1
00:00:04,960 --> 00:00:20,000
[Music]

2
00:00:20,640 --> 00:00:25,519
hello oh the mic works okay great hello

3
00:00:23,680 --> 00:00:29,199
everyone welcome to the lightning talk

4
00:00:25,519 --> 00:00:30,519
session for uh the dev oops track um

5
00:00:29,199 --> 00:00:32,680
this is our first

6
00:00:30,519 --> 00:00:34,440
um time running lightning talks for Dev

7
00:00:32,680 --> 00:00:38,200
oops so there's going to be a lot of

8
00:00:34,440 --> 00:00:39,120
oopsies um but it's going to be fun

9
00:00:38,200 --> 00:00:44,440
who's

10
00:00:39,120 --> 00:00:50,160
ready yeah let's go Okay cool so

11
00:00:44,440 --> 00:00:52,879
um we are all good okay so um how is

12
00:00:50,160 --> 00:00:54,800
this going to work so uh we're going to

13
00:00:52,879 --> 00:00:58,280
introduce who's up on deck which in this

14
00:00:54,800 --> 00:01:02,239
case is James uh and then who is the

15
00:00:58,280 --> 00:01:02,239
next upcoming speaker which will will

16
00:01:03,600 --> 00:01:09,560
be in one of these

17
00:01:06,119 --> 00:01:11,920
tabs I found the tab oh so next up will

18
00:01:09,560 --> 00:01:15,119
be Pat so when I say next up I want you

19
00:01:11,920 --> 00:01:18,439
to be ready to come on and um connect to

20
00:01:15,119 --> 00:01:20,000
the podium Podium has a HDMI cable um

21
00:01:18,439 --> 00:01:22,439
please connect as quickly as possible we

22
00:01:20,000 --> 00:01:26,320
only have one um

23
00:01:22,439 --> 00:01:27,799
Podium uh and so the quicker you do it

24
00:01:26,320 --> 00:01:30,680
the less awkward silence there will be

25
00:01:27,799 --> 00:01:32,439
and the less improv I'll have to do cuz

26
00:01:30,680 --> 00:01:35,399
I'm not that good at

27
00:01:32,439 --> 00:01:38,840
improv uh but maybe I'll get better in

28
00:01:35,399 --> 00:01:41,600
this hour uh

29
00:01:38,840 --> 00:01:44,880
okay that's the spirit yeah let's go

30
00:01:41,600 --> 00:01:47,320
okay um yes it's very important to keep

31
00:01:44,880 --> 00:01:50,399
to time because we only have an hour and

32
00:01:47,320 --> 00:01:51,960
we have eight amazing speakers uh and if

33
00:01:50,399 --> 00:01:53,600
it takes too long and you go over five

34
00:01:51,960 --> 00:02:00,280
minutes we're going to throw you off the

35
00:01:53,600 --> 00:02:01,680
stage um I will I won't be able to um

36
00:02:00,280 --> 00:02:05,560
okay uh so there'll be a timer for

37
00:02:01,680 --> 00:02:08,879
speakers which um Justin will be um

38
00:02:05,560 --> 00:02:11,480
showing you uh and in the last 10

39
00:02:08,879 --> 00:02:13,720
seconds there'll be um we want everyone

40
00:02:11,480 --> 00:02:17,120
to be doing finger claps like this CA

41
00:02:13,720 --> 00:02:19,160
show me yeah cool okay so that's last 10

42
00:02:17,120 --> 00:02:20,760
seconds and then the last 3 seconds

43
00:02:19,160 --> 00:02:23,200
we're going to do like a little hand

44
00:02:20,760 --> 00:02:24,720
Applause a little bit louder yeah yeah

45
00:02:23,200 --> 00:02:26,720
cool and then the last second we're

46
00:02:24,720 --> 00:02:28,640
going have really loud claps and and

47
00:02:26,720 --> 00:02:30,360
telling you we get off stage yeah nice

48
00:02:28,640 --> 00:02:33,920
nice nice okay and then

49
00:02:30,360 --> 00:02:36,599
okay awesome cool so bring it up for

50
00:02:33,920 --> 00:02:39,000
James who is going to be speaking

51
00:02:36,599 --> 00:02:42,080
about

52
00:02:39,000 --> 00:02:45,400
something which is Pixie plus UV modern

53
00:02:42,080 --> 00:02:45,400
python packaging take your away

54
00:02:45,640 --> 00:02:49,760
James thanks everyone uh pleasure to be

55
00:02:48,159 --> 00:02:52,560
here my name is James today we're

56
00:02:49,760 --> 00:02:55,360
talking about the modern state of python

57
00:02:52,560 --> 00:02:58,519
packaging so talking about two tools

58
00:02:55,360 --> 00:03:00,200
today uh one is Pixie and one is UV uh

59
00:02:58,519 --> 00:03:03,720
so could I get a raise of hand who's

60
00:03:00,200 --> 00:03:05,799
heard of UV before probably most or a

61
00:03:03,720 --> 00:03:07,280
good chunk of people it looks like if

62
00:03:05,799 --> 00:03:09,080
you if you've been on hacking news I

63
00:03:07,280 --> 00:03:10,959
guess at any point in the last 6 months

64
00:03:09,080 --> 00:03:13,319
you've probably seen it on there seems

65
00:03:10,959 --> 00:03:14,920
to be popping up quite frequently but

66
00:03:13,319 --> 00:03:17,440
for those who haven't this one's for you

67
00:03:14,920 --> 00:03:19,360
so usually these days when you do like

68
00:03:17,440 --> 00:03:21,400
some python packaging you're probably

69
00:03:19,360 --> 00:03:23,280
using well if you're old school you're

70
00:03:21,400 --> 00:03:24,319
using pip you're creating your virtual

71
00:03:23,280 --> 00:03:27,360
environments you have like a

72
00:03:24,319 --> 00:03:30,120
requirements txt uh maybe you use pip

73
00:03:27,360 --> 00:03:31,519
compile you have like a requirements.in

74
00:03:30,120 --> 00:03:34,560
thing and then you compile the

75
00:03:31,519 --> 00:03:36,560
requirements txt well UV is like here to

76
00:03:34,560 --> 00:03:38,439
sort of change all of this and it's

77
00:03:36,560 --> 00:03:40,799
essentially an improvement on a lot of

78
00:03:38,439 --> 00:03:44,080
where python packaging has gone from a

79
00:03:40,799 --> 00:03:45,959
pip install through to poetry there's

80
00:03:44,080 --> 00:03:48,280
like hatch

81
00:03:45,959 --> 00:03:50,239
PDM you know there's probably like 10

82
00:03:48,280 --> 00:03:51,519
other tools that exist currently so

83
00:03:50,239 --> 00:03:53,280
hopefully this is the tool that we can

84
00:03:51,519 --> 00:03:55,000
sort of get behind um so I'm just going

85
00:03:53,280 --> 00:03:59,879
to like quickly just rock in and do a

86
00:03:55,000 --> 00:04:02,400
little live demo so um well I'm going to

87
00:03:59,879 --> 00:04:05,599
make a thing called UV and just show

88
00:04:02,400 --> 00:04:08,159
around how it works so if I do UV in it

89
00:04:05,599 --> 00:04:11,239
now I have a project that is now my

90
00:04:08,159 --> 00:04:13,959
python project I can do um if I actually

91
00:04:11,239 --> 00:04:15,360
do a it's going to have I've got a

