language-creators.srt

1
00:00:00,530 --> 00:00:01,729
You guys ready some language

2
00:00:01,730 --> 00:00:02,730
creator.

3
00:00:07,700 --> 00:00:09,669
I have further to introduce

4
00:00:09,940 --> 00:00:10,940
Carol Willey.

5
00:00:11,420 --> 00:00:13,089
She is a python during counseling

6
00:00:13,090 --> 00:00:14,589
Project Jupiter Steering Council

7
00:00:14,590 --> 00:00:15,590
member.

8
00:00:15,610 --> 00:00:17,129
She blends the strings of pythonic.

9
00:00:17,140 --> 00:00:18,759
Jupiter notebook's to improve access

10
00:00:18,760 --> 00:00:19,929
to learning and education

11
00:00:20,730 --> 00:00:22,179
due to the worldwide foundation of

12
00:00:22,180 --> 00:00:22,989
the Pyfrom language.

13
00:00:22,990 --> 00:00:24,339
The hard work of the Jupiter team

14
00:00:24,410 --> 00:00:25,410
users worldwide.

15
00:00:25,570 --> 00:00:26,679
Carroll was awarded the Twenty

16
00:00:26,680 --> 00:00:28,659
Seventeen ACM Software System

17
00:00:28,660 --> 00:00:30,130
Award, recognizing the lasting

18
00:00:30,850 --> 00:00:32,228
influence of Jupiter, a broad

19
00:00:32,229 --> 00:00:34,149
collaboration to develop its open

20
00:00:34,150 --> 00:00:35,522
source tools for interactive

21
00:00:35,560 --> 00:00:37,226
computing with a language agnostic

22
00:00:37,840 --> 00:00:38,619
design.

23
00:00:38,620 --> 00:00:39,620
Thank you.

24
00:00:47,800 --> 00:00:49,630
Welcome, everybody, for being here.

25
00:00:49,720 --> 00:00:51,639
It is my absolute

26
00:00:51,850 --> 00:00:54,009
pleasure to share this evening.

27
00:00:54,040 --> 00:00:55,040
You and some

28
00:00:55,870 --> 00:00:57,249
of the people that actually

29
00:00:57,250 --> 00:00:59,289
influenced my career in software

30
00:00:59,290 --> 00:01:00,290
development.

31
00:01:01,180 --> 00:01:03,010
So let's get through this.

32
00:01:03,220 --> 00:01:04,641
And for those of you that are

33
00:01:04,810 --> 00:01:06,099
watching us on the lifespan.

34
00:01:06,520 --> 00:01:08,590
Hello and welcome and

35
00:01:08,770 --> 00:01:10,419
hope you have a wonderful evening

36
00:01:10,420 --> 00:01:11,420
with us.

37
00:01:12,160 --> 00:01:13,659
We haven't really distinguished

38
00:01:13,660 --> 00:01:15,969
panel of language creators.

39
00:01:16,060 --> 00:01:17,529
And I'd like to invite them to the

40
00:01:17,530 --> 00:01:19,359
stage as we read

41
00:01:19,360 --> 00:01:20,409
their bios. I'm going to keep the

42
00:01:20,410 --> 00:01:22,689
bios really short because

43
00:01:22,810 --> 00:01:24,519
I think most of them need no

44
00:01:24,520 --> 00:01:25,969
introduction whatsoever.

45
00:01:26,710 --> 00:01:27,886
And that'll save us some

46
00:01:28,540 --> 00:01:29,540
time more.

47
00:01:29,820 --> 00:01:31,094
Do you see language design

48
00:01:31,840 --> 00:01:32,840
stuff? OK.

49
00:01:33,190 --> 00:01:35,159
So first we have been

50
00:01:35,190 --> 00:01:36,190
Rassam

51
00:01:37,030 --> 00:01:39,009
to develop the Python language that

52
00:01:39,010 --> 00:01:40,235
touches pretty much every

53
00:01:40,870 --> 00:01:42,389
corner of society today in some

54
00:01:43,060 --> 00:01:43,949
way.

55
00:01:43,950 --> 00:01:45,819
And it's.

56
00:01:51,540 --> 00:01:53,510
He's the creator of Java,

57
00:01:53,570 --> 00:01:54,991
which is UK-EU, used across a

58
00:01:55,400 --> 00:01:57,349
broad variety of devices and

59
00:01:57,350 --> 00:01:58,700
deployments at scale.

60
00:01:59,590 --> 00:02:00,239
Well done.

61
00:02:00,240 --> 00:02:01,540
And please get a.

62
00:02:07,740 --> 00:02:09,719
Childcare has architected

63
00:02:09,720 --> 00:02:11,435
a number of languages over the past

64
00:02:11,520 --> 00:02:13,650
several decades, including

65
00:02:13,710 --> 00:02:16,079
Turbo Pascal, Delphi,

66
00:02:16,170 --> 00:02:17,885
C-sharp, and so you might know this

67
00:02:18,120 --> 00:02:19,469
little language typescript.

68
00:02:26,580 --> 00:02:27,580
Definitely not these.

69
00:02:27,700 --> 00:02:29,790
Larry Wall to combined

70
00:02:29,800 --> 00:02:31,025
his unique perspective on

71
00:02:31,750 --> 00:02:33,699
text and computing with

72
00:02:33,700 --> 00:02:35,589
his strong background in linguistics

73
00:02:35,620 --> 00:02:37,449
to Craig Pearl and its

74
00:02:37,450 --> 00:02:39,469
sister sibling Pearl 6.

75
00:02:39,820 --> 00:02:40,820
Welcome.

76
00:02:51,420 --> 00:02:52,420
So just a couple of

77
00:02:53,280 --> 00:02:55,139
housekeeping things about this

78
00:02:55,140 --> 00:02:56,140
evening's format.

79
00:02:57,000 --> 00:02:58,740
You're in for an engaging

80
00:02:58,950 --> 00:03:00,539
and hopefully very lively

81
00:03:00,540 --> 00:03:02,279
discussion. More likely than we were

82
00:03:02,280 --> 00:03:03,280
at dinner.

83
00:03:03,300 --> 00:03:04,440
I think that's possible, right?

84
00:03:04,910 --> 00:03:05,910
OK.

85
00:03:05,940 --> 00:03:06,769
The format.

86
00:03:06,770 --> 00:03:07,770
More beer.

87
00:03:08,870 --> 00:03:10,379
OK. We can do that.

88
00:03:11,280 --> 00:03:12,280
Keep it coming.

89
00:03:13,830 --> 00:03:15,055
OK. So the format will be

90
00:03:15,750 --> 00:03:17,729
two sessions with a fifteen minute

91
00:03:18,300 --> 00:03:19,300
intermission.

92
00:03:19,740 --> 00:03:21,539
What I'll do is I will start by

93
00:03:21,540 --> 00:03:23,550
asking one of the panelists a

94
00:03:23,580 --> 00:03:25,429
question, and hopefully

95
00:03:25,430 --> 00:03:27,089
a response will be like one to two

96
00:03:27,090 --> 00:03:29,009
minutes and then kind of throw

97
00:03:29,010 --> 00:03:30,480
it open for other panelists to

98
00:03:30,660 --> 00:03:31,885
respond and hopefully get

99
00:03:32,610 --> 00:03:34,169
a really good discussion going.

100
00:03:34,500 --> 00:03:35,921
And the more you discuss, the

101
00:03:36,360 --> 00:03:37,928
less I have to ask questions and

102
00:03:38,220 --> 00:03:40,349
the more in-depth we can go

103
00:03:40,410 --> 00:03:41,410
on the really cool stuff

104
00:03:43,620 --> 00:03:45,710
after the 15 minute intermission.

105
00:03:46,290 --> 00:03:47,290
We're going to take

106
00:03:48,120 --> 00:03:50,459
questions from the audience over

107
00:03:52,140 --> 00:03:53,789
a winner

108
00:03:54,210 --> 00:03:54,839
Twitter.

109
00:03:54,840 --> 00:03:56,789
OK. So tweet out your questions.

110
00:03:57,090 --> 00:03:59,390
Hashtag is happy bbf

111
00:03:59,400 --> 00:04:00,400
out.

112
00:04:00,570 --> 00:04:01,746
And if you can get those

113
00:04:02,580 --> 00:04:04,349
in by 8:00, we'll kind of collect

114
00:04:04,350 --> 00:04:05,399
them and

115
00:04:06,870 --> 00:04:08,339
have some really great audience

116
00:04:08,340 --> 00:04:09,779
questions about language creation

117
00:04:09,780 --> 00:04:10,780
and language design.

118
00:04:11,280 --> 00:04:13,439
So I'm going to

119
00:04:13,500 --> 00:04:15,629
invite you to sit back,

120
00:04:16,260 --> 00:04:18,299
enjoy a cup of strong Java

121
00:04:19,070 --> 00:04:20,910
Sea, start offering responses and

122
00:04:20,940 --> 00:04:23,100
insights, learned pearls of wisdom

123
00:04:23,670 --> 00:04:25,389
and embrace. The Python program

124
00:04:25,400 --> 00:04:26,400
makes it.

125
00:04:33,190 --> 00:04:34,856
We'll start the celebration with a

126
00:04:34,900 --> 00:04:36,566
question for Kito about principles

127
00:04:36,990 --> 00:04:38,019
of language design.

128
00:04:39,200 --> 00:04:40,621
You mentioned before, perhaps

129
00:04:41,170 --> 00:04:43,089
somewhat jokingly, the Harry

130
00:04:43,090 --> 00:04:44,829
Potter theory of language design.

131
00:04:45,490 --> 00:04:47,205
What is this theory and what do you

132
00:04:47,380 --> 00:04:48,729
see as the key principles of

133
00:04:48,730 --> 00:04:49,730
language design?

134
00:04:57,520 --> 00:04:58,520
I think you're good to me.

135
00:04:58,610 --> 00:04:59,610
Am I good?

136
00:04:59,620 --> 00:05:00,620
OK.

137
00:05:01,800 --> 00:05:03,515
The Harry Potter theory of language

138
00:05:03,720 --> 00:05:05,149
design was close to.

139
00:05:08,400 --> 00:05:10,709
I do know that I cannot believe

140
00:05:10,710 --> 00:05:11,710
that every detail

141
00:05:12,740 --> 00:05:14,455
that J.K. Rowling puts on the first

142
00:05:14,790 --> 00:05:16,559
Harry Potter book was

143
00:05:16,620 --> 00:05:18,480
intended as.

