Min-maxing_Software_Costs_-_Konstantin_Kudryashov_-_Laracon_EU_2015_Russian.srt

1
00:00:00,000 --> 00:00:23,042
alright everyone I do it so it's ok so
she asked me to talk at Larkin goes to

2
00:00:23,042 --> 00:00:28,064
the last year was obliged i enjoy to
throw out hope some of you that reappear

3
00:00:28,064 --> 00:00:34,121
in this year also enjoyed it and
basically would like but it's like local

4
00:00:35,021 --> 00:00:39,064
don't really care what you're talking
about but like what would be the talk

5
00:00:39,064 --> 00:00:49,963
under that so I wanted to basically I
can run my usual spill was like peace

6
00:00:50,539 --> 00:00:56,610
respect introduction or BDD or more than
by example like a basic theme not very

7
00:00:56,061 --> 00:01:01,152
connected to the Laurel and then we can
agreed on it and then a couple of next

8
00:01:02,052 --> 00:01:07,651
week's I was opening a checking it up
are going through the BDD hashtag and

9
00:01:08,119 --> 00:01:12,860
like seeing the street every now and
then which was like how do I unit test

10
00:01:12,086 --> 00:01:17,775
my eloquent model or how do I unit is
this controller how do I test drive that

11
00:01:18,549 --> 00:01:27,540
thing and it kind of hurts me that
there's a set of questions that we do

12
00:01:27,054 --> 00:01:31,125
have answers for between them speak
about them and whatever kind of Corey

13
00:01:32,025 --> 00:01:35,574
fight there are a lot of things that we
like using in terms of explaining things

14
00:01:35,799 --> 00:01:41,350
there's a context think about the things
in in terms of the context but never

15
00:01:41,035 --> 00:01:48,764
actually explain what this context is
about so I basically decided to talk

16
00:01:49,079 --> 00:01:58,540
about that and if you wish Taylor talk
for this conference I'm results

17
00:01:58,054 --> 00:02:02,057
or constantine career shops have
problems pronounce my last name all of

18
00:02:02,084 --> 00:02:13,088
them at a normal thing for russian noise
I'm and basically did 23 mistakes in my

19
00:02:13,088 --> 00:02:20,166
life big ones he had been respect caring
about the deity too much I work as a

20
00:02:21,066 --> 00:02:26,072
team to practice manager in the company
called in London and the things that I

21
00:02:27,026 --> 00:02:33,034
love are needed to related stuff
learning how to communicate inside and

22
00:02:33,034 --> 00:02:37,036
outside of the codebase how people
communicate between each other before

23
00:02:37,036 --> 00:02:39,885
the develops after how people
communicate with each other through the

24
00:02:40,209 --> 00:02:47,650
software and that's basically where my
wall tangent on DVD comes from working

25
00:02:47,065 --> 00:02:52,164
company called where people call me as a
bid to practice manager to do crazy

26
00:02:53,064 --> 00:02:58,118
things like that I'm preaching I like to
say that everything that I'm showing and

27
00:02:59,018 --> 00:03:04,023
everything I'm talking about his actual
things that we do it's not some

28
00:03:04,023 --> 00:03:07,122
arbitrary original stuff that I think of
in the corner of the company and then

29
00:03:08,022 --> 00:03:13,093
nobody in the company if you speak with
people in and we actually leave by the

30
00:03:13,093 --> 00:03:17,139
principles Woodbridge right

31
00:03:18,039 --> 00:03:24,080
so that's a bit about me just to give
you a context but what is this talk

32
00:03:24,008 --> 00:03:33,417
about right because software costs mean
muxing what is it about estimation heard

33
00:03:34,209 --> 00:03:40,160
some some some people are talking about
estimation no its not so what is it that

34
00:03:40,016 --> 00:03:41,027
right

35
00:03:41,027 --> 00:03:47,101
very tricky talk arm that larval facades
and flick among suspecting developers

36
00:03:48,001 --> 00:03:55,014
bad idea that is active records and
eloquent other frameworks superiority

37
00:03:55,014 --> 00:04:03,110
over larval killing kittens and other
subjective from destructive crap like

38
00:04:04,001 --> 00:04:16,020
that is not in stock right sorry does
somebody know

39
00:04:16,029 --> 00:04:22,114
well actually the stock is about this
talk is about introducing in trying to

40
00:04:23,014 --> 00:04:27,036
make sense of development costs that's a
topic that nobody talks about it and I

41
00:04:27,036 --> 00:04:30,071
think it's a shame because if you think
if you look at the software from

42
00:04:30,071 --> 00:04:36,124
communication and economical perspective
of things that you kind of didn't

43
00:04:37,024 --> 00:04:43,082
thought about before start making sense
they start clicking the stock is also

44
00:04:43,082 --> 00:04:50,157
highlighting the context of tools and
practices we use Active Record I would

45
00:04:51,057 --> 00:04:55,060
not I would definitely not criticize the
tools themselves but I will critically

46
00:04:55,006 --> 00:04:59,031
consider them and show the context in
which the stores are very small and the

47
00:04:59,085 --> 00:05:04,183
context in which the stores might not be
a good fit for this stock is involved

48
00:05:05,083 --> 00:05:10,087
seen the context for the tools and
trying to find and understand why for

49
00:05:10,087 --> 00:05:16,095
some people this particular tool works
amazingly does magic and wonder why for

50
00:05:16,095 --> 00:05:20,103
some other projects and other teams do
stories like is the most horrendous

51
00:05:21,003 --> 00:05:27,024
horrible thing they ever did the most
important beads in this introductory

52
00:05:27,024 --> 00:05:32,120
slide is this talk is about is based on
years of observations and experience the

53
00:05:33,002 --> 00:05:37,095
stock is based on me going through the
three different companies client

54
00:05:38,013 --> 00:05:42,034
companies of seven projects small
projects big projects complex projects

55
00:05:42,034 --> 00:05:48,113
obvious projects small team medium team
big teams with a different context and

56
00:05:49,013 --> 00:05:53,017
finding some patrons and some
repercussions what the stock is not it's

57
00:05:53,053 --> 00:05:59,061
not built based on the actual data
collection analysis I will show some

58
00:05:59,061 --> 00:06:03,093
graph that represents my observations
and experience those graphs do not

59
00:06:03,093 --> 00:06:08,093
represent a collection of data throw the
world and analysis of the date right

60
00:06:08,093 --> 00:06:15,116
it's purely talk from my contacts and I
hope you'll see some repercussions worse

61
00:06:16,016 --> 00:06:19,049
things that you do a new senior
day-to-day lives and if you do then

62
00:06:19,049 --> 00:06:21,072
means we were coming from the same place

63
00:06:21,072 --> 00:06:25,085
if you don't maybe you're coming from
the different context the one that I

64
00:06:25,085 --> 00:06:33,107
never seen before yet so just to
simplify another name for the stalk and

65
00:06:34,007 --> 00:06:44,066
the easier one is a context the talk
right software costs so when you when

66
00:06:44,066 --> 00:06:49,081
you talk about costs usually think about
things like money right that's a

67
00:06:49,081 --> 00:06:56,127
background there in the software costs
means slightly different thing used in

68
00:06:57,027 --> 00:07:01,056
the software will not talking about
costs were talking about time how long

69
00:07:01,056 --> 00:07:06,063
does it take to do this or how long does
it take to do this thing how long does