92
00:04:13,959 --> 00:04:16,799
python version here which people might

93
00:04:15,360 --> 00:04:19,239
have seen before so it's going to

94
00:04:16,799 --> 00:04:21,759
basically tell my IDE and whatever which

95
00:04:19,239 --> 00:04:23,000
python version I'm using um I don't have

96
00:04:21,759 --> 00:04:24,919
to actually activate into the

97
00:04:23,000 --> 00:04:28,479
environment so it's if I install a

98
00:04:24,919 --> 00:04:31,440
package like um let's just add EV ad py

99
00:04:28,479 --> 00:04:32,360
test so that was pretty pretty quick um

100
00:04:31,440 --> 00:04:35,520
don't think I've done that before so

101
00:04:32,360 --> 00:04:37,759
that's probably not either um anyway so

102
00:04:35,520 --> 00:04:40,320
now par so I don't have to like activate

103
00:04:37,759 --> 00:04:43,720
into this environment uh if I do like UV

104
00:04:40,320 --> 00:04:46,320
run I think it's hello.py is the default

105
00:04:43,720 --> 00:04:48,680
one yep there it is so UV run is like an

106
00:04:46,320 --> 00:04:50,520
UV ad all this kind of stuff kind of

107
00:04:48,680 --> 00:04:51,560
abstracts away the virtual environment

108
00:04:50,520 --> 00:04:54,520
so you don't have to worry about that

109
00:04:51,560 --> 00:04:56,199
anymore it's very fast uh it installs

110
00:04:54,520 --> 00:04:59,120
python for you so you don't have to have

111
00:04:56,199 --> 00:05:00,639
like Pi en or any of these other tools

112
00:04:59,120 --> 00:05:05,120
um this is of the direction where a lot

113
00:05:00,639 --> 00:05:07,440
of modern pip Python package development

114
00:05:05,120 --> 00:05:09,840
and all this kind of stuff is going so

115
00:05:07,440 --> 00:05:12,080
if you haven't tried it or done anything

116
00:05:09,840 --> 00:05:15,160
with it uh recently then I'd highly

117
00:05:12,080 --> 00:05:16,360
recommend um but the second tool which

118
00:05:15,160 --> 00:05:18,600
I'm not sure I'm tracking for time so

119
00:05:16,360 --> 00:05:22,120
we'll keep going but

120
00:05:18,600 --> 00:05:25,840
um okay great halfway that's that's good

121
00:05:22,120 --> 00:05:27,639
uh Pixie so raise your hand if you're

122
00:05:25,840 --> 00:05:29,919
like a condo

123
00:05:27,639 --> 00:05:31,639
person there's a couple of people yeah

124
00:05:29,919 --> 00:05:33,160
so it's probably much less but you know

125
00:05:31,639 --> 00:05:35,039
cond does have a place especially if

126
00:05:33,160 --> 00:05:37,319
you're in a certain I mean ourwork

127
00:05:35,039 --> 00:05:39,240
corporate environment you can't in

128
00:05:37,319 --> 00:05:41,720
sometimes install everything you want

129
00:05:39,240 --> 00:05:43,800
directly from from GitHub or wherever so

130
00:05:41,720 --> 00:05:47,319
cond is a good way to install some of

131
00:05:43,800 --> 00:05:49,520
these things um so condo is kind of the

132
00:05:47,319 --> 00:05:51,240
thing that was used um a while ago to

133
00:05:49,520 --> 00:05:52,840
install like things like numai and

134
00:05:51,240 --> 00:05:54,840
pandas because you have all these like C

135
00:05:52,840 --> 00:05:56,440
dependencies that are were hard to get

136
00:05:54,840 --> 00:05:58,479
nowadays you can use Wheels to do that

137
00:05:56,440 --> 00:06:00,759
but that didn't really exist for a while

138
00:05:58,479 --> 00:06:02,360
um and condo was really slow so people

139
00:06:00,759 --> 00:06:03,639
kind of were sick of it and then they

140
00:06:02,360 --> 00:06:05,319
they moved to this thing called Mamba

141
00:06:03,639 --> 00:06:06,800
which was like kind of like condo but it

142
00:06:05,319 --> 00:06:09,440
was a bit faster and it pretty much did

143
00:06:06,800 --> 00:06:10,880
all the same stuff nowadays uh probably

144
00:06:09,440 --> 00:06:12,680
in the last two years I think the the

145
00:06:10,880 --> 00:06:14,840
same guys who created Mamba have created

146
00:06:12,680 --> 00:06:16,919
this store called pixie um and it's kind

147
00:06:14,840 --> 00:06:20,080
of a similar thing to UV it's built with

148
00:06:16,919 --> 00:06:22,360
rust it's very fast it has a lot of

149
00:06:20,080 --> 00:06:23,720
different benefits so you can pretty

150
00:06:22,360 --> 00:06:26,199
much install whatever you want so if I

151
00:06:23,720 --> 00:06:29,840
do like pixie in it it's going to like

152
00:06:26,199 --> 00:06:32,120
create my my thing and then I can add uh

153
00:06:29,840 --> 00:06:34,319
Python 3 out2 and it's going to be

154
00:06:32,120 --> 00:06:36,720
pretty quick hopefully yep there you go

155
00:06:34,319 --> 00:06:38,240
okay great um you can do other stuff in

156
00:06:36,720 --> 00:06:39,840
here so you can have like a different

157
00:06:38,240 --> 00:06:40,800
environments within the same project

158
00:06:39,840 --> 00:06:43,440
well actually that's one of the main

159
00:06:40,800 --> 00:06:45,560
differences is like cond and Mambo are

160
00:06:43,440 --> 00:06:46,520
like you create this global environment

161
00:06:45,560 --> 00:06:49,720
you can sort of take it with you

162
00:06:46,520 --> 00:06:51,160
wherever you go but pixie is Project

163
00:06:49,720 --> 00:06:53,840
based which is similar to how a lot of

164
00:06:51,160 --> 00:06:56,199
other python tools work which is sort of

165
00:06:53,840 --> 00:06:58,960
more ergonomic from a developer

166
00:06:56,199 --> 00:07:00,000
perspective um you can yeah install

167
00:06:58,960 --> 00:07:01,400
whatever you want you have different

168
00:07:00,000 --> 00:07:02,800
environments so you can have like a test

169
00:07:01,400 --> 00:07:05,080
environment you want to run your tests

170
00:07:02,800 --> 00:07:06,720
with like optional packages that should

171
00:07:05,080 --> 00:07:08,919
only exist in the test environment

172
00:07:06,720 --> 00:07:10,639
things like that anyway very good if

173
00:07:08,919 --> 00:07:12,560
you're on cond and you haven't tried

174
00:07:10,639 --> 00:07:14,080
pixie already would highly recommend

175
00:07:12,560 --> 00:07:16,879
anyway I'm going to sort of wrap it up

176
00:07:14,080 --> 00:07:17,990
there I think we're nearly at time but

177
00:07:16,879 --> 00:07:26,540
there you go thanks

178
00:07:17,990 --> 00:07:26,540
[Applause]