144
00:05:20,530 --> 00:05:22,539
Some important

145
00:05:22,860 --> 00:05:24,100
Wolf point game

146
00:05:25,020 --> 00:05:26,550
part 6 or 7.

147
00:05:29,350 --> 00:05:31,239
And even if J.K.

148
00:05:31,240 --> 00:05:33,129
Rowling was such a genius that

149
00:05:33,130 --> 00:05:34,999
he hadn't actually owned land, that

150
00:05:35,180 --> 00:05:36,879
and that was

151
00:05:37,410 --> 00:05:39,439
who I think we found out the

152
00:05:39,670 --> 00:05:40,670
FUTHI was.

153
00:05:44,460 --> 00:05:45,460
But who won?

154
00:05:45,610 --> 00:05:46,982
It was just like a temporary

155
00:05:47,470 --> 00:05:49,149
diversion during the train ride.

156
00:05:49,200 --> 00:05:50,679
And otherwise, you didn't hear much

157
00:05:50,680 --> 00:05:52,839
about the rap with fingerprints

158
00:05:54,330 --> 00:05:55,099
in languages.

159
00:05:55,100 --> 00:05:56,989
I often.

160
00:05:57,250 --> 00:05:59,079
That's exactly how things go.

161
00:05:59,110 --> 00:06:00,110
You

162
00:06:01,260 --> 00:06:03,129
you choose some detail

163
00:06:03,760 --> 00:06:05,439
because you have to commit.

164
00:06:05,510 --> 00:06:06,510
You sort of

165
00:06:07,720 --> 00:06:09,509
you have to pick keywords.

166
00:06:09,970 --> 00:06:11,910
You have to choose a style of

167
00:06:12,640 --> 00:06:14,559
coding like you maybe

168
00:06:14,560 --> 00:06:16,749
you choose indentation

169
00:06:16,780 --> 00:06:18,340
instead of curly braces

170
00:06:18,720 --> 00:06:19,720
or maybe you don't.

171
00:06:22,220 --> 00:06:23,339
You're a special case,

172
00:06:25,330 --> 00:06:26,330
but

173
00:06:27,570 --> 00:06:28,620
whatever you do

174
00:06:29,670 --> 00:06:31,110
in some way, you're stuck with them

175
00:06:32,550 --> 00:06:34,499
and you find new

176
00:06:34,500 --> 00:06:36,070
uses of that.

177
00:06:36,950 --> 00:06:38,770
That detail that you picked

178
00:06:38,790 --> 00:06:40,809
before you knew how important

179
00:06:40,810 --> 00:06:42,829
that would be and

180
00:06:43,100 --> 00:06:45,509
the sort of the craft of designing

181
00:06:45,560 --> 00:06:47,620
the language is on one hand pick

182
00:06:47,640 --> 00:06:49,379
your initial set of

183
00:06:50,250 --> 00:06:51,690
choices so that

184
00:06:52,350 --> 00:06:54,269
there are a lot

185
00:06:54,330 --> 00:06:55,555
of possible continuations

186
00:06:56,520 --> 00:06:57,520
of the story.

187
00:06:57,990 --> 00:06:59,490
The other half of the

188
00:07:00,600 --> 00:07:02,399
of the art of language design

189
00:07:02,430 --> 00:07:04,229
is going back to your story

190
00:07:04,260 --> 00:07:05,260
and

191
00:07:06,420 --> 00:07:08,579
inventing creative ways

192
00:07:08,670 --> 00:07:09,670
of continuing

193
00:07:10,590 --> 00:07:12,600
it in a way that you had not.

194
00:07:15,420 --> 00:07:16,743
So I think each of you have

195
00:07:16,800 --> 00:07:18,959
languages that span multiple

196
00:07:18,960 --> 00:07:19,709
decades.

197
00:07:19,710 --> 00:07:21,479
So what are some of the principles

198
00:07:21,480 --> 00:07:22,480
that made,

199
00:07:24,090 --> 00:07:25,949
you know, so so

200
00:07:26,670 --> 00:07:28,649
Javal would sort of on in the.

201
00:07:30,000 --> 00:07:31,279
It didn't come out of like a

202
00:07:31,290 --> 00:07:32,613
personal passion project or

203
00:07:33,060 --> 00:07:34,079
something. It was.

204
00:07:36,870 --> 00:07:39,029
From trying to build a prototype,

205
00:07:39,810 --> 00:07:41,279
you know, I mean, we were trying to

206
00:07:41,280 --> 00:07:42,829
understand it because the.

207
00:07:48,840 --> 00:07:50,359
A lot of time talking to people

208
00:07:50,670 --> 00:07:52,139
who build software or embedded

209
00:07:52,140 --> 00:07:53,140
systems.

210
00:07:59,800 --> 00:08:01,679
And you know, what was

211
00:08:01,680 --> 00:08:03,599
it in the whole process?

212
00:08:10,920 --> 00:08:12,870
How did their systems

213
00:08:12,890 --> 00:08:14,179
fit into the universe?

214
00:08:14,390 --> 00:08:15,860
It wasn't because I was trying

215
00:08:16,220 --> 00:08:17,984
to worry when we were worrying about

216
00:08:18,080 --> 00:08:19,560
things outside of data centers.

217
00:08:22,920 --> 00:08:24,537
This was a project that had about

218
00:08:24,750 --> 00:08:26,980
a dozen people on it and

219
00:08:26,990 --> 00:08:27,990
my little.

220
00:08:28,450 --> 00:08:30,139
Yeah, but it was my

221
00:08:30,300 --> 00:08:32,010
my little piece was

222
00:08:32,500 --> 00:08:33,809
what it was was sort of like.

223
00:08:36,210 --> 00:08:37,450
You know, we we we we sort of

224
00:08:37,490 --> 00:08:39,949
figured out that part of the problem

225
00:08:40,580 --> 00:08:41,580
was that.

226
00:08:43,320 --> 00:08:45,739
See has issues.

227
00:08:48,260 --> 00:08:49,260
And and

228
00:08:50,700 --> 00:08:52,289
so I started out so.

229
00:08:52,340 --> 00:08:54,179
So out of this

230
00:08:54,180 --> 00:08:56,490
like large pie of a project,

231
00:08:56,550 --> 00:08:57,720
my slice

232
00:08:58,380 --> 00:08:59,380
was.

233
00:09:01,420 --> 00:09:03,460
To make things a little easier

234
00:09:03,470 --> 00:09:04,969
from a programing language point of

235
00:09:04,970 --> 00:09:06,440
view and fix the

236
00:09:06,980 --> 00:09:08,450
pain points that came from the

237
00:09:08,720 --> 00:09:09,919
programing language part.

238
00:09:10,010 --> 00:09:11,840
And it started out as

239
00:09:13,460 --> 00:09:14,809
kind of do a better see.

240
00:09:15,890 --> 00:09:17,350
And then it and it

241
00:09:18,320 --> 00:09:19,219
got out of control.

242
00:09:19,220 --> 00:09:20,389
The rest of the project

243
00:09:22,130 --> 00:09:23,509
really ended up just providing

244
00:09:23,510 --> 00:09:24,510
context.

245
00:09:25,050 --> 00:09:26,779
You know, the only thing out of that

246
00:09:26,780 --> 00:09:28,669
project that survived was

247
00:09:28,670 --> 00:09:29,670
was Java.

248
00:09:31,670 --> 00:09:34,009
But but but the fact that it was

249
00:09:35,210 --> 00:09:37,309
sort of directed at a set of pain

250
00:09:37,310 --> 00:09:38,310
points.

251
00:09:38,780 --> 00:09:40,201
But it happened to be it, you

252
00:09:40,660 --> 00:09:41,660
know, about

253
00:09:42,740 --> 00:09:44,089
people who were living outside of

254
00:09:44,090 --> 00:09:45,709
data centers and people who were.

255
00:09:49,050 --> 00:09:50,422
Getting shredded by problems

256
00:09:50,880 --> 00:09:53,640
with networking and security

257
00:09:53,730 --> 00:09:54,899
and reliability.

258
00:09:56,850 --> 00:09:58,271
They had to build things that

259
00:09:58,710 --> 00:10:00,510
ran in hostile environments

260
00:10:01,080 --> 00:10:02,549
like in the homes,

261
00:10:05,010 --> 00:10:06,010
which in any home

262
00:10:07,020 --> 00:10:08,550
with a child in it is a

263
00:10:09,090 --> 00:10:11,070
it is a hostile environment.

264
00:10:12,560 --> 00:10:14,639
Speaking of hostile environments,

265
00:10:15,820 --> 00:10:16,820
feral.

266
00:10:20,590 --> 00:10:21,962
What are some of the guiding

267
00:10:22,420 --> 00:10:23,860
principles for designing for a

268
00:10:24,260 --> 00:10:25,260
business?

269
00:10:28,150 --> 00:10:29,150
I love you, Larry.

270
00:10:40,880 --> 00:10:42,739
More and more coming

271
00:10:42,740 --> 00:10:43,740
out of this.

272
00:10:43,760 --> 00:10:44,899
A computer scientist.

273
00:10:45,360 --> 00:10:46,389
So I sort of actively

274
00:10:47,600 --> 00:10:48,972
ignored some of the computer

275
00:10:49,700 --> 00:10:51,289
science literature of the time and

276
00:10:51,290 --> 00:10:53,220
said, well, what can we just talk

277
00:10:53,240 --> 00:10:54,739
together? One part of the brain will

278
00:10:54,740 --> 00:10:56,419
work more like a natural language

279
00:10:58,010 --> 00:10:59,190
instead of. But

280
00:10:59,840 --> 00:11:00,840
instead of

281
00:11:02,180 --> 00:11:03,180
putting on a

282
00:11:05,210 --> 00:11:07,100
campus university campus and

283
00:11:07,190 --> 00:11:08,629
deciding world, the walkways would

284
00:11:08,630 --> 00:11:09,589
go, we're just going to see where

285
00:11:09,590 --> 00:11:10,590
people want to walk.

286
00:11:10,910 --> 00:11:12,139
And they put shortcuts in all those

287
00:11:12,140 --> 00:11:13,610
places and know that more as a

288
00:11:13,940 --> 00:11:15,739
network, not as a terribly

289
00:11:15,800 --> 00:11:18,049
orthogonal or computer

290
00:11:18,050 --> 00:11:19,580
science or mathematical thing.

