,0:21:10.375
so keep that

0:21:10.375,0:21:12.627
Set that up for other people.

0:21:12.627,0:21:15.310
You can also contribute to open source.

0:21:15.310,0:21:20.738
So say you find a problem in [br]someone else software or gem,

0:21:20.738,0:21:22.596
go ahead and let them know

0:21:22.596,0:21:26.996
So open up an issue on Github [br]if it doesn't already exist.

0:21:26.996,0:21:31.223
Let them know the step that [br]it takes to recreate this problem.

0:21:32.163,0:21:34.668
If you were able to write a test for it,

0:21:34.668,0:21:37.513
go ahead and submit a pull request with that test

0:21:37.513,0:21:40.976
Sometimes that's helpful even if you [br]don't know how to solve the problem.

0:21:40.976,0:21:42.591
That's helpful for maintainers to say

0:21:42.591,0:21:45.046
"Oh, this is the test that's failing."

0:21:45.046,0:21:46.088
And of course,

0:21:46.088,0:21:47.800
if you do know how how to solve the problem,

0:21:47.800,0:21:49.704
go ahead and submit a pull request for that

0:21:49.704,0:21:51.924
You just got some open source contribution

0:21:51.924,0:21:53.764
and that's awesome

0:21:54.984,0:21:58.774
You can also write about your experiences

0:22:00.464,0:22:03.420
Have you ever written something up and then

0:22:03.420,0:22:07.666
later down the road and you're like[br]googling for that same problem

0:22:07.666,0:22:10.437
and you come back and [br]you find your own blog post?

0:22:10.787,0:22:12.058
Yes. Yes.

0:22:12.058,0:22:16.218
And you're like "Man,[br]thank you so much old me. [br]

0:22:16.378,0:22:17.628
That was really helpful.

0:22:17.628,0:22:20.108
Thank you for writing that down."

0:22:20.108,0:22:22.775
So think about all those other people out there

0:22:22.775,0:22:25.632
that are getting the same reaction [br]when they find your blog post

0:22:25.632,0:22:28.428
and they're like "Man, that was really helpful blog post.

0:22:28.428,0:22:30.364
Thank you for writing that down."

0:22:30.364,0:22:33.080
So if you're not doing this already, [br]you should be doing this

0:22:33.080,0:22:34.784
Anyone can start a blog.

0:22:34.784,0:22:38.088
If you don't already have one, go start it.

0:22:39.522,0:22:42.061
You can also write failed experiment.

0:22:42.061,0:22:44.060
Don't be afraid to do that.

0:22:44.060,0:22:48.370
In grad school, we always joked about [br]starting a journal of failed experiments

0:22:48.370,0:22:51.671
That way we would publish a million papers that way

0:22:51.671,0:22:55.462
Like well, these are the million things[br]we tried and they didn't work

0:22:55.462,0:22:57.456
But that's OK. If you...

0:22:57.456,0:23:00.310
Sometimes that's also helpful

0:23:00.310,0:23:03.705
and maybe somebody comes along [br]and read this and can help you

0:23:03.705,0:23:05.085
and say "Hey,

0:23:05.085,0:23:07.095
you might try doing it this way.

0:23:07.095,0:23:09.387
That might solve your problem."

0:23:10.067,0:23:11.696
The most important part

0:23:11.696,0:23:16.005
about keeping a lab notebook[br]is sharing the the knowledge

0:23:16.005,0:23:18.489
In academia, it's all about

0:23:18.489,0:23:21.523
gaining knowledge and [br]then sharing that knowledge

0:23:21.523,0:23:24.958
either through publication [br]or through teaching

0:23:24.958,0:23:29.622
and I think developers can [br]really learn a lot from that

0:23:29.622,0:23:31.743
So as you're doing these experiments,

0:23:31.743,0:23:33.464
go ahead and share that information

0:23:33.464,0:23:35.417
Maybe tweet about it

0:23:35.417,0:23:38.510
or blog about it, like I said

0:23:38.510,0:23:42.124
or you can even come and present [br]a talk at a conference like this

0:23:42.124,0:23:44.361
"Hey, here's this weird [br]thing that I figured out.

0:23:44.361,0:23:46.368
Let me tell you about it."

0:23:47.057,0:23:49.817
So that's really the most important part

0:23:49.817,0:23:52.109
is sharing that knowledge with other people

0:23:52.109,0:23:54.211
and teaching it to other people

0:23:54.211,0:23:59.595
And so I'm really looking forward to [br]seeing all of your lab notebooks online.

0:23:59.595,0:24:01.542
Thanks very much