179
00:07:27,160 --> 00:07:35,720
folks next up is oh look it works okay

180
00:07:29,919 --> 00:07:39,120
great um awesome so up next we have uh

181
00:07:35,720 --> 00:07:40,919
GTH could you get ready please um but uh

182
00:07:39,120 --> 00:07:47,000
right now we have Pat who's going to be

183
00:07:40,919 --> 00:07:48,240
speaking about um ml Ops what how that's

184
00:07:47,000 --> 00:07:52,039
a very good

185
00:07:48,240 --> 00:07:56,240
question um all right let's see how

186
00:07:52,039 --> 00:07:56,240
quickly this works

187
00:08:06,159 --> 00:08:10,479
um cool take it away

188
00:08:10,680 --> 00:08:16,159
Pat

189
00:08:12,599 --> 00:08:19,759
um who here has heard of the term

190
00:08:16,159 --> 00:08:21,360
mlops yeah that's good that's maybe a

191
00:08:19,759 --> 00:08:26,039
third is a lot of people haven't all

192
00:08:21,360 --> 00:08:27,960
right who here has worked on some worked

193
00:08:26,039 --> 00:08:29,080
on a project being a data scientist done

194
00:08:27,960 --> 00:08:30,080
something which you'd say you'd have

195
00:08:29,080 --> 00:08:34,479
some sort of

196
00:08:30,080 --> 00:08:36,120
Ops okay like 10ish maybe okay that's

197
00:08:34,479 --> 00:08:37,599
cool I mean I was I wasn't really sure I

198
00:08:36,120 --> 00:08:39,599
was going to say I just like it's

199
00:08:37,599 --> 00:08:41,479
something I talk about a lot um this is

200
00:08:39,599 --> 00:08:43,159
a slide deck I put together like a year

201
00:08:41,479 --> 00:08:45,399
ago I just thought like I'll Chuck it up

202
00:08:43,159 --> 00:08:49,279
and see what it is so um emots machine

203
00:08:45,399 --> 00:08:52,800
learning operations uh it's yeah devops

204
00:08:49,279 --> 00:08:55,880
for machine learning um it's not

205
00:08:52,800 --> 00:08:58,040
something that even existed uh like six

206
00:08:55,880 --> 00:09:00,399
six years ago now I remember I first

207
00:08:58,040 --> 00:09:03,800
started working in data science in in

208
00:09:00,399 --> 00:09:06,000
2019 and like end of 2019 I some of us

209
00:09:03,800 --> 00:09:07,600
were saying we need we need to do this

210
00:09:06,000 --> 00:09:08,680
better this is this is a mess like we

211
00:09:07,600 --> 00:09:10,399
got to have something better I've heard

212
00:09:08,680 --> 00:09:13,839
there's this thing out there called

213
00:09:10,399 --> 00:09:16,720
mlops um and yeah like it's I think a

214
00:09:13,839 --> 00:09:19,200
lot of organizations in the the process

215
00:09:16,720 --> 00:09:22,399
of trying to turn create actual working

216
00:09:19,200 --> 00:09:23,839
data science products um that at the

217
00:09:22,399 --> 00:09:27,640
same of time are trying to figure this

218
00:09:23,839 --> 00:09:30,079
out it's an ongoing thing um this is

219
00:09:27,640 --> 00:09:32,399
probably the best and simplest

220
00:09:30,079 --> 00:09:34,760
explanation of it if I talk to like

221
00:09:32,399 --> 00:09:37,320
managers orone want this anyone wants

222
00:09:34,760 --> 00:09:38,839
this for exacts like it's really good so

223
00:09:37,320 --> 00:09:40,920
um especially if we're talking so you'll

224
00:09:38,839 --> 00:09:44,640
have a bit of familiar familiarity with

225
00:09:40,920 --> 00:09:47,320
software so super simplistic um can

226
00:09:44,640 --> 00:09:49,800
shoot it down but yeah traditional like

227
00:09:47,320 --> 00:09:52,800
devops will involve you know we've got a

228
00:09:49,800 --> 00:09:54,600
code base we're making changes to it for

229
00:09:52,800 --> 00:09:56,200
for whatever reasons features fixes

230
00:09:54,600 --> 00:09:57,839
whatever and there's a whole bunch of

231
00:09:56,200 --> 00:10:00,399
tests that will run on that before we

232
00:09:57,839 --> 00:10:01,959
deploy it um and they'll monitor it to

233
00:10:00,399 --> 00:10:03,160
make sure that it keeps on running okay

234
00:10:01,959 --> 00:10:05,320
or if it doesn't we find out what's

235
00:10:03,160 --> 00:10:07,160
going wrong and you know that's a that's

236
00:10:05,320 --> 00:10:09,320
part of a cicd system to make sure we

237
00:10:07,160 --> 00:10:12,600
can make change and iterate and improve

238
00:10:09,320 --> 00:10:13,480
you know in a reliable way and so

239
00:10:12,600 --> 00:10:16,440
machine

240
00:10:13,480 --> 00:10:19,079
learning lost of bit that's all right um

241
00:10:16,440 --> 00:10:21,360
is this plus the added comp like a whole

242
00:10:19,079 --> 00:10:24,720
extra dimension of complexity of

243
00:10:21,360 --> 00:10:26,519
changing data over time um and a model a

244
00:10:24,720 --> 00:10:27,920
machine learning model like there's all

245
00:10:26,519 --> 00:10:31,079
all these things neuron

246
00:10:27,920 --> 00:10:33,800
networks you know um so I could learn AI

247
00:10:31,079 --> 00:10:36,040
all that stuff like it's just a way of

248
00:10:33,800 --> 00:10:38,000
encapsulating information about a data

249
00:10:36,040 --> 00:10:40,160
set um and then try to use that

250
00:10:38,000 --> 00:10:41,800
information to make predictions about

251
00:10:40,160 --> 00:10:43,519
new data but it's all very heavily

252
00:10:41,800 --> 00:10:47,399
dependent on data and end of the day

253
00:10:43,519 --> 00:10:50,600
there's more testing more monitoring um

254
00:10:47,399 --> 00:10:53,720
more uh yeah more more you know like

255
00:10:50,600 --> 00:10:56,320
just extra tests um extra tracking that

256
00:10:53,720 --> 00:10:59,399
needs to be done as part of it

257
00:10:56,320 --> 00:11:01,480
um and so like a team that's going to do

258
00:10:59,399 --> 00:11:03,279
this is going to have to be more complex

259
00:11:01,480 --> 00:11:04,560
and more more skilled like have a WI a

260
00:11:03,279 --> 00:11:06,959
variety of data and other engineering

261
00:11:04,560 --> 00:11:09,079
skills in a regular team um and that's

262
00:11:06,959 --> 00:11:11,639
the main thing and and this is another

263
00:11:09,079 --> 00:11:14,560
famous um uh image that people might

264
00:11:11,639 --> 00:11:15,519
have seen two minutes go um they're like

265
00:11:14,560 --> 00:11:17,040
if you want to get to the point of

266
00:11:15,519 --> 00:11:19,880
having machine learning running

267
00:11:17,040 --> 00:11:21,519
application you know a lot of what gets

268
00:11:19,880 --> 00:11:23,160
taught in say when I did a data science

269
00:11:21,519 --> 00:11:24,399
course is how to make models the model

270
00:11:23,160 --> 00:11:26,639
and the data science the machine

271
00:11:24,399 --> 00:11:28,920
learning part is only a tiny little part

272
00:11:26,639 --> 00:11:30,720
of what's going on we need all the rest

273
00:11:28,920 --> 00:11:33,079
of like the traditional software

274
00:11:30,720 --> 00:11:35,079
application code that goes around that

275
00:11:33,079 --> 00:11:36,519
um and if we're hosting in cloud and

276
00:11:35,079 --> 00:11:38,320
doing everything there everything that

277
00:11:36,519 --> 00:11:40,800
goes on there plus a whole bunch of

278
00:11:38,320 --> 00:11:42,399
extra stuff around like managing the

279
00:11:40,800 --> 00:11:44,639
very complex data processes and

280
00:11:42,399 --> 00:11:46,639
monitoring that's going on there and so

281
