1
00:00:13,120 --> 00:00:18,880
hello pycon welcome back uh you're back

2
00:00:16,880 --> 00:00:22,000
with the education track

3
00:00:18,880 --> 00:00:25,439
and uh this talk is a pre-recorded talk

4
00:00:22,000 --> 00:00:28,800
but with a special super exciting twist

5
00:00:25,439 --> 00:00:31,119
um the fantastic andre roberge will be

6
00:00:28,800 --> 00:00:34,399
joining us uh for his talk it doesn't

7
00:00:31,119 --> 00:00:38,320
work uh and he will also be joining us

8
00:00:34,399 --> 00:00:39,920
in the chat um on the stream yard chat

9
00:00:38,320 --> 00:00:41,840
you will be able to interact with him

10
00:00:39,920 --> 00:00:44,640
during the talk and

11
00:00:41,840 --> 00:00:48,640
he'll join us live afterwards

12
00:00:44,640 --> 00:00:51,039
um for some questions so

13
00:00:48,640 --> 00:00:52,719
without further ado uh i should

14
00:00:51,039 --> 00:00:55,120
introduce him he's not here i'm

15
00:00:52,719 --> 00:00:57,280
forgetting my things okay so andre first

16
00:00:55,120 --> 00:00:59,840
learned programming by taking a course

17
00:00:57,280 --> 00:01:03,359
in fortran when he was a teenager but he

18
00:00:59,840 --> 00:01:05,040
never had much use for it uh in 2004

19
00:01:03,359 --> 00:01:07,040
while in his 40s he decided to pick up

20
00:01:05,040 --> 00:01:08,640
programming as a hobby and a creative

21
00:01:07,040 --> 00:01:10,880
outlet to distract him from his day job

22
00:01:08,640 --> 00:01:12,880
that sounds fantastic i love it

23
00:01:10,880 --> 00:01:14,960
since then he mostly focused on creating

24
00:01:12,880 --> 00:01:17,119
programs and programming environments

25
00:01:14,960 --> 00:01:18,479
intended to help beginners learn his

26
00:01:17,119 --> 00:01:21,920
favorite hobby

27
00:01:18,479 --> 00:01:24,720
these include uh rupple reeborgs world

28
00:01:21,920 --> 00:01:27,600
avant pai and a few others with the

29
00:01:24,720 --> 00:01:28,400
latest being friendly

30
00:01:27,600 --> 00:01:29,520
so

31
00:01:28,400 --> 00:01:32,320
i will

32
00:01:29,520 --> 00:01:33,920
pass over to the pre-recorded talk um

33
00:01:32,320 --> 00:01:36,479
think up some questions put them in the

34
00:01:33,920 --> 00:01:39,200
chat put them in the questions chat

35
00:01:36,479 --> 00:01:42,799
and we'll answer them all throughout and

36
00:01:39,200 --> 00:01:42,799
after thank you so much

37
00:01:44,399 --> 00:01:48,880
welcome to it doesn't work

38
00:01:46,880 --> 00:01:53,360
a talk given as part of the education

39
00:01:48,880 --> 00:01:55,759
track of pycon australia 2021

40
00:01:53,360 --> 00:01:58,320
my name is andre robert joining you

41
00:01:55,759 --> 00:02:01,200
today from canada where i live and have

42
00:01:58,320 --> 00:02:01,200
fun with python

43
00:02:01,439 --> 00:02:07,040
note that many slides shown during the

44
00:02:03,600 --> 00:02:08,560
stock include tiny xkcd cartoon panels

45
00:02:07,040 --> 00:02:10,399
too small for you to read during the

46
00:02:08,560 --> 00:02:12,160
presentation

47
00:02:10,399 --> 00:02:13,520
these are intended to provide additional

48
00:02:12,160 --> 00:02:16,640
entertainment to people that have

49
00:02:13,520 --> 00:02:16,640
downloaded the slides

50
00:02:18,000 --> 00:02:21,840
since this is the education track i'm

51
00:02:20,160 --> 00:02:23,840
assuming that you have taught or planned

52
00:02:21,840 --> 00:02:27,840
to teach python to someone else that i

53
00:02:23,840 --> 00:02:27,840
will refer to as your students

54
00:02:28,319 --> 00:02:34,080
your students have python including idle

55
00:02:31,440 --> 00:02:36,480
installed on a computer

56
00:02:34,080 --> 00:02:38,800
you decide to introduce them to python

57
00:02:36,480 --> 00:02:41,760
using the turtle module starting with a

58
00:02:38,800 --> 00:02:41,760
complete example

59
00:02:42,640 --> 00:02:46,400
you show them how to run a program

60
00:02:48,000 --> 00:02:52,160
success a square has been drawn

61
00:02:53,040 --> 00:02:58,560
your students copied a code from your

62
00:02:54,720 --> 00:02:58,560
example and attempt to run it

63
00:02:59,599 --> 00:03:02,560
it doesn't work

64
00:03:03,040 --> 00:03:08,319
they see virtual red ink in their

65
00:03:05,200 --> 00:03:10,480
experience red ink is not a good thing

66
00:03:08,319 --> 00:03:13,040
they take a closer look at the content

67
00:03:10,480 --> 00:03:15,760
and see a lot of unknown jargon

68
00:03:13,040 --> 00:03:20,000
traceback attribute error partially

69
00:03:15,760 --> 00:03:21,599
initialized module circular import etc

70
00:03:20,000 --> 00:03:23,599
at this point your students might get

71
00:03:21,599 --> 00:03:25,120
discouraged thinking that programming is

72
00:03:23,599 --> 00:03:27,360
not for them

73
00:03:25,120 --> 00:03:29,840
they feel inadequate and might just want

74
00:03:27,360 --> 00:03:29,840
to give up

75
00:03:30,879 --> 00:03:36,319
could this be made more user-friendly

76
00:03:34,080 --> 00:03:38,080
