Up next in 10
Code Reviews are Mostly Useless
0 views
Aug 6, 2025
Code reviews are mostly useless… and it's all your fault! Please litigate my syntax and formatting choices across fifteen separate comments on Github… for this five line-of-code change. Better yet, why don't you reject my pull request until I guess exactly what you're thinking the code should look like. Supportive, coöperative co-working can be an enormously beneficial process for improving our ability to deliver high quality work, but let's make our code review process as antagonistic and accusatory as possible! In this presentation, we'll gnash our teeth for a bit about how useless and pointless most code review processes tend to be… and when we eventually calm down, we'll try to talk about how to improve these processes, and provide a framing for how to conduct these as a useful quality control tool. 👉 Code Quality Conference 2024 https://codequalityconf.com/ 📺 CSharp TV - Dev Streaming Destination http://csharp.tv 🌎 C# Corner - Community of Software and Data Developers https://www.c-sharpcorner.com/ #CSharpTV #csharpcorner #CodeQualityConf
View Video Transcript
0:03
this is code reviews are useless or or
0:06
this is code reviews are useless or or
0:06
this is code reviews are useless or or maybe I should be a little bit fairer
0:08
maybe I should be a little bit fairer
0:08
maybe I should be a little bit fairer and say code reviews are mostly useless
0:11
and say code reviews are mostly useless
0:11
and say code reviews are mostly useless we are at code quality conference 2024
0:13
we are at code quality conference 2024
0:13
we are at code quality conference 2024 it's Thursday June 27th my name is James
0:16
it's Thursday June 27th my name is James
0:16
it's Thursday June 27th my name is James Powell you can follow me on LinkedIn
0:18
Powell you can follow me on LinkedIn
0:18
Powell you can follow me on LinkedIn there's also a QR code that you'll see
0:19
there's also a QR code that you'll see
0:19
there's also a QR code that you'll see in the presentation with a couple of
0:20
in the presentation with a couple of
0:20
in the presentation with a couple of links to some of our
0:23
links to some of our
0:23
links to some of our materials there's one disclaimer I want
0:25
materials there's one disclaimer I want
0:25
materials there's one disclaimer I want to share with you before we get started
0:28
to share with you before we get started
0:28
to share with you before we get started first I'm going to share with you a
0:29
first I'm going to share with you a
0:29
first I'm going to share with you a couple of very strong opinions about the
0:31
couple of very strong opinions about the
0:31
couple of very strong opinions about the process of code of code review and how
0:33
process of code of code review and how
0:33
process of code of code review and how we conceptualize code quality and where
0:37
we conceptualize code quality and where
0:37
we conceptualize code quality and where I choose to include examples they're
0:39
I choose to include examples they're
0:39
I choose to include examples they're going to be python examples and I know a
0:40
going to be python examples and I know a
0:40
going to be python examples and I know a lot of you are C programmers but I think
0:43
lot of you are C programmers but I think
0:43
lot of you are C programmers but I think you'll be able to squint a little bit
0:44
you'll be able to squint a little bit
0:44
you'll be able to squint a little bit and get a sense for what we're aiming
0:47
and get a sense for what we're aiming
0:47
and get a sense for what we're aiming for here's the question upon us are code
0:50
for here's the question upon us are code
0:50
for here's the question upon us are code reviews an a useful activity or are they
0:52
reviews an a useful activity or are they
0:52
reviews an a useful activity or are they completely useless and it's my view that
0:54
completely useless and it's my view that
0:54
completely useless and it's my view that they're pretty much useless here's my
0:56
they're pretty much useless here's my
0:57
they're pretty much useless here's my proof I'll start with a little bit of
0:58
proof I'll start with a little bit of
0:58
proof I'll start with a little bit of evidence my first piece of evidence is
1:02
evidence my first piece of evidence is
1:02
evidence my first piece of evidence is you've all had this experience where you
1:04
you've all had this experience where you
1:04
you've all had this experience where you were part of a code review and the
1:06
were part of a code review and the
1:06
were part of a code review and the change itself was five lines of code
1:08
change itself was five lines of code
1:08
change itself was five lines of code long and people litigated every single
1:11
long and people litigated every single
1:11
long and people litigated every single minute detail of that code change they
1:15
minute detail of that code change they
1:15
minute detail of that code change they said oh in pep 8 there's supposed to be
1:17
said oh in pep 8 there's supposed to be
1:17
said oh in pep 8 there's supposed to be two spaces between Class definitions oh
1:20
two spaces between Class definitions oh
1:20
two spaces between Class definitions oh you need to write this in this slight
1:22
you need to write this in this slight
1:22
you need to write this in this slight way over here oh don't use a
1:23
way over here oh don't use a
1:23
way over here oh don't use a comprehension or do use a comprehension
1:25
comprehension or do use a comprehension
1:25
comprehension or do use a comprehension and it became a long arduous process and
1:28
and it became a long arduous process and
1:28
and it became a long arduous process and then maybe a week later or 2 weeks later
1:30
then maybe a week later or 2 weeks later
1:30
then maybe a week later or 2 weeks later you were part of a code review where the
1:32
you were part of a code review where the
1:32
you were part of a code review where the change was more significant 10,000 lines
1:34
change was more significant 10,000 lines
1:34
change was more significant 10,000 lines of code and the response you got was
1:36
of code and the response you got was
1:36
of code and the response you got was looks good okay fine and you might think
1:39
looks good okay fine and you might think
1:39
looks good okay fine and you might think if code reviews are an actual useful
1:42
if code reviews are an actual useful
1:42
if code reviews are an actual useful real process surely the 10,000 line of
1:45
real process surely the 10,000 line of
1:45
real process surely the 10,000 line of code code review would be a much more
1:47
code code review would be a much more
1:47
code code review would be a much more engaged activity than the five line of
1:48
engaged activity than the five line of
1:48
engaged activity than the five line of code code review surely if we're
1:50
code code review surely if we're
1:50
code code review surely if we're litigating every single line in a five
1:52
litigating every single line in a five
1:53
litigating every single line in a five line of code change then there must be a
1:54
line of code change then there must be a
1:54
line of code change then there must be a lot that we can say about a 10,000 line
1:56
lot that we can say about a 10,000 line
1:56
lot that we can say about a 10,000 line of code change and yet in practice this
1:58
of code change and yet in practice this
1:58
of code change and yet in practice this doesn't seem to be the case
2:00
doesn't seem to be the case
2:00
doesn't seem to be the case the other piece of evidence I'll share
2:02
the other piece of evidence I'll share
2:02
the other piece of evidence I'll share with you is if we think of code reviews
2:04
with you is if we think of code reviews
2:04
with you is if we think of code reviews as a process by which we can assess the
2:07
as a process by which we can assess the
2:07
as a process by which we can assess the quality of some code and determine
2:09
quality of some code and determine
2:09
quality of some code and determine whether that code is ready for inclusion
2:12
whether that code is ready for inclusion
2:12
whether that code is ready for inclusion in some production system then well why
2:16
in some production system then well why
2:16
in some production system then well why do we need such a detailed code review
2:18
do we need such a detailed code review
2:18
do we need such a detailed code review process code quality is very easy to
2:21
process code quality is very easy to
2:21
process code quality is very easy to assess in fact I can tell you in the
2:24
assess in fact I can tell you in the
2:24
assess in fact I can tell you in the span of no more than one sentence the
2:26
span of no more than one sentence the
2:27
span of no more than one sentence the difference between good code and bad
2:28
difference between good code and bad
2:28
difference between good code and bad code as I have seen this in almost every
2:30
code as I have seen this in almost every
2:30
code as I have seen this in almost every code review I've been a part of and you
2:32
code review I've been a part of and you
2:32
code review I've been a part of and you can say the same thing we don't need
2:34
can say the same thing we don't need
2:34
can say the same thing we don't need books on what constitutes better code
2:36
books on what constitutes better code
2:36
books on what constitutes better code quality cleaner code we don't need long
2:39
quality cleaner code we don't need long
2:39
quality cleaner code we don't need long detailed arguments good code is code I
2:41
detailed arguments good code is code I
2:41
detailed arguments good code is code I wrote and bad code is code you wrote and
2:44
wrote and bad code is code you wrote and
2:44
wrote and bad code is code you wrote and unfortunately in a lot of code reviews
2:45
unfortunately in a lot of code reviews
2:45
unfortunately in a lot of code reviews this seems to be how it plays out and so
2:49
this seems to be how it plays out and so
2:49
this seems to be how it plays out and so given these two pieces of evidence I
2:50
given these two pieces of evidence I
2:50
given these two pieces of evidence I assert that code reviews are completely
2:52
assert that code reviews are completely
2:52
assert that code reviews are completely useless that's my proof however because
2:55
useless that's my proof however because
2:55
useless that's my proof however because we're at a code quality conference I
2:57
we're at a code quality conference I
2:57
we're at a code quality conference I think it makes sense that we should take
2:58
think it makes sense that we should take
2:58
think it makes sense that we should take a step back and we should should
2:59
a step back and we should should
2:59
a step back and we should should litigate these points a little bit more
3:02
litigate these points a little bit more
3:02
litigate these points a little bit more delicately we should go through each one
3:04
delicately we should go through each one
3:04
delicately we should go through each one of these points and really get a better
3:05
of these points and really get a better
3:05
of these points and really get a better understanding for well our code reviews
3:07
understanding for well our code reviews
3:07
understanding for well our code reviews useless inherently or is it a
3:09
useless inherently or is it a
3:09
useless inherently or is it a consequence of the surrounding
3:12
consequence of the surrounding
3:12
consequence of the surrounding environment why is it the case that a
3:14
environment why is it the case that a
3:14
environment why is it the case that a 10,000 line of code code review is
3:16
10,000 line of code code review is
3:16
10,000 line of code code review is considered to be a burden why is it the
3:19
considered to be a burden why is it the
3:19
considered to be a burden why is it the case that when you see that five line of
3:20
case that when you see that five line of
3:20
case that when you see that five line of code code review and it's Friday and
3:22
code code review and it's Friday and
3:22
code code review and it's Friday and it's 3 p.m. leading into a 3-day weekend
3:25
it's 3 p.m. leading into a 3-day weekend
3:25
it's 3 p.m. leading into a 3-day weekend you say well you know I only got two
3:27
you say well you know I only got two
3:27
you say well you know I only got two more hours I might as well site chapter
3:29
more hours I might as well site chapter
3:29
more hours I might as well site chapter inverse for all of things that are wrong
3:31
inverse for all of things that are wrong
3:31
inverse for all of things that are wrong in the small code review and really
3:33
in the small code review and really
3:33
in the small code review and really Grill the junior programmer who
3:34
Grill the junior programmer who
3:34
Grill the junior programmer who submitted it to me And yet when you get
3:36
submitted it to me And yet when you get
3:36
submitted it to me And yet when you get the 10,000 line code review you say you
3:38
the 10,000 line code review you say you
3:38
the 10,000 line code review you say you know it's it's 3M leading into a
3:41
know it's it's 3M leading into a
3:41
know it's it's 3M leading into a three-day weekend is this really what I
3:42
three-day weekend is this really what I
3:42
three-day weekend is this really what I want to be doing is this really where I
3:45
want to be doing is this really where I
3:45
want to be doing is this really where I want to spend my effort and you say
3:47
want to spend my effort and you say
3:47
want to spend my effort and you say looks good let me just head off early
3:50
looks good let me just head off early
3:50
looks good let me just head off early well if it's the case that the code
3:51
well if it's the case that the code
3:51
well if it's the case that the code review is seen as a burden then it's
3:54
review is seen as a burden then it's
3:54
review is seen as a burden then it's likely the case that's not a
3:55
likely the case that's not a
3:55
likely the case that's not a particularly useful process if you're
3:56
particularly useful process if you're
3:56
particularly useful process if you're regretting code reviews if you look at
3:58
regretting code reviews if you look at
3:58
regretting code reviews if you look at all of the work that you to do as part
4:00
all of the work that you to do as part
4:00
all of the work that you to do as part of a code review and it's something that
4:01
of a code review and it's something that
4:01
of a code review and it's something that you don't look forward to it's not an
4:03
you don't look forward to it's not an
4:03
you don't look forward to it's not an opportunity for improvement then it's
4:06
opportunity for improvement then it's
4:06
opportunity for improvement then it's probably the case that we're approaching
4:08
probably the case that we're approaching
4:08
probably the case that we're approaching this wrong if we think about code
4:10
this wrong if we think about code
4:10
this wrong if we think about code reviews and especially long code reviews
4:12
reviews and especially long code reviews
4:12
reviews and especially long code reviews as things that require a lot of reading
4:14
as things that require a lot of reading
4:14
as things that require a lot of reading or a lot of context or a lot of effort
4:16
or a lot of context or a lot of effort
4:16
or a lot of context or a lot of effort well it's definitely the case to 10,000
4:17
well it's definitely the case to 10,000
4:17
well it's definitely the case to 10,000 lines of code you need to read through
4:19
lines of code you need to read through
4:19
lines of code you need to read through every single line there you need to have
4:21
every single line there you need to have
4:21
every single line there you need to have all the implicit context of every change
4:22
all the implicit context of every change
4:23
all the implicit context of every change that was made it's a huge amount of
4:24
that was made it's a huge amount of
4:24
that was made it's a huge amount of effort and you're left asking yourself
4:27
effort and you're left asking yourself
4:27
effort and you're left asking yourself why what's the point well I think that
4:30
why what's the point well I think that
4:30
why what's the point well I think that code reviews are often rendered useless
4:32
code reviews are often rendered useless
4:32
code reviews are often rendered useless as a consequence of contextual factors
4:36
as a consequence of contextual factors
4:36
as a consequence of contextual factors specifically cultural factors in an
4:38
specifically cultural factors in an
4:38
specifically cultural factors in an organization that is to say with the
4:40
organization that is to say with the
4:40
organization that is to say with the lack of a culture of training a culture
4:43
lack of a culture of training a culture
4:43
lack of a culture of training a culture of learning a culture of knowledge and a
4:45
of learning a culture of knowledge and a
4:45
of learning a culture of knowledge and a culture of quality it is very difficult
4:47
culture of quality it is very difficult
4:48
culture of quality it is very difficult to make code reviews a useful practice
4:51
to make code reviews a useful practice
4:51
to make code reviews a useful practice let me go through each one of these one
4:53
let me go through each one of these one
4:53
let me go through each one of these one one by one and give you a sense of what
4:54
one by one and give you a sense of what
4:54
one by one and give you a sense of what I mean so what do we mean when we talk
4:56
I mean so what do we mean when we talk
4:56
I mean so what do we mean when we talk about a culture of training well here's
4:59
about a culture of training well here's
4:59
about a culture of training well here's an example
5:00
an example
5:00
an example let's say you have that five line of
5:02
let's say you have that five line of
5:02
let's say you have that five line of code code review and in that there is
5:04
code code review and in that there is
5:04
code code review and in that there is two lines that look something like this
5:06
two lines that look something like this
5:06
two lines that look something like this I said they would be python examples
5:08
I said they would be python examples
5:08
I said they would be python examples here and the person reviewing your code
5:10
here and the person reviewing your code
5:10
here and the person reviewing your code looks at that and says no I don't like
5:12
looks at that and says no I don't like
5:12
looks at that and says no I don't like that syntax that's too too new for me I
5:15
that syntax that's too too new for me I
5:15
that syntax that's too too new for me I I really didn't Learn Python after
5:16
I really didn't Learn Python after
5:16
I really didn't Learn Python after python 3.6 so the equal sign and the F
5:18
python 3.6 so the equal sign and the F
5:18
python 3.6 so the equal sign and the F string is something I'm not familiar
5:20
string is something I'm not familiar
5:20
string is something I'm not familiar with I don't like it you know figure out
5:22
with I don't like it you know figure out
5:22
with I don't like it you know figure out how to rewrite that and the exercise
5:25
how to rewrite that and the exercise
5:25
how to rewrite that and the exercise turns into one of those guess what I'm
5:26
turns into one of those guess what I'm
5:26
turns into one of those guess what I'm thinking types of exercises which are
5:28
thinking types of exercises which are
5:28
thinking types of exercises which are extraordinarily frustra tring you say
5:30
extraordinarily frustra tring you say
5:31
extraordinarily frustra tring you say you know if you don't like the way that
5:32
you know if you don't like the way that
5:32
you know if you don't like the way that I wrote my code if you're not a fan of
5:34
I wrote my code if you're not a fan of
5:34
I wrote my code if you're not a fan of the stylistic choices I've made don't
5:36
the stylistic choices I've made don't
5:36
the stylistic choices I've made don't make me guess what stylistic choices you
5:39
make me guess what stylistic choices you
5:39
make me guess what stylistic choices you would make why don't you just accept it
5:41
would make why don't you just accept it
5:41
would make why don't you just accept it if it's functionally correct and then go
5:42
if it's functionally correct and then go
5:42
if it's functionally correct and then go fix it yourself you can submit your own
5:46
fix it yourself you can submit your own
5:46
fix it yourself you can submit your own PR with any changes that you want the