291
00:11:20,190 --> 00:11:21,649
And then that turned out to be

292
00:11:23,420 --> 00:11:24,829
in the right place at the right time

293
00:11:24,840 --> 00:11:25,840
for bootstrapping a

294
00:11:26,660 --> 00:11:27,660
lot of the web

295
00:11:29,840 --> 00:11:30,840
and.

296
00:11:31,720 --> 00:11:33,249
It also got used a lot for system

297
00:11:33,250 --> 00:11:35,290
administration and so

298
00:11:37,270 --> 00:11:38,740
various principles that relate

299
00:11:39,190 --> 00:11:40,809
to trying to

300
00:11:41,590 --> 00:11:43,479
provide API eyes to everything.

301
00:11:45,200 --> 00:11:46,629
I'm trying to be both a good text

302
00:11:46,630 --> 00:11:47,630
processing language

303
00:11:48,580 --> 00:11:50,559
linguistically, but also

304
00:11:50,560 --> 00:11:51,649
a quick Google language.

305
00:11:51,920 --> 00:11:52,920
Here's why it was.

306
00:11:53,630 --> 00:11:54,860
Kind of very useful,

307
00:11:55,450 --> 00:11:57,709
was hasty and there was text

308
00:11:57,980 --> 00:12:00,080
and, you

309
00:12:00,300 --> 00:12:02,129
know, database seconds that's that

310
00:12:02,220 --> 00:12:03,109
needs.

311
00:12:03,110 --> 00:12:04,219
And so it was in the right place at

312
00:12:04,220 --> 00:12:05,220
the right time.

313
00:12:06,350 --> 00:12:07,790
We were very fortunate to

314
00:12:08,330 --> 00:12:09,770
have that time because

315
00:12:10,390 --> 00:12:11,615
in the 90s it became very

316
00:12:12,290 --> 00:12:14,509
stable, a lot of a lot of widespread

317
00:12:14,520 --> 00:12:16,250
use. Lot of people were

318
00:12:16,600 --> 00:12:18,090
you know, the language itself was

319
00:12:18,320 --> 00:12:19,447
stabilizing in the form

320
00:12:20,330 --> 00:12:21,330
that it was.

321
00:12:21,410 --> 00:12:22,759
But there were a lot of issues

322
00:12:23,240 --> 00:12:24,240
and salat.

323
00:12:24,620 --> 00:12:26,539
And so in the year 2000,

324
00:12:26,540 --> 00:12:27,618
we we took a step back

325
00:12:28,580 --> 00:12:29,580
and and

326
00:12:30,410 --> 00:12:32,220
and basically said, we're gonna

327
00:12:32,240 --> 00:12:33,619
break everything he's breaking.

328
00:12:33,650 --> 00:12:34,924
So we kind of did the same

329
00:12:35,750 --> 00:12:36,750
thing.

330
00:12:37,450 --> 00:12:39,150
You know, it's done the pipeline to

331
00:12:39,170 --> 00:12:40,934
the step, except instead of breaking

332
00:12:41,330 --> 00:12:43,159
a few things, we decided to

333
00:12:43,160 --> 00:12:44,369
break everything here.

334
00:12:45,200 --> 00:12:46,866
So we came up with a whole raft of

335
00:12:46,880 --> 00:12:48,799
design principles and 50 years since

336
00:12:48,800 --> 00:12:49,729
then.

337
00:12:49,730 --> 00:12:51,169
By the way, forensics did come up

338
00:12:51,170 --> 00:12:52,159
two years ago.

339
00:12:52,160 --> 00:12:53,160
It's getting faster.

340
00:12:54,640 --> 00:12:56,279
And it's

341
00:12:56,640 --> 00:12:58,269
suddenly at the beginning we said we

342
00:12:58,270 --> 00:12:59,149
were going to do something

343
00:12:59,150 --> 00:13:00,559
impossible and and

344
00:13:01,020 --> 00:13:02,629
fail at it could be a very big

345
00:13:02,630 --> 00:13:04,570
failure. And so far, this proved

346
00:13:04,600 --> 00:13:06,669
me to be that. But in the course

347
00:13:06,670 --> 00:13:08,059
of the fail.

348
00:13:08,340 --> 00:13:09,879
Yeah. In the course of

349
00:13:10,220 --> 00:13:12,049
that, we came up some really

350
00:13:12,050 --> 00:13:13,519
interesting list of about 60

351
00:13:13,520 --> 00:13:15,019
different design principles.

352
00:13:19,160 --> 00:13:21,120
When read, even though I think many

353
00:13:21,310 --> 00:13:22,970
of them have already, you know,

354
00:13:23,790 --> 00:13:25,407
you know, kill two birds with one

355
00:13:25,450 --> 00:13:27,190
stone, pick the right people, you

356
00:13:27,220 --> 00:13:28,220
know.

357
00:13:29,640 --> 00:13:30,940
It's possible like that.

358
00:13:31,070 --> 00:13:32,649
But we came up with some cute ones,

359
00:13:32,650 --> 00:13:34,559
like, you think that's cute today?

360
00:13:38,650 --> 00:13:40,299
Oh, oh,

361
00:13:41,030 --> 00:13:42,030
know.

362
00:13:43,210 --> 00:13:44,960
Conserve your brackets because

363
00:13:45,130 --> 00:13:46,649
SFE, even Unicode does not have

364
00:13:47,000 --> 00:13:48,000
enough brackets

365
00:13:51,260 --> 00:13:52,534
that don't reinvent object

366
00:13:53,240 --> 00:13:54,349
orientation queerly,

367
00:13:55,190 --> 00:13:57,199
which arguably provided

368
00:13:59,480 --> 00:14:00,901
no substitute strong features

369
00:14:01,640 --> 00:14:03,169
for weak buttons, late binding

370
00:14:03,170 --> 00:14:04,889
sometimes oposite programs be.

371
00:14:09,060 --> 00:14:10,830
But some of the major ones are,

372
00:14:11,230 --> 00:14:13,159
you know, a lot of the stuff is

373
00:14:13,250 --> 00:14:14,909
just perfect to screw over time.

374
00:14:15,360 --> 00:14:16,980
And, you know,

375
00:14:18,330 --> 00:14:19,649
there were a lot of weird magical

376
00:14:19,650 --> 00:14:21,509
variables I'm going

377
00:14:21,510 --> 00:14:22,510
to.

378
00:14:23,130 --> 00:14:24,990
And so.

379
00:14:27,280 --> 00:14:28,879
A great deal of the redesign was to

380
00:14:28,890 --> 00:14:30,360
say, OK, what is the right peg

381
00:14:30,820 --> 00:14:31,840
to hang everything on?

382
00:14:32,290 --> 00:14:33,370
Is it off to Korea?

383
00:14:33,520 --> 00:14:34,892
Is it something in the works

384
00:14:35,440 --> 00:14:37,450
for school or in a larger school?

385
00:14:37,570 --> 00:14:38,570
Or what

386
00:14:39,570 --> 00:14:41,460
is the right to have each

387
00:14:41,640 --> 00:14:42,970
piece of information on?

388
00:14:43,330 --> 00:14:44,709
And if we don't have that thing, how

389
00:14:44,710 --> 00:14:45,710
do we create it?

390
00:14:46,810 --> 00:14:47,810
Great question.

391
00:14:48,040 --> 00:14:49,363
And since it's such a great

392
00:14:50,080 --> 00:14:51,969
question, Anderson, the spot

393
00:14:52,410 --> 00:14:54,490
and Terrence Parr

394
00:14:54,670 --> 00:14:56,589
had a quiz about while

395
00:14:56,590 --> 00:14:58,540
programmers value simplicity,

396
00:14:58,740 --> 00:15:00,640
they more often value

397
00:15:01,030 --> 00:15:02,590
powerful functionality.

398
00:15:02,680 --> 00:15:04,330
An amazing one liners

399
00:15:04,750 --> 00:15:06,580
incurring the cost of complexity.

400
00:15:07,540 --> 00:15:09,200
So for express.

401
00:15:09,710 --> 00:15:10,499
Yes.

402
00:15:10,500 --> 00:15:12,460
So what are your thoughts about

403
00:15:12,970 --> 00:15:15,250
simplicity versus complexity

404
00:15:15,280 --> 00:15:16,539
and some of the principles that

405
00:15:16,540 --> 00:15:18,360
might guide you, as you know?

406
00:15:18,610 --> 00:15:19,610
Really?

407
00:15:22,850 --> 00:15:23,850
Suppose I.

408
00:15:26,930 --> 00:15:27,930
I was always late

409
00:15:29,280 --> 00:15:31,090
in doing the language, saying

410
00:15:31,100 --> 00:15:31,879
that I've done.

411
00:15:31,880 --> 00:15:33,320
Always tried to

412
00:15:33,770 --> 00:15:35,179
make it so that there was only one

413
00:15:35,180 --> 00:15:36,180
way.

414
00:15:37,010 --> 00:15:38,889
So a lot of languages have

415
00:15:38,920 --> 00:15:40,159
four different ways of doing

416
00:15:40,160 --> 00:15:41,509
something and you pick the wrong

417
00:15:41,510 --> 00:15:43,640
one. And then only

418
00:15:45,080 --> 00:15:46,729
later you realize you went down the

419
00:15:46,730 --> 00:15:48,396
wrong branch and now you've got it

420
00:15:48,580 --> 00:15:49,319
back up.

421
00:15:49,320 --> 00:15:50,549
And so.

422
00:15:50,880 --> 00:15:52,820
So it's not

423
00:15:52,970 --> 00:15:53,970
it's not always easy.

424
00:15:54,530 --> 00:15:55,706
And I think there's like

425
00:15:56,360 --> 00:15:58,279
often we're guilty of creating

426
00:15:58,280 --> 00:15:59,840
what I call

427
00:16:00,140 --> 00:16:01,279
this thing I call simplex.

428
00:16:02,470 --> 00:16:03,880
Take something from blacks,

429
00:16:04,640 --> 00:16:06,012
you grab a single wrapper on

430
00:16:06,570 --> 00:16:08,040
top of it. That's make the

431
00:16:08,050 --> 00:16:09,450
complexity go away, but you're

432
00:16:09,510 --> 00:16:11,119
really creating is simplicity.

433
00:16:11,220 --> 00:16:13,049
You know, it's just like a

434
00:16:13,050 --> 00:16:14,700
bad abstraction on top of

435
00:16:15,510 --> 00:16:17,100
another bad abstraction that is

436
00:16:17,580 --> 00:16:18,580
simpler.