before we see if this is possible

77
00:03:36,319 --> 00:03:42,080
let's review the type of information we

78
00:03:38,080 --> 00:03:42,080
can get from a normal python traceback

79
00:03:42,959 --> 00:03:47,760
we can see where in the code the error

80
00:03:44,879 --> 00:03:49,519
occurred

81
00:03:47,760 --> 00:03:51,920
we can see what type of exception was

82
00:03:49,519 --> 00:03:51,920
raised

83
00:03:52,480 --> 00:03:57,120
the message gives us an idea as to why

84
00:03:54,480 --> 00:03:59,040
the exception was raised

85
00:03:57,120 --> 00:04:01,360
of course it requires some experience to

86
00:03:59,040 --> 00:04:02,720
be able to decipher such tracebacks

87
00:04:01,360 --> 00:04:06,480
something that we cannot expect

88
00:04:02,720 --> 00:04:10,159
beginners to be able to do on their own

89
00:04:06,480 --> 00:04:10,159
what if there existed a better way

90
00:04:10,640 --> 00:04:15,599
on the left we have a normal traceback

91
00:04:13,599 --> 00:04:18,000
what if we could have like what is shown

92
00:04:15,599 --> 00:04:20,160
on the right side a modified version

93
00:04:18,000 --> 00:04:22,000
with shorter lines and some black and

94
00:04:20,160 --> 00:04:26,080
blue in addition to the red

95
00:04:22,000 --> 00:04:26,080
let's zoom in to this modified traceback

96
00:04:27,520 --> 00:04:31,199
this is the modified traceback

97
00:04:29,520 --> 00:04:33,280
student might perhaps recognize more

98
00:04:31,199 --> 00:04:35,360
easily the text in black as being the

99
00:04:33,280 --> 00:04:36,800
code they wrote

100
00:04:35,360 --> 00:04:39,600
at the bottom

101
00:04:36,800 --> 00:04:42,000
except for the word module

102
00:04:39,600 --> 00:04:44,080
what if a nearly jargon-free easy to

103
00:04:42,000 --> 00:04:46,160
understand sentence was added giving

104
00:04:44,080 --> 00:04:48,400
them a hint of what might be the source

105
00:04:46,160 --> 00:04:50,240
of their problem

106
00:04:48,400 --> 00:04:53,280
did you give your program the same name

107
00:04:50,240 --> 00:04:53,280
as a python module

108
00:04:54,639 --> 00:04:58,320
when using idle we have access to a

109
00:04:56,560 --> 00:05:00,720
python interpreter after running a

110
00:04:58,320 --> 00:05:00,720
program

111
00:05:00,960 --> 00:05:05,440
if students didn't know about python

112
00:05:02,479 --> 00:05:07,759
exceptions what if asking what could

113
00:05:05,440 --> 00:05:11,840
reveal the meaning of the exception in

114
00:05:07,759 --> 00:05:11,840
this case here an attribute error

115
00:05:12,160 --> 00:05:16,320
what if asking where would show the

116
00:05:14,080 --> 00:05:19,199
location of the error more explicitly

117
00:05:16,320 --> 00:05:21,199
than a traditional traceback would

118
00:05:19,199 --> 00:05:23,280
we see the part in the line highlighted

119
00:05:21,199 --> 00:05:26,000
in red indicating the exact location of

120
00:05:23,280 --> 00:05:27,759
the problematic code

121
00:05:26,000 --> 00:05:29,840
at the bottom we see a reference to

122
00:05:27,759 --> 00:05:32,080
turtle.py

123
00:05:29,840 --> 00:05:34,560
before continuing with this example

124
00:05:32,080 --> 00:05:38,280
let me show you a similar where output

125
00:05:34,560 --> 00:05:38,280
from a different example

126
00:05:38,880 --> 00:05:45,280
here at the bottom instead of turtle.py

127
00:05:41,600 --> 00:05:46,720
we see pythonlib turtle.py

128
00:05:45,280 --> 00:05:49,199
this tells us that the code is from

129
00:05:46,720 --> 00:05:50,800
python standards library and not code

130
00:05:49,199 --> 00:05:52,320
written by a student

131
00:05:50,800 --> 00:05:53,759
unlike what we saw with the previous

132
00:05:52,320 --> 00:05:57,039
example

133
00:05:53,759 --> 00:05:57,039
let's go back to that example

134
00:05:57,919 --> 00:06:02,000
what if asking why gave the student a

135
00:06:00,319 --> 00:06:03,280
more complete explanation as to what

136
00:06:02,000 --> 00:06:06,000
might have caused the problem they have

137
00:06:03,280 --> 00:06:06,880
encountered

138
00:06:06,000 --> 00:06:08,720
and

139
00:06:06,880 --> 00:06:12,240
if this was not enough what if there

140
00:06:08,720 --> 00:06:15,919
existed yet one more option

141
00:06:12,240 --> 00:06:18,400
what if typing www as a function

142
00:06:15,919 --> 00:06:20,400
used python's web browser module to do

143
00:06:18,400 --> 00:06:24,479
an internet search with the error

144
00:06:20,400 --> 00:06:24,479
message used as a search string

145
00:06:25,680 --> 00:06:30,160
all of this is possible using friendly

146
00:06:27,919 --> 00:06:34,240
or friendly traceback

147
00:06:30,160 --> 00:06:37,120
let's have a look at a second example

148
00:06:34,240 --> 00:06:39,120
during pycon australia 2018 much fun was

149
00:06:37,120 --> 00:06:41,120
had talking about something called a

150
00:06:39,120 --> 00:06:43,039
flip-flop rater

151
00:06:41,120 --> 00:06:45,440
the code shown on this slide was used in

152