70
00:07:06,063 --> 00:07:13,118
it take to deliver project that's a cost
of the project and if you're struct

71
00:07:14,018 --> 00:07:19,021
everything we do with our projects with
our products you can basically

72
00:07:19,048 --> 00:07:25,062
aggregated into three different chunks
of work there is a time it takes you to

73
00:07:25,062 --> 00:07:29,136
ride and toilet test code in the first
place and even if you're not using

74
00:07:30,036 --> 00:07:35,129
automation by the way you're still
testing I don't like to yourself you are

75
00:07:36,029 --> 00:07:42,081
writing code and you open in a browser
hey you're testing you are testing no

76
00:07:42,081 --> 00:07:46,162
matter which tools you use your always
destined the work after you did it

77
00:07:47,062 --> 00:07:52,107
that's a normal state of mind but in
addition to just like introducing things

78
00:07:53,007 --> 00:07:54,016
are writing things in the first place

79
00:07:54,097 --> 00:08:01,130
you do start after after a while feeling
the urge to change them right and this

80
00:08:02,003 --> 00:08:06,074
this phase where like your writing a new
stuff faces out and let's start phasing

81
00:08:07,001 --> 00:08:14,007
in is this needs to change the existing
things right time to change code and

82
00:08:14,061 --> 00:08:18,137
tests and again the same was test if you
are not doing automation it doesn't mean

83
00:08:19,037 --> 00:08:24,041
you're not doing your adopted us right
if you introduce an authentication

84
00:08:24,041 --> 00:08:28,087
mechanism in front of your website and
you're not doing any automation you

85
00:08:28,087 --> 00:08:31,066
still will do tests and it means you
need to update your manual test

86
00:08:31,849 --> 00:08:34,870
routine to go through this
authentication mechanism every time you

87
00:08:34,087 --> 00:08:39,186
need to test new feature and then there
is the third set of things are times

88
00:08:40,086 --> 00:08:45,425
that was spend a lot of people talk
about actually a lot of people talk

89
00:08:46,199 --> 00:08:51,207
about not a lot of people doing them
time to reflect her cold and tests and

90
00:08:51,279 --> 00:08:55,160
we're talking a lot about like oh you
need to reflect her at the centerpiece

91
00:08:55,016 --> 00:08:59,425
of TDD or chrome is like one when it
goes into actual delivery is very very

92
00:08:59,569 --> 00:09:03,590
hard to do these things and it's very
very hard to reason why we need in the

93
00:09:03,779 --> 00:09:08,790
first place so for the purpose of this
talk I will give this three different

94
00:09:08,079 --> 00:09:13,106
distinct sections names and we're going
to explore in time to time to write and

95
00:09:14,006 --> 00:09:17,011
test called I will call it from now on

96
00:09:17,011 --> 00:09:23,013
cost of introduction time to change cold
and tests is a cost of change and that's

97
00:09:23,013 --> 00:09:27,892
a cost that you most likely hear a lot
about when you're in a jail conferences

98
00:09:28,009 --> 00:09:33,220
or you're trying to talk with somebody
about the Legacy Project time to reflect

99
00:09:33,022 --> 00:09:41,631
our content as I will call it cost of
ownership rights a cost of only coat so

100
00:09:41,829 --> 00:09:45,430
we'll go through each one of them and I
will try to give you enough context

101
00:09:45,043 --> 00:09:49,074
information so that you can start making
sense of those forces and then we will

102
00:09:49,074 --> 00:09:54,166
go into seen how we can make this three
forces play with each other and how we

103
00:09:55,066 --> 00:09:59,129
can make sense of them so that we
develop software taken them into account

104
00:10:00,029 --> 00:10:06,188
which is the normal state of mind for
engineers the cost of introduction

105
00:10:06,449 --> 00:10:11,544
essentially is a time it takes to
introduce new natural independent the

106
00:10:12,399 --> 00:10:18,443
coupled application logic to the good
example of this might be if you're

107
00:10:18,839 --> 00:10:22,842
building a lot of application created
new controller of arbitrary complexity

108
00:10:23,139 --> 00:10:28,220
or creating new elegant model or
creating new form right so time it takes

109
00:10:28,949 --> 00:10:35,120
to create completely new controller or
completely new model that time is a cost

110
00:10:35,012 --> 00:10:38,088
of introducing the concept into the card
base

111
00:10:38,088 --> 00:10:43,142
so that your business model has is it
has very clear direct correlation to the

112
00:10:44,042 --> 00:10:49,044
business value the more controllers you
create more value there is to the

113
00:10:49,044 --> 00:10:54,070
project more models you create a more
vital areas to the project this cost

114
00:10:54,007 --> 00:10:58,048
also has direct relation to the lines of
code number of lines of code and it

115
00:10:59,011 --> 00:11:03,013
basically means the more lines of code
require more time to write them so the

116
00:11:03,013 --> 00:11:09,112
less lines of code means cheaper cost of
introduction those two first fighters

117
00:11:10,012 --> 00:11:15,047
basically mean that this cost is
relatedly easy to optimize for usually

118
00:11:15,047 --> 00:11:19,144
by generalization or convicted interdict
introduction of convenience there so it

119
00:11:20,044 --> 00:11:25,108
doesn't mean it means if your cost is
directly related with amount of lines of

120
00:11:26,008 --> 00:11:31,105
code and less lines of code you have the
cheaper it is to introduce thing it's

121
00:11:32,005 --> 00:11:35,071
very easy to optimizing just like you
makes some sense something introduced a

122
00:11:35,071 --> 00:11:41,075
ball with less lines in addition to
attribute this cost has very interesting

123
00:11:41,075 --> 00:11:48,154
dynamic associated to the first and the
key dynamics areas visible it is visible

124
00:11:49,054 --> 00:11:54,067
from the outset you start project and
the first and only cost you probably

125
00:11:54,067 --> 00:11:57,146
sees how long it takes you to create new
controller how long it takes you to

126
00:11:58,046 --> 00:12:03,051
create a new route and how long it takes
you to creating you for this very

127
00:12:03,051 --> 00:12:07,103
visible is very clear you can show it to
the business and after quite some time

128
00:12:08,003 --> 00:12:13,099
after working on a floral projects even
business can understand how much it

129
00:12:13,099 --> 00:12:21,114
costs to create a controller and Laurel
or any other framework these costs

130
00:12:22,014 --> 00:12:28,082
losses relevancy over the project
lifetime so the longer project goes the

131
00:12:28,082 --> 00:12:33,180
less important this coast is a good
example of this is you do not create

132
00:12:34,008 --> 00:12:38,060
controllers new controllers and forms
throughout the project life time there

133
00:12:39,032 --> 00:12:44,057
is a phase where you start creating less
and less controllers less and less forms

134
00:12:44,057 --> 00:12:47,063
and when you start doing is you start
changing the existing ones rather than

135
00:12:48,017 --> 00:12:52,064
trying to create a new new new
controller you can't logically more but

136
00:12:52,064 --> 00:12:58,083
I didn't you stuff on top and as I said
previously discussed a stable across

137
00:12:58,083 --> 00:13:01,177
multiple projects to the time it takes
to create a controller in this project

138
00:13:02,077 --> 00:13:06,130
in larval is exactly the same amount of
time it takes to create exactly same

139
00:13:07,003 --> 00:13:11,078
controller or exactly same complexity
level controller and another project