00:11:44,639 --> 00:11:48,560
that's kind of the space that we're

282
00:11:46,639 --> 00:11:52,120
working in trying to like figure that

283
00:11:48,560 --> 00:11:53,800
out as we go um there are tools and

284
00:11:52,120 --> 00:11:55,800
platforms out there that are supposed to

285
00:11:53,800 --> 00:11:58,320
be you know supposed to make a lot of

286
00:11:55,800 --> 00:12:00,440
this easier um I'm yet to use them I'm I

287
00:11:58,320 --> 00:12:02,720
was hoping it' be like Has anyone used

288
00:12:00,440 --> 00:12:04,079
any of these ml platforms if you have

289
00:12:02,720 --> 00:12:06,240
please come and talk to me I'd love to

290
00:12:04,079 --> 00:12:07,880
pick the brains of some people who have

291
00:12:06,240 --> 00:12:09,480
um but at the end of the day it still

292
00:12:07,880 --> 00:12:12,959
come it seems it still comes down to a

293
00:12:09,480 --> 00:12:14,360
problem of any machine learning data

294
00:12:12,959 --> 00:12:15,920
science project involves a lot of

295
00:12:14,360 --> 00:12:17,160
veryable spoke stuff we have a problem

296
00:12:15,920 --> 00:12:19,360
we're trying to fix we work at

297
00:12:17,160 --> 00:12:21,199
industrial settings we've got you know

298
00:12:19,360 --> 00:12:23,040
industrial machines and data sources

299
00:12:21,199 --> 00:12:24,639
which are very specific to that we need

300
00:12:23,040 --> 00:12:27,480
to write a whole bunch of very specific

301
00:12:24,639 --> 00:12:29,240
code to you know turn that into

302
00:12:27,480 --> 00:12:31,240
information we put into a model and use

303
00:12:29,240 --> 00:12:33,199
outputs and there's there's like always

304
00:12:31,240 --> 00:12:36,000
going to be an element of what's bespoke

305
00:12:33,199 --> 00:12:37,560
and what can be automated in terms of

306
00:12:36,000 --> 00:12:41,120
like scaling doing this sort of

307
00:12:37,560 --> 00:12:45,519
monitoring and improving over time

308
00:12:41,120 --> 00:12:48,320
um yeah cool I mean that's that's emops

309
00:12:45,519 --> 00:12:51,199
in a nutshell um I would love to talk to

310
00:12:48,320 --> 00:12:53,800
anyone who's you know engaged or doing

311
00:12:51,199 --> 00:12:53,800
anything in that

312
00:12:57,760 --> 00:13:08,440
space thank you so much that was great

313
00:13:00,880 --> 00:13:08,440
um so GTH you're on now y uh and

314
00:13:08,720 --> 00:13:16,560
in up next we

315
00:13:12,720 --> 00:13:21,000
have mysterious person called Anthony

316
00:13:16,560 --> 00:13:25,959
you're up next get ready please cool

317
00:13:21,000 --> 00:13:31,199
HDMI yes do you not have HDMI okay time

318
00:13:25,959 --> 00:13:34,399
for some improv um cool who has done ml

319
00:13:31,199 --> 00:13:37,600
Ops raise of hands can anyone USB see

320
00:13:34,399 --> 00:13:41,040
this uh okay look at that the audience

321
00:13:37,600 --> 00:13:42,279
is coming out to save us wooo thanny

322
00:13:41,040 --> 00:13:44,839
thank

323
00:13:42,279 --> 00:13:47,440
youny okay oh I haven't said what goth

324
00:13:44,839 --> 00:13:51,480
is speaking about uh when will python

325
00:13:47,440 --> 00:13:53,639
3.6 support end hasn't it already ended

326
00:13:51,480 --> 00:13:57,800
maybe goth will tell us about

327
00:13:53,639 --> 00:13:59,639
it uh okay but put put the laptop here

328
00:13:57,800 --> 00:14:03,600
so people can hear you

329
00:13:59,639 --> 00:14:07,199
okay screen ah because microphones yes

330
00:14:03,600 --> 00:14:10,519
look python 3.6 excellent oh there we

331
00:14:07,199 --> 00:14:13,399
go when will it

332
00:14:10,519 --> 00:14:16,959
end sorry that was my tick not me this

333
00:14:13,399 --> 00:14:18,519
is me long story come talk to me later

334
00:14:16,959 --> 00:14:21,800
uh anyway

335
00:14:18,519 --> 00:14:25,680
you that'll help right when will python

336
00:14:21,800 --> 00:14:27,480
3.6 support end I wonder allow

337
00:14:25,680 --> 00:14:29,560
rhetorically isn't that just ancient

338
00:14:27,480 --> 00:14:32,600
history

339
00:14:29,560 --> 00:14:37,040
it's so old that if you go to um endof

340
00:14:32,600 --> 00:14:39,399
life. whatever that it was handy website

341
00:14:37,040 --> 00:14:42,959
um you have to go all the way down and

342
00:14:39,399 --> 00:14:44,880
click show more unmaintained releases to

343
00:14:42,959 --> 00:14:47,240
find out and then we can go

344
00:14:44,880 --> 00:14:50,560
2018 was the last that we had to worry

345
00:14:47,240 --> 00:14:50,560
about python 3.6

346
00:14:50,639 --> 00:14:55,959
right because python that's that's when

347
00:14:53,480 --> 00:14:59,639
they stopped fixing

348
00:14:55,959 --> 00:15:02,040
bugs ah but some of you point out

349
00:14:59,639 --> 00:15:02,040
what about

350
00:15:03,279 --> 00:15:11,560
security real answer is 2021 because

351
00:15:07,440 --> 00:15:11,560
vules but that was still in the past

352
00:15:11,920 --> 00:15:16,600
right this is somewhat Complicated by

353
00:15:16,800 --> 00:15:22,880
the the role of distributions where

354
00:15:20,440 --> 00:15:24,519
they've taken software maintained by

355
00:15:22,880 --> 00:15:26,199
other people and then repackaged it and

356
00:15:24,519 --> 00:15:27,360
then taken over maintenance for for

357
00:15:26,199 --> 00:15:30,160
their

358
00:15:27,360 --> 00:15:33,800
customers and so for the people who

359
00:15:30,160 --> 00:15:37,480
have to Target these platforms they have

360
00:15:33,800 --> 00:15:39,360
to stick to this road map which is not

361
00:15:37,480 --> 00:15:42,759
Python's road

362
00:15:39,360 --> 00:15:44,720
map um now why is Fedora important well

363
00:15:42,759 --> 00:15:47,600
the reason is that in hidden in the user

364
00:15:44,720 --> 00:15:51,440
lib exec directory which many hands up

365
00:15:47,600 --> 00:15:55,319
who knows what user Li exec contains and

366
00:15:51,440 --> 00:15:55,319
why you lucky

367
00:15:55,560 --> 00:16:00,079
bastards some of us have to worry about

368
00:15:57,639 --> 00:16:01,959
this for a living um us a li exec is

369
00:16:00,079 --> 00:16:04,160
where you hide binaries that are used by

370
00:16:01,959 --> 00:16:05,560
other binaries they they are not really

371
00:16:04,160 --> 00:16:08,920
for your

372
00:16:05,560 --> 00:16:10,920
use um anyway so there's a copy of

373
00:16:08,920 --> 00:16:13,440
python in there it's called user Li ex

374
00:16:10,920 --> 00:16:16,160
platform python

375
00:16:13,440 --> 00:16:19,199
why what so dnf can use

376
00:16:16,160 --> 00:16:21,279
it which is the new name of yum their

377
00:16:19,199 --> 00:16:23,880
package manager so literally just to

378
00:16:21,279 --> 00:16:25,639
install packages you need a copy of this

379
00:16:23,880 --> 00:16:27,399
yeah you can see it hidden the upper

380
00:16:25,639 --> 00:16:31,639
where the little thing is use a liic

381
00:16:27,399 --> 00:16:33,800
platform python and that's a python 3.6

382
00:16:31,639 --> 00:16:35,279
which means that anytime that any python

383
00:16:33,800 --> 00:16:36,880
programmer is told hey you you have to

384
00:16:35,279 --> 00:16:42,199
Target this

385
00:16:36,880 --> 00:16:44,360
particular uh ancient Fedora um

386
00:16:42,199 --> 00:16:47,079
hypothetically um but you're not allowed