00:06:43,039 --> 00:06:47,440
an example

153
00:06:45,440 --> 00:06:49,120
this code includes a text

154
00:06:47,440 --> 00:06:50,720
i have had a great day with this large

155
00:06:49,120 --> 00:06:52,720
boy mr bubble

156
00:06:50,720 --> 00:06:57,759
you may also know him as masjid chunky

157
00:06:52,720 --> 00:06:57,759
bubble or mr loves to eat the food bobo

158
00:06:58,319 --> 00:07:02,319
here i show the result when i attempt to

159
00:07:00,160 --> 00:07:04,880
run this example using friendly in a

160
00:07:02,319 --> 00:07:08,080
terminal instead of with an idle

161
00:07:04,880 --> 00:07:10,240
i get a trace back with a name error

162
00:07:08,080 --> 00:07:13,759
the text in white at the bottom is a

163
00:07:10,240 --> 00:07:13,759
hint headed by friendly

164
00:07:14,639 --> 00:07:19,599
asking why gives us a fuller explanation

165
00:07:16,800 --> 00:07:21,360
as to what might have caused the problem

166
00:07:19,599 --> 00:07:24,720
it suggests that we might have forgotten

167
00:07:21,360 --> 00:07:26,639
to install the module from pipi

168
00:07:24,720 --> 00:07:29,039
note that friendly also subscribes to

169
00:07:26,639 --> 00:07:31,840
the notion that using the flips operator

170
00:07:29,039 --> 00:07:31,840
is a bad idea

171
00:07:32,639 --> 00:07:36,639
when i watch the talks about the flips

172
00:07:34,240 --> 00:07:38,479
operator i notice the use of the phrase

173
00:07:36,639 --> 00:07:40,160
music

174
00:07:38,479 --> 00:07:42,000
clearly this one example was an

175
00:07:40,160 --> 00:07:45,840
indication that there must be a thriving

176
00:07:42,000 --> 00:07:45,840
french-speaking community in australia

177
00:07:45,919 --> 00:07:50,000
assuming they are watching this talk

178
00:07:47,680 --> 00:07:51,680
those french-speaking pathonistas should

179
00:07:50,000 --> 00:07:53,199
be pleased to know that friendly can

180
00:07:51,680 --> 00:07:55,520
give them information in their preferred

181
00:07:53,199 --> 00:07:57,120
language

182
00:07:55,520 --> 00:07:59,840
this is the end of a rather long

183
00:07:57,120 --> 00:07:59,840
introduction

184
00:08:00,560 --> 00:08:03,759
for the rest of this talk i'd like to

185
00:08:02,080 --> 00:08:05,759
give you a brief historical overview of

186
00:08:03,759 --> 00:08:07,520
the evolution of friendly initially

187
00:08:05,759 --> 00:08:10,879
known as friendly traceback

188
00:08:07,520 --> 00:08:10,879
showing a few more examples

189
00:08:11,039 --> 00:08:14,879
then i will attempt to justify the

190
00:08:12,479 --> 00:08:16,639
approach i use in developing friendly by

191
00:08:14,879 --> 00:08:19,039
using some work from experts in the

192
00:08:16,639 --> 00:08:19,039
field

193
00:08:20,400 --> 00:08:25,039
the purpose of this talk is to motivate

194
00:08:22,240 --> 00:08:27,520
you to use friendly with your students

195
00:08:25,039 --> 00:08:29,360
my hope is that in return you will

196
00:08:27,520 --> 00:08:32,400
provide me with some feedback so that i

197
00:08:29,360 --> 00:08:32,400
can make it even better

198
00:08:33,120 --> 00:08:36,479
when i created friendly traceback my

199
00:08:34,800 --> 00:08:38,080
initial idea was to create a tool that

200
00:08:36,479 --> 00:08:39,680
could help students decipher normal

201
00:08:38,080 --> 00:08:41,200
python tracebacks

202
00:08:39,680 --> 00:08:43,279
so that they could eventually learn to

203
00:08:41,200 --> 00:08:45,279
extract the what where and why on their

204
00:08:43,279 --> 00:08:47,120
own

205
00:08:45,279 --> 00:08:49,600
initially i thought of simply taking the

206
00:08:47,120 --> 00:08:51,519
message included in the traceback

207
00:08:49,600 --> 00:08:53,200
reformulating it so that students could

208
00:08:51,519 --> 00:08:55,519
make better sense of it

209
00:08:53,200 --> 00:08:57,760
and make it possible to translate this

210
00:08:55,519 --> 00:09:00,560
reform loaded message into languages

211
00:08:57,760 --> 00:09:00,560
other than english

212
00:09:01,040 --> 00:09:04,000
so

213
00:09:01,839 --> 00:09:09,080
given an error like what is shown name

214
00:09:04,000 --> 00:09:09,080
error name misu is not defined

215
00:09:09,120 --> 00:09:13,760
friendly trace backwards to provide all

216
00:09:10,880 --> 00:09:15,440
the information shown on this slide

217
00:09:13,760 --> 00:09:17,360
what is a name error

218
00:09:15,440 --> 00:09:18,800
an explanation that might be useful if

219
00:09:17,360 --> 00:09:19,839
such an error is encountered for the

220
00:09:18,800 --> 00:09:21,920
first time

221
00:09:19,839 --> 00:09:23,839
but might become superfluous after a

222
00:09:21,920 --> 00:09:26,000
while

223
00:09:23,839 --> 00:09:29,279
why this error might have occurred

224
00:09:26,000 --> 00:09:30,800
and where exactly in the code

225
00:09:29,279 --> 00:09:33,519
this is the information provided by

226
00:09:30,800 --> 00:09:37,839
friendly traceback with overlaid boxes

227
00:09:33,519 --> 00:09:37,839