140
00:13:12,005 --> 00:13:20,076
business loves it right that's amazing
actually think I feel that most people

141
00:13:20,076 --> 00:13:25,091
focus on the first one and a third one
it is visible from the outside makes it

142
00:13:25,091 --> 00:13:29,104
very easy to sell it is stable across
projects means you optimize it for one

143
00:13:30,004 --> 00:13:36,102
project left to live for another project
also I reckoned

144
00:13:37,002 --> 00:13:42,008
attribute or dynamic part that you need
to focus most about is not this to this

145
00:13:42,008 --> 00:13:49,012
will cost of introduction losses
relevancy over the project last time the

146
00:13:49,012 --> 00:13:54,056
longer project goes the less you
benefits unless you care about this one

147
00:13:54,056 --> 00:14:04,097
so goes the construction is relatable
EEZ to optimize for right so there are

148
00:14:04,097 --> 00:14:08,133
usually two ways to optimize for
customer interaction with number one is

149
00:14:09,033 --> 00:14:13,117
convenience later so you introduced
convenience player in your framework on

150
00:14:14,017 --> 00:14:18,099
your tall set and that's how you
optimize for a good example of

151
00:14:18,099 --> 00:14:26,166
convenience is service locator as you
probably know in local sushi go with

152
00:14:27,066 --> 00:14:33,089
dependency injection and you will end up
with like was the code base that is like

153
00:14:33,089 --> 00:14:35,978
this right you need to inject the cash
service and then you need to use the

154
00:14:36,779 --> 00:14:44,430
service later on in comparison to just
use it just like one line that's a very

155
00:14:44,043 --> 00:14:47,137
good example convenience lear keep in
mind I'm not saying I'm not saying this

156
00:14:48,037 --> 00:14:51,126
is bad I'm not saying this is good I'm
just giving a context it's just two

157
00:14:52,026 --> 00:14:59,031
different ways to do the same job and
this one takes more lines another

158
00:14:59,076 --> 00:15:02,965
example from outside liable that's a
symphony so you have a base controller

159
00:15:03,649 --> 00:15:08,230
which also makes it easier to do trivial
tasks in one single line of code because

160
00:15:08,023 --> 00:15:10,111
you extend from the base controller and
just called this matter on the base one

161
00:15:11,011 --> 00:15:16,860
right another way to optimize for cost
of introduction is generalization so you

162
00:15:16,959 --> 00:15:20,972
take some problem which is very specific
for the particular the main or that

163
00:15:21,089 --> 00:15:25,176
particular business and you trivialize
it you generalize it to the point where

164
00:15:25,959 --> 00:15:31,984
it is reusable across domains a good
example of this is ActiveRecord rights

165
00:15:32,209 --> 00:15:37,920
activities you have a business object in
your code base and then you need to

166
00:15:37,092 --> 00:15:40,141
store information you need to store this
object somewhere in the persistence

167
00:15:41,041 --> 00:15:44,072
layer so usually like the manual way of
doing this is you introduce some

168
00:15:44,072 --> 00:15:48,080
persistence responsible object that will
store this thing manually maybe we'll

169
00:15:49,052 --> 00:15:56,061
connect sturdy ball to the database now
people recognize that there is a set of

170
00:15:56,061 --> 00:16:00,077
options that you doing most of the
projects so why do you need to create

171
00:16:00,077 --> 00:16:03,886
this again and again and again why can't
we just like colds this hole in the base

172
00:16:04,579 --> 00:16:10,350
class that will that will provide you
was like most of the 70 90% of the

173
00:16:10,035 --> 00:16:14,107
methods that you will need to use and
that's what ActiveRecord it takes less

174
00:16:15,007 --> 00:16:17,086
lines of codes takes less work

175
00:16:17,149 --> 00:16:25,920
convenience even despite was another
great example from outside of horrible

176
00:16:25,092 --> 00:16:26,158
and I think it's Tyler

177
00:16:27,058 --> 00:16:32,085
so if you don't know there is a pot
during a flight from Ganga for which

178
00:16:32,085 --> 00:16:36,143
subscriber an issue the way in
subscriber access you have an interface

179
00:16:37,043 --> 00:16:40,126
and you have some somebody that is
dependent on this interface or whenever

180
00:16:41,026 --> 00:16:43,035
something happens you just call this
method in interface that's how events

181
00:16:44,016 --> 00:16:47,085
tribal words somebody looked into the
subscriber and said look we need to

182
00:16:47,085 --> 00:16:50,103
create interfaces all the time and
implement them all the time that kind of

183
00:16:51,003 --> 00:16:54,077
a lot of muscle generalize it and create
events Pacha

184
00:16:54,077 --> 00:16:57,091
so that's another example of
generalizing on events described

185
00:16:57,091 --> 00:17:02,092
resulted in events Pacha even more
interesting one as dependency injection

186
00:17:02,092 --> 00:17:06,147
container that's a generalization on
dependency injection so you have a

187
00:17:07,047 --> 00:17:11,062
normal object-oriented code were you
need to create a service and you have a

188
00:17:11,062 --> 00:17:15,137
couple of defenses dependency possible
to manually and then somebody looked at

189
00:17:16,037 --> 00:17:23,053
it look it's too many lines of code is
too complex to buy the things around us

190
00:17:23,053 --> 00:17:26,069
generalize it was created dependency
injection container that will do all

191
00:17:26,069 --> 00:17:35,072
these things for us so the thing that
I'm trying to say here is very easy to

192
00:17:35,099 --> 00:17:41,110
optimize for the cost of introduction
and no matter what you think

193
00:17:42,001 --> 00:17:47,086
optimizing for custom introduction is
not inherently about that's what level

194
00:17:47,095 --> 00:17:50,178
does a lot and that's actually a good
thing but it's a good thing in the

195
00:17:51,078 --> 00:17:59,133
context and the context is this is a
cut-off point for of the product so kind

196
00:18:00,033 --> 00:18:04,119
of point is when you need to stop
working on the project or product when

197
00:18:05,019 --> 00:18:09,033
the product is ready and you will
probably never touching it again you

198
00:18:09,033 --> 00:18:10,088
will not updated

199
00:18:10,088 --> 00:18:14,117
not just any complex changes to it
you'll probably action there and

200
00:18:15,017 --> 00:18:19,110
simplify things from this point on no
new features will be added a good

201
00:18:20,001 --> 00:18:26,040
example of cut off base project is based
that's how first become the second base

202
00:18:26,049 --> 00:18:31,143
camp created they had a very low or
limited kind of point they created based

203
00:18:32,043 --> 00:18:35,109
company said no more features ever just
simplify what we have

204
00:18:36,009 --> 00:18:44,010
here if you have a project was really
short cut-off point it means the cost of

205
00:18:44,001 --> 00:18:47,053
introduction is probably the cost you
need to worry about and that probably is

206
00:18:47,062 --> 00:18:51,991
the only cause you to worry about
because when you kind of when this like

207
00:18:52,549 --> 00:18:57,590
introduction of new controller phases
out a project and it benefited from

208
00:18:57,059 --> 00:19:01,076
introducing control is really really
quickly introduce informs really really

209
00:19:01,076 --> 00:19:06,145
quick and then it kind of becomes
trickier but no problem project and

210
00:19:07,045 --> 00:19:12,059
other projects and that's something that
I feel many of you probably experienced

211
00:19:12,059 --> 00:19:17,388
and that's something that I experienced
a lot on the short projects right if the