437
00:16:19,140 --> 00:16:20,140
So I don't

438
00:16:21,030 --> 00:16:22,409
know. I mean, it's it's

439
00:16:24,110 --> 00:16:25,869
we think that as picking up correct

440
00:16:25,870 --> 00:16:27,299
default overly complex thing.

441
00:16:28,810 --> 00:16:30,609
Yeah. But but I mean,

442
00:16:31,180 --> 00:16:32,649
the thing about that, about language

443
00:16:32,650 --> 00:16:34,880
design is like like any any decision

444
00:16:35,670 --> 00:16:37,809
you have to live with forever.

445
00:16:38,500 --> 00:16:40,740
I mean that and in languages

446
00:16:40,790 --> 00:16:42,659
you you can always add,

447
00:16:42,700 --> 00:16:43,899
but you can never take away.

448
00:16:44,580 --> 00:16:46,269
And so you've got to actually, as a

449
00:16:46,270 --> 00:16:47,529
language sign up to be very

450
00:16:47,530 --> 00:16:48,530
judicious.

451
00:16:50,920 --> 00:16:52,929
About reasoning over not

452
00:16:52,930 --> 00:16:54,519
so much what to put in, but what not

453
00:16:54,520 --> 00:16:56,309
to put it, an army

454
00:16:56,350 --> 00:16:57,879
because people come to you all the

455
00:16:57,880 --> 00:16:59,259
time. Wouldn't it be great if you

456
00:16:59,260 --> 00:17:00,419
could do this with that?

457
00:17:00,460 --> 00:17:01,389
Are you going to?

458
00:17:01,390 --> 00:17:03,309
Well, yeah, but you can't

459
00:17:03,730 --> 00:17:05,259
fundamentally change the nature of

460
00:17:05,260 --> 00:17:06,519
the beast. If you created an

461
00:17:06,520 --> 00:17:08,209
empirical and an

462
00:17:08,800 --> 00:17:10,239
imperative programing language, you

463
00:17:10,240 --> 00:17:11,469
can't turn it into a functional

464
00:17:11,470 --> 00:17:12,299
program.

465
00:17:12,300 --> 00:17:14,289
Show me you can you can borrow from

466
00:17:14,290 --> 00:17:15,390
functional programing.

467
00:17:16,180 --> 00:17:17,809
But it is what it is.

468
00:17:17,839 --> 00:17:19,269
And you gotta stay true to the

469
00:17:19,270 --> 00:17:20,789
nature of the beast, or you gotta

470
00:17:20,800 --> 00:17:21,800
create a new beast

471
00:17:22,710 --> 00:17:23,710
so-to-speak.

472
00:17:24,990 --> 00:17:25,990
Right.

473
00:17:27,250 --> 00:17:28,250
Any other thoughts?

474
00:17:29,530 --> 00:17:30,530
Excuse me?

475
00:17:30,650 --> 00:17:32,359
Can we interrupt you or you may have

476
00:17:32,790 --> 00:17:34,490
to interrupt any time.

477
00:17:35,960 --> 00:17:36,960
So.

478
00:17:38,040 --> 00:17:39,040
This is sort

479
00:17:39,930 --> 00:17:41,819
of the point about what what

480
00:17:41,820 --> 00:17:42,820
do you leave out?

481
00:17:43,620 --> 00:17:45,479
And the what I

482
00:17:45,480 --> 00:17:47,309
remember in

483
00:17:47,310 --> 00:17:49,169
the early days of Python's design,

484
00:17:50,490 --> 00:17:52,320
there were so many people

485
00:17:52,410 --> 00:17:54,380
who thought they had a good idea.

486
00:17:58,060 --> 00:18:00,190
Is sufficiently critical

487
00:18:00,880 --> 00:18:02,079
to say no.

488
00:18:03,880 --> 00:18:04,958
And so there are a few

489
00:18:05,920 --> 00:18:06,920
things that

490
00:18:08,350 --> 00:18:10,150
didn't work out at the time,

491
00:18:10,230 --> 00:18:11,308
I I haven't learned to

492
00:18:12,190 --> 00:18:13,079
say no.

493
00:18:13,080 --> 00:18:15,039
And I quickly enough

494
00:18:15,040 --> 00:18:16,719
did learn to say no.

495
00:18:16,750 --> 00:18:18,190
And I sort of felt

496
00:18:19,030 --> 00:18:20,030
like several lines

497
00:18:21,460 --> 00:18:23,319
of defense. Like the first line of

498
00:18:23,320 --> 00:18:25,299
defense is, hey,

499
00:18:25,330 --> 00:18:26,799
you think you need a new language

500
00:18:26,800 --> 00:18:28,330
feature, but actually you can

501
00:18:28,390 --> 00:18:30,109
already write that typo

502
00:18:31,930 --> 00:18:33,579
if you needed the locked right to

503
00:18:33,580 --> 00:18:34,720
function or a module.

504
00:18:35,740 --> 00:18:37,210
Well, if they say, yeah, but

505
00:18:37,270 --> 00:18:38,649
everybody needs that.

506
00:18:38,800 --> 00:18:40,900
Well nowadays we say put it on the

507
00:18:41,590 --> 00:18:42,913
package repository in those

508
00:18:43,630 --> 00:18:45,880
days. I said well maybe

509
00:18:46,480 --> 00:18:47,919
you can propose a new standard

510
00:18:47,920 --> 00:18:48,789
library module.

511
00:18:48,790 --> 00:18:50,019
That's a lot cheaper for the

512
00:18:50,020 --> 00:18:51,196
language design than the

513
00:18:51,910 --> 00:18:52,960
new language feature.

514
00:18:54,250 --> 00:18:55,250
And

515
00:18:56,170 --> 00:18:58,299
another line of defense was,

516
00:18:58,330 --> 00:18:59,920
well, you can actually write

517
00:18:59,980 --> 00:19:00,980
extensions in C

518
00:19:01,900 --> 00:19:03,999
or Fortran if you really care to.

519
00:19:04,930 --> 00:19:06,549
And you can help yourself that you

520
00:19:06,550 --> 00:19:07,824
can modify the behavior of

521
00:19:08,470 --> 00:19:10,059
the language in ways that

522
00:19:10,810 --> 00:19:12,329
aren't accessible when you just

523
00:19:12,910 --> 00:19:14,739
write the pure language.

524
00:19:14,770 --> 00:19:16,029
But you can still

525
00:19:16,600 --> 00:19:17,678
sort of you can extend

526
00:19:18,700 --> 00:19:19,876
it in a way that doesn't

527
00:19:19,960 --> 00:19:21,609
fundamentally alter

528
00:19:22,060 --> 00:19:23,260
the core language.

529
00:19:23,980 --> 00:19:25,839
And if you've tried all those

530
00:19:25,840 --> 00:19:27,970
things and and you failed,

531
00:19:29,230 --> 00:19:31,060
then maybe you could

532
00:19:31,150 --> 00:19:32,865
argue for, well, you have to change

533
00:19:33,160 --> 00:19:34,539
something in the core language.

534
00:19:36,140 --> 00:19:37,859
Hi, I'm interrupting.

535
00:19:39,060 --> 00:19:40,109
I find that

536
00:19:41,070 --> 00:19:42,739
whenever I get feedback on

537
00:19:43,230 --> 00:19:45,009
whatever whatever it is, program

538
00:19:45,130 --> 00:19:46,349
makers, I've been working on it.

539
00:19:46,450 --> 00:19:48,269
It's people come to

540
00:19:48,270 --> 00:19:50,460
you with a with a particular

541
00:19:50,520 --> 00:19:52,589
instance of a problem.

542
00:19:52,710 --> 00:19:54,229
And it's your job as a language

543
00:19:54,240 --> 00:19:56,459
assignment to tease out the class

544
00:19:56,550 --> 00:19:58,170
problem they're talking about

545
00:19:58,650 --> 00:20:00,539
and then try to understand whether

546
00:20:00,540 --> 00:20:02,579
there is a feature there underneath

547
00:20:02,580 --> 00:20:03,779
that you could put in place

548
00:20:04,440 --> 00:20:06,299
that is broadly applicable,

549
00:20:06,960 --> 00:20:08,819
not just to that one crawl.

550
00:20:13,770 --> 00:20:16,200
You're putting in place a class,

551
00:20:16,350 --> 00:20:17,390
not an instance suit.

552
00:20:17,520 --> 00:20:19,219
So that damn slide.

553
00:20:20,010 --> 00:20:21,676
And if all you're talking about is

554
00:20:21,750 --> 00:20:23,009
an instance, then you're probably

555
00:20:23,010 --> 00:20:24,010
better off.

556
00:20:28,800 --> 00:20:29,999
Save your money for power.

557
00:20:32,700 --> 00:20:33,809
Actually found this.

558
00:20:35,320 --> 00:20:37,089
In the early design of the Perl 6,

559
00:20:37,090 --> 00:20:38,829
we asked our CS I expected.

560
00:20:40,020 --> 00:20:41,829
Since we got 361,

561
00:20:42,150 --> 00:20:43,630
so we have this in spades

562
00:20:44,670 --> 00:20:45,670
lately.

563
00:20:45,720 --> 00:20:47,189
And, you know, it's completely

564
00:20:47,190 --> 00:20:48,190
overwhelming.

565
00:20:48,260 --> 00:20:49,260
Until I

566
00:20:50,120 --> 00:20:51,660
hit upon this principle,

567
00:20:53,130 --> 00:20:54,809
which Anderson alluded to,

568
00:20:55,170 --> 00:20:56,689
which is basically a lot of the

569
00:20:56,940 --> 00:20:57,949
proposed solution.

570
00:20:58,180 --> 00:20:59,999
Yeah, but there is a pain point

571
00:21:00,000 --> 00:21:01,000
underneath.

572
00:21:01,140 --> 00:21:02,849
And if you look at more enough of

573
00:21:02,870 --> 00:21:04,769
these RFD and they haven't

574
00:21:04,770 --> 00:21:06,485
seen pain point, there is something

575
00:21:06,960 --> 00:21:07,960
you should address.

576
00:21:08,160 --> 00:21:09,809
There's some fundamental unification

577
00:21:09,810 --> 00:21:11,309
you should do underneath or

578
00:21:11,760 --> 00:21:13,589
something that is is

579
00:21:14,220 --> 00:21:15,779
clunky that could be fixed.