on the slide to make it easier to parse

228
00:09:38,000 --> 00:09:42,000
of course in reality

229
00:09:39,600 --> 00:09:44,959
these boxes are not present and we see a

230
00:09:42,000 --> 00:09:44,959
big chunk of text

231
00:09:45,839 --> 00:09:50,080
here instead of using friendly traceback

232
00:09:47,920 --> 00:09:52,000
i used friendly which does some

233
00:09:50,080 --> 00:09:53,680
additional formatting including adding

234
00:09:52,000 --> 00:09:55,519
colors

235
00:09:53,680 --> 00:09:58,480
these added colors should make it easier

236
00:09:55,519 --> 00:10:00,560
to parse the information

237
00:09:58,480 --> 00:10:02,640
still having all information shown at

238
00:10:00,560 --> 00:10:06,959
once especially with more complex

239
00:10:02,640 --> 00:10:09,360
examples can lead to a huge wall of text

240
00:10:06,959 --> 00:10:11,279
this led me to add the functions what

241
00:10:09,360 --> 00:10:13,440
where and why that can be used

242
00:10:11,279 --> 00:10:17,640
interactively to obtain only parts of

243
00:10:13,440 --> 00:10:17,640
the information when required

244
00:10:18,880 --> 00:10:23,279
for example why

245
00:10:21,040 --> 00:10:25,120
in your program no object with the name

246
00:10:23,279 --> 00:10:28,320
must you exist

247
00:10:25,120 --> 00:10:30,160
i have no additional information for you

248
00:10:28,320 --> 00:10:32,480
originally from the traceback did not

249
00:10:30,160 --> 00:10:33,200
add i have no additional information for

250
00:10:32,480 --> 00:10:35,680
you

251
00:10:33,200 --> 00:10:36,720
since that was always the case

252
00:10:35,680 --> 00:10:38,640
however

253
00:10:36,720 --> 00:10:41,680
that has since changed as i will show in

254
00:10:38,640 --> 00:10:41,680
a different example later

255
00:10:42,880 --> 00:10:46,480
of course this approach of explaining

256
00:10:44,640 --> 00:10:48,560
what a given exception means and what

257
00:10:46,480 --> 00:10:50,800
might have caused the error simply based

258
00:10:48,560 --> 00:10:52,880
on the text provided by python

259
00:10:50,800 --> 00:10:56,079
is not exactly helpful when the message

260
00:10:52,880 --> 00:10:58,399
is simply invalid syntax

261
00:10:56,079 --> 00:11:00,399
so a little over two years ago i figured

262
00:10:58,399 --> 00:11:02,320
out a way to automatically analyze a few

263
00:11:00,399 --> 00:11:04,240
invalid syntax cases

264
00:11:02,320 --> 00:11:08,399
and wrote a blog post inviting other

265
00:11:04,240 --> 00:11:08,399
people to suggest additional examples

266
00:11:08,800 --> 00:11:13,120
it did not take long for the first

267
00:11:10,320 --> 00:11:15,279
suggestion to come in

268
00:11:13,120 --> 00:11:17,920
over time a few more suggestions were

269
00:11:15,279 --> 00:11:19,760
made encouraging me to pursue further on

270
00:11:17,920 --> 00:11:22,480
my own in attempt to analyze more and

271
00:11:19,760 --> 00:11:22,480
more cases

272
00:11:23,760 --> 00:11:28,640
here's an example done while using new

273
00:11:26,720 --> 00:11:31,519
arguably the most beginner-friendly

274
00:11:28,640 --> 00:11:33,680
python editor that exists

275
00:11:31,519 --> 00:11:36,880
i first start by importing everything

276
00:11:33,680 --> 00:11:39,440
from friendly.new

277
00:11:36,880 --> 00:11:40,640
after importing the turtle module i try

278
00:11:39,440 --> 00:11:43,839
to call pen

279
00:11:40,640 --> 00:11:43,839
which results in a traceback

280
00:11:44,320 --> 00:11:47,839
friendly sand at the bottom includes two

281
00:11:46,480 --> 00:11:49,920
suggestions

282
00:11:47,839 --> 00:11:53,120
either a typo was made and i meant to

283
00:11:49,920 --> 00:11:55,920
write the built-in lan instead of pen

284
00:11:53,120 --> 00:11:58,399
or since span is an attribute of turtle

285
00:11:55,920 --> 00:12:00,240
a module that was previously imported

286
00:11:58,399 --> 00:12:02,639
perhaps a reference to that module is

287
00:12:00,240 --> 00:12:02,639
missing

288
00:12:04,240 --> 00:12:08,079
mu has different themes

289
00:12:06,399 --> 00:12:09,120
switching to the knight theme changes

290
00:12:08,079 --> 00:12:10,880
the color

291
00:12:09,120 --> 00:12:13,279
and the output provided by friendly no

292
00:12:10,880 --> 00:12:15,360
micro blends in

293
00:12:13,279 --> 00:12:18,560
when using friendly with mu in a night

294
00:12:15,360 --> 00:12:20,160
theme a student or teacher can type in

295
00:12:18,560 --> 00:12:22,000
knight as a function

296
00:12:20,160 --> 00:12:24,639
so that the information displayed

297
00:12:22,000 --> 00:12:26,560
afterwards by friendly matches the theme

298
00:12:24,639 --> 00:12:27,839
used by mu

299
00:12:26,560 --> 00:12:30,079
asking y

300
00:12:27,839 --> 00:12:32,000
results in a more detailed explanation

301
00:12:30,079 --> 00:12:35,200
than that given by the hint added at the

302
00:12:32,000 --> 00:12:37,200
bottom of the traceback by friendly

303
00:12:35,200 --> 00:12:39,360
in the previous example you might have