212
00:19:17,919 --> 00:19:21,700
product life is short enough to not
encounter loss of cost of introduction

213
00:19:21,007 --> 00:19:29,048
relevancy than the cost of introduction
it only cost optimized but keep in mind

214
00:19:30,011 --> 00:19:36,110
Kenyans based projects either die a hero
or live long enough to see themselves

215
00:19:37,001 --> 00:19:50,048
and that's doable and I cost cost of
change and that's the reason why

216
00:19:50,057 --> 00:19:58,266
comedians based projects if they live
long enough to change the time it takes

217
00:19:58,779 --> 00:20:03,817
to adapt existing application logic your
existing controllers your existing forms

218
00:20:04,159 --> 00:20:12,130
to new business realities on the ongoing
basis and also has set up attributes so

219
00:20:12,013 --> 00:20:15,074
it also has direct correlation to
business value the more changes you do

220
00:20:15,074 --> 00:20:19,096
the more value to the business and has
no direct relation two lines of code

221
00:20:19,096 --> 00:20:21,665
whatsoever

222
00:20:22,529 --> 00:20:27,210
introducing hundred lines according to
my dad doesn't necessary removing

223
00:20:27,021 --> 00:20:30,035
hundred lines of court might add value
not necessary

224
00:20:30,035 --> 00:20:34,038
replacing one hundred lines of code was
dumb mines of God my action a huge

225
00:20:34,038 --> 00:20:40,082
amount of value there is no correlation
cost of change also very much affected

226
00:20:40,082 --> 00:20:43,176
by generalization

227
00:20:44,076 --> 00:20:48,097
the worst thing about the cost of
changes that in and it isn't visible

228
00:20:48,097 --> 00:20:51,190
from the outside its complete opposite
of course introduction when you just

229
00:20:52,009 --> 00:20:56,076
start a project you have no idea what
cost of changes the problem was cost of

230
00:20:57,057 --> 00:21:02,080
changes that it gains relevancy during
the project last time and it starts

231
00:21:02,008 --> 00:21:07,089
exponentially increasing so there is the
point where you stop introducing your

232
00:21:08,061 --> 00:21:12,090
new controllers and then you need to
update the existing ones and forms and

233
00:21:12,009 --> 00:21:12,108
entities

234
00:21:13,089 --> 00:21:17,160
skating really rehired know is getting
really really hard no I can't do it

235
00:21:18,006 --> 00:21:26,087
anymore right that's a cost of change
change increases exponentially over time

236
00:21:27,041 --> 00:21:31,079
I want to give you a very short example
let's say would have this like very

237
00:21:31,079 --> 00:21:39,116
simplistic search mechanism it searches
packages on arbitrary package management

238
00:21:40,016 --> 00:21:46,093
system let's say you don't know what's
written down that's kind of part of the

239
00:21:46,093 --> 00:21:49,125
point so you need to learn what what is
there before you can do anything so

240
00:21:50,025 --> 00:21:57,111
let's say we want to tax support into
this thing to raise your hands if you

241
00:21:58,011 --> 00:22:08,102
think that it will take up to a day
lower less than a day I know you will go

242
00:22:09,002 --> 00:22:13,011
back and that's exactly what you will
say to your project manager desde ok two

243
00:22:13,092 --> 00:22:14,178
days two days

244
00:22:15,078 --> 00:22:24,165
today's going three days alright so two
days somewhere ok you know let's not do

245
00:22:25,065 --> 00:22:30,068
this situation let's wait right let's do
something else there's some something

246
00:22:30,068 --> 00:22:34,069
else more important than tax so you're
working for a couple of situations maybe

247
00:22:34,069 --> 00:22:40,107
a monster to you know what let's not
talk now

248
00:22:41,007 --> 00:22:46,446
it's the same search right same time
same amount of functionality nothing

249
00:22:46,509 --> 00:22:50,535
changed about requirements raise your
hand if you still think it's two days

250
00:22:50,769 --> 00:22:54,580
raise your hand if you think it's four
days

251
00:22:54,058 --> 00:23:04,066
ok five days right you know what let's
not do it now let's do it later

252
00:23:05,038 --> 00:23:12,053
but here's the thing

253
00:23:12,053 --> 00:23:14,151
exponential increase of course the
changes not inherently a program of

254
00:23:15,051 --> 00:23:19,139
every project and that's a challenge was
discussed at the challenge was talking

255
00:23:20,039 --> 00:23:22,093
about the cost of changing saying like
you need to worry about the cost of

256
00:23:22,093 --> 00:23:25,119
change the reality is not every project
needs to worry about the cost of change

257
00:23:26,019 --> 00:23:32,072
quite well because you have a cut-off
planes again if your project

258
00:23:32,072 --> 00:23:39,088
stuff's just before you need to change
things start changing thanks to care

259
00:23:39,088 --> 00:23:45,179
about the cost of change if the product
life is long enough to encounter

260
00:23:46,079 --> 00:23:50,148
exponential growth of changes in the
cost of change is the only cost

261
00:23:51,048 --> 00:23:58,050
optimizing for cost of introduction has
a very short time line right it's two

262
00:23:58,068 --> 00:24:02,121
months three months project action is
really cool if its one-year project

263
00:24:03,021 --> 00:24:07,095
those three months mean nothing at the
beginning at the end of those year this

264
00:24:07,095 --> 00:24:13,161
year the only thing you care about is
how long it takes to change things and

265
00:24:14,061 --> 00:24:18,130
the key here is we're getting more and
more complex projects holidays

266
00:24:19,003 --> 00:24:23,021
making a difference in the world
requires more and more complexity of the

267
00:24:23,048 --> 00:24:28,134
software right so the projects naturally
growing timeline if you want to change

268
00:24:29,034 --> 00:24:33,038
the world with your product the chances
are you have a complex project in your

269
00:24:33,038 --> 00:24:39,054
head on your hands and it will take long
and the changes the primary thing that

270
00:24:39,054 --> 00:24:46,063
you need to worry about a thing about
worrying about changes as I said the

271
00:24:47,044 --> 00:24:51,070
changes affected by generalization
optimizing for cost reduction in most

272
00:24:51,007 --> 00:24:56,079
cases has a negative effect on the coast
of change right you make things to be

273
00:24:57,042 --> 00:25:01,044
describable an implementable was less
lines of code that comes at a cost of

274
00:25:01,044 --> 00:25:05,130
flexibility something takes just one
line to introduce some complex concept X

275
00:25:06,003 --> 00:25:11,059
192 and reuse whenever you need to
change the complex context that's not a

276
00:25:11,086 --> 00:25:18,089
street anymore and that's why some
engineers actually alot of Engineers

277
00:25:18,089 --> 00:25:20,151
tried to increase the cost of
introduction in attempt to affect the

278
00:25:21,051 --> 00:25:22,114
cost of change

279
00:25:23,014 --> 00:25:27,061
ok so if decreasing costs of
introduction have a negative effect on

280
00:25:27,061 --> 00:25:32,116
the cost of change why don't we all over
engineer everything upfront with the

281
00:25:33,016 --> 00:25:37,019
perfect design in France use hexagonal
architecture using common bus everywhere

282
00:25:37,019 --> 00:25:38,064
all these things right

283
00:25:38,064 --> 00:25:43,141
change will be easy here's what happens
to you have your graph you have a