387
00:16:44,360 --> 00:16:47,079
to install

388
00:16:47,839 --> 00:16:54,800
anything well there is a python on there

389
00:16:51,440 --> 00:16:59,480
and if you don't use it you have to use

390
00:16:54,800 --> 00:16:59,480
bash or Pearl four

391
00:17:01,360 --> 00:17:07,520
so from that perspective you could argue

392
00:17:04,480 --> 00:17:11,600
that the real end of life for python 3.6

393
00:17:07,520 --> 00:17:13,720
was 2018 because

394
00:17:11,600 --> 00:17:15,880
UTA but it's not that

395
00:17:13,720 --> 00:17:20,000
easy red

396
00:17:15,880 --> 00:17:21,480
hat red hat tooker and turned it into

397
00:17:20,000 --> 00:17:25,000
release for

398
00:17:21,480 --> 00:17:28,640
eight right so the real answer is 2029

399
00:17:25,000 --> 00:17:28,640
because red hat

400
00:17:29,919 --> 00:17:34,039
or 2032 because

401
00:17:37,120 --> 00:17:41,080
Enterprise but what will it be because

402
00:17:39,440 --> 00:17:43,280
you and and I'm asking this because

403
00:17:41,080 --> 00:17:46,160
there might be people who maintain stuff

404
00:17:43,280 --> 00:17:49,600
that I would like to rely on as a target

405
00:17:46,160 --> 00:17:51,600
of these platforms um and and so really

406
00:17:49,600 --> 00:17:54,280
it comes down to whether or not you

407
00:17:51,600 --> 00:17:57,159
think I'm an edge

408
00:17:54,280 --> 00:18:00,320
case or as I prefer to put it a real

409
00:17:57,159 --> 00:18:01,640
life actual person yeah I am a fellow

410
00:18:00,320 --> 00:18:04,880
python

411
00:18:01,640 --> 00:18:09,320
programmer so the hero stick I would

412
00:18:04,880 --> 00:18:11,120
like to urge you to consider is um one

413
00:18:09,320 --> 00:18:13,280
minute thank you just to consider where

414
00:18:11,120 --> 00:18:15,960
you are between the

415
00:18:13,280 --> 00:18:17,679
platform and and the user I guess at the

416
00:18:15,960 --> 00:18:20,159
other end the

417
00:18:17,679 --> 00:18:21,880
application and the closer you are to

418
00:18:20,159 --> 00:18:25,640
the

419
00:18:21,880 --> 00:18:26,799
fun definitely use 3113 full of great

420
00:18:25,640 --> 00:18:28,840
stuff right but the closer you are to

421
00:18:26,799 --> 00:18:30,039
the platform if you could support these

422
00:18:28,840 --> 00:18:32,640
the the people closer to the metal

423
00:18:30,039 --> 00:18:34,679
that'd be great so and in particular if

424
00:18:32,640 --> 00:18:38,799
you're supporting build

425
00:18:34,679 --> 00:18:40,840
tooling uh packaging like UV um I've

426
00:18:38,799 --> 00:18:43,159
already fixed 1 3.6 bug in I broke it

427
00:18:40,840 --> 00:18:45,720
again so I have to go back and fix it um

428
00:18:43,159 --> 00:18:47,520
fundamental packages like requests like

429
00:18:45,720 --> 00:18:49,480
if if you're big and bad enough to be

430
00:18:47,520 --> 00:18:51,280
maintaining a package which is used by

431
00:18:49,480 --> 00:18:53,159
literally everybody how about supporting

432
00:18:51,280 --> 00:18:57,240
it for literally

433
00:18:53,159 --> 00:18:58,799
everybody thanks um and uh yeah like

434
00:18:57,240 --> 00:19:03,520
like I said close you out of the app

435
00:18:58,799 --> 00:19:03,520
doesn't matter and please do so until

436
00:19:04,120 --> 00:19:07,559
2032 because the

437
00:19:12,000 --> 00:19:18,120
Enterprise thank you very much who's

438
00:19:14,559 --> 00:19:22,880
next uh good question let's see

439
00:19:18,120 --> 00:19:27,200
Anthony's up on deck and up next Once my

440
00:19:22,880 --> 00:19:30,000
laptop logs in you can do it

441
00:19:27,200 --> 00:19:32,880
laptop no

442
00:19:30,000 --> 00:19:36,159
oh up next we have uh

443
00:19:32,880 --> 00:19:39,480
Warwick cool

444
00:19:36,159 --> 00:19:43,039
uh cool Anthony's going to speak about

445
00:19:39,480 --> 00:19:48,640
also Python and releases but bug F bug

446
00:19:43,039 --> 00:19:51,400
fixes how and why uh almost

447
00:19:48,640 --> 00:19:56,360
there to

448
00:19:51,400 --> 00:19:59,600
come bug fix yay okay great is the

449
00:19:56,360 --> 00:20:04,799
window gone cool take it away Anthony

450
00:19:59,600 --> 00:20:07,679
okay ah so uh sorry I'm my glasses are

451
00:20:04,799 --> 00:20:07,679
fogging up

452
00:20:07,760 --> 00:20:15,640
uh oh good good point um okay python bug

453
00:20:12,799 --> 00:20:18,720
fix releases why do we have them and how

454
00:20:15,640 --> 00:20:21,280
do we do do them so I'm Anthony I used

455
00:20:18,720 --> 00:20:23,440
to be the python release manager uh of

456
00:20:21,280 --> 00:20:25,720
all those versions people want to do

457
00:20:23,440 --> 00:20:26,960
quick math that's 5 and a half years and

458
00:20:25,720 --> 00:20:29,840
three and a half years where I was the

459
00:20:26,960 --> 00:20:32,760
only release manager

460
00:20:29,840 --> 00:20:35,080
um B so the short version of this talk

461
00:20:32,760 --> 00:20:37,360
one slide pep 6 please read it it'll

462
00:20:35,080 --> 00:20:39,520
make you happy will make me happy these

463
00:20:37,360 --> 00:20:41,919
are examples of a bug fix release you

464
00:20:39,520 --> 00:20:44,080
know the third digit uh some people call

465
00:20:41,919 --> 00:20:46,200
them Point fix releases they're wrong

466
00:20:44,080 --> 00:20:49,080
it's pep six says they bug fix releases

467
00:20:46,200 --> 00:20:52,320
and it's right so what's what's in a bug

468
00:20:49,080 --> 00:20:54,320
fix release only bug fixes except

469
00:20:52,320 --> 00:20:58,039
there's a security buug then then we can

470
00:20:54,320 --> 00:21:00,640
change things why because true and false

471
00:20:58,039 --> 00:21:01,640
and 2 3 and 2.2 yes we're talking about

472
00:21:00,640 --> 00:21:04,400
the Dark

473
00:21:01,640 --> 00:21:08,480
Ages true and false were unexpectedly

474
00:21:04,400 --> 00:21:10,840
backported to python 2.2 in a point

475
00:21:08,480 --> 00:21:14,039
release it was originally added in two

476
00:21:10,840 --> 00:21:15,360
three and again unexpected back backport

477
00:21:14,039 --> 00:21:18,600
so why is this

478
00:21:15,360 --> 00:21:21,360
important because of this if you look at

479
00:21:18,600 --> 00:21:24,279
old python 2 code you'll often find this

480
00:21:21,360 --> 00:21:26,559
at the top of every single python

481
00:21:24,279 --> 00:21:28,760
file because you had to you didn't know

482
00:21:26,559 --> 00:21:31,440
you couldn't use Python 22 does it have

483
00:21:28,760 --> 00:21:34,400
true and false I don't

484
00:21:31,440 --> 00:21:36,840
know so I had some words with GTO some

485
00:21:34,400 --> 00:21:38,880
of them were rude and said said let's

486
00:21:36,840 --> 00:21:41,320
not do that again and I sort of took

487
00:21:38,880 --> 00:21:43,440
over doing the maintenance releases then

488
00:21:41,320 --> 00:21:45,960
uh so there was a big problem there the

489
00:21:43,440 --> 00:21:48,080
vendors as GTH was just saying didn't

490
00:21:45,960 --> 00:21:51,159
trust us the distributions didn't trust

491
00:21:48,080 --> 00:21:53,279
us which was a pain in the butt because

492
00:21:51,159 --> 00:21:55,120
if we break functionality they don't

493
00:21:53,279 --> 00:21:58,080
trust us cuz they've got customers they'