304
00:12:37,200 --> 00:12:41,440
noticed that i use a star import which

305
00:12:39,360 --> 00:12:43,040
is generally frowned upon usually for

306
00:12:41,440 --> 00:12:44,639
good reasons

307
00:12:43,040 --> 00:12:47,040
i don't have enough time in this talk to

308
00:12:44,639 --> 00:12:47,920
explain why i nonetheless suggest to do

309
00:12:47,040 --> 00:12:49,120
so

310
00:12:47,920 --> 00:12:50,959
i invite you to consult the

311
00:12:49,120 --> 00:12:52,959
documentation for the detailed

312
00:12:50,959 --> 00:12:55,600
explanation and for suggested

313
00:12:52,959 --> 00:12:57,200
alternatives specific to friendly

314
00:12:55,600 --> 00:12:59,680
such as the one i will show in the next

315
00:12:57,200 --> 00:12:59,680
example

316
00:13:00,959 --> 00:13:05,279
here's an example of a normal traceback

317
00:13:03,360 --> 00:13:07,360
created when using the latest version of

318
00:13:05,279 --> 00:13:09,839
jupiter lab

319
00:13:07,360 --> 00:13:12,399
the details of the code are important

320
00:13:09,839 --> 00:13:14,079
however i want to draw your attention to

321
00:13:12,399 --> 00:13:16,240
the weird file name we see in a

322
00:13:14,079 --> 00:13:18,000
traceback

323
00:13:16,240 --> 00:13:21,240
how is a student supposed to make sense

324
00:13:18,000 --> 00:13:21,240
of this information

325
00:13:22,000 --> 00:13:25,839
let's use friendly

326
00:13:23,600 --> 00:13:27,839
this time by importing a single object

327
00:13:25,839 --> 00:13:31,440
from friendly.jupiter

328
00:13:27,839 --> 00:13:33,200
as an alternative to doing a star import

329
00:13:31,440 --> 00:13:34,800
while it is not needed i adjusted the

330
00:13:33,200 --> 00:13:36,639
friendliest theme to better match the

331
00:13:34,800 --> 00:13:38,240
jupiter light theme

332
00:13:36,639 --> 00:13:40,079
i also set the width of friendly's

333
00:13:38,240 --> 00:13:43,199
output so that it made for better screen

334
00:13:40,079 --> 00:13:43,199
captures for the stock

335
00:13:44,560 --> 00:13:49,760
i now run the same code as before

336
00:13:47,440 --> 00:13:52,320
notice how instead of some weird file

337
00:13:49,760 --> 00:13:55,279
names the traceback include references

338
00:13:52,320 --> 00:13:59,199
to code blocks which is the terminology

339
00:13:55,279 --> 00:13:59,199
used in jupiter's documentation

340
00:13:59,920 --> 00:14:03,600
if i want to know the cause of the

341
00:14:01,279 --> 00:14:06,000
problem according to friendly this time

342
00:14:03,600 --> 00:14:08,959
i cannot simply type the function y

343
00:14:06,000 --> 00:14:11,279
since i did not use a star import

344
00:14:08,959 --> 00:14:15,800
instead i can use the y method of the

345
00:14:11,279 --> 00:14:15,800
friendly object to achieve the same goal

346
00:14:16,399 --> 00:14:20,000
i've mentioned friendly and friendly

347
00:14:18,320 --> 00:14:21,360
trace pack

348
00:14:20,000 --> 00:14:23,760
initially the project was called

349
00:14:21,360 --> 00:14:26,160
friendly traceback and later renamed

350
00:14:23,760 --> 00:14:26,160
friendly

351
00:14:26,560 --> 00:14:30,959
at the end of june of this year the

352
00:14:28,399 --> 00:14:32,880
project was split into two projects move

353
00:14:30,959 --> 00:14:35,040
from my personal github site to a new

354
00:14:32,880 --> 00:14:36,800
friendly traceback organization i

355
00:14:35,040 --> 00:14:38,720
created

356
00:14:36,800 --> 00:14:41,199
friendly traceback does the analysis

357
00:14:38,720 --> 00:14:43,839
required to provide answers to the what

358
00:14:41,199 --> 00:14:46,160
where and why i mentioned before

359
00:14:43,839 --> 00:14:48,480
friendly takes care of providing the

360
00:14:46,160 --> 00:14:49,839
required formatting in a variety of

361
00:14:48,480 --> 00:14:51,920
environments

362
00:14:49,839 --> 00:14:55,120
interpreters using terminals editors

363
00:14:51,920 --> 00:14:58,240
such as idle and mu ipython console

364
00:14:55,120 --> 00:15:00,639
jupyter notebooks etc

365
00:14:58,240 --> 00:15:02,560
it builds upon friendly traceback

366
00:15:00,639 --> 00:15:05,199
friendly is also the program most users

367
00:15:02,560 --> 00:15:07,040
should install

368
00:15:05,199 --> 00:15:09,760
some third party projects

369
00:15:07,040 --> 00:15:11,199
such as the websites hack in science and

370
00:15:09,760 --> 00:15:12,959
future quarter

371
00:15:11,199 --> 00:15:17,199
use information provided by friendly

372
00:15:12,959 --> 00:15:17,199
traceback and do their own formatting

373
00:15:18,720 --> 00:15:22,720
the split was done on the request of one

374
00:15:20,399 --> 00:15:24,160
of those third-party users who found the

375
00:15:22,720 --> 00:15:25,920
additional dependencies to be

376
00:15:24,160 --> 00:15:28,240
detrimental to the performance of their

377
00:15:25,920 --> 00:15:28,240
site

378
00:15:29,199 --> 00:15:32,480
the approach i use in developing

379
00:15:30,480 --> 00:15:33,440