284
00:25:44,041 --> 00:25:46,140
support you you have an introduction of
payment

285
00:25:47,004 --> 00:25:53,005
let's say you using cashier
newly-released spark and there is like

286
00:25:53,041 --> 00:25:57,047
general costs and then later and you
will need to introduce additional

287
00:25:58,001 --> 00:26:02,092
payment method and that cost increases
the cost of change so what happens if

288
00:26:02,092 --> 00:26:06,140
you increase the cost of supporting
payments in the first place so let's say

289
00:26:07,004 --> 00:26:11,049
we don't want to share you want to read
implement the payment gateway ourselves

290
00:26:11,085 --> 00:26:14,179
to make sure that it's like it supports
change from our perspective in the

291
00:26:15,079 --> 00:26:21,141
future so you move it up right to say
okay that will be a bit more expensive

292
00:26:22,041 --> 00:26:28,090
but now the cost of the cost of change
changes drastically doesn't really does

293
00:26:28,009 --> 00:26:38,009
change this way you delayed delayed
increase of course the change so what

294
00:26:38,009 --> 00:26:42,024
happens is like you're saying OK let's
let's over design that payment gateway

295
00:26:43,005 --> 00:26:47,098
and then I didn't you payments will be
easier and true I don't first payments

296
00:26:47,098 --> 00:26:50,197
easy here but in the second payment
still complex and the third one is more

297
00:26:51,097 --> 00:26:56,123
complex are you still have this
exponential increase and that's what up

298
00:26:57,023 --> 00:27:02,054
from design is a key a wonderful that's
an illusion that one can control costs

299
00:27:02,054 --> 00:27:02,135
of change

300
00:27:03,035 --> 00:27:07,130
pipeline enough analysis approached it
doesn't work

301
00:27:08,003 --> 00:27:13,096
and the reason why doesn't work is
because the nature and the type of

302
00:27:14,023 --> 00:27:20,037
change is unpredictable and the closer
you are to the beginning of the project

303
00:27:20,037 --> 00:27:24,085
more unpredictable this where you can
predict some changes like to three

304
00:27:24,085 --> 00:27:28,156
months ahead you can't predict what
needs to happen with this product

305
00:27:29,056 --> 00:27:34,143
year from now so you can't predict if
you can't predict the changes you can't

306
00:27:35,043 --> 00:27:44,069
adapt to them and this is where cost of
ownership comes the cost of ownership is

307
00:27:44,069 --> 00:27:47,167
a time it takes to maintain dawn
application logic to support its ongoing

308
00:27:48,067 --> 00:27:56,135
change such a bit of time of soaring
cost of ownership it's an intermediate

309
00:27:57,035 --> 00:28:01,070
between cost of introduction cost of
change it has no direct correlation to

310
00:28:01,007 --> 00:28:07,065
business value and that's a big problem
of this cost maintaining things they

311
00:28:08,028 --> 00:28:10,115
don't add any value to the business

312
00:28:11,015 --> 00:28:15,041
refactoring doesn't add any value to the
business test anybody to the business

313
00:28:15,041 --> 00:28:20,048
advantage you make you to be able to
change things but there's no value to

314
00:28:20,048 --> 00:28:25,135
the business and that's why it's so hard
to sell them it does have there a

315
00:28:26,035 --> 00:28:30,057
correlation to and of course the more
lines of code you test your reflector

316
00:28:30,057 --> 00:28:33,150
the more expensive for this right so the
more lines of code the more expensive it

317
00:28:34,005 --> 00:28:40,032
is kind of the same as was cost of
interaction it is always advisable for

318
00:28:40,077 --> 00:28:45,141
business at least it is always relevant
throughout the project life time and the

319
00:28:46,041 --> 00:28:50,078
interesting thing about this it's almost
exactly the same as cost of introduction

320
00:28:50,078 --> 00:28:51,139
stable over time

321
00:28:52,039 --> 00:28:56,138
never increases its a stable costs but
it adds up so the more things you own

322
00:28:57,038 --> 00:29:00,129
the more expensive for example if you
just do am going to factor into five

323
00:29:01,029 --> 00:29:04,068
classes that will be that cost if you do
I'm going to factor in 10 classes that

324
00:29:04,068 --> 00:29:10,156
will be double and the easiest way to
think about the cost of ownership costs

325
00:29:11,056 --> 00:29:16,149
of interaction across the changes like
one of payment I want this controller I

326
00:29:17,049 --> 00:29:18,061
want this change

327
00:29:18,061 --> 00:29:26,500
cost of ownership is just going for you
pay almost like on a monthly basis cost

328
00:29:27,049 --> 00:29:31,690
of ownership is the cost you pay for the
right to change particular part module

329
00:29:31,069 --> 00:29:38,083
class method whatever application
continuously and sustainably and it's

330
00:29:38,083 --> 00:29:43,092
essentially testing even manual testing
that's a cost of ownership right your

331
00:29:44,073 --> 00:29:47,121
testers run through entire system all
the time it doesn't add any value to the

332
00:29:48,021 --> 00:29:53,104
project except you're predicting the
bugs it's a cost you pay just because

333
00:29:54,004 --> 00:29:58,101
you're on the project is because you do
this even more clear ownership unit test

334
00:29:59,001 --> 00:30:05,034
unit tests in particular class is a
direct statement only in that class

335
00:30:05,034 --> 00:30:10,122
method and the closest one to actual
clearest terms of ownership is

336
00:30:11,022 --> 00:30:15,100
reflecting when you change across
directly with the reason to support the

337
00:30:16,000 --> 00:30:23,073
future change and reducing cost of
ownership allows you to balance two

338
00:30:23,073 --> 00:30:29,087
other costs that's a core concept of it
that's the reason for the existence in

339
00:30:29,087 --> 00:30:34,116
the first place here so it works we
still have our support payments and

340
00:30:35,016 --> 00:30:40,057
payments payment due payment method as
increasing cost of change so we'll see

341
00:30:40,057 --> 00:30:45,060
what happens if we introduce ongoing
payments reflector interesting start

342
00:30:45,087 --> 00:30:49,129
being discussed this fee basically every
time we're touching anything near the

343
00:30:50,029 --> 00:30:54,100
payments will look at the payments and
it's just more and what can we refer to

344
00:30:55,000 --> 00:30:59,061
make it easier in the future I'm going
payments so what happens if you do when

345
00:30:59,061 --> 00:31:00,500
you do this

346
00:31:01,049 --> 00:31:05,900
cost of change changes drastically even
though he didn't do anything was cost of

347
00:31:05,009 --> 00:31:11,063
introduction of change instead of being
exponential stars been something like

348
00:31:12,044 --> 00:31:15,773
this becomes corrupt so now

349
00:31:16,169 --> 00:31:18,251
because you're constantly reflected
because you constantly look at your

350
00:31:18,989 --> 00:31:24,004
payments and you think ok how can I
support the future changes you make to

351
00:31:24,139 --> 00:31:27,600
all the future changes that much easier
to the thing that i dont showed

352
00:31:27,006 --> 00:31:31,075
yesterday were before introducing the
day before actually I did this value of

353
00:31:31,669 --> 00:31:37,714
the change you look at the consulate
really hard to change now what if I just

354
00:31:38,119 --> 00:31:43,840
things in bits and then I introduced the
change to the squad from cans back so

355
00:31:43,084 --> 00:31:49,543
that's been cost of ownership item back
then decided okay well paid now to make