494
00:21:55,120 --> 00:22:00,760
got Downstream vendors or Downstream dys

495
00:21:58,080 --> 00:22:03,200
and so won't they won't put in bug

496
00:22:00,760 --> 00:22:05,120
fixers there they'll do delays of

497
00:22:03,200 --> 00:22:08,000
getting the bug fixers out there'll be

498
00:22:05,120 --> 00:22:10,240
vendors back porting bugs fixes which is

499
00:22:08,000 --> 00:22:14,360
also

500
00:22:10,240 --> 00:22:16,200
bad uh I didn't swear um duplicate this

501
00:22:14,360 --> 00:22:19,559
the problem with this was that we ended

502
00:22:16,200 --> 00:22:20,840
up with duplicate bugs in in our tracker

503
00:22:19,559 --> 00:22:22,799
and we go no that's already fixed you

504
00:22:20,840 --> 00:22:24,840
just need to upgrade and they go well I

505
00:22:22,799 --> 00:22:27,640
can't upgrade or whatever um it makes

506
00:22:24,840 --> 00:22:30,520
this look bad when there's fit bugs we

507
00:22:27,640 --> 00:22:32,480
know about and people like see and there

508
00:22:30,520 --> 00:22:35,520
might be unpatched security fixes

509
00:22:32,480 --> 00:22:38,080
security bugs bad again so how do we do

510
00:22:35,520 --> 00:22:40,919
this well you have a python release

511
00:22:38,080 --> 00:22:43,120
manager who watches the uh maintenance

512
00:22:40,919 --> 00:22:44,720
maintenance branch and looks for things

513
00:22:43,120 --> 00:22:47,000
that change functionality and then you

514
00:22:44,720 --> 00:22:49,919
yell at someone and and if necessary you

515
00:22:47,000 --> 00:22:53,440
roll their fix their change back so

516
00:22:49,919 --> 00:22:55,000
again read pep 6 it's not very long it's

517
00:22:53,440 --> 00:22:58,200
you know dates back a little bit but not

518
00:22:55,000 --> 00:22:59,840
too not that okay dates back to 2001 um

519
00:22:58,200 --> 00:23:02,000
so

520
00:22:59,840 --> 00:23:04,080
if you want to do an update please do it

521
00:23:02,000 --> 00:23:08,080
refers to things like CVS and Source

522
00:23:04,080 --> 00:23:09,960
Forge um it really it really needs an

523
00:23:08,080 --> 00:23:11,360
update and you can also do a bit of

524
00:23:09,960 --> 00:23:13,679
archaeology because it also has a list

525
00:23:11,360 --> 00:23:15,840
of who was the maintainer for various

526
00:23:13,679 --> 00:23:19,559
branches so that might be fun and if

527
00:23:15,840 --> 00:23:22,600
you've never never done a push a a patch

528
00:23:19,559 --> 00:23:24,919
this is a good excuse anyway please

529
00:23:22,600 --> 00:23:25,880
please don't break up don't break bug

530
00:23:24,919 --> 00:23:29,880
fix

531
00:23:25,880 --> 00:23:29,880
releases anyway

532
00:23:32,120 --> 00:23:38,400
thank you

533
00:23:33,679 --> 00:23:42,360
Anthony hold on uh okay say lightning

534
00:23:38,400 --> 00:23:43,240
so uh yes wck is up on stage up next we

535
00:23:42,360 --> 00:23:47,520
have

536
00:23:43,240 --> 00:23:50,440
Kara uh sorry oh good walk is going to

537
00:23:47,520 --> 00:23:53,840
speak about I would actually appreciate

538
00:23:50,440 --> 00:23:57,919
a oh no no no I don't have a something

539
00:23:53,840 --> 00:23:59,279
about HDMX yes awesome okay are we

540
00:23:57,919 --> 00:24:02,919
plugged in are you

541
00:23:59,279 --> 00:24:05,240
ready one two three oh

542
00:24:02,919 --> 00:24:09,640
okay one two

543
00:24:05,240 --> 00:24:12,000
three okay uh oh it worked okay that was

544
00:24:09,640 --> 00:24:16,000
some time delay cool okay take it away

545
00:24:12,000 --> 00:24:18,279
thank you okay so um I've just got a

546
00:24:16,000 --> 00:24:20,120
ajango project that I just scaffold

547
00:24:18,279 --> 00:24:24,159
together right now uh it's essentially

548
00:24:20,120 --> 00:24:27,279
got one view uh which returns this work

549
00:24:24,159 --> 00:24:30,000
function which is just uh sleeps for 3

550
00:24:27,279 --> 00:24:31,600
seconds and returns a u ID uh so

551
00:24:30,000 --> 00:24:34,640
anything that's

552
00:24:31,600 --> 00:24:37,559
complicated um on you know your database

553
00:24:34,640 --> 00:24:41,240
or some calculation so the page

554
00:24:37,559 --> 00:24:43,120
therefore takes a while to load uh so I

555
00:24:41,240 --> 00:24:44,640
want to speed that up or at least make

556
00:24:43,120 --> 00:24:48,840
it better for the

557
00:24:44,640 --> 00:24:52,000
user uh so first thing obviously we can

558
00:24:48,840 --> 00:24:59,120
just go and uh in the view we can just

559
00:24:52,000 --> 00:25:01,200
cache it um so cool um so now it'll

560
00:24:59,120 --> 00:25:03,159
take a while to load but when I do

561
00:25:01,200 --> 00:25:04,679
subsequent loads uh the problem is that

562
00:25:03,159 --> 00:25:06,600
you've got the cold start so the first

563
00:25:04,679 --> 00:25:11,080
person to get to your website is always

564
00:25:06,600 --> 00:25:15,679
going to get uh that crap um loading

565
00:25:11,080 --> 00:25:18,480
experience uh so has anyone here used

566
00:25:15,679 --> 00:25:22,360
hmx uh for those who haven't it's it's a

567
00:25:18,480 --> 00:25:24,039
really great way to uh make look also

568
00:25:22,360 --> 00:25:25,559
Twitter's Twitter is a dumpster fire

569
00:25:24,039 --> 00:25:27,120
right now but he's well worth a follow

570
00:25:25,559 --> 00:25:29,600
he's like the one thing that's still

571
00:25:27,120 --> 00:25:32,080
good there um it's a really simple

572
00:25:29,600 --> 00:25:34,520
JavaScript library to return tiny little

573
00:25:32,080 --> 00:25:36,679
page Snippets over the wire but use HTML

574
00:25:34,520 --> 00:25:39,440
instead of Json so you don't need to F

575
00:25:36,679 --> 00:25:41,200
around with view or react or any kind of

576
00:25:39,440 --> 00:25:43,320
client side templating you can literally

577
00:25:41,200 --> 00:25:48,960
return HTML from Jango so I've got a

578
00:25:43,320 --> 00:25:51,799
second view here uh so um

579
00:25:48,960 --> 00:25:53,840
slor uh and my index.html this is all I

580
00:25:51,799 --> 00:25:57,240
need to do I just need to add HTM X and

581
00:25:53,840 --> 00:25:59,279
I just say HX get from the work URL and

582
00:25:57,240 --> 00:26:02,200
Trigger it on load

583
00:25:59,279 --> 00:26:05,399
uh so that then brings up the page

584
00:26:02,200 --> 00:26:10,520
snippet so now I get an instant load uh

585
00:26:05,399 --> 00:26:12,679
but I've still got that time there um so

586
00:26:10,520 --> 00:26:14,279
better cuz you're still you're getting a

587
00:26:12,679 --> 00:26:16,039
immediate load and you're lazy loading

588
00:26:14,279 --> 00:26:17,039
the page component uh in production I

589
00:26:16,039 --> 00:26:20,120
actually use this for far more

590
00:26:17,039 --> 00:26:23,520
sophisticated things um but we can do

591
00:26:20,120 --> 00:26:27,840
better so uh I want to mix and match a

592
00:26:23,520 --> 00:26:32,640
little bit um so on the uh do work I've

593
00:26:27,840 --> 00:26:35,039
now put the caching on again uh so cool

594
00:26:32,640 --> 00:26:38,320
uh the only problem here is I'm still

595
00:26:35,039 --> 00:26:40,480
doing that extra Network request so if I

596
00:26:38,320 --> 00:26:43,640
change this and delay just to simulate

597
00:26:40,480 --> 00:26:47,039
the network uh you'll see um just doing

598
00:26:43,640 --> 00:26:48,440
the full load this time but we'll do um

599
00:26:47,039 --> 00:26:51,799
so it's still doing Network request

600
00:26:48,440 --> 00:26:53,559
every time uh but we've already got this

601
00:26:51,799 --> 00:26:54,840
data and I'm I'm a bit traditional I

602
00:26:53,559 --> 00:26:56,159
feel like if you're loading a page from

603
00:26:54,840 --> 00:27:00,000
the server you just want to load the

604
00:26:56,159 --> 00:27:01,320
whole thing if you can uh so so I have

605
00:27:00,000 --> 00:27:04,840
this hacky

606
00:27:01,320 --> 00:27:06,679
solution um that I [ __ ]