friendly is to add new examples one at a

380
00:15:32,480 --> 00:15:35,120
time

381
00:15:33,440 --> 00:15:37,360
trying to find a way to automatically

382
00:15:35,120 --> 00:15:39,199
identify the relevant information

383
00:15:37,360 --> 00:15:41,839
and providing explanations that can be

384
00:15:39,199 --> 00:15:41,839
easily understood

385
00:15:42,079 --> 00:15:45,519
currently there are more than 300

386
00:15:43,759 --> 00:15:47,839
individual cases for which friendly can

387
00:15:45,519 --> 00:15:50,800
provide an explanation

388
00:15:47,839 --> 00:15:55,839
those cases were either some of my own

389
00:15:50,800 --> 00:15:55,839
or very often suggested by other users

390
00:15:57,120 --> 00:16:01,040
while i keep adding new examples i

391
00:15:58,800 --> 00:16:03,279
rarely go back and revisit explanations

392
00:16:01,040 --> 00:16:05,040
given in older examples

393
00:16:03,279 --> 00:16:07,279
so i encourage all of you to look at the

394
00:16:05,040 --> 00:16:09,040
examples which are all available in the

395
00:16:07,279 --> 00:16:11,040
documentation site

396
00:16:09,040 --> 00:16:13,440
and invite you to either suggest new

397
00:16:11,040 --> 00:16:16,240
examples or better explanations for the

398
00:16:13,440 --> 00:16:16,240
existing ones

399
00:16:17,440 --> 00:16:22,480
when using an interactive interpreter it

400
00:16:19,839 --> 00:16:25,199
is unfortunately very easy to make typos

401
00:16:22,480 --> 00:16:27,360
causing spurious race-backs

402
00:16:25,199 --> 00:16:30,160
can an interactive python interpreter be

403
00:16:27,360 --> 00:16:30,160
user-friendly

404
00:16:30,800 --> 00:16:34,880
here i have some code that caused a

405
00:16:32,320 --> 00:16:34,880
traceback

406
00:16:35,279 --> 00:16:39,279
i want to find out why

407
00:16:37,600 --> 00:16:41,600
but i made a typo in writing the

408
00:16:39,279 --> 00:16:43,279
function name at the top

409
00:16:41,600 --> 00:16:46,480
this should be easy to correct given the

410
00:16:43,279 --> 00:16:49,040
hint provided by friendly

411
00:16:46,480 --> 00:16:50,880
i typed it correctly this time

412
00:16:49,040 --> 00:16:52,160
but it does not give me the explanation

413
00:16:50,880 --> 00:16:53,759
i needed

414
00:16:52,160 --> 00:16:56,560
it told me that i made a mistake in

415
00:16:53,759 --> 00:16:58,160
typing out y which i already knew and

416
00:16:56,560 --> 00:17:00,720
has nothing to do with the first

417
00:16:58,160 --> 00:17:00,720
traceback

418
00:17:01,199 --> 00:17:07,280
by using the history function i can view

419
00:17:03,519 --> 00:17:07,280
all the exceptions raised so far

420
00:17:07,839 --> 00:17:13,120
i can go back in history

421
00:17:09,600 --> 00:17:13,120
thus erasing the latest mistake

422
00:17:13,360 --> 00:17:16,799
i can now ask why

423
00:17:15,280 --> 00:17:20,880
and get an explanation for the first

424
00:17:16,799 --> 00:17:20,880
traceback as i wanted in the first place

425
00:17:22,880 --> 00:17:26,240
since recording a first version of this

426
00:17:24,400 --> 00:17:28,079
talk i've added a new way to use

427
00:17:26,240 --> 00:17:30,160
friendly within jupiter notebooks and

428
00:17:28,079 --> 00:17:32,400
similar environments

429
00:17:30,160 --> 00:17:34,960
instead of having to type what where and

430
00:17:32,400 --> 00:17:36,720
why the user can simply click on buttons

431
00:17:34,960 --> 00:17:37,919
to show or hide the corresponding

432
00:17:36,720 --> 00:17:40,480
explanation

433
00:17:37,919 --> 00:17:43,600
thus eliminating typo related problems

434
00:17:40,480 --> 00:17:43,600
in these environments

435
00:17:45,280 --> 00:17:49,520
in trying to improve friendly there's

436
00:17:47,039 --> 00:17:52,880
always a question at the back of my mind

437
00:17:49,520 --> 00:17:54,799
am i making the right choices

438
00:17:52,880 --> 00:17:57,200
so far i have had some feedback from

439
00:17:54,799 --> 00:17:58,799
individual users who are advanced python

440
00:17:57,200 --> 00:18:00,320
programmers

441
00:17:58,799 --> 00:18:04,400
but there has been no systematic

442
00:18:00,320 --> 00:18:04,400
evaluation done with actual beginners

443
00:18:05,280 --> 00:18:09,200
in the absence of such studies i

444
00:18:07,280 --> 00:18:11,600
recently read a paper on the topic after

445
00:18:09,200 --> 00:18:13,679
i saw a link to it on twitter

446
00:18:11,600 --> 00:18:15,280
on designing programming error messages

447
00:18:13,679 --> 00:18:19,919
for novices

448
00:18:15,280 --> 00:18:19,919
readability and its constituent factors

449
00:18:21,760 --> 00:18:25,360
while reading this paper i noted that

450
00:18:23,440 --> 00:18:27,679
the problems associated with bad error

451
00:18:25,360 --> 00:18:31,520
messages have been discussed for a few

452
00:18:27,679 --> 00:18:31,520
decades without being solved

453
00:18:32,080 --> 00:18:36,320
according to that paper there are two

454
00:18:34,240 --> 00:18:37,760
important factors for readability of