5:49
PR with any changes that you want the
5:49
PR with any changes that you want the guess what I'm thinking style of code
5:50
guess what I'm thinking style of code
5:50
guess what I'm thinking style of code review is definitely one that all of us
5:52
review is definitely one that all of us
5:52
review is definitely one that all of us have experienced the one that we
5:53
have experienced the one that we
5:53
have experienced the one that we definitely want to avoid what might be a
5:55
definitely want to avoid what might be a
5:55
definitely want to avoid what might be a more pleasant experience is the let me
5:57
more pleasant experience is the let me
5:57
more pleasant experience is the let me show you how I'd like it to be
5:58
show you how I'd like it to be
5:58
show you how I'd like it to be experienced and and if it's the case
6:00
experienced and and if it's the case
6:00
experienced and and if it's the case that the code review isn't a well you
6:01
that the code review isn't a well you
6:01
that the code review isn't a well you wrote the code in this way but let me
6:03
wrote the code in this way but let me
6:03
wrote the code in this way but let me show you the style that we use in our
6:04
show you the style that we use in our
6:04
show you the style that we use in our codebase let me show you the practices
6:07
codebase let me show you the practices
6:07
codebase let me show you the practices that we have or let me show you what
6:09
that we have or let me show you what
6:09
that we have or let me show you what constitutes consistency across all of
6:11
constitutes consistency across all of
6:11
constitutes consistency across all of the other code that's been written well
6:13
the other code that's been written well
6:13
the other code that's been written well maybe that's a better code review
6:14
maybe that's a better code review
6:14
maybe that's a better code review practice they might say you know what
6:17
practice they might say you know what
6:17
practice they might say you know what instead of that comprehension write this
6:18
instead of that comprehension write this
6:18
instead of that comprehension write this as a for Loop in this formulation you
6:20
as a for Loop in this formulation you
6:20
as a for Loop in this formulation you know saving yourself one line of code
6:22
know saving yourself one line of code
6:22
know saving yourself one line of code really doesn't benefit anybody and just
6:25
really doesn't benefit anybody and just
6:25
really doesn't benefit anybody and just using syntax because it's there is
6:27
using syntax because it's there is
6:27
using syntax because it's there is definitely not a benefit however what's
6:29
definitely not a benefit however what's
6:29
definitely not a benefit however what's missing here is the culture of training
6:32
missing here is the culture of training
6:32
missing here is the culture of training when we see things in a code review we
6:33
when we see things in a code review we
6:34
when we see things in a code review we really need to use these as an
6:35
really need to use these as an
6:35
really need to use these as an opportunity to train especially Junior
6:37
opportunity to train especially Junior
6:37
opportunity to train especially Junior programmers in what is acceptable within
6:39
programmers in what is acceptable within
6:39
programmers in what is acceptable within our organization what approaches that we
6:41
our organization what approaches that we
6:41
our organization what approaches that we want to use and how to use these tools
6:43
want to use and how to use these tools
6:43
want to use and how to use these tools better a better form of code review
6:45
better a better form of code review
6:45
better a better form of code review should be let me explain to you how I'd
6:47
should be let me explain to you how I'd
6:47
should be let me explain to you how I'd like it to be done let me explain to you
6:49
like it to be done let me explain to you
6:49
like it to be done let me explain to you the foundational concepts for why it is
6:51
the foundational concepts for why it is
6:51
the foundational concepts for why it is the way things are done the way they are
6:53
the way things are done the way they are
6:53
the way things are done the way they are and so in the case of this oneline
6:56
and so in the case of this oneline
6:56
and so in the case of this oneline complaint we could say well you know
6:58
complaint we could say well you know
6:58
complaint we could say well you know what if we think a little bit about how
6:59
what if we think a little bit about how
6:59
what if we think a little bit about how python Works a comprehension is supposed
7:02
python Works a comprehension is supposed
7:02
python Works a comprehension is supposed to express to the reader of this code
7:04
to express to the reader of this code
7:04
to express to the reader of this code that it's building something in this
7:06
that it's building something in this
7:06
that it's building something in this case a list comprehension built a list
7:08
case a list comprehension built a list
7:08
case a list comprehension built a list whereas a for Loop is to express to the
7:09
whereas a for Loop is to express to the
7:09
whereas a for Loop is to express to the reader that it's performing an action
7:12
reader that it's performing an action
7:12
reader that it's performing an action and if we think about the change that
7:13
and if we think about the change that
7:13
and if we think about the change that you had you were printing things to the
7:14
you had you were printing things to the
7:14
you had you were printing things to the screen you're not building something
7:16
screen you're not building something
7:16
screen you're not building something you're building a list of nonone values
7:17
you're building a list of nonone values
7:17
you're building a list of nonone values nobody wants to use that you're actually
7:19
nobody wants to use that you're actually
7:19
nobody wants to use that you're actually performing an action and through this
7:21
performing an action and through this
7:21
performing an action and through this explanation we can share with them not
7:23
explanation we can share with them not
7:23
explanation we can share with them not only why it is that the code that
7:25
only why it is that the code that
7:25
only why it is that the code that they've submitted is unacceptable but
7:27
they've submitted is unacceptable but
7:27
they've submitted is unacceptable but also what they can do to generalize that
7:29
also what they can do to generalize that
7:29
also what they can do to generalize that learning to make sure that they approach
7:32
learning to make sure that they approach
7:32
learning to make sure that they approach the problem better next time instead of
7:34
the problem better next time instead of
7:34
the problem better next time instead of always trying to guess what it is you
7:35
always trying to guess what it is you
7:36
always trying to guess what it is you want until they've figured out how to
7:37
want until they've figured out how to
7:37
want until they've figured out how to mimic you they can understand the core
7:39
mimic you they can understand the core
7:39
mimic you they can understand the core principles that you're using in order to
7:41
principles that you're using in order to
7:41
principles that you're using in order to deliberate whether this code is good or
7:42
deliberate whether this code is good or
7:42
deliberate whether this code is good or bad and if it's the case that such
7:44
bad and if it's the case that such
7:44
bad and if it's the case that such principles do not exist maybe you're
7:46
principles do not exist maybe you're
7:46
principles do not exist maybe you're just nitpicking and maybe it's the case
7:48
just nitpicking and maybe it's the case
7:48
just nitpicking and maybe it's the case that as part of the code review process
7:50
that as part of the code review process
7:50
that as part of the code review process you shouldn't really be nitpicking these
7:52
you shouldn't really be nitpicking these
7:52
you shouldn't really be nitpicking these small stylistic
7:53
small stylistic
7:53
small stylistic details now some of you might try to
7:55
details now some of you might try to
7:55
details now some of you might try to take it maybe a little bit of a step
7:57
take it maybe a little bit of a step
7:57
take it maybe a little bit of a step further and say oh well in an even
7:59
further and say oh well in an even
7:59
further and say oh well in an even better organization could we make this a
8:02
better organization could we make this a
8:02
better organization could we make this a discussion where we discuss how
8:03
discussion where we discuss how
8:03
discussion where we discuss how something should be done rather than
8:05
something should be done rather than
8:05
something should be done rather than having the code reviewer explain to the
8:07
having the code reviewer explain to the
8:07
having the code reviewer explain to the person under review and unfortunately I
8:09
person under review and unfortunately I
8:09
person under review and unfortunately I do think that that is likely to be
8:10
do think that that is likely to be
8:11
do think that that is likely to be unrealistic especially in the case of
8:12
unrealistic especially in the case of
8:12
unrealistic especially in the case of Junior programmers it is very much the
8:14
Junior programmers it is very much the
8:14
Junior programmers it is very much the case that as part of a code review
8:15
case that as part of a code review
8:15
case that as part of a code review process where we are establishing that
8:17
process where we are establishing that
8:17
process where we are establishing that there is some hierarchy there's one
8:18
there is some hierarchy there's one
8:18
there is some hierarchy there's one person who reviews and one person who is
8:19
person who reviews and one person who is
8:19
person who reviews and one person who is being reviewed even in the case of a
8:21
being reviewed even in the case of a
8:21
being reviewed even in the case of a peer code review and it's very likely to
8:23
peer code review and it's very likely to
8:23
peer code review and it's very likely to be the case that if it's a junior code
8:25
be the case that if it's a junior code
8:25
be the case that if it's a junior code review well the junior might be missing
8:28
review well the junior might be missing
8:28
review well the junior might be missing some of the important content or this
8:30
some of the important content or this
8:30
some of the important content or this important perspective it's often the
8:31
important perspective it's often the
8:31
important perspective it's often the case that we see Junior programmers will
8:33
case that we see Junior programmers will
8:33
case that we see Junior programmers will say oh you know what let's just fix
8:35
say oh you know what let's just fix
8:35
say oh you know what let's just fix everything and you say well hold on a
8:36
everything and you say well hold on a
8:36
everything and you say well hold on a second we tried that 10 years ago you
8:38
second we tried that 10 years ago you
8:38
second we tried that 10 years ago you know that's not quite the direction the
8:39
know that's not quite the direction the
8:39
know that's not quite the direction the business is taking right now there's a
8:40
business is taking right now there's a
8:40
business is taking right now there's a lot of details that they may need to
8:42
lot of details that they may need to
8:42
lot of details that they may need to know in order to be able to effectively
8:45
know in order to be able to effectively
8:45
know in order to be able to effectively process what something should look like
8:47
process what something should look like
8:47
process what something should look like and so I think it is appropriate even
8:49
and so I think it is appropriate even
8:49
and so I think it is appropriate even within a culture of training to still be
8:51
within a culture of training to still be
8:51
within a culture of training to still be a little bit prescriptive about this is
8:53
a little bit prescriptive about this is
8:53
a little bit prescriptive about this is what we want rather than having to make
8:55
what we want rather than having to make
8:55
what we want rather than having to make every single part of this discussion
8:56
every single part of this discussion
8:56
every single part of this discussion since it is unlikely that some some some
9:00
since it is unlikely that some some some
9:00
since it is unlikely that some some some code reviews would actually be able to
9:01
code reviews would actually be able to
9:01
code reviews would actually be able to turn into a discussion where both sides
9:03
turn into a discussion where both sides
9:03
turn into a discussion where both sides have equal say in how the code will look
9:06
have equal say in how the code will look
9:06
have equal say in how the code will look now there are some cases where when we
9:08
now there are some cases where when we
9:08
now there are some cases where when we take a look at code reviews we can see
9:10
take a look at code reviews we can see
9:10
take a look at code reviews we can see that there's a lot of nuance that needs
9:12
that there's a lot of nuance that needs
9:12
that there's a lot of nuance that needs to be added there and this is important
9:14
to be added there and this is important
9:14
to be added there and this is important in that culture of training the reason
9:15
in that culture of training the reason
9:15
in that culture of training the reason we want to train people is because we
9:16
we want to train people is because we
9:16
we want to train people is because we don't want them to just memorize
9:18
don't want them to just memorize
9:18
don't want them to just memorize formulations we like or to rely on
9:20
formulations we like or to rely on
9:20
formulations we like or to rely on standards documents oh it's already
9:22
standards documents oh it's already
9:22
standards documents oh it's already documented here this is how we write our
9:24
documented here this is how we write our
9:24
documented here this is how we write our code because for example if we rely on
9:26
code because for example if we rely on
9:26
code because for example if we rely on pep eight which is largely an erone ly
9:29
pep eight which is largely an erone ly
9:29
pep eight which is largely an erone ly assumed to be a standardized way to
9:31
assumed to be a standardized way to
9:31
assumed to be a standardized way to write python it's not it's how python
9:34
write python it's not it's how python
9:34
write python it's not it's how python largely should mostly be written if
9:35
largely should mostly be written if
9:35
largely should mostly be written if you're contributing to python itself or
9:37
you're contributing to python itself or
9:37
you're contributing to python itself or the python standard Library it will give
9:39
the python standard Library it will give
9:39
the python standard Library it will give us some guidance where that guidance is
9:40
us some guidance where that guidance is
9:40
us some guidance where that guidance is wrong for example pep8 says if we want
9:42
wrong for example pep8 says if we want
9:43
wrong for example pep8 says if we want to determine if a structure is empty or
9:44
to determine if a structure is empty or
9:44
to determine if a structure is empty or not we shouldn't use the first two
9:46
not we shouldn't use the first two
9:46
not we shouldn't use the first two formulations checking the length we
9:47
formulations checking the length we
9:47
formulations checking the length we should use the last formulation just
9:49
should use the last formulation just
9:49
should use the last formulation just converting this thing to a Boolean value
9:51
converting this thing to a Boolean value
9:51
converting this thing to a Boolean value but this is actually bad advice because
9:53
but this is actually bad advice because
9:53
but this is actually bad advice because if we have a list yeah sure this if
9:55
if we have a list yeah sure this if
9:55
if we have a list yeah sure this if Clause the code contained within will be
9:57
Clause the code contained within will be
9:57
Clause the code contained within will be triggered if this thing is not empty but
9:59
triggered if this thing is not empty but
9:59
triggered if this thing is not empty but if we have a numpy and Dr array this
10:01
if we have a numpy and Dr array this
10:01
if we have a numpy and Dr array this codee's going to break it's going to
10:02
codee's going to break it's going to
10:02
codee's going to break it's going to raise an exception it's going to tell us
10:03
raise an exception it's going to tell us
10:03
raise an exception it's going to tell us that the truth value of an array with
10:04
that the truth value of an array with
10:04
that the truth value of an array with more than one element is ambiguous the
10:07
more than one element is ambiguous the
10:07
more than one element is ambiguous the guidance that pep 8 gives us is not
10:08
guidance that pep 8 gives us is not
10:08
guidance that pep 8 gives us is not correct guidance in this per in this
10:10
correct guidance in this per in this
10:10
correct guidance in this per in this case in fact one of the other
10:11
case in fact one of the other
10:11
case in fact one of the other formulations would be more nearly
10:13
formulations would be more nearly
10:13
formulations would be more nearly correct now as you can see we're
10:15
correct now as you can see we're
10:15
correct now as you can see we're beginning to enter a very delicate
10:17
beginning to enter a very delicate
10:17
beginning to enter a very delicate territory because there's an enormous
10:18
territory because there's an enormous
10:18
territory because there's an enormous amount of nuance to this if we think
10:20
amount of nuance to this if we think
10:20
amount of nuance to this if we think about the difference between a list and
10:21
about the difference between a list and
10:21
about the difference between a list and a numpy and Dray in Python well a list
10:24
a numpy and Dray in Python well a list
10:24
a numpy and Dray in Python well a list is some potentially nested grouped
10:26
is some potentially nested grouped
10:26
is some potentially nested grouped collection of elements whereas A numpy
10:27
collection of elements whereas A numpy
10:27
collection of elements whereas A numpy and D is a mathematical structure so if
10:29
and D is a mathematical structure so if
10:29
and D is a mathematical structure so if we go back to this this area of
10:31
we go back to this this area of
10:32
we go back to this this area of disagreement in our code reviews pep at
10:33
disagreement in our code reviews pep at
10:33
disagreement in our code reviews pep at the guide that we want to stick to or
10:35
the guide that we want to stick to or
10:35
the guide that we want to stick to or not and we say well well this line of
10:37
not and we say well well this line of
10:37
not and we say well well this line of code here that's following Pepe is
10:39
code here that's following Pepe is
10:40
code here that's following Pepe is asking does this list this container of
10:42
asking does this list this container of
10:42
asking does this list this container of entities have any stuff in it whereas if
10:45
entities have any stuff in it whereas if
10:45
entities have any stuff in it whereas if we were already have an MP and D array
10:47
we were already have an MP and D array
10:47
we were already have an MP and D array this is
10:48
this is
10:48
this is asking if this were to work does this
10:50
asking if this were to work does this
10:50
asking if this were to work does this thing have dimensions that's not a very
10:52
thing have dimensions that's not a very
10:52
thing have dimensions that's not a very meaningful question to ask for an NP and
10:54
meaningful question to ask for an NP and
10:54
meaningful question to ask for an NP and d and of course any subsequent operation
10:56
d and of course any subsequent operation
10:56
d and of course any subsequent operation that we perform on that NP and D will
10:58
that we perform on that NP and D will
10:58
that we perform on that NP and D will not broadcast you obviously the
11:00
not broadcast you obviously the
11:00
not broadcast you obviously the broadcasting rules Won't Let You
11:01
broadcasting rules Won't Let You
11:01
broadcasting rules Won't Let You broadcast against something with no
11:02
broadcast against something with no
11:02