607
00:27:04,840 --> 00:27:09,760
sorry I

608
00:27:06,679 --> 00:27:11,840
know oh um that's my fault I'm using

609
00:27:09,760 --> 00:27:13,600
Source tree specifically because I knew

610
00:27:11,840 --> 00:27:16,919
that trying to do this in the command

611
00:27:13,600 --> 00:27:20,240
line would break my brain uh doing great

612
00:27:16,919 --> 00:27:23,080
then um so what I've done is this uh

613
00:27:20,240 --> 00:27:25,720
essentially taken D jango's existing C

614
00:27:23,080 --> 00:27:27,320
template tag uh copied it and just

615
00:27:25,720 --> 00:27:29,559
modified it so that it will never

616
00:27:27,320 --> 00:27:34,480
actually write to the C

617
00:27:29,559 --> 00:27:38,480
uh so in the work snippet I use the old

618
00:27:34,480 --> 00:27:41,360
C tag uh which does the complicated

619
00:27:38,480 --> 00:27:44,320
thing uh writes it into the case in my

620
00:27:41,360 --> 00:27:46,159
index though I use my tag which will

621
00:27:44,320 --> 00:27:48,919
only get from the cas and won't write to

622
00:27:46,159 --> 00:27:50,039
it uh the fallback is to then use the

623
00:27:48,919 --> 00:27:52,440
HTM X

624
00:27:50,039 --> 00:27:57,360
trigger

625
00:27:52,440 --> 00:28:00,600
so uh we now get slow page instant page

626
00:27:57,360 --> 00:28:03,360
and if I even go and change

627
00:28:00,600 --> 00:28:07,399
this yeah

628
00:28:03,360 --> 00:28:07,399
good uh to use an HTM X

629
00:28:08,960 --> 00:28:15,720
delay

630
00:28:10,799 --> 00:28:18,080
should there we go so yep um so that's

631
00:28:15,720 --> 00:28:19,760
that's basically it um I've been

632
00:28:18,080 --> 00:28:22,519
migrating a very large project as part

633
00:28:19,760 --> 00:28:25,240
of my work uh and um that's just a handy

634
00:28:22,519 --> 00:28:28,320
little tidbit I learned along the way um

635
00:28:25,240 --> 00:28:30,720
so if anyone hasn't tried HDMX uh do go

636
00:28:28,320 --> 00:28:34,279
check it out um it's really great for D

637
00:28:30,720 --> 00:28:35,880
Jango uh just means you you can you can

638
00:28:34,279 --> 00:28:38,240
write websites the good old way with

639
00:28:35,880 --> 00:28:42,360
HTML and no Jason and I really like that

640
00:28:38,240 --> 00:28:42,360
so um thank you uh

641
00:28:43,919 --> 00:28:47,960
yeah cool thank you very

642
00:28:46,200 --> 00:28:52,039
much

643
00:28:47,960 --> 00:28:55,120
um cool we've got casara on

644
00:28:52,039 --> 00:28:58,600
stage and up next

645
00:28:55,120 --> 00:29:02,200
Susanna is Susanna here

646
00:28:58,600 --> 00:29:05,799
okay great uh well not great but

647
00:29:02,200 --> 00:29:08,120
um yes kar's on stage is going to talk

648
00:29:05,799 --> 00:29:10,240
about how Docker best practices can

649
00:29:08,120 --> 00:29:12,279
haunt you are you going to do scary

650
00:29:10,240 --> 00:29:15,480
things are you going to make this slide

651
00:29:12,279 --> 00:29:19,000
thing appear that's kind of scary

652
00:29:15,480 --> 00:29:22,440
cry it's like a jump scare but with

653
00:29:19,000 --> 00:29:22,440
slides okay take

654
00:29:22,480 --> 00:29:29,799
itara hello my name is Kara and um I'm a

655
00:29:27,039 --> 00:29:31,679
Pon near you first

656
00:29:29,799 --> 00:29:35,000
timer thank

657
00:29:31,679 --> 00:29:37,880
you um so I'm a software dual fav

658
00:29:35,000 --> 00:29:41,240
profession and kind of forc into this

659
00:29:37,880 --> 00:29:44,880
whole Dev oops thing because working for

660
00:29:41,240 --> 00:29:48,200
a small organization with in a small

661
00:29:44,880 --> 00:29:52,159
team you have to do

662
00:29:48,200 --> 00:29:56,880
everything so we had this bunch of uh

663
00:29:52,159 --> 00:29:59,679
all services running on U bareal

664
00:29:56,880 --> 00:30:03,919
servers and and we tried

665
00:29:59,679 --> 00:30:05,600
to get them into the cloud and so first

666
00:30:03,919 --> 00:30:07,760
step was to

667
00:30:05,600 --> 00:30:11,320
create put them in a

668
00:30:07,760 --> 00:30:13,360
Docker images and create Docker images

669
00:30:11,320 --> 00:30:15,480
so I was following the best practices

670
00:30:13,360 --> 00:30:18,799
I'm trying to choose the

671
00:30:15,480 --> 00:30:22,120
proper Docker image from a verified

672
00:30:18,799 --> 00:30:24,000
publisher and Docker best practices say

673
00:30:22,120 --> 00:30:28,679
don't install anything that you don't

674
00:30:24,000 --> 00:30:32,399
want so using the slim package

675
00:30:28,679 --> 00:30:35,159
and yeah everything's good up until the

676
00:30:32,399 --> 00:30:37,240
time that you think you don't have to

677
00:30:35,159 --> 00:30:41,080
connect to the production

678
00:30:37,240 --> 00:30:42,720
server but sometimes you do when you do

679
00:30:41,080 --> 00:30:46,760
when you're using slim packages you

680
00:30:42,720 --> 00:30:50,799
might not have tools like PS or

681
00:30:46,760 --> 00:30:50,799
top and now in

682
00:30:51,080 --> 00:30:57,639
production you don't have the tools that

683
00:30:53,279 --> 00:31:02,120
you need the second do one of the other

684
00:30:57,639 --> 00:31:05,000
do best practice is to run Services as a

685
00:31:02,120 --> 00:31:08,000
unprivileged user how many of you run

686
00:31:05,000 --> 00:31:10,760
things as root on

687
00:31:08,000 --> 00:31:16,480
Docker yes

688
00:31:10,760 --> 00:31:18,399
SP so this hor to when because all right

689
00:31:16,480 --> 00:31:21,760
I'm running my services as a

690
00:31:18,399 --> 00:31:24,080
unprivileged user now I can't be the

691
00:31:21,760 --> 00:31:28,039
root on the production

692
00:31:24,080 --> 00:31:31,559
server so I can't install anything

693