455
00:18:36,320 --> 00:18:42,320
error messages

456
00:18:37,760 --> 00:18:43,919
message length and use of jargon

457
00:18:42,320 --> 00:18:46,400
ideally one should

458
00:18:43,919 --> 00:18:49,679
remove jargon in acronyms

459
00:18:46,400 --> 00:18:52,080
write messages in complete sentences

460
00:18:49,679 --> 00:18:54,559
use simple vocabulary

461
00:18:52,080 --> 00:18:58,320
use an economy of words

462
00:18:54,559 --> 00:18:58,320
so complete but short sentences

463
00:18:59,039 --> 00:19:03,840
the paper focused on readability which

464
00:19:01,679 --> 00:19:07,120
the authors deem to be more fundamental

465
00:19:03,840 --> 00:19:07,120
than usability

466
00:19:07,360 --> 00:19:12,080
by usability they mean how useful is the

467
00:19:09,919 --> 00:19:15,200
information provided in finding a

468
00:19:12,080 --> 00:19:15,200
solution to the problem

469
00:19:15,840 --> 00:19:20,240
the others are groups of students

470
00:19:17,360 --> 00:19:22,000
evaluating error messages from java c

471
00:19:20,240 --> 00:19:24,160
and python

472
00:19:22,000 --> 00:19:26,559
i expected the results to be that error

473
00:19:24,160 --> 00:19:28,880
messages from java were two verbose

474
00:19:26,559 --> 00:19:32,880
those from c to cryptic

475
00:19:28,880 --> 00:19:32,880
and those from python just right

476
00:19:32,960 --> 00:19:36,799
i was wrong

477
00:19:34,400 --> 00:19:39,600
error messages from python were rated as

478
00:19:36,799 --> 00:19:41,280
the worst of the three

479
00:19:39,600 --> 00:19:44,080
let's consider why that might be the

480
00:19:41,280 --> 00:19:47,120
case using the first example shown at

481
00:19:44,080 --> 00:19:47,120
the beginning of this talk

482
00:19:48,080 --> 00:19:53,039
prior to version 3.8

483
00:19:50,400 --> 00:19:54,480
the error message provided by python was

484
00:19:53,039 --> 00:19:59,120
relatively short

485
00:19:54,480 --> 00:19:59,120
module turtle has no attribute forward

486
00:20:00,240 --> 00:20:03,760
this message did not contain any

487
00:20:02,080 --> 00:20:06,760
additional information about the actual

488
00:20:03,760 --> 00:20:06,760
cause

489
00:20:07,360 --> 00:20:11,200
then more information was added

490
00:20:09,679 --> 00:20:12,400
potentially useful to more advanced

491
00:20:11,200 --> 00:20:14,240
programmers

492
00:20:12,400 --> 00:20:17,440
but full of jargon that cannot be

493
00:20:14,240 --> 00:20:17,440
understood by beginners

494
00:20:18,720 --> 00:20:22,559
however when using friendly with any

495
00:20:21,120 --> 00:20:24,880
python version

496
00:20:22,559 --> 00:20:25,840
an additional hint is provided which

497
00:20:24,880 --> 00:20:29,600
should be

498
00:20:25,840 --> 00:20:29,600
more easily understood by beginners

499
00:20:30,400 --> 00:20:34,240
to be fair i must note that some recent

500
00:20:32,559 --> 00:20:36,240
improvement for python where some

501
00:20:34,240 --> 00:20:38,960
invalid syntax messages have been

502
00:20:36,240 --> 00:20:40,159
replaced by something more useful

503
00:20:38,960 --> 00:20:41,679
for example

504
00:20:40,159 --> 00:20:43,039
errors due to forgetting to close a

505
00:20:41,679 --> 00:20:45,280
parenthesis

506
00:20:43,039 --> 00:20:48,280
are now identified correctly in python

507
00:20:45,280 --> 00:20:48,280
3.10

508
00:20:48,960 --> 00:20:52,559
here's the same example run with

509
00:20:50,559 --> 00:20:54,320
friendly which gives the same correct

510
00:20:52,559 --> 00:20:55,679
explanation for all supported python

511
00:20:54,320 --> 00:20:58,480
versions

512
00:20:55,679 --> 00:21:01,720
even if you only get an invalid syntax

513
00:20:58,480 --> 00:21:01,720
error message

514
00:21:03,360 --> 00:21:06,240
python caters to all kinds of

515
00:21:05,039 --> 00:21:08,559
programmers

516
00:21:06,240 --> 00:21:10,480
from beginners just learning the basics

517
00:21:08,559 --> 00:21:12,559
to advanced programmers using python in

518
00:21:10,480 --> 00:21:14,640
their daily job

519
00:21:12,559 --> 00:21:16,480
when it comes to error messages

520
00:21:14,640 --> 00:21:18,880
it would seem that python seeks a middle

521
00:21:16,480 --> 00:21:21,360
ground and occasionally uses jargon not

522
00:21:18,880 --> 00:21:23,919
accessible to beginners

523
00:21:21,360 --> 00:21:26,880
by contrast friendly attempts to cater

524
00:21:23,919 --> 00:21:26,880
to such beginners

525
00:21:27,679 --> 00:21:31,760
let's have a look at a different example

526
00:21:29,840 --> 00:21:34,640
where python gives an error message that

527
00:21:31,760 --> 00:21:39,360
is rather obscure for beginners

528
00:21:34,640 --> 00:21:39,360
none type object is not subscriptable

529
00:21:40,480 --> 00:21:44,480
using friendly and asking why

530
00:21:42,640 --> 00:21:46,880
gives us an explanation

531
00:21:44,480 --> 00:21:49,039
but no actual clue as to what the object

532
00:21:46,880 --> 00:21:50,799
causing the problem is

533
00:21:49,039 --> 00:21:54,159