broadcast against something with no dimensions and so it's unlikely that you
11:04
dimensions and so it's unlikely that you
11:04
dimensions and so it's unlikely that you would see code that looks like this in
11:06
would see code that looks like this in
11:06
would see code that looks like this in generalized numerical Computing
11:08
generalized numerical Computing
11:08
generalized numerical Computing analytical or machine learning code and
11:10
analytical or machine learning code and
11:10
analytical or machine learning code and so the Nuance here is you know pep baate
11:13
so the Nuance here is you know pep baate
11:13
so the Nuance here is you know pep baate is wrong strictly but it's actually
11:15
is wrong strictly but it's actually
11:15
is wrong strictly but it's actually mostly right in terms of how we use
11:17
mostly right in terms of how we use
11:17
mostly right in terms of how we use these things and no simple sloganeering
11:21
these things and no simple sloganeering
11:21
these things and no simple sloganeering no simple oneline message write it this
11:24
no simple oneline message write it this
11:24
no simple oneline message write it this way is going to be able to provide that
11:26
way is going to be able to provide that
11:26
way is going to be able to provide that context and that opportunity for growth
11:28
context and that opportunity for growth
11:28
context and that opportunity for growth for somebody who's under review this
11:30
for somebody who's under review this
11:30
for somebody who's under review this context is important to how we do our
11:33
context is important to how we do our
11:33
context is important to how we do our work and it makes everything that we do
11:35
work and it makes everything that we do
11:35
work and it makes everything that we do in our work much richer it makes our
11:38
in our work much richer it makes our
11:38
in our work much richer it makes our understanding of the problem domain and
11:39
understanding of the problem domain and
11:39
understanding of the problem domain and the tools that we use far greater and it
11:41
the tools that we use far greater and it
11:41
the tools that we use far greater and it allows us to use these tools more
11:43
allows us to use these tools more
11:43
allows us to use these tools more effectively in terms of using them in a
11:45
effectively in terms of using them in a
11:45
effectively in terms of using them in a more efficient way and also using them
11:47
more efficient way and also using them
11:47
more efficient way and also using them in a more precise way and in a culture
11:50
in a more precise way and in a culture
11:50
in a more precise way and in a culture of training we have that discussion we
11:51
of training we have that discussion we
11:52
of training we have that discussion we have that larger context of well why is
11:54
have that larger context of well why is
11:54
have that larger context of well why is it the way that we might do these things
11:55
it the way that we might do these things
11:56
it the way that we might do these things now in some cases we may cut it short
11:57
now in some cases we may cut it short
11:57
now in some cases we may cut it short because we don't want to spend two weeks
11:59
because we don't want to spend two weeks
11:59
because we don't want to spend two weeks teaching somebody every one of the
12:01
teaching somebody every one of the
12:01
teaching somebody every one of the necessary details behind why we do
12:03
necessary details behind why we do
12:03
necessary details behind why we do things in one place in some cases we
12:05
things in one place in some cases we
12:05
things in one place in some cases we might simplify these but we should
12:06
might simplify these but we should
12:06
might simplify these but we should always have some willingness to kind of
12:09
always have some willingness to kind of
12:09
always have some willingness to kind of say okay you know this is why things are
12:11
say okay you know this is why things are
12:11
say okay you know this is why things are the way they are I'm willing to show
12:12
the way they are I'm willing to show
12:12
the way they are I'm willing to show that to you and we might then simplify
12:15
that to you and we might then simplify
12:15
that to you and we might then simplify these in the form of General guidances
12:18
these in the form of General guidances
12:18
these in the form of General guidances but we shouldn't confuse those guidances
12:19
but we shouldn't confuse those guidances
12:19
but we shouldn't confuse those guidances for the training process we might
12:20
for the training process we might
12:20
for the training process we might provide a guidance like well Library
12:22
provide a guidance like well Library
12:22
provide a guidance like well Library code in which the user out of
12:24
code in which the user out of
12:24
code in which the user out of convenience might provide an unpay that
12:26
convenience might provide an unpay that
12:26
convenience might provide an unpay that represents a collection of elements
12:27
represents a collection of elements
12:27
represents a collection of elements should be handled slightly differently
12:28
should be handled slightly differently
12:28
should be handled slightly differently than Library code which uses a nump and
12:30
than Library code which uses a nump and
12:30
than Library code which uses a nump and d as a mathematical structure and in
12:32
d as a mathematical structure and in
12:32
d as a mathematical structure and in some cases we might we might go against
12:34
some cases we might we might go against
12:34
some cases we might we might go against pep eight in other cases we might not
12:36
pep eight in other cases we might not
12:36
pep eight in other cases we might not and through that context even if those
12:39
and through that context even if those
12:39
and through that context even if those lines of code find themselves excised
12:42
lines of code find themselves excised
12:42
lines of code find themselves excised from the codebase one pull request later
12:45
from the codebase one pull request later
12:45
from the codebase one pull request later the code review process is still a very
12:46
the code review process is still a very
12:46
the code review process is still a very useful one because both parties have had
12:48
useful one because both parties have had
12:48
useful one because both parties have had an opportunity to communicate or to
12:50
an opportunity to communicate or to
12:50
an opportunity to communicate or to learn more about the underlying
12:53
learn more about the underlying
12:53
learn more about the underlying Technologies and the underlying system
12:55
Technologies and the underlying system
12:55
Technologies and the underlying system irrespective of whether the code that
12:56
irrespective of whether the code that
12:56
irrespective of whether the code that they've written is code that will be
12:58
they've written is code that will be
12:58
they've written is code that will be included in the fin product or not other
13:00
included in the fin product or not other
13:00
included in the fin product or not other guidance you might provide is to be
13:02
guidance you might provide is to be
13:02
guidance you might provide is to be aware that EQ may not compose with bu
13:04
aware that EQ may not compose with bu
13:04
aware that EQ may not compose with bu and Python and these are cases where you
13:06
and Python and these are cases where you
13:06
and Python and these are cases where you might give somebody some guidance on
13:08
might give somebody some guidance on
13:08
might give somebody some guidance on errors that they might see for example
13:09
errors that they might see for example
13:09
errors that they might see for example if you had some code like if X's equals
13:11
if you had some code like if X's equals
13:11
if you had some code like if X's equals y's in Python you didn't know what x's
13:13
y's in Python you didn't know what x's
13:13
y's in Python you didn't know what x's and y's are which is very much a library
13:15
and y's are which is very much a library
13:15
and y's are which is very much a library concern not really a user code concern
13:17
concern not really a user code concern
13:17
concern not really a user code concern well this is actually strictly wrong and
13:19
well this is actually strictly wrong and
13:19
well this is actually strictly wrong and in fact you even see this incorrectly in
13:21
in fact you even see this incorrectly in
13:21
in fact you even see this incorrectly in the python standard Library cases where
13:23
the python standard Library cases where
13:23
the python standard Library cases where you might for example try to create an
13:25
you might for example try to create an
13:25
you might for example try to create an enum value that represents cardinal
13:27
enum value that represents cardinal
13:27
enum value that represents cardinal directions and you might want these to
13:28
directions and you might want these to
13:28
directions and you might want these to be the offset that's an R an r squar and
13:31
be the offset that's an R an r squar and
13:31
be the offset that's an R an r squar and if you try and create an enum where the
13:32
if you try and create an enum where the
13:32
if you try and create an enum where the values are nump and D arrays somewhere
13:34
values are nump and D arrays somewhere
13:34
values are nump and D arrays somewhere in the code you see something looks like
13:36
in the code you see something looks like
13:36
in the code you see something looks like this and you get the truth value of
13:37
this and you get the truth value of
13:37
this and you get the truth value of array with more than one element is
13:38
array with more than one element is
13:38
array with more than one element is ambiguous use any or all and it breaks I
13:40
ambiguous use any or all and it breaks I
13:40
ambiguous use any or all and it breaks I actually brought this up to a python
13:42
actually brought this up to a python
13:42
actually brought this up to a python core developer and they gave me the
13:43
core developer and they gave me the
13:43
core developer and they gave me the feedback that numpy is wrong let's go
13:45
feedback that numpy is wrong let's go
13:45
feedback that numpy is wrong let's go change numpy and this is why in some
13:47
change numpy and this is why in some
13:47
change numpy and this is why in some cases we may need to be willing to have
13:50
cases we may need to be willing to have
13:50
cases we may need to be willing to have the conversation but also say well hold
13:52
the conversation but also say well hold
13:53
the conversation but also say well hold on a second we're not going to read
13:54
on a second we're not going to read
13:55
on a second we're not going to read numpy numpy was
13:56
numpy numpy was
13:56
numpy numpy was created now almost 18 years ago it's
14:00
created now almost 18 years ago it's
14:00
created now almost 18 years ago it's very unlikely that we can make such a
14:02
very unlikely that we can make such a
14:02
very unlikely that we can make such a substantive change and with the
14:04
substantive change and with the
14:04
substantive change and with the perspective of knowing that these tools
14:06
perspective of knowing that these tools
14:06
perspective of knowing that these tools are so wides spread we have to sometimes
14:08
are so wides spread we have to sometimes
14:08
are so wides spread we have to sometimes just accept that things are the way they
14:10
just accept that things are the way they
14:10
just accept that things are the way they are but we should never shy away from
14:12
are but we should never shy away from
14:12
are but we should never shy away from part of that culture of training people
14:14
part of that culture of training people
14:14
part of that culture of training people to at least give them a little bit of
14:15
to at least give them a little bit of
14:16
to at least give them a little bit of context and a little bit of
14:17
context and a little bit of
14:17
context and a little bit of understanding for why things are the way
14:18
understanding for why things are the way
14:18
understanding for why things are the way they are so what is a culture of
14:20
they are so what is a culture of
14:20
they are so what is a culture of training well a culture of training is
14:22
training well a culture of training is
14:22
training well a culture of training is instilling in People a sense of
14:23
instilling in People a sense of
14:23
instilling in People a sense of deliberateness that the actions that we
14:25
deliberateness that the actions that we
14:25
deliberateness that the actions that we take are deliberate choices we're not
14:27
take are deliberate choices we're not
14:27
take are deliberate choices we're not just writing the code for the sake of WR
14:29
just writing the code for the sake of WR
14:29
just writing the code for the sake of WR the code but we make choices in every
14:30
the code but we make choices in every
14:31
the code but we make choices in every single line of code that we make and
14:32
single line of code that we make and
14:32
single line of code that we make and there's some reasoning behind it even if
14:34
there's some reasoning behind it even if
14:34
there's some reasoning behind it even if it's the case that we don't relitigate
14:36
it's the case that we don't relitigate
14:36
it's the case that we don't relitigate that reasoning every single time or we
14:37
that reasoning every single time or we
14:37
that reasoning every single time or we don't explicate that reasoning every
14:39
don't explicate that reasoning every
14:39
don't explicate that reasoning every single time it is a willingness to share
14:41
single time it is a willingness to share
14:42
single time it is a willingness to share knowledge understanding that the code
14:43
knowledge understanding that the code
14:43
knowledge understanding that the code review process is an opportunity for you
14:44
review process is an opportunity for you
14:44
review process is an opportunity for you to share knowledge it's a training
14:45
to share knowledge it's a training
14:45
to share knowledge it's a training opportunity for you to work with your co
14:47
opportunity for you to work with your co
14:47
opportunity for you to work with your co co colleagues even in a peer rview and
14:49
co colleagues even in a peer rview and
14:49
co colleagues even in a peer rview and to share with them why is that things
14:51
to share with them why is that things
14:51
to share with them why is that things are the way they are it is not an
14:53
are the way they are it is not an
14:53
are the way they are it is not an opportunity for you merely to denigrate
14:55
opportunity for you merely to denigrate
14:55
opportunity for you merely to denigrate or to say you know this codee's not good
14:57
or to say you know this codee's not good
14:57
or to say you know this codee's not good enough try and make it better make it as
14:59
enough try and make it better make it as
14:59
enough try and make it better make it as good as I want it make it to my
15:01
good as I want it make it to my
15:01
good as I want it make it to my satisfaction a culture of training is
15:04
satisfaction a culture of training is
15:04
satisfaction a culture of training is seeing that code review is a mechanism
15:06
seeing that code review is a mechanism
15:06
seeing that code review is a mechanism and one of the most important and most
15:07
and one of the most important and most
15:07
and one of the most important and most common mechanisms by which we share this
15:09
common mechanisms by which we share this
15:09
common mechanisms by which we share this knowledge and hopefully capturing and
15:12
knowledge and hopefully capturing and
15:12
knowledge and hopefully capturing and recording shared knowledge in a code
15:14
recording shared knowledge in a code
15:14
recording shared knowledge in a code review this is one thing that you can
15:15
review this is one thing that you can
15:15
review this is one thing that you can see from failed code review processes
15:17
see from failed code review processes
15:17
see from failed code review processes well does anybody ever go back to
15:19
well does anybody ever go back to
15:19
well does anybody ever go back to previous reviews and look through them
15:21
previous reviews and look through them
15:21
previous reviews and look through them to understand why decisions were made
15:23
to understand why decisions were made
15:23
to understand why decisions were made one way or the other if it's the case
15:25
one way or the other if it's the case
15:25
one way or the other if it's the case that they don't then it's likely that
15:26
that they don't then it's likely that
15:26
that they don't then it's likely that the code review process is useless
15:29
the code review process is useless
15:29
the code review process is useless now let's talk about a culture of
15:31
now let's talk about a culture of
15:31
now let's talk about a culture of learning because here we can see another
15:33
learning because here we can see another
15:33
learning because here we can see another place where cultural misalignment will
15:36
place where cultural misalignment will
15:36
place where cultural misalignment will lead to a code review process that is
15:37
lead to a code review process that is
15:37
lead to a code review process that is not particularly useful in your
15:39
not particularly useful in your
15:39
not particularly useful in your organization here's an example we I do a
15:42
organization here's an example we I do a
15:42
organization here's an example we I do a lot of corporate training as part of my
15:43
lot of corporate training as part of my
15:43
lot of corporate training as part of my work we do a lot of Consulting we do a
15:44
work we do a lot of Consulting we do a
15:44
work we do a lot of Consulting we do a lot of corporate training and as part of
15:45
lot of corporate training and as part of
15:45
lot of corporate training and as part of corporate training especially in Python
15:47
corporate training especially in Python
15:47
corporate training especially in Python we want to teach people not only how to
15:49
we want to teach people not only how to
15:49
we want to teach people not only how to use Python but how to use Python the way
15:52
use Python but how to use Python the way
15:52
use Python but how to use Python the way that it was designed to be used unlike
15:55
that it was designed to be used unlike
15:55
that it was designed to be used unlike many other tools that are very non
15:57
many other tools that are very non
15:57
many other tools that are very non opinionated about how to be used C++ is
16:00
opinionated about how to be used C++ is
16:00
opinionated about how to be used C++ is a great example you know every
16:02
a great example you know every
16:02
a great example you know every organization uses their own unique
16:04
organization uses their own unique
16:04
organization uses their own unique subset of the C++ functionality and says
16:06
subset of the C++ functionality and says
16:06
subset of the C++ functionality and says these are the things you should use and
16:07
these are the things you should use and
16:07
these are the things you should use and then these are the things you should
16:08
then these are the things you should
16:08
then these are the things you should never touch and then the next place you
16:10
never touch and then the next place you
16:10
never touch and then the next place you you work on a C++ code base they reverse
16:13
you work on a C++ code base they reverse
16:13
you work on a C++ code base they reverse the things you should never use and the
16:14
the things you should never use and the
16:14
the things you should never use and the things you should always use whereas in
16:16
things you should always use whereas in
16:16
things you should always use whereas in Python it's a much more holistically
16:18
Python it's a much more holistically
16:18
Python it's a much more holistically designed language and there's a
16:19
designed language and there's a
16:19
designed language and there's a particular opinionated way to go about
16:21
particular opinionated way to go about
16:21
particular opinionated way to go about using Python and being a good python
16:23
using Python and being a good python
16:23
using Python and being a good python programmer is understanding that that
16:25
programmer is understanding that that
16:25
programmer is understanding that that way is relatively consistent across all