580
00:21:16,120 --> 00:21:17,519
And especially if you're doing a

581
00:21:17,520 --> 00:21:18,929
complete redesign that you can think

582
00:21:18,930 --> 00:21:20,170
about the sort of.

583
00:21:34,630 --> 00:21:35,630
To fix it.

584
00:21:36,610 --> 00:21:38,029
How do you know?

585
00:21:38,060 --> 00:21:40,279
There were a lot of problems

586
00:21:40,360 --> 00:21:41,360
where.

587
00:22:00,830 --> 00:22:02,509
Difference that just

588
00:22:02,920 --> 00:22:04,829
won a victory with themselves.

589
00:22:05,170 --> 00:22:06,839
You get like 40 different.

590
00:22:35,350 --> 00:22:36,350
Released.

591
00:22:36,520 --> 00:22:37,870
It was really clear in the way

592
00:22:38,690 --> 00:22:40,569
that this was like a big

593
00:22:40,570 --> 00:22:42,730
issue and

594
00:22:43,090 --> 00:22:45,189
you feel dry and I came

595
00:22:45,190 --> 00:22:46,758
about as close to spilling blood

596
00:22:47,140 --> 00:22:49,029
on this topic as I've

597
00:22:49,030 --> 00:22:50,030
got.

598
00:22:51,330 --> 00:22:52,408
And, you know, I would

599
00:22:53,220 --> 00:22:54,220
much rather.

600
00:22:55,700 --> 00:22:56,700
Do something

601
00:22:57,690 --> 00:22:59,140
much, rather leave something out

602
00:22:59,630 --> 00:23:00,880
there, do something stupid.

603
00:23:02,440 --> 00:23:04,319
One of our principals planned to

604
00:23:04,320 --> 00:23:05,320
visit Margaret later.

605
00:23:05,690 --> 00:23:07,430
Yeah. Well, well,

606
00:23:07,780 --> 00:23:09,940
what that means is that that means

607
00:23:11,410 --> 00:23:12,539
the opening answer is no.

608
00:23:12,720 --> 00:23:13,669
No.

609
00:23:13,670 --> 00:23:14,089
Right.

610
00:23:14,090 --> 00:23:15,409
And one of the things that happened

611
00:23:15,410 --> 00:23:17,179
with like a whole bunch of different

612
00:23:17,260 --> 00:23:19,000
outfits and their job

613
00:23:19,550 --> 00:23:21,167
was that they kind of turned into

614
00:23:21,170 --> 00:23:22,170
competitions

615
00:23:23,710 --> 00:23:25,819
and generics

616
00:23:25,850 --> 00:23:26,950
and closer to work both.

617
00:23:27,110 --> 00:23:28,110
Probably the

618
00:23:29,420 --> 00:23:31,219
most hard fought competition.

619
00:23:31,990 --> 00:23:33,349
People were writing pretty steep

620
00:23:33,350 --> 00:23:34,639
pieces on the topic.

621
00:23:35,120 --> 00:23:36,739
And I was like trying them all out.

622
00:23:38,420 --> 00:23:40,160
I mean, some had no shortage

623
00:23:40,730 --> 00:23:42,619
of smart people at the time, but

624
00:23:43,040 --> 00:23:44,390
it's really hard to be

625
00:23:45,320 --> 00:23:47,269
a planet full

626
00:23:47,270 --> 00:23:48,470
of grad students,

627
00:23:49,330 --> 00:23:51,140
even grad students.

628
00:23:51,350 --> 00:23:52,659
At some point, I think most kids who

629
00:23:52,660 --> 00:23:53,519
were in grad school.

630
00:23:53,520 --> 00:23:54,520
That's correct.

631
00:23:54,860 --> 00:23:56,859
And in linguistics and linguistics.

632
00:23:58,790 --> 00:23:59,790
But it's.

633
00:24:00,480 --> 00:24:01,640
Oh, probably

634
00:24:02,360 --> 00:24:04,519
20, 30 years ago to say

635
00:24:04,940 --> 00:24:06,650
I'm going to write my own language.

636
00:24:07,310 --> 00:24:09,169
And people are very different.

637
00:24:09,410 --> 00:24:10,910
Back then, the

638
00:24:11,390 --> 00:24:12,390
Internet was sort

639
00:24:13,220 --> 00:24:14,209
of starting.

640
00:24:14,210 --> 00:24:16,769
Maybe not that they're single.

641
00:24:17,300 --> 00:24:18,460
I think we have computers.

642
00:24:18,470 --> 00:24:20,700
We have cell phones.

643
00:24:20,930 --> 00:24:22,819
Nobody cared, but nobody

644
00:24:22,820 --> 00:24:23,930
cared. What's all that about?

645
00:24:24,980 --> 00:24:26,599
Nobody cared what you did with the

646
00:24:26,600 --> 00:24:27,600
eighth that.

647
00:24:31,630 --> 00:24:32,630
Decades.

648
00:24:35,180 --> 00:24:37,499
But these are languages

649
00:24:37,530 --> 00:24:38,849
that persisted.

650
00:24:39,300 --> 00:24:41,380
Many changes and

651
00:24:42,010 --> 00:24:44,109
they're on different hardware.

652
00:24:45,300 --> 00:24:47,099
When you started

653
00:24:47,370 --> 00:24:49,210
designing your language, what

654
00:24:49,230 --> 00:24:50,669
was the key goal that you were

655
00:24:50,670 --> 00:24:52,739
trying to meet versus

656
00:24:52,740 --> 00:24:54,640
using some other extinguishment?

657
00:24:56,570 --> 00:24:57,844
Well, I mean, for me, it's

658
00:24:58,460 --> 00:25:00,309
never been people.

659
00:25:01,730 --> 00:25:03,980
And one of the secrets of success

660
00:25:04,010 --> 00:25:05,010
is try to solve as

661
00:25:05,930 --> 00:25:07,749
many problems at once as you can.

662
00:25:08,690 --> 00:25:10,279
And, you know, if there's just

663
00:25:10,550 --> 00:25:12,859
one problem, there's probably

664
00:25:12,860 --> 00:25:13,940
an easier way.

665
00:25:17,350 --> 00:25:18,350
Right.

666
00:25:19,270 --> 00:25:20,270
Right.

667
00:25:20,640 --> 00:25:22,749
You think that way you'll never

668
00:25:22,750 --> 00:25:23,760
have a new program.

669
00:25:25,380 --> 00:25:26,609
Well, there's always a better

670
00:25:26,610 --> 00:25:27,950
solution in the 300.

671
00:25:27,990 --> 00:25:29,970
Well, it depends on how much

672
00:25:30,000 --> 00:25:31,000
you pay.

673
00:25:31,050 --> 00:25:32,339
Chasing down memory,

674
00:25:32,950 --> 00:25:33,950
corruption.

675
00:25:34,050 --> 00:25:35,050
But

676
00:25:35,930 --> 00:25:37,529
the three virtues of the program

677
00:25:37,530 --> 00:25:39,510
are. Blasi This impatience

678
00:25:39,540 --> 00:25:40,589
and cupids

679
00:25:42,350 --> 00:25:43,409
takes Kupers.

680
00:25:46,290 --> 00:25:47,579
The world needs another program

681
00:25:47,920 --> 00:25:48,920
like.

682
00:25:51,150 --> 00:25:52,150
Is that a code?

683
00:25:53,160 --> 00:25:54,299
But the thing about programing

684
00:25:54,300 --> 00:25:56,009
language is that I think a lot of

685
00:25:56,010 --> 00:25:57,989
people don't realize that every

686
00:25:57,990 --> 00:25:59,730
programing language is about

687
00:26:00,240 --> 00:26:01,580
90 percent the same

688
00:26:02,220 --> 00:26:03,869
and maybe 10 percent

689
00:26:04,590 --> 00:26:05,590
if you're lucky.

690
00:26:06,630 --> 00:26:08,249
And a lot of people get very focused

691
00:26:08,250 --> 00:26:09,959
on the 10 percent new and forget

692
00:26:09,960 --> 00:26:11,390
about the 90 percent of the same.

693
00:26:12,060 --> 00:26:13,559
That makes it at great length.

694
00:26:13,770 --> 00:26:14,159
Right.

695
00:26:14,160 --> 00:26:16,140
I mean, there's a lot of hard

696
00:26:16,230 --> 00:26:17,896
or you work in creating programing

697
00:26:18,060 --> 00:26:19,760
language, loss, semantics that

698
00:26:19,910 --> 00:26:21,059
it's got to worry about.

699
00:26:21,120 --> 00:26:23,009
And like everyone and loves

700
00:26:23,010 --> 00:26:24,239
to talk about syntax

701
00:26:25,760 --> 00:26:27,839
and syntax is the easy part.

702
00:26:27,900 --> 00:26:29,849
Semantics is the hard part.

703
00:26:30,540 --> 00:26:31,979
You know, like how do the types of

704
00:26:31,980 --> 00:26:33,940
work and one of one only dogbone

705
00:26:35,130 --> 00:26:36,689
supported promotions and one of the

706
00:26:36,690 --> 00:26:38,729
different kinds of type constructors

707
00:26:38,730 --> 00:26:40,319
that the language, as etc.

708
00:26:40,380 --> 00:26:41,999
said are these are the hard things

709
00:26:42,000 --> 00:26:44,229
to do to design.

710
00:26:44,330 --> 00:26:46,019
But, but but not the things that

711
00:26:46,020 --> 00:26:48,089
people people love to bike

712
00:26:48,090 --> 00:26:49,539
shed on the syntax, you know,

713
00:26:49,560 --> 00:26:51,510
shouldn't be a colon or or

714
00:26:51,630 --> 00:26:52,239
comma.

715
00:26:52,240 --> 00:26:53,690
You know, it's like, oh my God,

716
00:26:54,160 --> 00:26:55,859
it's have a long thread about that.

717
00:26:58,360 --> 00:27:00,250
So talking about

718
00:27:00,640 --> 00:27:02,799
people having opinions and

719
00:27:02,800 --> 00:27:03,999
syntax and

720
00:27:04,700 --> 00:27:05,700
typing.

721
00:27:08,560 --> 00:27:10,570
These languages don't all

722
00:27:11,140 --> 00:27:13,119
take the same approach to typing.

723
00:27:14,200 --> 00:27:15,359
Maybe we'll start with veto

724
00:27:16,060 --> 00:27:18,309
and talk about typing in

725
00:27:18,760 --> 00:27:20,619
Python and then

726
00:27:20,620 --> 00:27:22,509
kind of work our way around.