the note at the bottom explain what the

534
00:21:50,799 --> 00:21:54,159
non-type jargon means

535
00:21:56,080 --> 00:22:01,600
asking where gives us more information

536
00:21:59,840 --> 00:22:03,520
friendly traceback always attempts to

537
00:22:01,600 --> 00:22:04,960
isolate the part of the code that causes

538
00:22:03,520 --> 00:22:08,480
problems and

539
00:22:04,960 --> 00:22:11,039
identify the value of each object

540
00:22:08,480 --> 00:22:12,880
here we have a dict x containing other

541
00:22:11,039 --> 00:22:15,440
python digs

542
00:22:12,880 --> 00:22:18,480
at the very bottom we can see that x

543
00:22:15,440 --> 00:22:20,480
bracket 42 bracket 1 bracket 2 has the

544
00:22:18,480 --> 00:22:22,880
value of none which is the cause of the

545
00:22:20,480 --> 00:22:22,880
problem

546
00:22:24,240 --> 00:22:28,960
this almost concludes this introductory

547
00:22:26,080 --> 00:22:30,880
talk to friendly and friendly traceback

548
00:22:28,960 --> 00:22:32,000
to find out more you could use a search

549
00:22:30,880 --> 00:22:33,520
engine to find out friendly's

550
00:22:32,000 --> 00:22:35,679
documentation

551
00:22:33,520 --> 00:22:38,960
but there's a better way

552
00:22:35,679 --> 00:22:38,960
first install friendly

553
00:22:39,600 --> 00:22:42,880
start the friendly console

554
00:22:43,280 --> 00:22:47,760
typing in friendly provides some basic

555
00:22:45,440 --> 00:22:47,760
help

556
00:22:50,159 --> 00:22:54,640
to find out even more type www as a

557
00:22:53,360 --> 00:22:56,080
function

558
00:22:54,640 --> 00:22:58,559
assuming you write this function call

559
00:22:56,080 --> 00:23:00,320
without making any typo no exception

560
00:22:58,559 --> 00:23:02,480
will have been raised and instead of

561
00:23:00,320 --> 00:23:03,600
doing an internet search that function

562
00:23:02,480 --> 00:23:06,400
will take you to friendly's

563
00:23:03,600 --> 00:23:08,400
documentation

564
00:23:06,400 --> 00:23:10,480
as in addition to unit tests i record

565
00:23:08,400 --> 00:23:11,919
almost all the example tracebacks for

566
00:23:10,480 --> 00:23:13,760
different python versions in a

567
00:23:11,919 --> 00:23:15,280
documentation

568
00:23:13,760 --> 00:23:17,760
this occasionally helps to spot

569
00:23:15,280 --> 00:23:19,520
regressions it is also useful to review

570
00:23:17,760 --> 00:23:21,600
the explanation given

571
00:23:19,520 --> 00:23:24,960
i encourage you to do so and make

572
00:23:21,600 --> 00:23:26,880
suggestions to improve them

573
00:23:24,960 --> 00:23:30,159
you might also come across some notices

574
00:23:26,880 --> 00:23:32,000
about design choices

575
00:23:30,159 --> 00:23:33,919
these are two record choices i make for

576
00:23:32,000 --> 00:23:36,320
friendly traceback inviting you to

577
00:23:33,919 --> 00:23:38,240
comment about their suitability perhaps

578
00:23:36,320 --> 00:23:40,799
there are better ways to present the

579
00:23:38,240 --> 00:23:40,799
information

580
00:23:41,679 --> 00:23:46,000
of course using pip to install packages

581
00:23:43,840 --> 00:23:48,480
such as friendly is often something that

582
00:23:46,000 --> 00:23:51,840
beginners find difficult ideally there

583
00:23:48,480 --> 00:23:51,840
should be a better way

584
00:23:52,559 --> 00:23:56,559
perhaps friendly will be installed by

585
00:23:54,159 --> 00:23:59,039
default in a future version of new as

586
00:23:56,559 --> 00:24:03,720
has been recorded on mu's github repo by

587
00:23:59,039 --> 00:24:03,720
nicholas tallarvy muse creator

588
00:24:04,320 --> 00:24:08,320
one of the main ipython and jupiter

589
00:24:06,240 --> 00:24:10,320
developer has expressed support for

590
00:24:08,320 --> 00:24:12,799
enabling friendly by default if it is

591
00:24:10,320 --> 00:24:12,799
installed

592
00:24:14,640 --> 00:24:18,640
in conclusion

593
00:24:16,000 --> 00:24:21,440
i invite you to try out friendly

594
00:24:18,640 --> 00:24:23,360
read the documentation and give feedback

595
00:24:21,440 --> 00:24:25,440
so that it can hopefully make friendly

596
00:24:23,360 --> 00:24:27,919
even more useful for you and for your

597
00:24:25,440 --> 00:24:27,919
students

598
00:24:28,880 --> 00:24:32,559
and of course i will come any question

599
00:24:31,279 --> 00:24:34,960
you might have

600
00:24:32,559 --> 00:24:34,960
thank you

601
00:24:37,760 --> 00:24:43,120
um immediately after lunch we have

602
00:24:40,000 --> 00:24:45,200
a fantastic talk by steven and allison

603
00:24:43,120 --> 00:24:48,159
on at the turn of the tide teaching

604
00:24:45,200 --> 00:24:51,200
python in real time at scale so make

605
00:24:48,159 --> 00:24:54,720
sure you all join us for that one in the

606
00:24:51,200 --> 00:24:56,960
meantime andre thank you once again uh

607
00:24:54,720 --> 00:24:58,640
have a fantastic lunch everyone and

608
00:24:56,960 --> 00:25:01,840
we'll see you soon

609
00:24:58,640 --> 00:25:01,840
cheers thank you