16:28
way is relatively consistent across all
16:28
way is relatively consistent across all uses of python at least all good uses of
16:29
uses of python at least all good uses of
16:30
uses of python at least all good uses of python one case of that is encouraging
16:32
python one case of that is encouraging
16:32
python one case of that is encouraging people in Python to use iteration
16:34
people in Python to use iteration
16:34
people in Python to use iteration helpers so you provide them with some
16:36
helpers so you provide them with some
16:36
helpers so you provide them with some code that looks like this that's looking
16:37
code that looks like this that's looking
16:37
code that looks like this that's looking at some values and it's looking at these
16:38
at some values and it's looking at these
16:38
at some values and it's looking at these values piece-wise it's trying to
16:39
values piece-wise it's trying to
16:40
values piece-wise it's trying to determine what's the pairwise difference
16:41
determine what's the pairwise difference
16:41
determine what's the pairwise difference you know from -2 to 94 that changed by
16:44
you know from -2 to 94 that changed by
16:44
you know from -2 to 94 that changed by 96 from 94 to 7 and you ask them is this
16:47
96 from 94 to 7 and you ask them is this
16:47
96 from 94 to 7 and you ask them is this code readable is this code
16:48
code readable is this code
16:48
code readable is this code understandable and maybe they might say
16:52
understandable and maybe they might say
16:52
understandable and maybe they might say well not really but I can rewrite that
16:54
well not really but I can rewrite that
16:54
well not really but I can rewrite that code very slightly I can make the code
16:56
code very slightly I can make the code
16:56
code very slightly I can make the code look like this and so it's a little bit
16:57
look like this and so it's a little bit
16:57
look like this and so it's a little bit easier to see that we're looking at
16:58
easier to see that we're looking at
16:58
easier to see that we're looking at previous values and current values
17:00
previous values and current values
17:00
previous values and current values offset by one and we're looking at
17:01
offset by one and we're looking at
17:01
offset by one and we're looking at overlapping windows but what you really
17:03
overlapping windows but what you really
17:03
overlapping windows but what you really want to teach them is well in Python
17:06
want to teach them is well in Python
17:06
want to teach them is well in Python there are iteration helpers things in
17:07
there are iteration helpers things in
17:07
there are iteration helpers things in the built-in like Zip and enumerate or
17:09
the built-in like Zip and enumerate or
17:09
the built-in like Zip and enumerate or things in the iter tools module and
17:10
things in the iter tools module and
17:10
things in the iter tools module and instead of all of these other
17:11
instead of all of these other
17:11
instead of all of these other contortions we have direct ways in which
17:14
contortions we have direct ways in which
17:14
contortions we have direct ways in which to write some code that very very
17:17
to write some code that very very
17:17
to write some code that very very clearly expresses the idea look at
17:19
clearly expresses the idea look at
17:19
clearly expresses the idea look at things in overlapping pairs look at the
17:21
things in overlapping pairs look at the
17:21
things in overlapping pairs look at the previous value and the current value
17:22
previous value and the current value
17:22
previous value and the current value then the next previous value and the
17:23
then the next previous value and the
17:23
then the next previous value and the next current value with all of that
17:25
next current value with all of that
17:25
next current value with all of that without all of that administrative noise
17:27
without all of that administrative noise
17:27
without all of that administrative noise of managing that Loop state
17:29
of managing that Loop state
17:29
of managing that Loop state now when we show that to people what
17:32
now when we show that to people what
17:32
now when we show that to people what often happens is you know we ask is this
17:34
often happens is you know we ask is this
17:34
often happens is you know we ask is this better code a a and by a metric of how
17:37
better code a a and by a metric of how
17:37
better code a a and by a metric of how many lines of code it is it's shorter
17:40
many lines of code it is it's shorter
17:41
many lines of code it is it's shorter but a lot of people very legitimately
17:43
but a lot of people very legitimately
17:43
but a lot of people very legitimately say you know what the earlier
17:44
say you know what the earlier
17:45
say you know what the earlier formulations are what I'm more used to
17:46
formulations are what I'm more used to
17:46
formulations are what I'm more used to from writing code in C or C++ or Java
17:49
from writing code in C or C++ or Java
17:49
from writing code in C or C++ or Java I've never used iteration helpers it's a
17:51
I've never used iteration helpers it's a
17:51
I've never used iteration helpers it's a very pythonic approach for solving
17:53
very pythonic approach for solving
17:53
very pythonic approach for solving problems and as a consequence I'm a
17:55
problems and as a consequence I'm a
17:55
problems and as a consequence I'm a little bit more comfortable with the
17:56
little bit more comfortable with the
17:56
little bit more comfortable with the other approaches even if maybe you're
17:58
other approaches even if maybe you're
17:58
other approaches even if maybe you're saying that they're not better in fact
18:00
saying that they're not better in fact
18:00
saying that they're not better in fact we even get feedback that says things
18:02
we even get feedback that says things
18:02
we even get feedback that says things like yeah sure that's great that's the
18:05
like yeah sure that's great that's the
18:05
like yeah sure that's great that's the way to do it on python but I kind of
18:07
way to do it on python but I kind of
18:07
way to do it on python but I kind of have to know what iter tools. pairwise
18:10
have to know what iter tools. pairwise
18:10
have to know what iter tools. pairwise is and that it exists and as a
18:13
is and that it exists and as a
18:13
is and that it exists and as a consequence well we try and say well you
18:15
consequence well we try and say well you
18:15
consequence well we try and say well you know if you're learning about python
18:16
know if you're learning about python
18:16
know if you're learning about python then you ought to learn about Python and
18:18
then you ought to learn about Python and
18:18
then you ought to learn about Python and this is the opportunity for you to learn
18:19
this is the opportunity for you to learn
18:19
this is the opportunity for you to learn that these things exist and how to use
18:21
that these things exist and how to use
18:21
that these things exist and how to use them and I share with them a very
18:24
them and I share with them a very
18:24
them and I share with them a very genuine rubric that I use for
18:26
genuine rubric that I use for
18:26
genuine rubric that I use for determining what I should include in my
18:28
determining what I should include in my
18:28
determining what I should include in my code and for answering that question of
18:30
code and for answering that question of
18:30
code and for answering that question of is this code too complex or not is this
18:33
is this code too complex or not is this
18:33
is this code too complex or not is this code worth putting in front of a junior
18:35
code worth putting in front of a junior
18:35
code worth putting in front of a junior person or not because you know anytime
18:37
person or not because you know anytime
18:37
person or not because you know anytime somebody says to me I I don't use fancy
18:40
somebody says to me I I don't use fancy
18:40
somebody says to me I I don't use fancy features that my co-workers may not know
18:41
features that my co-workers may not know
18:41
features that my co-workers may not know maybe they're all Junior that feels too
18:43
maybe they're all Junior that feels too
18:43
maybe they're all Junior that feels too wishy-washy and too subjective for me to
18:46
wishy-washy and too subjective for me to
18:46
wishy-washy and too subjective for me to really be able to say oh this is code
18:48
really be able to say oh this is code
18:48
really be able to say oh this is code that's acceptable and this is code
18:49
that's acceptable and this is code
18:49
that's acceptable and this is code that's not acceptable and so I
18:52
that's not acceptable and so I
18:52
that's not acceptable and so I understand because originally when we
18:53
understand because originally when we
18:53
understand because originally when we show this examples iter tools. pair wise
18:55
show this examples iter tools. pair wise
18:55
show this examples iter tools. pair wise didn't exist so we show this to them in
18:56
didn't exist so we show this to them in
18:57
didn't exist so we show this to them in the context of an N wise that we
18:58
the context of an N wise that we
18:58
the context of an N wise that we structed ourselves and we and we get
19:00
structed ourselves and we and we get
19:00
structed ourselves and we and we get where you know if you actually have to
19:02
where you know if you actually have to
19:02
where you know if you actually have to look at how all of this stuff works it
19:03
look at how all of this stuff works it
19:03
look at how all of this stuff works it gets a little bit confusing but I think
19:07
gets a little bit confusing but I think
19:07
gets a little bit confusing but I think in a code review you know complexity can
19:10
in a code review you know complexity can
19:10
in a code review you know complexity can be boiled down to a fairly simple rubric
19:12
be boiled down to a fairly simple rubric
19:12
be boiled down to a fairly simple rubric that we can litigate I think that what's
19:15
that we can litigate I think that what's
19:15
that we can litigate I think that what's fair game in code is anything that takes
19:18
fair game in code is anything that takes
19:18
fair game in code is anything that takes me less than 90 seconds to explain to
19:20
me less than 90 seconds to explain to
19:20
me less than 90 seconds to explain to somebody or takes me less than 5 seconds
19:22
somebody or takes me less than 5 seconds
19:22
somebody or takes me less than 5 seconds to explain to somebody or takes me less
19:24
to explain to somebody or takes me less
19:24
to explain to somebody or takes me less than the space of one cocktail napkin to
19:27
than the space of one cocktail napkin to
19:27
than the space of one cocktail napkin to write out an explanation and if I can
19:29
write out an explanation and if I can
19:29
write out an explanation and if I can explain something under these under
19:32
explain something under these under
19:32
explain something under these under these three restrictions then maybe it's
19:33
these three restrictions then maybe it's
19:33
these three restrictions then maybe it's at a sufficient level of complexity to
19:36
at a sufficient level of complexity to
19:36
at a sufficient level of complexity to be included and maybe we can remove the
19:38
be included and maybe we can remove the
19:38
be included and maybe we can remove the subjectivity of an argument is this too
19:39
subjectivity of an argument is this too
19:39
subjectivity of an argument is this too complex or not and in the case of it
19:42
complex or not and in the case of it
19:42
complex or not and in the case of it tools. pawise I can do that ER tools.
19:44
tools. pawise I can do that ER tools.
19:44
tools. pawise I can do that ER tools. pairwise looks at something you could
19:45
pairwise looks at something you could
19:45
pairwise looks at something you could iterate over in overlapping Windows here
19:47
iterate over in overlapping Windows here
19:47
iterate over in overlapping Windows here I could write this on a cocktail napkin
19:49
I could write this on a cocktail napkin
19:49
I could write this on a cocktail napkin and I can explain this in two sentences
19:50
and I can explain this in two sentences
19:50
and I can explain this in two sentences and 30 seconds what it
19:52
and 30 seconds what it
19:52
and 30 seconds what it does but if we think about the
19:55
does but if we think about the
19:55
does but if we think about the generalized problem lurking behind all
19:57
generalized problem lurking behind all
19:57
generalized problem lurking behind all of this really seems to be the case is a
20:01
of this really seems to be the case is a
20:01
of this really seems to be the case is a lot of people look at code and they say
20:03
lot of people look at code and they say
20:03
lot of people look at code and they say I don't like that code and in my code
20:05
I don't like that code and in my code
20:05
I don't like that code and in my code review I'm going to express my
20:07
review I'm going to express my
20:07
review I'm going to express my displeasure on this code I don't like it
20:09
displeasure on this code I don't like it
20:09
displeasure on this code I don't like it it's too complicated they might even
20:12
it's too complicated they might even
20:12
it's too complicated they might even pejoratively say it's stupid because
20:14
pejoratively say it's stupid because
20:14
pejoratively say it's stupid because it's too complicated and I know they
20:15
it's too complicated and I know they
20:15
it's too complicated and I know they might say that because I know I've said
20:17
might say that because I know I've said
20:17
might say that because I know I've said that before I've looked at some code and
20:19
that before I've looked at some code and
20:19
that before I've looked at some code and said oh I hate this I hate this with all
20:21
said oh I hate this I hate this with all
20:21
said oh I hate this I hate this with all of my being it's far too complicated but
20:23
of my being it's far too complicated but
20:23
of my being it's far too complicated but what I really meant was I hate this with
20:25
what I really meant was I hate this with
20:25
what I really meant was I hate this with all of my being because I don't already
20:28
all of my being because I don't already
20:28
all of my being because I don't already know it
20:29
know it
20:29
know it I I hate this with all of my being
20:31
I I hate this with all of my being
20:31
I I hate this with all of my being because well I'm not getting older I
20:32
because well I'm not getting older I
20:32
because well I'm not getting older I sure am feeling older these days and I'm
20:35
sure am feeling older these days and I'm
20:35
sure am feeling older these days and I'm actually kind of scared of being left
20:36
actually kind of scared of being left
20:36
actually kind of scared of being left behind I'm scared of being confronted
20:37
behind I'm scared of being confronted
20:37
behind I'm scared of being confronted with things that I don't know and I'm
20:39
with things that I don't know and I'm
20:39
with things that I don't know and I'm not already familiar with I I'm used to
20:42
not already familiar with I I'm used to
20:42
not already familiar with I I'm used to the way that python used to be written
20:43
the way that python used to be written
20:43
the way that python used to be written 10 years ago and I'm kind of scared
20:45
10 years ago and I'm kind of scared
20:45
10 years ago and I'm kind of scared because the ways that people are writing
20:46
because the ways that people are writing
20:46
because the ways that people are writing it these ways are very different and I
20:48
it these ways are very different and I
20:48
it these ways are very different and I know that happens because it happens to
20:50
know that happens because it happens to
20:50
know that happens because it happens to me all the time well when we think about
20:53
me all the time well when we think about
20:53
me all the time well when we think about a culture of learning distinct from a
20:55
a culture of learning distinct from a
20:55
a culture of learning distinct from a culture of training this is one of the
20:57
culture of training this is one of the
20:58
culture of training this is one of the important and critical aspects of an
20:59
important and critical aspects of an
20:59
important and critical aspects of an effective code review culture
21:01
effective code review culture
21:01
effective code review culture specifically a culture of learning is
21:03
specifically a culture of learning is
21:03
specifically a culture of learning is understanding that learning never stops
21:05
understanding that learning never stops
21:05
understanding that learning never stops I know every single day I might not look
21:07
I know every single day I might not look
21:07
I know every single day I might not look older but I feel older and I understand
21:09
older but I feel older and I understand
21:09
older but I feel older and I understand that you know for all of the years that
21:11
that you know for all of the years that
21:11
that you know for all of the years that I've spent learning all of my tools it
21:13
I've spent learning all of my tools it
21:13
I've spent learning all of my tools it could very well be the case that
21:15
could very well be the case that
21:15
could very well be the case that tomorrow python goes away and I need to
21:17
tomorrow python goes away and I need to
21:18
tomorrow python goes away and I need to learn I need to learn rust rust goes
21:21
learn I need to learn rust rust goes
21:21
learn I need to learn rust rust goes away and I need to learn Zig Zig goes
21:23
away and I need to learn Zig Zig goes
21:23
away and I need to learn Zig Zig goes away and I need to learn the next thing
21:24
away and I need to learn the next thing
21:24
away and I need to learn the next thing and I should be aware of and come to
21:26
and I should be aware of and come to
21:26
and I should be aware of and come to terms with an idea that the learning
21:28
terms with an idea that the learning
21:28
terms with an idea that the learning never stops it's a consistent thing it's
21:30
never stops it's a consistent thing it's
21:30
never stops it's a consistent thing it's a constant thing as part of our work
21:32
a constant thing as part of our work
21:32
a constant thing as part of our work that there are things that I'm not going
21:34
that there are things that I'm not going
21:34
that there are things that I'm not going to know and it's okay that there's
21:35
to know and it's okay that there's
21:35
to know and it's okay that there's things that I'm not going to know and
21:36
things that I'm not going to know and
21:36
things that I'm not going to know and there's things even within my area of
21:39
there's things even within my area of
21:39
there's things even within my area of expertise that I won't know and it's a
21:41
expertise that I won't know and it's a
21:41
expertise that I won't know and it's a very human thing for me to especially
21:44
very human thing for me to especially
21:44
very human thing for me to especially with my co-workers kind of brush off and
21:45
with my co-workers kind of brush off and
21:45
with my co-workers kind of brush off and say oh yeah that was added recently oh
21:48
say oh yeah that was added recently oh
21:48
say oh yeah that was added recently oh yeah I knew that even if I didn't know
21:49
yeah I knew that even if I didn't know
21:50
yeah I knew that even if I didn't know that and that's a human failing and it's
21:52
that and that's a human failing and it's
21:52
that and that's a human failing and it's a fault that I am aware that I have but
21:54
a fault that I am aware that I have but
21:54
a fault that I am aware that I have but it's something which when I take a step
21:56
it's something which when I take a step
21:56
it's something which when I take a step back and I say well you know what I'm
21:57
back and I say well you know what I'm
21:57
back and I say well you know what I'm not going to know everything thing
21:58
not going to know everything thing
21:58
not going to know everything thing there's always going to be some detail I
21:59
there's always going to be some detail I
21:59
there's always going to be some detail I don't know this is an opportunity for me
22:01
don't know this is an opportunity for me
22:01
don't know this is an opportunity for me to learn something new a culture of