727
00:27:30,700 --> 00:27:32,619
I mean, I just

728
00:27:32,620 --> 00:27:34,059
remember that when

729
00:27:35,640 --> 00:27:37,900
Pyfrom first happened,

730
00:27:38,770 --> 00:27:40,630
it was not a class.

731
00:27:41,620 --> 00:27:43,510
It was a little conversion function.

732
00:27:43,780 --> 00:27:45,069
There was an internal

733
00:27:45,760 --> 00:27:47,739
object type, which was really

734
00:27:47,740 --> 00:27:49,630
just kind of a Zite table

735
00:27:50,890 --> 00:27:53,109
that represented integer objects.

736
00:27:55,250 --> 00:27:57,400
And there was a built in function

737
00:27:57,440 --> 00:27:59,660
if you needed to convert the string

738
00:27:59,690 --> 00:28:01,219
to an integer and that

739
00:28:02,000 --> 00:28:03,323
was we had a bunch of those

740
00:28:03,890 --> 00:28:04,890
functions

741
00:28:06,350 --> 00:28:08,420
and we realized that we

742
00:28:09,710 --> 00:28:10,999
we had made a mistake.

743
00:28:11,030 --> 00:28:12,304
We had given users classes

744
00:28:13,040 --> 00:28:14,930
that were

745
00:28:16,040 --> 00:28:17,363
different from the built in

746
00:28:17,900 --> 00:28:19,000
object types.

747
00:28:20,500 --> 00:28:22,439
And for a long time, it was like,

748
00:28:22,560 --> 00:28:23,560
oh, well, like the

749
00:28:24,420 --> 00:28:25,645
real stuff is implemented

750
00:28:26,460 --> 00:28:27,460
in C and the user

751
00:28:28,680 --> 00:28:30,199
writes a little bit of blue top

752
00:28:30,660 --> 00:28:32,009
of that. And when

753
00:28:32,970 --> 00:28:34,489
when I found out that we had 80

754
00:28:35,130 --> 00:28:38,040
different competing web frameworks,

755
00:28:42,270 --> 00:28:44,270
it was sort of time to

756
00:28:44,510 --> 00:28:46,225
to realize that people were writing

757
00:28:46,470 --> 00:28:48,920
much larger programs.

758
00:28:49,020 --> 00:28:51,299
And that's a different approach.

759
00:28:51,300 --> 00:28:52,680
Two types was needed.

760
00:28:52,770 --> 00:28:54,420
And that's where we sort of

761
00:28:55,170 --> 00:28:57,299
reinvented the whole approach

762
00:28:57,300 --> 00:28:58,550
to types in Python.

763
00:29:01,050 --> 00:29:02,680
And there was a bunch of cleanup

764
00:29:02,700 --> 00:29:04,569
that didn't happen until 5 3,

765
00:29:04,570 --> 00:29:05,529
actually.

766
00:29:05,530 --> 00:29:07,799
But one of the things we introduced

767
00:29:07,890 --> 00:29:08,968
and we were lucky that

768
00:29:09,720 --> 00:29:11,700
there weren't actually despite those

769
00:29:11,760 --> 00:29:13,319
80 web frameworks, there weren't

770
00:29:13,380 --> 00:29:15,420
enough users that

771
00:29:16,080 --> 00:29:17,969
we were completely stymied by

772
00:29:17,970 --> 00:29:19,289
backwards compatibility.

773
00:29:19,450 --> 00:29:21,029
Yet like we are now.

774
00:29:22,660 --> 00:29:24,081
So we just one day we changed

775
00:29:24,790 --> 00:29:26,529
the function into

776
00:29:26,940 --> 00:29:28,700
a designated for the class.

777
00:29:30,790 --> 00:29:32,569
And sort of

778
00:29:32,780 --> 00:29:34,660
followed that calling, the class

779
00:29:34,710 --> 00:29:36,789
would be constructing

780
00:29:36,790 --> 00:29:39,009
an instance of the class

781
00:29:39,460 --> 00:29:41,589
so that if you had simple

782
00:29:41,590 --> 00:29:43,480
code that wrote in

783
00:29:43,570 --> 00:29:45,419
left brand some expression right.

784
00:29:47,080 --> 00:29:49,149
And would work exactly the same way

785
00:29:49,600 --> 00:29:51,150
that we have the same effect.

786
00:29:51,190 --> 00:29:53,079
But the workings

787
00:29:53,170 --> 00:29:54,970
inside were completely different.

788
00:29:55,040 --> 00:29:57,010
There was one particular file

789
00:29:57,040 --> 00:29:58,420
that's sort of

790
00:29:59,560 --> 00:30:00,900
implemented the

791
00:30:01,720 --> 00:30:03,309
that's sort of the basic

792
00:30:03,520 --> 00:30:04,941
functionality of types in the

793
00:30:05,170 --> 00:30:06,759
language and it

794
00:30:07,030 --> 00:30:08,402
in for one to it was like 50

795
00:30:09,010 --> 00:30:10,010
lines of code.

796
00:30:10,090 --> 00:30:12,279
And at some point I rewrote

797
00:30:12,280 --> 00:30:13,770
it and it was five thousand by.

798
00:30:16,560 --> 00:30:17,560
That's great.

799
00:30:18,590 --> 00:30:20,359
What about Titus in Dublin?

800
00:30:26,970 --> 00:30:28,779
I got this this this one

801
00:30:28,800 --> 00:30:31,089
history of caring about things

802
00:30:31,090 --> 00:30:32,090
like.

803
00:30:37,820 --> 00:30:39,410
Building robot software

804
00:30:41,340 --> 00:30:43,499
often that that

805
00:30:44,390 --> 00:30:45,900
that comes out about, you know, I'm

806
00:30:45,980 --> 00:30:47,450
much more worried about

807
00:30:47,960 --> 00:30:49,579
what it takes to

808
00:30:50,030 --> 00:30:51,829
build production quality software

809
00:30:52,400 --> 00:30:54,109
than about what it takes, just like

810
00:30:54,770 --> 00:30:55,940
you like a quick thing

811
00:30:56,830 --> 00:30:58,447
and chuckles not a great language

812
00:30:58,940 --> 00:30:59,789
for.

813
00:30:59,790 --> 00:31:00,819
Things, but it's it's

814
00:31:01,610 --> 00:31:03,450
OK, it's not

815
00:31:03,570 --> 00:31:05,609
for me. One of the one of the things

816
00:31:05,610 --> 00:31:07,359
that I love to do and

817
00:31:08,160 --> 00:31:09,240
maybe I'm weird, but

818
00:31:10,740 --> 00:31:12,119
being able to do automatic or

819
00:31:12,150 --> 00:31:14,009
improving on hunks of cup

820
00:31:15,070 --> 00:31:16,420
and and.

821
00:31:17,080 --> 00:31:19,179
Type system is

822
00:31:19,180 --> 00:31:20,950
a really great place

823
00:31:22,000 --> 00:31:24,069
to be, one of the foundations

824
00:31:24,630 --> 00:31:26,080
there begin they're improving

825
00:31:26,470 --> 00:31:27,470
it for.

826
00:31:28,440 --> 00:31:30,329
Four months of code turns out to be

827
00:31:30,330 --> 00:31:31,979
really useful for things like

828
00:31:32,550 --> 00:31:33,550
building optimized.

829
00:31:35,340 --> 00:31:37,109
And doing it

830
00:31:38,040 --> 00:31:39,059
ahead of time.

831
00:31:43,280 --> 00:31:45,299
To be able to hear

832
00:31:45,300 --> 00:31:47,720
improvable way as many

833
00:31:48,080 --> 00:31:50,220
things as possible, so

834
00:31:50,370 --> 00:31:51,650
satellite like like,

835
00:31:52,400 --> 00:31:54,289
you know, one of the not well-known

836
00:31:54,350 --> 00:31:56,329
things about about Java is

837
00:31:56,330 --> 00:31:58,190
that, you know, in the jobless pack,

838
00:31:58,310 --> 00:32:00,019
you know, it says Ouray, subscript

839
00:32:00,020 --> 00:32:01,900
checking is always on.

840
00:32:04,430 --> 00:32:05,930
But, you know, it's only

841
00:32:05,960 --> 00:32:07,490
conceptually all of his all

842
00:32:08,750 --> 00:32:10,279
right. But the truth is

843
00:32:10,660 --> 00:32:12,319
that there are there's more than on

844
00:32:12,320 --> 00:32:14,060
my books or

845
00:32:14,150 --> 00:32:15,470
the compiler

846
00:32:16,970 --> 00:32:18,859
to improve away

847
00:32:18,890 --> 00:32:20,690
almost all index checks.

848
00:32:22,100 --> 00:32:23,521
And same thing with like null

849
00:32:23,810 --> 00:32:25,329
pointer checks and all kinds of

850
00:32:25,640 --> 00:32:27,679
things that look

851
00:32:27,860 --> 00:32:29,029
like they're heavy weight,

852
00:32:30,170 --> 00:32:31,170
but they're really cheap,

853
00:32:32,180 --> 00:32:33,769
you know, so it's a one of the

854
00:32:33,770 --> 00:32:35,289
things that I like.

855
00:32:36,180 --> 00:32:37,400
I really cared about it.

856
00:32:37,430 --> 00:32:39,499
The time was that A plus

857
00:32:39,500 --> 00:32:41,150
B should almost always be

858
00:32:42,280 --> 00:32:43,790
at most one instruction

859
00:32:44,690 --> 00:32:46,779
bonus points for zero instructions.

860
00:32:49,700 --> 00:32:50,700
What about after?

861
00:32:51,450 --> 00:32:52,849
Well, the thing about zero

862
00:32:52,850 --> 00:32:54,619
instructions is often you can like

863
00:32:54,620 --> 00:32:56,749
fold them into some weird addressing

864
00:32:56,750 --> 00:32:57,750
mode.

865
00:32:57,980 --> 00:32:58,980
Right.

866
00:32:59,200 --> 00:33:00,200
And.

867
00:33:00,750 --> 00:33:02,769
And so

868
00:33:02,810 --> 00:33:04,500
in kind of the universe, I'm

869
00:33:04,650 --> 00:33:05,890
right, I tend to live.

870
00:33:07,180 --> 00:33:09,069
Being able to look

871
00:33:09,070 --> 00:33:11,019
at A plus B

872
00:33:11,020 --> 00:33:12,739
and realize it's

873
00:33:12,880 --> 00:33:14,839
that destruction that all

874
00:33:14,860 --> 00:33:16,589
feeds off of the type system.