356
00:31:50,299 --> 00:31:58,372
the changes and that's how it works like
based on my observations and outcome of

357
00:31:59,029 --> 00:32:01,100
this process is something that people
call emerging design which is an

358
00:32:01,739 --> 00:32:05,700
opposite thing to the print design
design emerges because you're constantly

359
00:32:05,007 --> 00:32:09,436
look at the sting in the US yourself ok
wish changes would I need to do next

360
00:32:10,129 --> 00:32:15,820
isn't simple enough couple write your
own in the sting outcome of only easier

361
00:32:15,082 --> 00:32:19,110
simplified everything and you create
design as an outcome of this that's

362
00:32:20,001 --> 00:32:25,060
where emerging design comes from its
usual result side-effect of ongoing

363
00:32:25,069 --> 00:32:32,878
ownership so let's do it everywhere

364
00:32:33,499 --> 00:32:39,586
the entire freakin thing if it's so
amazing if only thing is so incredibly

365
00:32:40,369 --> 00:32:45,440
great let's let's play through us from
the very beginning of the project was

366
00:32:45,044 --> 00:32:49,483
like every time before we even
introducing every morning going to the

367
00:32:49,879 --> 00:32:56,470
what and we read through the entire code
base and see how it can optimize it like

368
00:32:56,047 --> 00:33:02,366
that probably will just lower the cost
of change legislation but the problem

369
00:33:02,789 --> 00:33:08,877
with this is that still just need to add
payments need to change the entire

370
00:33:09,669 --> 00:33:15,701
system this caused all this remark
during all this cleaning up around

371
00:33:15,989 --> 00:33:17,070
demons

372
00:33:17,007 --> 00:33:21,049
doesn't cause all you do is change
payments that's what business wants

373
00:33:21,049 --> 00:33:29,060
doesn't want you to change how your
notifications work right so you doing

374
00:33:29,006 --> 00:33:35,095
this things and then question he doubts
the thing that does matter just payments

375
00:33:36,049 --> 00:33:51,328
but you pay so much more without any
clear benefit for anyone that's reality

376
00:33:51,769 --> 00:33:56,780
and that's what you do when you test
your system only wasn't too and

377
00:33:56,879 --> 00:34:03,944
end-to-end test is only in your entire
freakin thing anyone here has 45 minutes

378
00:34:04,529 --> 00:34:14,790
running tests that's that's what you're
doing your own the entire thing and all

379
00:34:14,079 --> 00:34:19,125
this maintenance 16 rerun interest
because of flickering spending this for

380
00:34:20,025 --> 00:34:26,098
25 Minister on them that you only lets
you pay this fee ok

381
00:34:26,098 --> 00:34:33,131
but the worse thing even more so if you
can imagine is this right when you're

382
00:34:34,031 --> 00:34:37,058
doing gonna factoring and testing of
everything except the payments logic and

383
00:34:37,058 --> 00:34:40,907
then you need to our dependence so you
have like you pay this incredible cost

384
00:34:41,429 --> 00:34:46,452
every morning and then you also have
exponential cost of change what's

385
00:34:46,659 --> 00:34:54,270
happening I a morning thing well that's
what you do when you do exclusive

386
00:34:54,027 --> 00:34:58,065
end-to-end testing exclusive and to
understand is only when you don't have

387
00:34:58,065 --> 00:35:02,146
unit testing your system is not is not
only the right thing your cost of change

388
00:35:03,046 --> 00:35:08,140
will still increase and your subpoena
ownership ownership would not help if

389
00:35:09,004 --> 00:35:22,021
you own the wrong thing everything
that's the way to use that cost of

390
00:35:22,057 --> 00:35:25,104
ownership is on things that are worth
owning

391
00:35:26,004 --> 00:35:31,085
which brings really interesting moral
dilemma you don't know what's was only

392
00:35:31,085 --> 00:35:39,139
the beginning of the project just a
simple recover so you have this will

393
00:35:40,039 --> 00:35:42,045
cost you have cost of introduction

394
00:35:42,045 --> 00:35:47,076
doesn't increase its really wanted the
beginning of the project is beginning of

395
00:35:47,076 --> 00:35:52,121
the project and it's very easy to
optimize stable across multiple projects

396
00:35:53,021 --> 00:35:58,069
there's a change which is exponential
relevant everywhere except the beginning

397
00:35:58,069 --> 00:36:02,155
of the project is very hard to optimize
for you can't generalize for cost of

398
00:36:03,055 --> 00:36:09,114
change because of changes sold tied to
your business to change cost of

399
00:36:10,014 --> 00:36:17,043
ownership is also a leaner it's relevant
throughout and also or only the wrong

400
00:36:17,043 --> 00:36:23,071
thing is bad and also in too much as bad
and you can just doesn't make sense like

401
00:36:23,071 --> 00:36:26,970
this will cost now a dog that no I can
make sense of them know we do understand

402
00:36:27,609 --> 00:36:31,900
what this happens it's like if I and two
interests everything apparently paid the

403
00:36:31,009 --> 00:36:35,013
cost of ownership for everything and it
doesn't help the cost of change but then

404
00:36:35,094 --> 00:36:37,136
if i dont everything

405
00:36:38,036 --> 00:36:51,102
but what do I do with it you gave it I
like any other costs and what all do now

406
00:36:52,002 --> 00:36:57,051
is I will give you a very simple
principles followed by a set of very

407
00:36:57,051 --> 00:37:02,051
simple obstruct instructions you can
follow in order to gain the street costs

408
00:37:02,051 --> 00:37:08,145
and make them work in your favor rather
than against you so the principles are

409
00:37:09,045 --> 00:37:15,093
all on the logic you need to change you
don't want to own things that are not

410
00:37:15,093 --> 00:37:25,114
changing you don't want to own a stack
you don't need to change that Taylor

411
00:37:26,014 --> 00:37:33,045
needs you don't if you just need basic
payment gateway right just use spark

412
00:37:33,045 --> 00:37:37,136
because you probably would not need to
change it

413
00:37:38,036 --> 00:37:49,071
Taylor will even go even further right
on the logic you needs to own that's

414
00:37:49,071 --> 00:37:53,149
like that's even trickier because like a
only means testing during so like tried

415
00:37:54,049 --> 00:37:57,052
to minimize the amount of things that I
mind of things that you need to own

416
00:37:57,052 --> 00:38:03,106
which is easier just like don't test
which is I'm doing anyways the trick

417
00:38:04,006 --> 00:38:09,042
there is right on the logic unit
everything else

418
00:38:09,042 --> 00:38:23,135
outsourced to Taylor is amazing with it
everything you write I have this report

419
00:38:24,035 --> 00:38:30,104
repeating question people ask me is like
to Utd everything my latest answer is

420
00:38:31,004 --> 00:38:37,037
yes I duty every single line of code I
write every single one of them

421
00:38:37,037 --> 00:38:45,048
the problem is I'm not trying to write
anything I'm trying not to writing when

422
00:38:45,048 --> 00:38:50,124
I'm looking at the problem what the
first thing I'm asking is can I reuse

423
00:38:51,024 --> 00:38:57,046
something that you know just you some
module some plugin will it help me to

424
00:38:57,046 --> 00:39:04,085
solve this business problem can't there
is no need to write to me does drive it

425
00:39:04,085 --> 00:39:11,123
because I need to own that nobody else
will on everything you write try not to