22:03
to learn something new a culture of
22:03
to learn something new a culture of learning is understanding that the
22:04
learning is understanding that the
22:05
learning is understanding that the learning process is not always easy even
22:07
learning process is not always easy even
22:07
learning process is not always easy even in fields where you have expertise it
22:09
in fields where you have expertise it
22:09
in fields where you have expertise it can be very difficult it can be very
22:10
can be very difficult it can be very
22:10
can be very difficult it can be very intimidating but it's worth doing no
22:12
intimidating but it's worth doing no
22:12
intimidating but it's worth doing no matter what in fact if we think about it
22:14
matter what in fact if we think about it
22:15
matter what in fact if we think about it continuous Improvement continuous
22:16
continuous Improvement continuous
22:16
continuous Improvement continuous learning is our responsibility it's
22:18
learning is our responsibility it's
22:18
learning is our responsibility it's worth doing no matter how difficult how
22:20
worth doing no matter how difficult how
22:20
worth doing no matter how difficult how intimidating even though it's something
22:22
intimidating even though it's something
22:22
intimidating even though it's something that never stops and even though I'm
22:25
that never stops and even though I'm
22:25
that never stops and even though I'm only going to discover more things that
22:27
only going to discover more things that
22:27
only going to discover more things that I don't know in fact some people may go
22:30
I don't know in fact some people may go
22:30
I don't know in fact some people may go so far as to say that learning is above
22:33
so far as to say that learning is above
22:33
so far as to say that learning is above all other human Pursuits that it is the
22:35
all other human Pursuits that it is the
22:35
all other human Pursuits that it is the greatest human Pursuit and honestly I
22:37
greatest human Pursuit and honestly I
22:37
greatest human Pursuit and honestly I wouldn't go that far I would however say
22:40
wouldn't go that far I would however say
22:40
wouldn't go that far I would however say that knowledge and the S the the the
22:43
that knowledge and the S the the the
22:43
that knowledge and the S the the the search for knowledge is genuinely above
22:45
search for knowledge is genuinely above
22:45
search for knowledge is genuinely above all other human Pursuits which leads us
22:48
all other human Pursuits which leads us
22:48
all other human Pursuits which leads us into a culture of knowledge and one
22:50
into a culture of knowledge and one
22:50
into a culture of knowledge and one thing that is critically missing and
22:52
thing that is critically missing and
22:52
thing that is critically missing and unfortunately one thing that unlike a
22:54
unfortunately one thing that unlike a
22:54
unfortunately one thing that unlike a culture of learning and unlike a culture
22:55
culture of learning and unlike a culture
22:55
culture of learning and unlike a culture of teaching where we can try through
22:58
of teaching where we can try through
22:58
of teaching where we can try through some liberate process develop these the
22:59
some liberate process develop these the
22:59
some liberate process develop these the culture of knowledge is actually quite
23:01
culture of knowledge is actually quite
23:01
culture of knowledge is actually quite difficult to instill in an organization
23:03
difficult to instill in an organization
23:03
difficult to instill in an organization if it's not already present here's what
23:05
if it's not already present here's what
23:05
if it's not already present here's what we mean by a culture of knowledge well
23:07
we mean by a culture of knowledge well
23:07
we mean by a culture of knowledge well one of my biggest pet peeves when we're
23:09
one of my biggest pet peeves when we're
23:09
one of my biggest pet peeves when we're talking about design decisions in a
23:11
talking about design decisions in a
23:11
talking about design decisions in a design or code review is when people say
23:13
design or code review is when people say
23:13
design or code review is when people say things like it's a series of trade-offs
23:16
things like it's a series of trade-offs
23:16
things like it's a series of trade-offs okay yeah but what in life isn't a
23:18
okay yeah but what in life isn't a
23:18
okay yeah but what in life isn't a series of trade-offs everything's a
23:19
series of trade-offs everything's a
23:20
series of trade-offs everything's a series of trade-offs that added no
23:21
series of trade-offs that added no
23:21
series of trade-offs that added no knowledge no information or when they
23:23
knowledge no information or when they
23:23
knowledge no information or when they say well use this feature under these
23:26
say well use this feature under these
23:26
say well use this feature under these circumstances or you say when do you use
23:29
circumstances or you say when do you use
23:29
circumstances or you say when do you use a generator versus a generator co-
23:30
a generator versus a generator co-
23:30
a generator versus a generator co- routine versus some other feature in
23:31
routine versus some other feature in
23:31
routine versus some other feature in Python they say it depends and I'm just
23:33
Python they say it depends and I'm just
23:33
Python they say it depends and I'm just thinking yeah it depends on what the
23:36
thinking yeah it depends on what the
23:36
thinking yeah it depends on what the knowledge here is giving me some
23:37
knowledge here is giving me some
23:37
knowledge here is giving me some guidance for how to methodically apply
23:40
guidance for how to methodically apply
23:40
guidance for how to methodically apply some theoretical constructs to reach an
23:42
some theoretical constructs to reach an
23:42
some theoretical constructs to reach an end goal not to just give me some Pat
23:44
end goal not to just give me some Pat
23:44
end goal not to just give me some Pat statement saying I don't know it depends
23:46
statement saying I don't know it depends
23:46
statement saying I don't know it depends sometimes it's this way sometimes it's
23:47
sometimes it's this way sometimes it's
23:47
sometimes it's this way sometimes it's the other way and carving out within
23:50
the other way and carving out within
23:50
the other way and carving out within what is often times a holy unknowable
23:53
what is often times a holy unknowable
23:53
what is often times a holy unknowable field the regions of knowledge that we
23:55
field the regions of knowledge that we
23:55
field the regions of knowledge that we can Sur you can find another thing that
23:58
can Sur you can find another thing that
23:58
can Sur you can find another thing that is are a humongous pet peeve for me is
24:00
is are a humongous pet peeve for me is
24:00
is are a humongous pet peeve for me is when I see circumstances where people
24:03
when I see circumstances where people
24:03
when I see circumstances where people are Rel litigating in the sense that it
24:07
are Rel litigating in the sense that it
24:07
are Rel litigating in the sense that it is an interesting controversial
24:09
is an interesting controversial
24:09
is an interesting controversial questions questions that are ultimately
24:11
questions questions that are ultimately
24:11
questions questions that are ultimately and utterly meaningless and are in fact
24:14
and utterly meaningless and are in fact
24:14
and utterly meaningless and are in fact not controversial and not particularly
24:15
not controversial and not particularly
24:15
not controversial and not particularly interesting but are instead almost and
24:18
interesting but are instead almost and
24:18
interesting but are instead almost and completely obvious in their answer once
24:20
completely obvious in their answer once
24:20
completely obvious in their answer once you understand some foundational Theory
24:22
you understand some foundational Theory
24:22
you understand some foundational Theory or some foundational knowledge I
24:24
or some foundational knowledge I
24:24
or some foundational knowledge I recently saw online some people going
24:26
recently saw online some people going
24:26
recently saw online some people going back and forth on should they parse the
24:27
back and forth on should they parse the
24:27
back and forth on should they parse the output of Less in a shell script and I
24:30
output of Less in a shell script and I
24:30
output of Less in a shell script and I just I just thought this is hours of
24:32
just I just thought this is hours of
24:32
just I just thought this is hours of human effort and the answer is so
24:34
human effort and the answer is so
24:34
human effort and the answer is so blindingly obvious this is not an
24:36
blindingly obvious this is not an
24:36
blindingly obvious this is not an interesting or controversial question
24:38
interesting or controversial question
24:38
interesting or controversial question that should be getting that much
24:39
that should be getting that much
24:39
that should be getting that much attention it's really just a
24:41
attention it's really just a
24:41
attention it's really just a understanding of the Core Concepts and
24:42
understanding of the Core Concepts and
24:43
understanding of the Core Concepts and then application of those Core Concepts
24:44
then application of those Core Concepts
24:44
then application of those Core Concepts and then we move on with their life but
24:46
and then we move on with their life but
24:46
and then we move on with their life but to give you an example I want to show
24:47
to give you an example I want to show
24:47
to give you an example I want to show you a little bit of code that I was
24:49
you a little bit of code that I was
24:49
you a little bit of code that I was looking at recently a representative
24:50
looking at recently a representative
24:50
looking at recently a representative example that came up in some code review
24:52
example that came up in some code review
24:52
example that came up in some code review that has an enormously subtle bug this
24:54
that has an enormously subtle bug this
24:54
that has an enormously subtle bug this is one of the most subtle bugs that I've
24:57
is one of the most subtle bugs that I've
24:57
is one of the most subtle bugs that I've seen in code in a while and probably the
24:59
seen in code in a while and probably the
24:59
seen in code in a while and probably the only case where we would make use of a
25:01
only case where we would make use of a
25:01
only case where we would make use of a fairly recent feature that was added in
25:03
fairly recent feature that was added in
25:03
fairly recent feature that was added in I want to say python 3.7 or 3.8 that
25:05
I want to say python 3.7 or 3.8 that
25:05
I want to say python 3.7 or 3.8 that otherwise doesn't have a lot of use
25:07
otherwise doesn't have a lot of use
25:07
otherwise doesn't have a lot of use cases you're not going to be able to
25:08
cases you're not going to be able to
25:08
cases you're not going to be able to spot the bug but you can at least get a
25:10
spot the bug but you can at least get a
25:10
spot the bug but you can at least get a sense of what this code does the
25:12
sense of what this code does the
25:12
sense of what this code does the person's using some kind of functional
25:13
person's using some kind of functional
25:13
person's using some kind of functional programming in Python to write a
25:15
programming in Python to write a
25:15
programming in Python to write a function that performs or calls another
25:16
function that performs or calls another
25:16
function that performs or calls another function some number of times and allows
25:18
function some number of times and allows
25:18
function some number of times and allows you to pass in vartic positional
25:20
you to pass in vartic positional
25:20
you to pass in vartic positional arguments vartic keyword arguments that
25:21
arguments vartic keyword arguments that
25:21
arguments vartic keyword arguments that are just forwarded through and the bug
25:24
are just forwarded through and the bug
25:24
are just forwarded through and the bug here is a design flaw the first two
25:27
here is a design flaw the first two
25:27
here is a design flaw the first two arguments need to be pass position only
25:29
arguments need to be pass position only
25:29
arguments need to be pass position only because if they're not pass positionally
25:31
because if they're not pass positionally
25:31
because if they're not pass positionally only then if the function that is then
25:34
only then if the function that is then
25:34
only then if the function that is then being repeated here contains or has an
25:37
being repeated here contains or has an
25:37
being repeated here contains or has an argument called n that argument n cannot
25:40
argument called n that argument n cannot
25:40
argument called n that argument n cannot be passed by key without an ambiguity
25:42
be passed by key without an ambiguity
25:42
be passed by key without an ambiguity between wait wait was that the N that
25:44
between wait wait was that the N that
25:44
between wait wait was that the N that I'm passing to the function that I'm
25:46
I'm passing to the function that I'm
25:46
I'm passing to the function that I'm calling or is that the end that I'm
25:47
calling or is that the end that I'm
25:47
calling or is that the end that I'm passing to represent the number of times
25:49
passing to represent the number of times
25:49
passing to represent the number of times I'm calling that function and that
25:51
I'm calling that function and that
25:51
I'm calling that function and that ambiguity means that the first of these
25:53
ambiguity means that the first of these
25:53
ambiguity means that the first of these two calls will work but the second one
25:55
two calls will work but the second one
25:55
two calls will work but the second one won't and it's a very hard bug to find
25:58
won't and it's a very hard bug to find
25:58
won't and it's a very hard bug to find unless you know how to look for it
26:00
unless you know how to look for it
26:00
unless you know how to look for it unless you dig on to the underlying
26:02
unless you dig on to the underlying
26:02
unless you dig on to the underlying Theory now if we take a look at this we
26:04
Theory now if we take a look at this we
26:04
Theory now if we take a look at this we could actually go a little bit further
26:05
could actually go a little bit further
26:05
could actually go a little bit further and say you know what we shouldn't be
26:07
and say you know what we shouldn't be
26:07
and say you know what we shouldn't be doing this eagerly we shouldn't be
26:08
doing this eagerly we shouldn't be
26:08
doing this eagerly we shouldn't be collecting this in a list because who
26:09
collecting this in a list because who
26:09
collecting this in a list because who knows if the end user really wants to
26:10
knows if the end user really wants to
26:10
knows if the end user really wants to collect this list we should make this a
26:12
collect this list we should make this a
26:12
collect this list we should make this a generator and and just yield these and
26:14
generator and and just yield these and
26:14
generator and and just yield these and you know if we're making this a
26:15
you know if we're making this a
26:15
you know if we're making this a generator why not just represent this as
26:17
generator why not just represent this as
26:17
generator why not just represent this as an infinite sequence of function calls
26:19
an infinite sequence of function calls
26:19
an infinite sequence of function calls and then give them the ability to
26:21
and then give them the ability to
26:21
and then give them the ability to control modalities externally because
26:22
control modalities externally because
26:22
control modalities externally because what if they want to be able to run this
26:24
what if they want to be able to run this
26:24
what if they want to be able to run this function until it gets a particular
26:25
function until it gets a particular
26:25
function until it gets a particular output which they wouldn't do in the
26:27
output which they wouldn't do in the
26:27
output which they wouldn't do in the first formulation without adding
26:28
first formulation without adding
26:28
first formulation without adding modalities and so instead we should
26:30
modalities and so instead we should
26:30
modalities and so instead we should write this as an infinite Loop and then
26:31
write this as an infinite Loop and then
26:31
write this as an infinite Loop and then use it tools. I slice or it tools. take
26:34
use it tools. I slice or it tools. take
26:34
use it tools. I slice or it tools. take while in order to control the modality
26:36
while in order to control the modality
26:36
while in order to control the modality of when it stops and if we're doing that
26:38
of when it stops and if we're doing that
26:38
of when it stops and if we're doing that maybe we shouldn't even do this in the
26:39
maybe we shouldn't even do this in the
26:39
maybe we shouldn't even do this in the first place we should just let them
26:40
first place we should just let them
26:40
first place we should just let them write our list comprehension or a
26:41
write our list comprehension or a
26:41
write our list comprehension or a generator expression if they want and
26:43
generator expression if they want and
26:43
generator expression if they want and this function doesn't even need to exist
26:45
this function doesn't even need to exist
26:45
this function doesn't even need to exist often times the use of core theory that
26:49
often times the use of core theory that
26:49
often times the use of core theory that has been derived by really searching for
26:51
has been derived by really searching for
26:51
has been derived by really searching for the knowledge the meaning behind the
26:53
the knowledge the meaning behind the
26:53
the knowledge the meaning behind the structures that we write leads to code
26:55
structures that we write leads to code
26:55
structures that we write leads to code reviews that look like this where we
26:56
reviews that look like this where we
26:56
reviews that look like this where we start by litigating small details in a
27:00
start by litigating small details in a
27:00
start by litigating small details in a useful sense because it gives us an
27:01
useful sense because it gives us an
27:01
useful sense because it gives us an opportunity to better understand some of
27:02
opportunity to better understand some of
27:02
opportunity to better understand some of the constraints that we're operating
27:03
the constraints that we're operating
27:03
the constraints that we're operating under and some of the details of our
27:05
under and some of the details of our
27:05
under and some of the details of our tools and then melts away the complexity
27:07
tools and then melts away the complexity
27:07
tools and then melts away the complexity until we discover oh actually we didn't
27:09
until we discover oh actually we didn't
27:09
until we discover oh actually we didn't even to do this at all this this thing
27:11
even to do this at all this this thing
27:11
even to do this at all this this thing didn't have to exist the code review is
27:12
didn't have to exist the code review is
27:12
didn't have to exist the code review is moot we shouldn't even be going in this
27:14
moot we shouldn't even be going in this
27:15
moot we shouldn't even be going in this direction but I want to back up a little
27:16
direction but I want to back up a little
27:16
direction but I want to back up a little bit to the original bug and the original
27:19
bit to the original bug and the original
27:19
bit to the original bug and the original bug is probably one of the most obvious
27:21
bug is probably one of the most obvious
27:21
bug is probably one of the most obvious bugs and yet appears all over the place
27:24
bugs and yet appears all over the place
27:24
bugs and yet appears all over the place throughout the design of so many tools
27:26
throughout the design of so many tools
27:26
throughout the design of so many tools it's a classical inband versus out of
27:28
it's a classical inband versus out of
27:28
it's a classical inband versus out of Bandon coding problem where if we think
27:31
Bandon coding problem where if we think
27:31
Bandon coding problem where if we think about what an inband and coding is it's
27:33