875
00:33:17,890 --> 00:33:19,779
And sometimes you can you

876
00:33:19,780 --> 00:33:21,969
can do this with sort of optimistic

877
00:33:22,940 --> 00:33:24,810
just-in-time coupon compilation,

878
00:33:26,110 --> 00:33:27,609
depending on how far you want to

879
00:33:27,610 --> 00:33:29,169
push that. So like the

880
00:33:29,860 --> 00:33:31,989
the JavaScript engine

881
00:33:31,990 --> 00:33:33,819
in Chrome is

882
00:33:33,980 --> 00:33:36,220
is absolutely astonishing

883
00:33:36,360 --> 00:33:37,910
for the the

884
00:33:38,440 --> 00:33:40,479
kind of core that they go

885
00:33:40,480 --> 00:33:42,940
through too optimistically.

886
00:33:45,120 --> 00:33:47,249
JavaScript code, but

887
00:33:48,600 --> 00:33:50,609
it's also very hard to do those

888
00:33:50,610 --> 00:33:52,178
kinds of tricks if you're trying

889
00:33:52,650 --> 00:33:54,229
to get into smaller

890
00:33:54,570 --> 00:33:55,570
devices.

891
00:33:56,340 --> 00:33:58,619
So, you know, there are

892
00:33:58,650 --> 00:34:00,240
there are Java compilers that

893
00:34:00,540 --> 00:34:02,130
work for machines that only have

894
00:34:02,160 --> 00:34:03,160
like.

895
00:34:06,660 --> 00:34:08,819
You know, to do that kind of compact

896
00:34:08,980 --> 00:34:10,352
distillation, you need every

897
00:34:11,110 --> 00:34:12,999
kind of hope that gives you

898
00:34:13,030 --> 00:34:14,919
every little every last

899
00:34:14,920 --> 00:34:16,030
drop of information.

900
00:34:16,270 --> 00:34:18,189
And the earlier, you

901
00:34:18,190 --> 00:34:20,079
know, what better job you can do.

902
00:34:20,489 --> 00:34:22,519
Right. So speaking

903
00:34:22,540 --> 00:34:24,339
of lots of information that one

904
00:34:24,340 --> 00:34:25,409
gets, typescript

905
00:34:26,860 --> 00:34:28,539
gives you a lot of flexibility

906
00:34:29,110 --> 00:34:30,110
and power.

907
00:34:30,730 --> 00:34:31,809
What's your general view that

908
00:34:31,810 --> 00:34:32,810
typing?

909
00:34:34,659 --> 00:34:36,227
Well, let me actually it's funny

910
00:34:36,610 --> 00:34:37,650
that you mention

911
00:34:38,550 --> 00:34:39,550
cycles and counting.

912
00:34:39,639 --> 00:34:40,928
How many? And whenever I do, I

913
00:34:40,929 --> 00:34:42,460
remember when I wrote Turbo Pascal,

914
00:34:42,489 --> 00:34:44,110
which was all written in C-A-T

915
00:34:44,260 --> 00:34:45,099
assembly code.

916
00:34:45,100 --> 00:34:46,389
And back in those days,

917
00:34:46,960 --> 00:34:48,428
you could literally look up the

918
00:34:48,429 --> 00:34:50,589
intel manual on the sidewalk

919
00:34:50,590 --> 00:34:51,819
or whatever, you know, and see how

920
00:34:51,820 --> 00:34:53,529
many clock cycles every construction

921
00:34:53,530 --> 00:34:55,239
took. And actually it would work out

922
00:34:55,600 --> 00:34:56,600
just like that.

923
00:34:56,650 --> 00:34:57,850
I remember now

924
00:34:58,690 --> 00:35:00,013
everything takes zero clock

925
00:35:00,730 --> 00:35:02,469
cycles except when it takes a

926
00:35:02,470 --> 00:35:03,470
thousand clocks.

927
00:35:05,200 --> 00:35:06,719
The memory and it is absolutely

928
00:35:07,480 --> 00:35:08,269
impossible.

929
00:35:08,270 --> 00:35:10,059
The reason about what I've seen you

930
00:35:10,060 --> 00:35:11,440
use execute your code today.

931
00:35:11,560 --> 00:35:12,820
That that's just one.

932
00:35:13,300 --> 00:35:14,979
Well, it's not impossible,

933
00:35:15,480 --> 00:35:16,699
but it's much harder.

934
00:35:17,090 --> 00:35:18,969
Yeah, it is a complete pain in

935
00:35:18,970 --> 00:35:19,970
the ass.

936
00:35:21,400 --> 00:35:22,400
And they don't give

937
00:35:23,260 --> 00:35:24,129
you manuals.

938
00:35:24,130 --> 00:35:25,130
Not good at all.

939
00:35:25,140 --> 00:35:26,049
You don't want to actually

940
00:35:26,050 --> 00:35:26,949
understand that.

941
00:35:26,950 --> 00:35:28,209
If you need to get the chip

942
00:35:28,210 --> 00:35:29,210
diagrams.

943
00:35:29,330 --> 00:35:31,659
But but with respect to types,

944
00:35:31,660 --> 00:35:33,369
I've always worked on

945
00:35:33,580 --> 00:35:35,295
programing languages that have type

946
00:35:35,440 --> 00:35:35,979
systems.

947
00:35:35,980 --> 00:35:37,156
But it's interesting how

948
00:35:38,200 --> 00:35:40,479
it's gone from being type systems

949
00:35:40,480 --> 00:35:41,480
far

950
00:35:42,400 --> 00:35:43,919
for the code generators sake or

951
00:35:44,260 --> 00:35:46,079
type systems for four

952
00:35:46,380 --> 00:35:48,549
for, you know, generating

953
00:35:48,550 --> 00:35:49,550
errors too.

954
00:35:50,080 --> 00:35:51,158
Now I almost view type

955
00:35:52,150 --> 00:35:53,277
systems as as a tooling

956
00:35:54,340 --> 00:35:56,109
feature, and that's really sort of

957
00:35:56,320 --> 00:35:57,869
the thing that has been interesting.

958
00:35:57,870 --> 00:35:59,290
High automate for typescript.

959
00:35:59,350 --> 00:36:00,880
It is, you know,

960
00:36:01,660 --> 00:36:03,081
first of all, starting with a

961
00:36:03,100 --> 00:36:04,449
dynamically typed programing

962
00:36:04,450 --> 00:36:05,799
language like JavaScript and then

963
00:36:05,800 --> 00:36:07,869
trying to add a type system on top

964
00:36:07,870 --> 00:36:09,000
of it that

965
00:36:10,320 --> 00:36:11,692
that faithfully reflects the

966
00:36:12,040 --> 00:36:13,363
semantics of the programing

967
00:36:13,480 --> 00:36:14,409
language.

968
00:36:14,410 --> 00:36:16,179
And the reason for doing it is

969
00:36:16,180 --> 00:36:18,159
actually not because we think type

970
00:36:18,160 --> 00:36:19,924
systems are interesting, but because

971
00:36:19,960 --> 00:36:21,381
if you think about what it is

972
00:36:21,820 --> 00:36:24,219
that powers programmer productivity

973
00:36:24,220 --> 00:36:25,689
today, like everyone loves their

974
00:36:25,690 --> 00:36:26,690
I.T. like.

975
00:36:26,710 --> 00:36:28,359
Whatever you're using, I hope it's

976
00:36:28,360 --> 00:36:29,289
Visual Studio code.

977
00:36:29,290 --> 00:36:31,179
But putting that if it's

978
00:36:31,180 --> 00:36:33,249
not, you know, I am sure

979
00:36:33,250 --> 00:36:35,289
you're all like accustomed to things

980
00:36:35,290 --> 00:36:36,515
like state completion and

981
00:36:36,610 --> 00:36:37,933
refactoring code navigation

982
00:36:39,010 --> 00:36:40,629
and go to definition and so forth.

983
00:36:40,630 --> 00:36:42,129
And if you think about it, the

984
00:36:42,130 --> 00:36:44,169
things that the thing that powers

985
00:36:44,170 --> 00:36:45,444
that is semantic knowledge

986
00:36:46,300 --> 00:36:47,529
of your code. And the thing that

987
00:36:47,530 --> 00:36:49,196
provides the semantic knowledge of

988
00:36:49,420 --> 00:36:51,820
your code is a compiler

989
00:36:51,850 --> 00:36:52,850
with a type system.

990
00:36:53,830 --> 00:36:55,480
And once you add types,

991
00:36:55,900 --> 00:36:57,321
you can actually dramatically

992
00:36:58,090 --> 00:36:59,919
increase productivity, which

993
00:37:00,220 --> 00:37:01,690
in some ways seems counter

994
00:37:02,230 --> 00:37:03,099
counter intuitive.

995
00:37:03,100 --> 00:37:03,399
Right.

996
00:37:03,400 --> 00:37:04,929
I thought like dynamic languages

997
00:37:04,930 --> 00:37:06,010
were easier to

998
00:37:07,120 --> 00:37:08,590
to to approach because you got

999
00:37:09,040 --> 00:37:10,479
rid of the types which was like a

1000
00:37:10,480 --> 00:37:11,429
bother all the time.

1001
00:37:11,430 --> 00:37:13,139
Well, look, when it turns out that

1002
00:37:13,360 --> 00:37:15,190
you can actually be more productive

1003
00:37:15,220 --> 00:37:16,249
by by adding types if

1004
00:37:17,170 --> 00:37:19,179
you do it in a non-intrusive manner

1005
00:37:19,210 --> 00:37:20,800
and if you work hard on doing

1006
00:37:21,130 --> 00:37:22,959
good type inference and so forth.

1007
00:37:23,590 --> 00:37:24,590
So, so,

1008
00:37:25,960 --> 00:37:27,249
so, so, so, so.

1009
00:37:27,280 --> 00:37:28,989
So I want to jump in here.

1010
00:37:30,690 --> 00:37:32,610
And then I take

1011
00:37:32,640 --> 00:37:33,640
over.

1012
00:37:34,530 --> 00:37:35,853
So I really, really, really

1013
00:37:36,670 --> 00:37:38,385
believe in that point in the in the

1014
00:37:38,580 --> 00:37:39,580
in the power tools

1015
00:37:40,600 --> 00:37:41,800
for power geeks.