00:31:28,039 --> 00:31:35,399
and pretty much done so the only Escape

694
00:31:31,559 --> 00:31:37,880
is to create another release using a

695
00:31:35,399 --> 00:31:42,080
different package and

696
00:31:37,880 --> 00:31:44,159
um have all the tools that you

697
00:31:42,080 --> 00:31:47,760
need

698
00:31:44,159 --> 00:31:50,600
um another thing that might haunt you

699
00:31:47,760 --> 00:31:54,720
with do best practices is

700
00:31:50,600 --> 00:31:58,240
uh so you kind of pin to a version uh on

701
00:31:54,720 --> 00:32:01,720
your docka image but over the years

702
00:31:58,240 --> 00:32:04,480
years pass by and you forget about which

703
00:32:01,720 --> 00:32:09,320
which Linux version that you're using it

704
00:32:04,480 --> 00:32:12,440
could be uh End of Line like all the

705
00:32:09,320 --> 00:32:14,760
security updates are gone and you might

706
00:32:12,440 --> 00:32:17,559
be using something really really

707
00:32:14,760 --> 00:32:20,639
stale and nothing tells you about that

708
00:32:17,559 --> 00:32:24,039
because it's hidden inside you you don't

709
00:32:20,639 --> 00:32:27,240
really look at your Docker files

710
00:32:24,039 --> 00:32:31,200
minutes so the takeaways from my

711
00:32:27,240 --> 00:32:33,720
experience with do do best practices is

712
00:32:31,200 --> 00:32:37,039
sometimes slim package is not the best

713
00:32:33,720 --> 00:32:39,799
you need you might need something tools

714
00:32:37,039 --> 00:32:42,480
like top P top

715
00:32:39,799 --> 00:32:45,639
PS

716
00:32:42,480 --> 00:32:48,120
um when you're starting a project always

717
00:32:45,639 --> 00:32:51,240
start doing uh start running it as an

718
00:32:48,120 --> 00:32:55,000
unprivileged user because trying to get

719
00:32:51,240 --> 00:32:58,360
that run get something that's that you

720
00:32:55,000 --> 00:33:00,840
have already developed to run as an

721
00:32:58,360 --> 00:33:04,200
unprivileged user is

722
00:33:00,840 --> 00:33:07,480
hard um unprivileged users don't have

723
00:33:04,200 --> 00:33:10,799
access to log directory um there are

724
00:33:07,480 --> 00:33:12,840
files that service can't read so it's

725
00:33:10,799 --> 00:33:18,080
better to

726
00:33:12,840 --> 00:33:19,720
um sort those out in the start of your

727
00:33:18,080 --> 00:33:21,080
building pipeline so you don't have to

728
00:33:19,720 --> 00:33:25,679
worry about those

729
00:33:21,080 --> 00:33:28,440
anymore one minute and yeah keep an eye

730
00:33:25,679 --> 00:33:30,480
on the uh Linux version and all the

731
00:33:28,440 --> 00:33:33,639
packages that you use on do Docker

732
00:33:30,480 --> 00:33:36,440
images because they can go get stale so

733
00:33:33,639 --> 00:33:38,480
maybe have some kind of a ci5 plan that

734
00:33:36,440 --> 00:33:41,679
tells you about any of the outdate

735
00:33:38,480 --> 00:33:41,679
packages thank

736
00:33:44,399 --> 00:33:54,639
you wow thank you so much okay uh up

737
00:33:48,600 --> 00:33:55,960
next we have the uh closing for devops

738
00:33:54,639 --> 00:33:57,799
cuz that's all we have time for for

739
00:33:55,960 --> 00:34:00,080
lightning talks I want to say thank

740
00:33:57,799 --> 00:34:02,000
thank you so much to everyone who uh

741
00:34:00,080 --> 00:34:04,200
applied and everyone else has come

742
00:34:02,000 --> 00:34:07,679
because we let everyone know very last

743
00:34:04,200 --> 00:34:09,119
minute they would be presenting and um

744
00:34:07,679 --> 00:34:12,079
they scrambling right at the last

745
00:34:09,119 --> 00:34:14,280
section second to get the slides going

746
00:34:12,079 --> 00:34:17,200
um okay we got

747
00:34:14,280 --> 00:34:20,839
it okay I'm going to do some more improv

748
00:34:17,200 --> 00:34:22,520
until Justin sets up the laptop which is

749
00:34:20,839 --> 00:34:25,359
going to involve me

750
00:34:22,520 --> 00:34:27,839
dancing uh you don't want to see me

751
00:34:25,359 --> 00:34:30,679
dance I have three left feet it's not

752
00:34:27,839 --> 00:34:30,679
good

753
00:34:31,240 --> 00:34:35,240
um cool how how did everyone find

754
00:34:33,960 --> 00:34:39,440
lightning

755
00:34:35,240 --> 00:34:40,720
talks yeah who's ready for uh tomorrow's

756
00:34:39,440 --> 00:34:43,280
lightning

757
00:34:40,720 --> 00:34:45,560
talks yeah nice okay they're not going

758
00:34:43,280 --> 00:34:50,000
to be like this they're going to be more

759
00:34:45,560 --> 00:34:51,760
uh open generic and so forth who is very

760
00:34:50,000 --> 00:34:55,000
keen to apply for tomorrow's lightning

761
00:34:51,760 --> 00:34:58,680
talks and submit a talk have we got one

762
00:34:55,000 --> 00:35:00,079
no one else really two three four five

763
00:34:58,680 --> 00:35:03,200
wow so many okay great I want to see

764
00:35:00,079 --> 00:35:05,480
more hands up uh as you saw it's very

765
00:35:03,200 --> 00:35:07,320
easy uh you just get up on stage and you

766
00:35:05,480 --> 00:35:08,960
talk you don't have to have slides you

767
00:35:07,320 --> 00:35:10,960
can just talk you don't have to make

768
00:35:08,960 --> 00:35:12,400
fancy slides because you won't enough

769
00:35:10,960 --> 00:35:14,839
you won't have enough

770
00:35:12,400 --> 00:35:18,440
time and now I'm getting kicked off

771
00:35:14,839 --> 00:35:18,440
that's my end of time thank you very

772
00:35:18,680 --> 00:35:22,800
much yes please Round of Applause for

773
00:35:20,920 --> 00:35:24,440
Evan for organizing the lightning talks

774
00:35:22,800 --> 00:35:26,640
uh he did that at the last minute and it

775
00:35:24,440 --> 00:35:32,200
is most

776
00:35:26,640 --> 00:35:32,200
excellent thank you so uh now