about what an inband and coding is it's
27:33
about what an inband and coding is it's any case that we could describe as
27:34
any case that we could describe as
27:34
any case that we could describe as having a single Channel through which
27:36
having a single Channel through which
27:36
having a single Channel through which we're encoding distinct data or distinct
27:38
we're encoding distinct data or distinct
27:38
we're encoding distinct data or distinct data and metadata whereas outof band
27:40
data and metadata whereas outof band
27:40
data and metadata whereas outof band would mean using distinct channels for
27:42
would mean using distinct channels for
27:42
would mean using distinct channels for encoding data and
27:44
encoding data and
27:44
encoding data and metadata and whether we're talking about
27:47
metadata and whether we're talking about
27:47
metadata and whether we're talking about two data signals or we're talking about
27:49
two data signals or we're talking about
27:49
two data signals or we're talking about a data and metadata signal really just
27:51
a data and metadata signal really just
27:51
a data and metadata signal really just comes down to the the case but it really
27:53
comes down to the the case but it really
27:53
comes down to the the case but it really is do we need to encode two things in
27:55
is do we need to encode two things in
27:55
is do we need to encode two things in the space of one or do we keep them
27:56
the space of one or do we keep them
27:56
the space of one or do we keep them separate and this is a problem that
27:58
separate and this is a problem that
27:58
separate and this is a problem that shows up all over the place in very
28:01
shows up all over the place in very
28:01
shows up all over the place in very distinct manifestations and yet is
28:04
distinct manifestations and yet is
28:04
distinct manifestations and yet is underlying the same problem for example
28:06
underlying the same problem for example
28:06
underlying the same problem for example all of you have written a shell script
28:07
all of you have written a shell script
28:07
all of you have written a shell script somewhere that used find in order to
28:10
somewhere that used find in order to
28:10
somewhere that used find in order to find the files that existed on disk and
28:13
find the files that existed on disk and
28:13
find the files that existed on disk and then you said oh I want to do an
28:14
then you said oh I want to do an
28:14
then you said oh I want to do an operation on each one of these files so
28:16
operation on each one of these files so
28:16
operation on each one of these files so you pipe that to exarc let me find out
28:18
you pipe that to exarc let me find out
28:18
you pipe that to exarc let me find out how much disk space is being used by
28:20
how much disk space is being used by
28:20
how much disk space is being used by each of these files except one of the
28:22
each of these files except one of the
28:22
each of these files except one of the files has a space in its name and this
28:25
files has a space in its name and this
28:25
files has a space in its name and this blows up in your face and if you went to
28:27
blows up in your face and if you went to
28:27
blows up in your face and if you went to your user users and you said well just
28:29
your user users and you said well just
28:29
your user users and you said well just don't have spaces in the names of your
28:31
don't have spaces in the names of your
28:31
don't have spaces in the names of your files that's probably the wrong answer
28:33
files that's probably the wrong answer
28:33
files that's probably the wrong answer because you made a mistake here and the
28:35
because you made a mistake here and the
28:35
because you made a mistake here and the mistake in the design of this is an
28:37
mistake in the design of this is an
28:37
mistake in the design of this is an inband encoding mistake fundamentally if
28:39
inband encoding mistake fundamentally if
28:39
inband encoding mistake fundamentally if we think about how the system shell
28:41
we think about how the system shell
28:41
we think about how the system shell works on Linux well shell scripting on
28:44
works on Linux well shell scripting on
28:44
works on Linux well shell scripting on Linux has a single binary Channel by
28:47
Linux has a single binary Channel by
28:47
Linux has a single binary Channel by which it transmits all information
28:49
which it transmits all information
28:49
which it transmits all information between processes the standard out on
28:51
between processes the standard out on
28:51
between processes the standard out on one side and the standard in the other
28:52
one side and the standard in the other
28:52
one side and the standard in the other side and as a consequence the connection
28:55
side and as a consequence the connection
28:55
side and as a consequence the connection between find and xargs needs to trans
28:58
between find and xargs needs to trans
28:58
between find and xargs needs to trans MIT over that single binary channel two
29:00
MIT over that single binary channel two
29:00
MIT over that single binary channel two pieces of information the names of the
29:02
pieces of information the names of the
29:02
pieces of information the names of the file and the delimiters between the
29:04
file and the delimiters between the
29:04
file and the delimiters between the files the standard on on these systems
29:08
files the standard on on these systems
29:08
files the standard on on these systems the standard with these tools is to
29:09
the standard with these tools is to
29:10
the standard with these tools is to space separate these or to new line
29:11
space separate these or to new line
29:11
space separate these or to new line separate these and so we have an
29:13
separate these and so we have an
29:13
separate these and so we have an ambiguity is the space a space
29:16
ambiguity is the space a space
29:16
ambiguity is the space a space separating the different file names
29:18
separating the different file names
29:18
separating the different file names metadata or is that space part of the
29:21
metadata or is that space part of the
29:21
metadata or is that space part of the file name data classical inband encoding
29:24
file name data classical inband encoding
29:24
file name data classical inband encoding ambiguity and every time we see an
29:27
ambiguity and every time we see an
29:27
ambiguity and every time we see an inband cting ambiguity there's pretty
29:29
inband cting ambiguity there's pretty
29:29
inband cting ambiguity there's pretty much the same approaches that we take to
29:30
much the same approaches that we take to
29:30
much the same approaches that we take to disambiguate that in the case of this we
29:32
disambiguate that in the case of this we
29:32
disambiguate that in the case of this we can either partition we can prefix we
29:34
can either partition we can prefix we
29:34
can either partition we can prefix we can escape or quote all zero in on
29:36
can escape or quote all zero in on
29:36
can escape or quote all zero in on partitioning what that generally means
29:38
partitioning what that generally means
29:38
partitioning what that generally means is you partition the space of values and
29:39
is you partition the space of values and
29:39
is you partition the space of values and indicate that some values can only ever
29:41
indicate that some values can only ever
29:41
indicate that some values can only ever be data and some values can only ever be
29:43
be data and some values can only ever be
29:43
be data and some values can only ever be metadata in which case for the shell
29:45
metadata in which case for the shell
29:45
metadata in which case for the shell scripting example will you say uh
29:48
scripting example will you say uh
29:48
scripting example will you say uh backslash and a null bite can never
29:49
backslash and a null bite can never
29:49
backslash and a null bite can never appear in a file name on Linux a
29:52
appear in a file name on Linux a
29:52
appear in a file name on Linux a backslash may be able to appear in a
29:53
backslash may be able to appear in a
29:53
backslash may be able to appear in a file name on other file systems but as
29:55
file name on other file systems but as
29:55
file name on other file systems but as far as I know there are no file systems
29:57
far as I know there are no file systems
29:57
far as I know there are no file systems ever that that allow a null bite to
29:59
ever that that allow a null bite to
29:59
ever that that allow a null bite to appear in the file name which means if
30:01
appear in the file name which means if
30:01
appear in the file name which means if we delimit them if we partition the
30:02
we delimit them if we partition the
30:02
we delimit them if we partition the space of all the characters that we have
30:04
space of all the characters that we have
30:04
space of all the characters that we have in that singular Channel such that
30:05
in that singular Channel such that
30:06
in that singular Channel such that anything that's not a null bite is data
30:07
anything that's not a null bite is data
30:07
anything that's not a null bite is data and anything that is a null bite is
30:09
and anything that is a null bite is
30:09
and anything that is a null bite is metadata the delimiter well we can
30:12
metadata the delimiter well we can
30:12
metadata the delimiter well we can actually unambiguously delimit these
30:14
actually unambiguously delimit these
30:14
actually unambiguously delimit these file names if we take a look at the
30:16
file names if we take a look at the
30:16
file names if we take a look at the actual tools we use we'll see that they
30:17
actual tools we use we'll see that they
30:17
actual tools we use we'll see that they implement this find has a print zero
30:18
implement this find has a print zero
30:18
implement this find has a print zero argument XRS has A- zero argument and
30:21
argument XRS has A- zero argument and
30:21
argument XRS has A- zero argument and we've solved the problem but the point
30:23
we've solved the problem but the point
30:23
we've solved the problem but the point of this example is not to show you how
30:25
of this example is not to show you how
30:25
of this example is not to show you how to fix a shell scripting problem but to
30:26
to fix a shell scripting problem but to
30:26
to fix a shell scripting problem but to show you where Foundation knowledge ends
30:28
show you where Foundation knowledge ends
30:28
show you where Foundation knowledge ends up being much more powerful than you
30:30
up being much more powerful than you
30:30
up being much more powerful than you think and why the search for
30:32
think and why the search for
30:32
think and why the search for foundational knowledge is a critical
30:33
foundational knowledge is a critical
30:33
foundational knowledge is a critical thing in your code review process
30:34
thing in your code review process
30:34
thing in your code review process because if we think about things like
30:36
because if we think about things like
30:36
because if we think about things like floating Point numbers i e 754 and we
30:39
floating Point numbers i e 754 and we
30:39
floating Point numbers i e 754 and we look at all the bit patterns of our i e
30:41
look at all the bit patterns of our i e
30:41
look at all the bit patterns of our i e 754 we will see that certain bit
30:43
754 we will see that certain bit
30:43
754 we will see that certain bit patterns have been carved out some
30:46
patterns have been carved out some
30:46
patterns have been carved out some represent actual values and some
30:47
represent actual values and some
30:47
represent actual values and some represent metadata Infinities nans and
30:50
represent metadata Infinities nans and
30:50
represent metadata Infinities nans and payload n and as a consequence i e 754
30:53
payload n and as a consequence i e 754
30:53
payload n and as a consequence i e 754 floating Point has an inband
30:55
floating Point has an inband
30:55
floating Point has an inband partitioning where it can partition this
30:57
partitioning where it can partition this
30:57
partitioning where it can partition this is actually ual data and this is
30:58
is actually ual data and this is
30:58
is actually ual data and this is metadata specifically Nan representing
31:00
metadata specifically Nan representing
31:00
metadata specifically Nan representing the absence of a measurement which is
31:02
the absence of a measurement which is
31:03
the absence of a measurement which is why if we happen to be Panda users one
31:05
why if we happen to be Panda users one
31:05
why if we happen to be Panda users one of the things that we've almost
31:06
of the things that we've almost
31:06
of the things that we've almost certainly run into in earlier versions
31:08
certainly run into in earlier versions
31:08
certainly run into in earlier versions of pandas was if we had a pandas series
31:10
of pandas was if we had a pandas series
31:10
of pandas was if we had a pandas series a pandas data frame or even a pandas
31:11
a pandas data frame or even a pandas
31:11
a pandas data frame or even a pandas array and we wanted to represent there
31:13
array and we wanted to represent there
31:13
array and we wanted to represent there was missing data we could do that if
31:15
was missing data we could do that if
31:15
was missing data we could do that if it's floating Point data because
31:16
it's floating Point data because
31:16
it's floating Point data because floating Point has that partitioning but
31:18
floating Point has that partitioning but
31:18
floating Point has that partitioning but we couldn't do that if it was integer
31:19
we couldn't do that if it was integer
31:20
we couldn't do that if it was integer data because integer data well how do we
31:21
data because integer data well how do we
31:22
data because integer data well how do we distinguish is a zero value an actual
31:23
distinguish is a zero value an actual
31:24
distinguish is a zero value an actual zero or a zero representing something
31:25
zero or a zero representing something
31:25
zero or a zero representing something that's missing and as a consequence we
31:27
that's missing and as a consequence we
31:27
that's missing and as a consequence we need either some sort of inband
31:30
need either some sort of inband
31:30
need either some sort of inband disambiguation mechanism of which there
31:31
disambiguation mechanism of which there
31:31
disambiguation mechanism of which there really isn't a good one or we need an
31:33
really isn't a good one or we need an
31:33
really isn't a good one or we need an outof band encoding and we could think
31:35
outof band encoding and we could think
31:35
outof band encoding and we could think about how an capital I N64 D type in a
31:39
about how an capital I N64 D type in a
31:39
about how an capital I N64 D type in a pandas in pandas is that out of Bandon
31:42
pandas in pandas is that out of Bandon
31:42
pandas in pandas is that out of Bandon coding it's something very similar to a
31:43
coding it's something very similar to a
31:43
coding it's something very similar to a numpy mast array which represents the
31:46
numpy mast array which represents the
31:46
numpy mast array which represents the missing values separate from the values
31:48
missing values separate from the values
31:48
missing values separate from the values themselves it's a classical outof band
31:49
themselves it's a classical outof band
31:49
themselves it's a classical outof band encoding and despite being in a
31:52
encoding and despite being in a
31:52
encoding and despite being in a completely different domain with a
31:54
completely different domain with a
31:54
completely different domain with a completely different programming
31:55
completely different programming
31:55
completely different programming language this is literally the exact
31:57
language this is literally the exact
31:57
language this is literally the exact same problem the inband and coding
31:59
same problem the inband and coding
31:59
same problem the inband and coding ambiguity problem and the mechanisms by
32:00
ambiguity problem and the mechanisms by
32:00
ambiguity problem and the mechanisms by which we disambiguate either within an
32:02
which we disambiguate either within an
32:02
which we disambiguate either within an inband and coding or the reasoning by
32:04
inband and coding or the reasoning by
32:04
inband and coding or the reasoning by which we push ourselves towards an out
32:05
which we push ourselves towards an out
32:05
which we push ourselves towards an out of Bandon coding a as the shell
32:07
of Bandon coding a as the shell
32:07
of Bandon coding a as the shell scripting example two very different
32:08
scripting example two very different
32:08
scripting example two very different domains two very different problems end
32:11
domains two very different problems end
32:11
domains two very different problems end up being the underlying same problem
32:14
up being the underlying same problem
32:14
up being the underlying same problem that you can unify and take from
32:16
that you can unify and take from
32:16
that you can unify and take from something that's relatively
32:17
something that's relatively
32:17
something that's relatively controversial to something that's just
32:18
controversial to something that's just
32:18
controversial to something that's just oh yeah this's an ined En coding
32:19
oh yeah this's an ined En coding
32:19
oh yeah this's an ined En coding ambiguity here are our choices and
32:21
ambiguity here are our choices and
32:21
ambiguity here are our choices and methodically work through them in fact
32:23
methodically work through them in fact
32:23
methodically work through them in fact even if we look at the design of
32:25
even if we look at the design of
32:25
even if we look at the design of different tools like for example our
32:27
different tools like for example our
32:27
different tools like for example our system path on on Linux which is colon
32:30
system path on on Linux which is colon
32:30
system path on on Linux which is colon separated or tools like Docker where if
32:32
separated or tools like Docker where if
32:32
separated or tools like Docker where if you're mounting a volume on Docker you
32:34
you're mounting a volume on Docker you
32:34
you're mounting a volume on Docker you have the source and the destination
32:36
have the source and the destination
32:36
have the source and the destination separated by a colon when you look at
32:38
separated by a colon when you look at
32:38
separated by a colon when you look at this you say well that's going to be an
32:40
this you say well that's going to be an
32:40
this you say well that's going to be an inandon coding ambiguity because what if
32:41
inandon coding ambiguity because what if
32:41
inandon coding ambiguity because what if there's a colon in the directory name
32:45
there's a colon in the directory name
32:45
there's a colon in the directory name well if there is a colon in the
32:46
well if there is a colon in the
32:46
well if there is a colon in the directory name you can't mount it in a
32:49
directory name you can't mount it in a
32:49
directory name you can't mount it in a Docker container Docker a tool that was
32:52
Docker container Docker a tool that was
32:52
Docker container Docker a tool that was written very recently by very
32:54
written very recently by very
32:54
written very recently by very knowledgeable people has exactly the
32:56
knowledgeable people has exactly the
32:56
knowledgeable people has exactly the same flaw as all these other tools
32:57
same flaw as all these other tools
32:58
same flaw as all these other tools because we didn't search for that
32:59
because we didn't search for that
32:59
because we didn't search for that foundational knowledge explicate that
33:01
foundational knowledge explicate that
33:01
foundational knowledge explicate that foundational knowledge and bring that to
33:02
foundational knowledge and bring that to
33:02
foundational knowledge and bring that to light because once you see these
33:04
light because once you see these
33:04
light because once you see these problems you can never unse them and
33:06
problems you can never unse them and
33:06
problems you can never unse them and then you look at the design of better
33:07
then you look at the design of better
33:07
then you look at the design of better tools tools like BP and you say oh
33:09
tools tools like BP and you say oh
33:09
tools tools like BP and you say oh that's why BP has separate Arguments for