426
00:39:12,023 --> 00:39:16,098
write and try to not write anything
reuse everything else

427
00:39:16,098 --> 00:39:21,164
repeat after me im serious im dead
serious

428
00:39:22,064 --> 00:39:35,157
own everything you write tried to not
write anything reused everything else

429
00:39:36,057 --> 00:39:42,130
so here's how you do it you get new
requirements you had to feature first

430
00:39:43,003 --> 00:39:47,039
thing you do you always do this document
the need I don't care if you use behind

431
00:39:47,066 --> 00:39:48,115
our Google or anything else

432
00:39:49,015 --> 00:39:53,057
document what you've been asked to
exactly and the best way to this is

433
00:39:53,057 --> 00:39:56,101
usually example said yes can you give me
an example of how this thing should work

434
00:39:57,001 --> 00:40:00,055
you write example down I don't care how
you write in which were much you write

435
00:40:00,055 --> 00:40:04,068
it you just write it down because you
will forget they will forget you will

436
00:40:04,068 --> 00:40:11,156
leave the company after use be screwed
after you document the need to spike

437
00:40:12,056 --> 00:40:17,117
experiment with tours available you try
to find tools that do this thing for you

438
00:40:18,017 --> 00:40:25,040
and get you like 90% of the way you are
able to use out of the box tools of

439
00:40:25,004 --> 00:40:31,082
horrible right and you try to get there
and then when you got there

440
00:40:32,018 --> 00:40:36,069
you start changing it but always when
you're changing the things test the

441
00:40:36,069 --> 00:40:42,085
constraints right does it fit so this
new payment methodology read this new

442
00:40:42,085 --> 00:40:45,158
group management thing doesn't really
fit spark

443
00:40:46,058 --> 00:40:49,065
despite becoming a problem for your
particular situation for your particular

444
00:40:49,065 --> 00:40:55,099
context and if it does you throw this
thing away and you can do this because

445
00:40:55,099 --> 00:40:58,117
you didn't spend too much time building
it in the first place you just use

446
00:40:59,017 --> 00:41:02,041
something and you stabilize it

447
00:41:02,041 --> 00:41:10,070
you claim to ownership of the freakin
thing if the solution grows outside of

448
00:41:10,007 --> 00:41:15,099
the boundaries of the tool available for
you you remove the tool and rewrite the

449
00:41:16,062 --> 00:41:23,159
thing in the proper way and the way to
do this in the proper way you isolate

450
00:41:24,059 --> 00:41:30,066
religious so unit testing your eloquent
controllers or so eloquent models will

451
00:41:31,029 --> 00:41:35,034
not help because eloquent models are not
yours you're not only

452
00:41:35,079 --> 00:41:39,163
seventy percent of eloquent morale is
Taylor's work dealers in communities

453
00:41:40,063 --> 00:41:45,130
work right seventy percent of things
that this thing does even though you're

454
00:41:46,003 --> 00:41:51,041
really it is your class that extends it
is not something that you own is not

455
00:41:51,068 --> 00:41:57,132
something that you will change something
that community will change right so what

456
00:41:58,032 --> 00:42:03,094
you do is you in order to take an
ownership you isolate the thing that you

457
00:42:03,094 --> 00:42:10,151
do you want to learn from the things
that you want others to do so again you

458
00:42:11,051 --> 00:42:15,146
don't need to spike it implemented in
the easiest way possible if there is a

459
00:42:16,046 --> 00:42:19,097
module available for it if not tough
luck you jump to that step number for

460
00:42:19,097 --> 00:42:25,143
you document changes and constrains that
this thing to initial documents and if

461
00:42:26,043 --> 00:42:31,045
this thing doesn't work anymore for your
business needs you throw it away

462
00:42:31,045 --> 00:42:36,132
everything except the communication and
the learnings that you had documented

463
00:42:37,032 --> 00:42:42,075
needs and documented changes and you
rewrite the state according to these

464
00:42:42,075 --> 00:42:47,140
documents according to this needs
according to the requirements right and

465
00:42:48,004 --> 00:42:55,103
you isolate religiously you cannot own
thing I just extending and testing it

466
00:42:56,039 --> 00:43:00,045
that's not only phone is that's exactly
your thing it doesn't expect from

467
00:43:00,045 --> 00:43:10,138
anything you own 100% of called their
unit testing is only

468
00:43:11,038 --> 00:43:19,113
factoring is owning test driven
development is an ownership technique

469
00:43:20,013 --> 00:43:33,025
that make it past reflector vest make it
past reflector test refactor those are

470
00:43:33,025 --> 00:43:40,087
two key factors of phone in the same
what did he does it helps you to pay

471
00:43:40,087 --> 00:43:46,171
costs of ownership on the ongoing basis
in a very comfortable way that's all

472
00:43:47,071 --> 00:43:50,142
there is to D that's driven development
is ongoing

473
00:43:51,042 --> 00:43:58,055
payment of ownership it helps you to
reflectors things to support the future

474
00:43:58,055 --> 00:44:03,124
changes but it costs time and the time
it costs has no value to the business to

475
00:44:04,024 --> 00:44:09,045
stop stop selling it to the business the
value that it adds as for you so you can

476
00:44:09,045 --> 00:44:12,110
change things for the business in the
future and we share which things you

477
00:44:13,001 --> 00:44:22,049
need to own it is kind of your call to
recover document everything you do

478
00:44:22,058 --> 00:44:28,116
spike and stabilize you can google for
it there is really cool technique and

479
00:44:29,016 --> 00:44:33,031
used to do for stabilization when you're
stabilizing DVDs the best tool for this

480
00:44:33,031 --> 00:44:40,110
thing ever used all those pictures in
the slides if you didn't notice thank

481
00:44:41,001 --> 00:45:05,097
you very much

482
00:45:06,006 --> 00:46:13,099
so you don't you don't you might want
but I definitely don't want to write

483
00:46:13,099 --> 00:46:19,146
anything I want to pull things in and
then there is this there is this

484
00:46:20,046 --> 00:46:27,050
question when does this thing available
it is very clear clear on someone like

485
00:46:27,005 --> 00:46:30,014
you don't have a plugin to use writer
just like you just need to write it

486
00:46:30,059 --> 00:46:34,085
would need to write a need to all meant
if there is a plugin but like what to do

487
00:46:34,085 --> 00:46:37,087
if there is a plug-in if my chest to 70%
of your needs

488
00:46:38,005 --> 00:46:44,010
what if it matches to like 50 percent of
your needs and I think there is no

489
00:46:44,001 --> 00:46:50,047
predetermined answer to this is
something that your team needs to do but

490
00:46:50,056 --> 00:46:52,147
no matter what you decide there is
literally two options for you to go

491
00:46:53,047 --> 00:46:58,115
about it first option is you take it in
you write this thing but you don't take

492
00:46:59,015 --> 00:47:04,057
ownership and you encourage potential
cost of change later on and another

493
00:47:04,057 --> 00:47:08,073
option used you're taking it in you
isolate religiously and you take

494
00:47:08,073 --> 00:47:10,132
ownership of the part that this

495
00:47:10,789 --> 00:47:15,813
does not provide I would I would prefer
person in the second option and the

496
00:47:16,029 --> 00:47:19,128
team's I'm working with would probably
prefer the second option but I don't

497
00:47:20,019 --> 00:47:28,025
think it's like right or wrong answer
follow-up question what about their