33:12
that's why BP has separate Arguments for
33:12
that's why BP has separate Arguments for the source and the destination when
33:14
the source and the destination when
33:14
the source and the destination when you're doing a a mounting here because
33:17
you're doing a a mounting here because
33:17
you're doing a a mounting here because there's an out abandon coding that is
33:19
there's an out abandon coding that is
33:19
there's an out abandon coding that is created by the shells by the Shell's uh
33:22
created by the shells by the Shell's uh
33:22
created by the shells by the Shell's uh globbing mechanism and the shell
33:23
globbing mechanism and the shell
33:23
globbing mechanism and the shell delineating these into different
33:25
delineating these into different
33:25
delineating these into different arguments of your arcv
33:28
arguments of your arcv
33:28
arguments of your arcv when you find that foundational
33:29
when you find that foundational
33:29
when you find that foundational knowledge it's very much the case that
33:30
knowledge it's very much the case that
33:30
knowledge it's very much the case that once you see these things you can never
33:32
once you see these things you can never
33:32
once you see these things you can never unsee them and this is the value of
33:33
unsee them and this is the value of
33:33
unsee them and this is the value of finding that foundational knowledge this
33:35
finding that foundational knowledge this
33:35
finding that foundational knowledge this is the value of making your code review
33:37
is the value of making your code review
33:37
is the value of making your code review processes a an opportunity to search for
33:39
processes a an opportunity to search for
33:40
processes a an opportunity to search for that knowledge itself so that you're not
33:41
that knowledge itself so that you're not
33:41
that knowledge itself so that you're not just litigating some tiny detail about
33:44
just litigating some tiny detail about
33:44
just litigating some tiny detail about one piece of code in one system at one
33:47
one piece of code in one system at one
33:47
one piece of code in one system at one point in time but searching for
33:48
point in time but searching for
33:48
point in time but searching for knowledge that tells you about all
33:51
knowledge that tells you about all
33:51
knowledge that tells you about all instances of patterns similar like that
33:53
instances of patterns similar like that
33:53
instances of patterns similar like that across all projects across all time it
33:57
across all projects across all time it
33:57
across all projects across all time it is often the case that the the search
33:59
is often the case that the the search
33:59
is often the case that the the search for that foundational knowledge that
34:00
for that foundational knowledge that
34:00
for that foundational knowledge that foundational underlying Theory takes
34:02
foundational underlying Theory takes
34:02
foundational underlying Theory takes things that were once controversial
34:03
things that were once controversial
34:03
things that were once controversial questions that turned into Flores in
34:05
questions that turned into Flores in
34:05
questions that turned into Flores in your code review process with one person
34:06
your code review process with one person
34:06
your code review process with one person saying one thing the other person saying
34:08
saying one thing the other person saying
34:08
saying one thing the other person saying the other thing and everybody collecting
34:10
the other thing and everybody collecting
34:10
the other thing and everybody collecting evidence on both sides but nobody coming
34:12
evidence on both sides but nobody coming
34:12
evidence on both sides but nobody coming to a conclusion and it turns them into
34:14
to a conclusion and it turns them into
34:14
to a conclusion and it turns them into just obvious answers here's the here's
34:17
just obvious answers here's the here's
34:17
just obvious answers here's the here's the methodical approach for
34:19
the methodical approach for
34:19
the methodical approach for understanding this problem it leads to
34:20
understanding this problem it leads to
34:20
understanding this problem it leads to these conclusions maybe both people were
34:22
these conclusions maybe both people were
34:22
these conclusions maybe both people were right if we take a look at it through
34:24
right if we take a look at it through
34:24
right if we take a look at it through this lens and through this framework the
34:26
this lens and through this framework the
34:26
this lens and through this framework the underlying theory is of often lurking in
34:28
underlying theory is of often lurking in
34:28
underlying theory is of often lurking in plain sight and yet there seems to be
34:30
plain sight and yet there seems to be
34:30
plain sight and yet there seems to be this resistance in the code review
34:31
this resistance in the code review
34:31
this resistance in the code review process to really dig into what that
34:32
process to really dig into what that
34:32
process to really dig into what that underlying theory is for example one of
34:36
underlying theory is for example one of
34:36
underlying theory is for example one of the other hugely controversial things
34:39
the other hugely controversial things
34:39
the other hugely controversial things that I see all over the place that comes
34:41
that I see all over the place that comes
34:41
that I see all over the place that comes up in code reviews all the time and yet
34:43
up in code reviews all the time and yet
34:43
up in code reviews all the time and yet it's not a particularly controversial
34:44
it's not a particularly controversial
34:44
it's not a particularly controversial question is in this code review should
34:47
question is in this code review should
34:47
question is in this code review should this thing be a function and people say
34:48
this thing be a function and people say
34:49
this thing be a function and people say things like slogans don't repeat
34:50
things like slogans don't repeat
34:50
things like slogans don't repeat yourself any time you're in a code
34:52
yourself any time you're in a code
34:52
yourself any time you're in a code review and you're not talking about some
34:54
review and you're not talking about some
34:54
review and you're not talking about some foundational knowledge you're not
34:55
foundational knowledge you're not
34:55
foundational knowledge you're not searching for that foundational Theory
34:57
searching for that foundational Theory
34:57
searching for that foundational Theory but you're instead relying on simple
34:59
but you're instead relying on simple
34:59
but you're instead relying on simple slogans there's a very good chance that
35:01
slogans there's a very good chance that
35:01
slogans there's a very good chance that you're misleading yourself because there
35:03
you're misleading yourself because there
35:04
you're misleading yourself because there is theory behind this we don't have to
35:05
is theory behind this we don't have to
35:05
is theory behind this we don't have to rely on silly things like don't repeat
35:07
rely on silly things like don't repeat
35:07
rely on silly things like don't repeat yourself we could reach back to theory
35:09
yourself we could reach back to theory
35:09
yourself we could reach back to theory that already exists like the theory of
35:10
that already exists like the theory of
35:10
that already exists like the theory of normalization where in the theory of
35:13
normalization where in the theory of
35:13
normalization where in the theory of database normalization there's a topic
35:15
database normalization there's a topic
35:15
database normalization there's a topic of why we why we normalized and one of
35:18
of why we why we normalized and one of
35:18
of why we why we normalized and one of the reasons we normalized to eliminate
35:19
the reasons we normalized to eliminate
35:19
the reasons we normalized to eliminate update anomalies you know you perform an
35:21
update anomalies you know you perform an
35:21
update anomalies you know you perform an update in one part of the system but
35:22
update in one part of the system but
35:22
update in one part of the system but that update doesn't flow to another part
35:23
that update doesn't flow to another part
35:24
that update doesn't flow to another part of the system you update somebody's
35:25
of the system you update somebody's
35:25
of the system you update somebody's address in this table but the mail still
35:26
address in this table but the mail still
35:26
address in this table but the mail still goes to the wrong address because there
35:28
goes to the wrong address because there
35:28
goes to the wrong address because there was another system and that leads us to
35:30
was another system and that leads us to
35:30
was another system and that leads us to say oh oh the way we do that is we
35:31
say oh oh the way we do that is we
35:31
say oh oh the way we do that is we create a single source of Truth well if
35:33
create a single source of Truth well if
35:33
create a single source of Truth well if we think of what a function is a
35:34
we think of what a function is a
35:34
we think of what a function is a function is a single source of truth
35:36
function is a single source of truth
35:36
function is a single source of truth it's a single implementation of some me
35:39
it's a single implementation of some me
35:39
it's a single implementation of some me some mechanism such that if that
35:41
some mechanism such that if that
35:41
some mechanism such that if that mechanism changes it changes in one
35:43
mechanism changes it changes in one
35:43
mechanism changes it changes in one place there's no risk of update anomaly
35:45
place there's no risk of update anomaly
35:45
place there's no risk of update anomaly well that's really what a function is
35:47
well that's really what a function is
35:47
well that's really what a function is and all of these arguments back and
35:49
and all of these arguments back and
35:49
and all of these arguments back and forth should this be our function or not
35:50
forth should this be our function or not
35:50
forth should this be our function or not some of they may be legitimate but in
35:52
some of they may be legitimate but in
35:52
some of they may be legitimate but in many cases we're just dancing around the
35:54
many cases we're just dancing around the
35:54
many cases we're just dancing around the underlying theory of normalization the
35:56
underlying theory of normalization the
35:56
underlying theory of normalization the underlying theory of rep repetition
35:58
underlying theory of rep repetition
35:58
underlying theory of rep repetition being either coincidental two pieces of
36:00
being either coincidental two pieces of
36:00
being either coincidental two pieces of code just happen to be the same at some
36:02
code just happen to be the same at some
36:02
code just happen to be the same at some point in time or intentional two pieces
36:04
point in time or intentional two pieces
36:04
point in time or intentional two pieces of code are the same because they must
36:06
of code are the same because they must
36:06
of code are the same because they must always move in Tandem and they must
36:08
always move in Tandem and they must
36:08
always move in Tandem and they must always be the same if we dig just a
36:11
always be the same if we dig just a
36:11
always be the same if we dig just a little bit beneath existing theory of
36:14
little bit beneath existing theory of
36:14
little bit beneath existing theory of normalization we can discover that
36:16
normalization we can discover that
36:16
normalization we can discover that there's just one tiny amendment to this
36:18
there's just one tiny amendment to this
36:18
there's just one tiny amendment to this the idea that repetition could be
36:20
the idea that repetition could be
36:20
the idea that repetition could be coincidental that two pieces of code
36:22
coincidental that two pieces of code
36:22
coincidental that two pieces of code that look the same might just happen to
36:24
that look the same might just happen to
36:24
that look the same might just happen to look the same because at some point in
36:25
look the same because at some point in
36:25
look the same because at some point in time somebody wanted them to be the same
36:27
time somebody wanted them to be the same
36:27
time somebody wanted them to be the same but they could rightfully and very
36:29
but they could rightfully and very
36:29
but they could rightfully and very readily diverge or two pieces of code
36:32
readily diverge or two pieces of code
36:32
readily diverge or two pieces of code must always be the same and if one
36:34
must always be the same and if one
36:34
must always be the same and if one changes the other must change if you
36:36
changes the other must change if you
36:36
changes the other must change if you have intentional repetition then you
36:39
have intentional repetition then you
36:39
have intentional repetition then you unify and you definitely want to write a
36:40
unify and you definitely want to write a
36:40
unify and you definitely want to write a function if you have coincidental
36:43
function if you have coincidental
36:43
function if you have coincidental repetition then it's likely that you've
36:45
repetition then it's likely that you've
36:45
repetition then it's likely that you've assessed it to be coincidental
36:47
assessed it to be coincidental
36:47
assessed it to be coincidental repetition by assessing that it's likely
36:49
repetition by assessing that it's likely
36:49
repetition by assessing that it's likely to diverge over time which means that
36:51
to diverge over time which means that
36:51
to diverge over time which means that managing that Divergence over time will
36:53
managing that Divergence over time will
36:53
managing that Divergence over time will cause modalities to sprawl in the code
36:55
cause modalities to sprawl in the code
36:55
cause modalities to sprawl in the code likely deep in the code then requiring
36:57
likely deep in the code then requiring
36:58
likely deep in the code then requiring to revalidate things the two pieces of
36:59
to revalidate things the two pieces of
36:59
to revalidate things the two pieces of code just happen to be the same at one
37:01
code just happen to be the same at one
37:01
code just happen to be the same at one point in time somebody changes this but
37:03
point in time somebody changes this but
37:03
point in time somebody changes this but they've already been unified into a
37:04
they've already been unified into a
37:04
they've already been unified into a function so somebody add a keyword
37:05
function so somebody add a keyword
37:06
function so somebody add a keyword argument to this function but they have
37:07
argument to this function but they have
37:07
argument to this function but they have to thread that keyword argument three
37:08
to thread that keyword argument three
37:08
to thread that keyword argument three functions deep they have to retest
37:10
functions deep they have to retest
37:10
functions deep they have to retest almost the entire framework to make sure
37:12
almost the entire framework to make sure
37:12
almost the entire framework to make sure this didn't disrupt anybody else's work
37:14
this didn't disrupt anybody else's work
37:14
this didn't disrupt anybody else's work and in this case we can see a deliberate
37:17
and in this case we can see a deliberate
37:17
and in this case we can see a deliberate piece of found Theory that's not
37:19
piece of found Theory that's not
37:19
piece of found Theory that's not particularly formalized that gives us
37:22
particularly formalized that gives us
37:22
particularly formalized that gives us methodical guidance as part of our code
37:24
methodical guidance as part of our code
37:24
methodical guidance as part of our code review process for what we should do
37:27
review process for what we should do
37:27
review process for what we should do removing all of the subjectivity and all
37:29
removing all of the subjectivity and all
37:30
removing all of the subjectivity and all of the controversy the unnatural
37:32
of the controversy the unnatural
37:32
of the controversy the unnatural unnecessary controversy that might arise
37:34
unnecessary controversy that might arise
37:34
unnecessary controversy that might arise all of the flamor that might arise as
37:36
all of the flamor that might arise as
37:36
all of the flamor that might arise as this what is a culture of knowledge well
37:38
this what is a culture of knowledge well
37:38
this what is a culture of knowledge well a culture of knowledge is about avoiding
37:40
a culture of knowledge is about avoiding
37:40
a culture of knowledge is about avoiding these intellectual shortcuts avoiding
37:42
these intellectual shortcuts avoiding
37:42
these intellectual shortcuts avoiding these slogans always be willing to ask
37:44
these slogans always be willing to ask
37:44
these slogans always be willing to ask why when somebody says don't repeat
37:46
why when somebody says don't repeat
37:46
why when somebody says don't repeat yourself say well actually I do repeat
37:49
yourself say well actually I do repeat
37:49
yourself say well actually I do repeat myself sometimes and sometimes I do that
37:51
myself sometimes and sometimes I do that
37:51
myself sometimes and sometimes I do that and it's worthwhile why is it that I'm
37:53
and it's worthwhile why is it that I'm
37:53
and it's worthwhile why is it that I'm not taking my own advice self-reflect
37:56
not taking my own advice self-reflect
37:56
not taking my own advice self-reflect and try and figure out well maybe I'm
37:59
and try and figure out well maybe I'm
37:59
and try and figure out well maybe I'm right I'm just phrasing this wrong don't
38:02
right I'm just phrasing this wrong don't
38:02
right I'm just phrasing this wrong don't repeat
38:03
repeat
38:03
repeat yourself intentionally but do repeat
38:06
yourself intentionally but do repeat
38:06
yourself intentionally but do repeat yourself
38:07
yourself
38:07
yourself coincidentally always dig for knowledge
38:10
coincidentally always dig for knowledge
38:10
coincidentally always dig for knowledge and that knowledge and know that that
38:11
and that knowledge and know that that
38:11
and that knowledge and know that that knowledge is often hiding in plain sight
38:13
knowledge is often hiding in plain sight
38:13
knowledge is often hiding in plain sight even though it may take you a while to
38:14
even though it may take you a while to
38:14
even though it may take you a while to dig for it when you take a look at a
38:16
dig for it when you take a look at a
38:16
dig for it when you take a look at a code riew really ask
38:17
code riew really ask
38:17
code riew really ask yourself what is the reasoning behind
38:21
yourself what is the reasoning behind
38:21
yourself what is the reasoning behind the advice that I'm given and is this
38:23
the advice that I'm given and is this
38:23
the advice that I'm given and is this generalized reasoning how do I collect
38:25
generalized reasoning how do I collect
38:25
generalized reasoning how do I collect all the evidence across all of the
38:26
all the evidence across all of the
38:26
all the evidence across all of the things that I do or don't do and unify
38:29
things that I do or don't do and unify
38:29
things that I do or don't do and unify those into One Singular piece of theory
38:31
those into One Singular piece of theory
38:31
those into One Singular piece of theory how we go about doing this is actually
38:33
how we go about doing this is actually
38:33
how we go about doing this is actually quite difficult and I can't really give
38:34
quite difficult and I can't really give
38:34
quite difficult and I can't really give you a lot of advice for how to develop
38:36
you a lot of advice for how to develop
38:36
you a lot of advice for how to develop this culture of knowledge or how to
38:37
this culture of knowledge or how to
38:37
this culture of knowledge or how to build that knowledge but I can give you
38:38
build that knowledge but I can give you
38:38
build that knowledge but I can give you one critical
38:40
one critical
38:40
one critical pointer typically the education that
38:43
pointer typically the education that
38:43
pointer typically the education that most programmers are subjected to is
38:46
most programmers are subjected to is
38:46
most programmers are subjected to is wholly quantitative they are hired for
38:48
wholly quantitative they are hired for
38:48
wholly quantitative they are hired for their quantitative skill they are hired
38:50
their quantitative skill they are hired
38:50
their quantitative skill they are hired because they did well in computer
38:51
because they did well in computer
38:51
because they did well in computer science courses in engineering courses
38:53
science courses in engineering courses
38:53
science courses in engineering courses in math courses in courses of that
38:56
in math courses in courses of that
38:56
in math courses in courses of that nature and they don't understand that
38:59
nature and they don't understand that
38:59
nature and they don't understand that knowledge can take many forms including
39:01
knowledge can take many forms including
39:02
knowledge can take many forms including non-quantifiable forms there is such
39:05
non-quantifiable forms there is such
39:05
non-quantifiable forms there is such thing as qualified objective knowledge
39:08
thing as qualified objective knowledge
39:08
thing as qualified objective knowledge and understanding that sometimes while
39:10
and understanding that sometimes while
39:10
and understanding that sometimes while we may not be able to say everything
39:13
we may not be able to say everything
39:13
we may not be able to say everything definitively about a particular question
39:16
definitively about a particular question
39:16
definitively about a particular question we may not be able to enumerate facts
39:18
we may not be able to enumerate facts
39:18
we may not be able to enumerate facts and figures or create formula we can
39:20
and figures or create formula we can
39:20
and figures or create formula we can still form formulate
39:23
still form formulate
39:23
still form formulate rigorous
39:25
rigorous
39:25
rigorous thorough arguments about something
39:28
thorough arguments about something
39:28
thorough arguments about something making use of qualif qualified knowledge
39:30
making use of qualif qualified knowledge
39:30
making use of qualif qualified knowledge one thing that I would strongly
39:32
one thing that I would strongly
39:32
one thing that I would strongly encourage every programmer who went
39:34
encourage every programmer who went
39:34
encourage every programmer who went through a science technology engineering
39:37
through a science technology engineering
39:37
through a science technology engineering or mathematical background is take
39:39
or mathematical background is take
39:39
or mathematical background is take liberal arts classes learn about history
39:43
liberal arts classes learn about history
39:43
liberal arts classes learn about history learn about literature learn about art
39:45
learn about literature learn about art
39:45
learn about literature learn about art because the kinds of reasoning that are
39:47
because the kinds of reasoning that are
39:47
because the kinds of reasoning that are used in those fields are very foreign to
39:50
used in those fields are very foreign to
39:50
used in those fields are very foreign to us but they are part of the key to
39:52
us but they are part of the key to
39:53
us but they are part of the key to understanding the fundamental
39:54
understanding the fundamental
39:54
understanding the fundamental epistemology that allows us to really
39:56
epistemology that allows us to really
39:56
epistemology that allows us to really search for knowledge and perform code
39:59
search for knowledge and perform code
40:00
search for knowledge and perform code reviews that are not just litigating
40:01
reviews that are not just litigating
40:01
reviews that are not just litigating tiny details or relying on shortcuts or
40:05
tiny details or relying on shortcuts or
40:05
tiny details or relying on shortcuts or slogans finally one thing that we are
40:08
slogans finally one thing that we are
40:08
slogans finally one thing that we are often missing that leads to code reviews
40:09
often missing that leads to code reviews
40:10
often missing that leads to code reviews not being particularly useful is a
40:11
not being particularly useful is a
40:11
not being particularly useful is a culture of quality and just like the
40:13
culture of quality and just like the
40:13
culture of quality and just like the culture of knowledge it's very difficult
40:15
culture of knowledge it's very difficult
40:15
culture of knowledge it's very difficult for us to instill a culture of quality
40:16
for us to instill a culture of quality
40:16
for us to instill a culture of quality where none exists it's difficult for us
40:18
where none exists it's difficult for us
40:18
where none exists it's difficult for us to create rules or create guidances or
40:21
to create rules or create guidances or
40:21
to create rules or create guidances or write memos that say everybody focus on
40:24
write memos that say everybody focus on
40:24
write memos that say everybody focus on quality if that culture of quality
40:26
quality if that culture of quality
40:26
quality if that culture of quality doesn't exist I'm not sure how we create
40:27
doesn't exist I'm not sure how we create
40:27
doesn't exist I'm not sure how we create it but I do know that I've seen it exist
40:29
it but I do know that I've seen it exist
40:29
it but I do know that I've seen it exist in some places and I've seen it fail to
40:31
in some places and I've seen it fail to
40:31
in some places and I've seen it fail to exist in some places but before we talk
40:33
exist in some places but before we talk
40:33
exist in some places but before we talk about the cultural quality one important
40:35
about the cultural quality one important
40:35
about the cultural quality one important thing we should ask ourselves is does
40:37
thing we should ask ourselves is does
40:37
thing we should ask ourselves is does code quality really matter because many
40:40
code quality really matter because many
40:40
code quality really matter because many of you are going to say yes of course it
40:42
of you are going to say yes of course it
40:42
of you are going to say yes of course it matters bad code quality leads to bad
40:44
matters bad code quality leads to bad
40:44
matters bad code quality leads to bad code that has security faults that leads
40:46
code that has security faults that leads
40:46
code that has security faults that leads to you losing money that leads to you
40:48
to you losing money that leads to you
40:48
to you losing money that leads to you losing customers but some of you are
40:50
losing customers but some of you are
40:50
losing customers but some of you are also going to say well actually I have
40:52
also going to say well actually I have
40:52
also going to say well actually I have some colleagues and they work at a
40:54
some colleagues and they work at a
40:54
some colleagues and they work at a company and their code is atrocious and
40:57
company and their code is atrocious and
40:57
company and their code is atrocious and yet they make money hand over fist in
41:00
yet they make money hand over fist in
41:00
yet they make money hand over fist in fact between you and me we do quite a
41:03
fact between you and me we do quite a
41:03
fact between you and me we do quite a bit of Consulting at some very big tech
41:05
bit of Consulting at some very big tech
41:05
bit of Consulting at some very big tech companies in the US and you may be
41:08
companies in the US and you may be
41:08
companies in the US and you may be surprised to learn that at some of the
41:10
surprised to learn that at some of the
41:10
surprised to learn that at some of the top tech companies the code quality is
41:12
top tech companies the code quality is
41:12
top tech companies the code quality is not that good and they still make money
41:15
not that good and they still make money
41:15
not that good and they still make money hands over fist and if you were to give
41:18
hands over fist and if you were to give
41:18
hands over fist and if you were to give me the choice between having $1 billion
41:21
me the choice between having $1 billion
41:21
me the choice between having $1 billion or really good
41:22
or really good
41:22
or really good code I'm not sure that I would pick the
41:25
code I'm not sure that I would pick the
41:25
code I'm not sure that I would pick the really good code and so maybe we should
41:28
really good code and so maybe we should
41:28
really good code and so maybe we should ask the question does code quality
41:29
ask the question does code quality
41:29
ask the question does code quality really matter well to a particular
41:31
really matter well to a particular
41:31
really matter well to a particular audience to who does code quality matter
41:33
audience to who does code quality matter
41:33
audience to who does code quality matter to
41:34
to
41:34
to employers well you almost certainly have
41:37
employers well you almost certainly have
41:37
employers well you almost certainly have seen circumstances where a colleague of
41:39
seen circumstances where a colleague of
41:39
seen circumstances where a colleague of yours doesn't really know what they're
41:40
yours doesn't really know what they're
41:40
yours doesn't really know what they're doing but is able to talk very
41:42
doing but is able to talk very
41:42
doing but is able to talk very convincing and they're able to get a job
41:44
convincing and they're able to get a job
41:44
convincing and they're able to get a job and it's very difficult to measure code
41:46
and it's very difficult to measure code
41:46
and it's very difficult to measure code quality in an interview circumstance so
41:49
quality in an interview circumstance so
41:50
quality in an interview circumstance so maybe it's the case that two employers
41:51
maybe it's the case that two employers
41:51
maybe it's the case that two employers they kind of want a proxy for code
41:52
they kind of want a proxy for code
41:52
they kind of want a proxy for code quality but but also they kind of want
41:54
quality but but also they kind of want
41:54
quality but but also they kind of want somebody who will just do what they're
41:55
somebody who will just do what they're
41:55
somebody who will just do what they're told and will just ship
41:58
told and will just ship
41:58
told and will just ship features does code quality matter to the
42:00
features does code quality matter to the
42:00
features does code quality matter to the market you know we could say that if
42:02
market you know we could say that if
42:02
market you know we could say that if there's poor code quality that leads to
42:04
there's poor code quality that leads to
42:04
there's poor code quality that leads to security issues and security issues lead
42:05
security issues and security issues lead
42:05
security issues and security issues lead to bad press you might lose customers
42:07
to bad press you might lose customers
42:07
to bad press you might lose customers but the unfortunate truth about security
42:09
but the unfortunate truth about security
42:09
but the unfortunate truth about security issues is a lot of companies have had
42:12
issues is a lot of companies have had
42:12
issues is a lot of companies have had some very embarrassing security
42:16
some very embarrassing security
42:16
some very embarrassing security circumstances and don't seem to have
42:18
circumstances and don't seem to have
42:18
circumstances and don't seem to have lost much money at all maybe a temporary
42:20
lost much money at all maybe a temporary
42:20
lost much money at all maybe a temporary dip and they recover from that and so
42:23
dip and they recover from that and so
42:23
dip and they recover from that and so often times if we look at the market we
42:24
often times if we look at the market we
42:24
often times if we look at the market we can say you know quality does seem to
42:27
can say you know quality does seem to
42:27
can say you know quality does seem to matter there are better Goods that do
42:28
matter there are better Goods that do
42:28
matter there are better Goods that do better than worse Goods there are better
42:30
better than worse Goods there are better
42:30
better than worse Goods there are better design products that do better than
42:31
design products that do better than
42:31
design products that do better than worse design products there's also a lot
42:33
worse design products there's also a lot
42:33
worse design products there's also a lot of junk out there that seems to do very
42:35
of junk out there that seems to do very
42:35
of junk out there that seems to do very well so it's not certain to me whether
42:36
well so it's not certain to me whether
42:36
well so it's not certain to me whether or not code quality can be said to
42:40
or not code quality can be said to
42:40
or not code quality can be said to uniformly be of importance to the market
42:42
uniformly be of importance to the market
42:42
uniformly be of importance to the market or not that does it matter to the
42:44
or not that does it matter to the
42:44
or not that does it matter to the project well I would say yes but often
42:48
project well I would say yes but often
42:48
project well I would say yes but often times when I'm working on a project I'm
42:50
times when I'm working on a project I'm
42:50
times when I'm working on a project I'm working to a certain standard I'm
42:51
working to a certain standard I'm
42:51
working to a certain standard I'm working under certain deadlines and I
42:53
working under certain deadlines and I
42:53
working under certain deadlines and I know that the project evolves over time
42:55
know that the project evolves over time
42:55
know that the project evolves over time so I know if I ship bad code today it
42:57
so I know if I ship bad code today it
42:57
so I know if I ship bad code today it may probably be the case in 3 months I'm
42:59
may probably be the case in 3 months I'm
42:59
may probably be the case in 3 months I'm going to rewrite it anyway so why should
43:01
going to rewrite it anyway so why should
43:01
going to rewrite it anyway so why should I put all that effort into shipping the
43:02
I put all that effort into shipping the
43:02
I put all that effort into shipping the best thing I possibly can knowing that
43:04
best thing I possibly can knowing that
43:04
best thing I possibly can knowing that it's going to be Rewritten does it
43:06
it's going to be Rewritten does it
43:06
it's going to be Rewritten does it matter to my colleagues well I don't
43:07
matter to my colleagues well I don't
43:07
matter to my colleagues well I don't want to be cleaning after I don't want
43:09
want to be cleaning after I don't want
43:09
want to be cleaning after I don't want to be chasing after somebody else's code
43:11
to be chasing after somebody else's code
43:11
to be chasing after somebody else's code and trying to clean up after that so I
43:12
and trying to clean up after that so I
43:12
and trying to clean up after that so I guess it kind of does but but also it's
43:14
guess it kind of does but but also it's
43:14
guess it kind of does but but also it's the case that you know a lot of the code
43:15
the case that you know a lot of the code
43:15
the case that you know a lot of the code that I write is really not that
43:16
that I write is really not that
43:16
that I write is really not that important and as long as it's
43:17
important and as long as it's
43:17
important and as long as it's functionally correct if it's a little
43:19
functionally correct if it's a little
43:19
functionally correct if it's a little bit sloppy you know that just kind of
43:21
bit sloppy you know that just kind of
43:21
bit sloppy you know that just kind of fix it if there's a bug we'll fix it I
43:23
fix it if there's a bug we'll fix it I
43:23
fix it if there's a bug we'll fix it I think when we talk about quality and I
43:25
think when we talk about quality and I
43:25
think when we talk about quality and I think the key to a culture of quality is
43:27
think the key to a culture of quality is
43:27
think the key to a culture of quality is really asking ourselves does quality
43:30
really asking ourselves does quality
43:30
really asking ourselves does quality matter to you and for all of the other
43:33
matter to you and for all of the other
43:33
matter to you and for all of the other audiences sometimes it matters sometimes
43:35
audiences sometimes it matters sometimes
43:35
audiences sometimes it matters sometimes it doesn't and we can find evidence and
43:37
it doesn't and we can find evidence and
43:37
it doesn't and we can find evidence and arguments that it does matter or it
43:38
arguments that it does matter or it
43:38
arguments that it does matter or it doesn't matter and we can find very
43:39
doesn't matter and we can find very
43:39
doesn't matter and we can find very compelling arguments both ways but when
43:42
compelling arguments both ways but when
43:42
compelling arguments both ways but when I think about it deep in my heart
43:45
I think about it deep in my heart
43:45
I think about it deep in my heart quality matters to me I want to do my
43:47
quality matters to me I want to do my
43:47
quality matters to me I want to do my work to a standard I want to do that
43:49
work to a standard I want to do that
43:49
work to a standard I want to do that work to a standard that I think I can
43:51
work to a standard that I think I can
43:51
work to a standard that I think I can achieve and I want to push that standard
43:53
achieve and I want to push that standard
43:53
achieve and I want to push that standard forward and if I don't do good quality
43:55
forward and if I don't do good quality
43:55
forward and if I don't do good quality work I don't feel like I have done a
43:57
work I don't feel like I have done a
43:57
work I don't feel like I have done a good job and I don't feel good about
43:59
good job and I don't feel good about
43:59
good job and I don't feel good about myself and I think that if we're talking
44:01
myself and I think that if we're talking
44:01
myself and I think that if we're talking about a culture of quality and if we're
44:02
about a culture of quality and if we're
44:02
about a culture of quality and if we're talking about caring about quality it
44:04
talking about caring about quality it
44:04
talking about caring about quality it really is something that has to be built
44:06
really is something that has to be built
44:06
really is something that has to be built culturally from within you need to find
44:08
culturally from within you need to find
44:08
culturally from within you need to find people who fundamentally care about
44:10
people who fundamentally care about
44:10
people who fundamentally care about doing good work for the sake of doing
44:12
doing good work for the sake of doing
44:12
doing good work for the sake of doing good work so when we talk about a
44:14
good work so when we talk about a
44:14
good work so when we talk about a culture of quality well you know it's
44:16
culture of quality well you know it's
44:16
culture of quality well you know it's about working to a standard and wanting
44:18
about working to a standard and wanting
44:18
about working to a standard and wanting to head us set a high standard for
44:20
to head us set a high standard for
44:20
to head us set a high standard for yourself now if we think about it in the
44:23
yourself now if we think about it in the
44:23
yourself now if we think about it in the absence of such a culture of quality in
44:26
absence of such a culture of quality in
44:26
absence of such a culture of quality in the absence of a culture of knowledge or
44:28
the absence of a culture of knowledge or
44:28
the absence of a culture of knowledge or a culture of learning or a culture of
44:29
a culture of learning or a culture of
44:29
a culture of learning or a culture of training what are we really doing in a
44:31
training what are we really doing in a
44:31
training what are we really doing in a code review in fact I would argue that
44:33
code review in fact I would argue that
44:33
code review in fact I would argue that in those
44:34
in those
44:34
in those circumstances code
44:36
circumstances code
44:36
circumstances code review probably is
44:39
review probably is
44:39
review probably is useless I'm James Powell thank you all
44:41
useless I'm James Powell thank you all
44:41
useless I'm James Powell thank you all for joining me for this presentation
44:42
for joining me for this presentation
44:42
for joining me for this presentation it's been an absolute pleasure
44:45
it's been an absolute pleasure
44:45
it's been an absolute pleasure [Music]