498
00:47:28,079 --> 00:47:34,125
practice if we talk about stripe in this
case for instance partial list as a

499
00:47:34,539 --> 00:47:38,615
great package on strike but I don't
really need all the functions of that

500
00:47:39,299 --> 00:47:48,140
but it really say there's one package to
70% and another two hundred and forty

501
00:47:48,014 --> 00:47:54,223
percent of the beaches I need the way to
go about that too I a bill of eighty

502
00:47:54,349 --> 00:48:00,438
percent want to just extend that to my
owner or a different question altogether

503
00:48:01,239 --> 00:48:05,296
not really tired but you generally
considered it costs of having that thing

504
00:48:05,809 --> 00:48:10,813
and I would usually answer that why
would you care how much features it has

505
00:48:10,849 --> 00:48:14,948
if it doesn't force you to use that
feature those features right it doesn't

506
00:48:15,839 --> 00:48:17,934
force you those features are still
hanging in there and you just not using

507
00:48:18,789 --> 00:48:25,833
them you're not taken ownership of that
thing using them if it does some

508
00:48:26,229 --> 00:48:31,640
instability problems does slow you down
then you're going into all different

509
00:48:31,064 --> 00:48:34,065
conversations like does it worth
actually owned in this thing on your own

510
00:48:34,065 --> 00:48:37,634
and this is a consideration that I would
have with the team and discussion that I

511
00:48:38,219 --> 00:48:44,284
would have a good example actually from
from from the life we worked on our team

512
00:48:44,869 --> 00:48:49,911
worked worked in the Drupal project and
the guys you some import script from

513
00:48:50,289 --> 00:48:53,388
gentle because they were trying to
integrate two systems and the used out

514
00:48:54,279 --> 00:48:57,650
of the box Drupal module for that
purpose like some of the community

515
00:48:57,065 --> 00:48:59,139
modules and the like

516
00:49:00,039 --> 00:49:04,208
the end-to-end tested the whole thing
just like through the critical path just

517
00:49:04,559 --> 00:49:08,940
to make sure that it works that went
crazy but they basically didn't take

518
00:49:08,094 --> 00:49:11,163
much ownership of the thing and then
they started flickering so it's like

519
00:49:12,063 --> 00:49:18,482
basically important every second time by
the clock but it didn't work every time

520
00:49:19,049 --> 00:49:24,380
so what they did they had a discussion
as a team with a product on there and

521
00:49:24,038 --> 00:49:29,126
they said look it's very risky doesn't
really work in the way we do so what do

522
00:49:30,026 --> 00:49:33,635
we do and they decided to rewrite the
thing to the drop the module and a row

523
00:49:33,869 --> 00:49:36,964
the roads their own logic for importing
the couple from Drupal and then they

524
00:49:37,819 --> 00:49:41,841
introduced the model that integrates it
with Drupal so it took them like up

525
00:49:42,039 --> 00:49:45,112
three more days instead of just like
using out of the box thing but then they

526
00:49:45,769 --> 00:49:50,852
have a very stable import and as a side
benefit it was work in like 10 times

527
00:49:51,599 --> 00:49:55,697
faster than the Drupal module of us so
that's just an example for you can go

528
00:49:56,579 --> 00:50:00,608
about it there is no strict answers just
like here is the thing that you always

529
00:50:00,869 --> 00:50:07,630
will need to do in this case which is
not answering granted but the best one I

530
00:50:07,063 --> 00:50:18,088
can give you a very good answer
questions

531
00:50:18,088 --> 00:50:30,977
laboriously assuming the evening meeting
interface and adapters to logic from the

532
00:50:31,769 --> 00:50:36,848
dependencies so you know that there are
many ways you can isolate things the

533
00:50:37,559 --> 00:50:43,670
classical pathway is a breach by doing
so you introduce an interface which you

534
00:50:43,067 --> 00:50:46,256
used to define what this thing needs to
do and then you implement this interface

535
00:50:46,859 --> 00:50:51,878
as like as a boundary ports in a doctor
is another kind of tangent on this that

536
00:50:52,049 --> 00:50:56,052
you can read into a doctor's bridge
butter and reintroduce two things that

537
00:50:56,079 --> 00:51:00,640
will give you enough information to know
how to isolate religiously but that's

538
00:51:00,064 --> 00:51:02,085
the thing you need to do if you want to
on anything

539
00:51:02,085 --> 00:51:13,139
and you kind of want to own everything
that you write your questions first

540
00:51:14,039 --> 00:51:20,078
question how do you manage to factor in
the risk of all the dependencies not

541
00:51:20,078 --> 00:51:25,126
necessarily having long-term support or
even then not taking ownership and you

542
00:51:26,026 --> 00:51:30,103
suddenly you know being exposed to that
so that that's a very great question

543
00:51:31,003 --> 00:51:36,049
thank you for this one thing I didn't
touch upon and that's like something

544
00:51:36,049 --> 00:51:42,080
that I didn't like in the previous
question is do we need to use the plugin

545
00:51:42,008 --> 00:51:47,040
is that module good enough that's the
wall separate discussion for you and

546
00:51:48,012 --> 00:51:50,088
your team and look for me personally

547
00:51:50,088 --> 00:51:54,114
anything that doesn't have tests in it
speak spectral there are testable there

548
00:51:55,014 --> 00:52:01,059
is not a good model for me it by default
means that this model does not exist in

549
00:52:01,059 --> 00:52:06,150
my eyes right because I like simply
because I cannot take on the risk of

550
00:52:07,005 --> 00:52:10,082
that module being broken and nobody
knows except including its outer alter

551
00:52:11,027 --> 00:52:15,116
that is broken it's easier for me to
take an ownership on that responsibility

552
00:52:16,016 --> 00:52:21,048
or find a better mental stability is the
important thing you need to take into

553
00:52:21,048 --> 00:52:28,087
consideration same as speed and the
general performance hit so those three

554
00:52:28,087 --> 00:52:35,090
things you keep in mind when you are
taking upon module and basically the

555
00:52:35,009 --> 00:52:41,095
threshold you set yourself and the team
thing and it's basic trishul it it's

556
00:52:42,076 --> 00:52:44,169
worth as a team when you're starting a
project to agree on that threshold

557
00:52:45,069 --> 00:52:49,086
that's what what is a set of
requirements for every module were

558
00:52:49,086 --> 00:52:54,122
integrated into this project right and
you can say was the team if it doesn't

559
00:52:55,022 --> 00:53:00,114
meet the threshold to a pretend it
doesn't exist and you make an agreement

560
00:53:01,014 --> 00:53:05,069
so we are cases agreement is like if it
doesn't he needs just pretend it's not

561
00:53:05,069 --> 00:53:09,073
there and then another module and if
there is no another module just write it

562
00:53:09,073 --> 00:53:13,147
and we tried to open source probably
open source said so that others do

563
00:53:14,047 --> 00:53:25,050
have dilemma that we have the answer we
don't have time for the questions we can

564
00:53:25,077 --> 00:53:38,094
do one more question I'm here till the
end of the day so find me talk with me

565
00:53:38,094 --> 00:53:42,143
about this thing I find it fascinating
so shared with me please leave feedback

566
00:53:43,043 --> 00:53:44,076
on the joins in

567
00:53:44,076 --> 00:53:46,143
thank you very much again for coming to
Sprint in an hour