1016
00:37:44,440 --> 00:37:46,659
And one of the things that drives

1017
00:37:46,660 --> 00:37:48,639
me nuts is

1018
00:37:48,640 --> 00:37:50,579
the real man usvi

1019
00:37:51,600 --> 00:37:52,600
movement.

1020
00:37:52,860 --> 00:37:54,669
And, you know, it's it's really

1021
00:37:55,490 --> 00:37:57,219
I just want to punch people who

1022
00:37:57,580 --> 00:37:57,979
are.

1023
00:37:57,980 --> 00:38:00,039
But I'm really inside state.

1024
00:38:00,200 --> 00:38:01,670
I I'm a real developer because

1025
00:38:02,260 --> 00:38:03,309
I use v.i.

1026
00:38:03,420 --> 00:38:05,080
And it's like, you know,

1027
00:38:05,530 --> 00:38:06,530
I do it the hard way.

1028
00:38:06,700 --> 00:38:07,700
I think.

1029
00:38:07,900 --> 00:38:10,270
I think I make language developers

1030
00:38:10,300 --> 00:38:11,300
lazy.

1031
00:38:20,190 --> 00:38:21,219
IBRD Let me get a lot

1032
00:38:22,170 --> 00:38:23,909
more done a lot faster,

1033
00:38:24,100 --> 00:38:25,849
OK? I mean, I'm not I

1034
00:38:26,590 --> 00:38:27,590
I mean, I'm really

1035
00:38:28,560 --> 00:38:30,239
not into proving my manhood.

1036
00:38:30,240 --> 00:38:31,240
I'm good

1037
00:38:32,290 --> 00:38:33,780
at getting things done.

1038
00:38:36,260 --> 00:38:37,949
Case in point, but I

1039
00:38:38,180 --> 00:38:39,869
take the devil's advocate here first

1040
00:38:39,890 --> 00:38:41,899
said he meant something like.

1041
00:38:43,670 --> 00:38:45,360
Yeah, but

1042
00:38:45,490 --> 00:38:47,599
you their notebooks, a lot

1043
00:38:47,600 --> 00:38:48,969
of science people.

1044
00:38:49,010 --> 00:38:51,249
Data scientists get a lot done.

1045
00:38:51,970 --> 00:38:54,519
It's actually a pretty simplistic

1046
00:38:55,410 --> 00:38:56,789
I need

1047
00:38:57,890 --> 00:38:59,270
a dynamic language.

1048
00:38:59,300 --> 00:39:00,300
Most of the time.

1049
00:39:00,820 --> 00:39:02,490
So you could get a lot more than

1050
00:39:03,510 --> 00:39:04,510
that.

1051
00:39:05,600 --> 00:39:06,600
I don't know.

1052
00:39:06,940 --> 00:39:08,309
No, but I think typeset.

1053
00:39:08,420 --> 00:39:09,739
I mean, it's not just

1054
00:39:11,350 --> 00:39:12,669
a type system can help you.

1055
00:39:12,680 --> 00:39:14,479
First of all, if you're uninitiated

1056
00:39:14,700 --> 00:39:16,549
and you want to know here, here's my

1057
00:39:16,550 --> 00:39:18,069
food and now I say food dot and

1058
00:39:18,510 --> 00:39:19,739
then the ivy can show you.

1059
00:39:19,760 --> 00:39:20,909
What can I type next?

1060
00:39:20,930 --> 00:39:22,694
Right. As opposed to I gotta go read

1061
00:39:22,910 --> 00:39:24,400
manuals and figure it out or know

1062
00:39:24,410 --> 00:39:26,329
it. All right. I mean, your original

1063
00:39:26,330 --> 00:39:28,309
developer of whatever piece of code

1064
00:39:28,340 --> 00:39:29,418
might know it all, but

1065
00:39:30,320 --> 00:39:32,179
then people move

1066
00:39:32,180 --> 00:39:33,829
on and new people come in.

1067
00:39:34,190 --> 00:39:35,419
You know, there's here's this

1068
00:39:35,420 --> 00:39:36,409
project.

1069
00:39:36,410 --> 00:39:38,300
There was there was no documentation

1070
00:39:38,330 --> 00:39:39,169
written about it.

1071
00:39:39,170 --> 00:39:40,729
And if you think about it, types or

1072
00:39:40,730 --> 00:39:42,649
documentation to write, I

1073
00:39:42,650 --> 00:39:44,267
mean, and then there just there's

1074
00:39:44,630 --> 00:39:45,630
so many things about

1075
00:39:46,550 --> 00:39:47,550
like

1076
00:39:48,680 --> 00:39:50,329
like adding them that

1077
00:39:50,840 --> 00:39:52,579
down the line give you

1078
00:39:52,670 --> 00:39:54,070
increased productivity.

1079
00:39:54,880 --> 00:39:56,540
So there's productivity

1080
00:39:57,030 --> 00:39:58,309
and then that's that's

1081
00:39:58,310 --> 00:39:59,599
maintainability.

1082
00:40:00,080 --> 00:40:01,080
Oh, well both.

1083
00:40:01,100 --> 00:40:02,570
And maybe

1084
00:40:03,440 --> 00:40:05,340
any different thoughts about how

1085
00:40:05,350 --> 00:40:06,739
you get to talk about types.

1086
00:40:07,110 --> 00:40:08,110
Well, talk.

1087
00:40:09,630 --> 00:40:10,820
The colonel types out.

1088
00:40:11,450 --> 00:40:13,329
Well, that story is very

1089
00:40:13,330 --> 00:40:14,949
different for profile purposes.

1090
00:40:16,920 --> 00:40:18,586
As you know, Girlfight, to sort of

1091
00:40:18,610 --> 00:40:20,679
rural over time and the whole idea

1092
00:40:20,870 --> 00:40:22,179
was sort of particularly idea.

1093
00:40:22,570 --> 00:40:23,619
Everything you can pretend

1094
00:40:23,620 --> 00:40:25,059
everything is a straight even if

1095
00:40:25,060 --> 00:40:26,800
it's a number or floating point

1096
00:40:27,010 --> 00:40:28,909
internally, it's stalled, you

1097
00:40:28,920 --> 00:40:29,920
know, interchangeable.

1098
00:40:30,340 --> 00:40:32,229
And that works out for

1099
00:40:32,620 --> 00:40:34,059
bootstrapping people into a

1100
00:40:34,060 --> 00:40:35,739
language. They're quite nicely.

1101
00:40:37,240 --> 00:40:39,099
And it's it's a feature that we

1102
00:40:39,100 --> 00:40:40,209
wanted to keep in Perl 6.

1103
00:40:40,210 --> 00:40:41,889
But what we discovered in Perl 5 is

1104
00:40:41,890 --> 00:40:43,720
part of the redesign was

1105
00:40:45,010 --> 00:40:46,749
it's it's fine if the

1106
00:40:47,080 --> 00:40:48,999
user is confused about the

1107
00:40:49,000 --> 00:40:50,699
interchangeability to tell more for

1108
00:40:50,720 --> 00:40:52,179
the Muslim to

1109
00:40:52,550 --> 00:40:54,579
take the care of

1110
00:40:54,590 --> 00:40:55,249
your standard.

1111
00:40:55,250 --> 00:40:57,099
Tikes But it's not so

1112
00:40:57,100 --> 00:40:58,809
good if the computer is

1113
00:40:59,170 --> 00:41:00,170
confused about.

1114
00:41:01,930 --> 00:41:02,959
You know, for all the

1115
00:41:03,820 --> 00:41:05,859
girl wanted, it was written way

1116
00:41:05,860 --> 00:41:06,860
back in the dark ages.

1117
00:41:06,960 --> 00:41:08,319
You had a schuberg, a bunch of stuff

1118
00:41:08,320 --> 00:41:09,830
in, and it

1119
00:41:10,600 --> 00:41:11,600
cheated a lot.

1120
00:41:11,980 --> 00:41:14,199
So part of the redesign

1121
00:41:14,350 --> 00:41:16,059
when we did girls school Perl 6

1122
00:41:16,060 --> 00:41:17,170
thing we wanted to do

1123
00:41:18,160 --> 00:41:19,930
object oriented programing

1124
00:41:20,000 --> 00:41:21,568
that is that these languages and

1125
00:41:22,000 --> 00:41:22,959
we wanted to be functional

1126
00:41:22,960 --> 00:41:24,249
programing that are the most

1127
00:41:24,250 --> 00:41:25,449
functional programing languages.

1128
00:41:26,260 --> 00:41:28,149
To do that you have to have a

1129
00:41:28,420 --> 00:41:30,450
fundamental very sound type

1130
00:41:30,460 --> 00:41:32,439
system of a sound object

1131
00:41:32,890 --> 00:41:33,890
model underneath.

1132
00:41:34,330 --> 00:41:36,189
And you really have

1133
00:41:36,190 --> 00:41:38,169
to take seriously these slogans

1134
00:41:38,170 --> 00:41:39,429
like everything is an object.

1135
00:41:39,830 --> 00:41:40,929
Everything is like closure.

1136
00:41:41,250 --> 00:41:43,149
Every feel the process,

1137
00:41:43,390 --> 00:41:45,539
even blocks are closures.

1138
00:41:46,300 --> 00:41:48,189
And he has got to

1139
00:41:48,200 --> 00:41:49,579
optimize. Hard to get up, get them

1140
00:41:49,690 --> 00:41:50,690
away from there.

1141
00:41:50,720 --> 00:41:52,629
But but I I also

1142
00:41:52,630 --> 00:41:54,850
agree with this idea

1143
00:41:54,880 --> 00:41:56,379
that the tikes also

1144
00:41:56,710 --> 00:41:57,710
are cultural.

1145
00:41:58,540 --> 00:41:59,900
I talked about Pei's hanging things

1146
00:41:59,920 --> 00:42:01,629
on pace. They're one of the one of

1147
00:42:01,630 --> 00:42:02,919
the picks. We didn't have something

1148
00:42:02,920 --> 00:42:04,149
to hang on the prettified.

1149
00:42:04,510 --> 00:42:06,149
And now we we can Pangle.

1150
00:42:06,190 --> 00:42:07,869
But all of that information

1151
00:42:08,070 --> 00:42:09,430
and then just run a little program

1152
00:42:10,070 --> 00:42:11,349
there. And then you have to have an

1153
00:42:11,350 --> 00:42:13,090
idea. They say, well, this object.

1154
00:42:13,920 --> 00:42:15,159
You know what methods.

1155
00:42:15,190 --> 00:42:16,719
But what you know, they could do

1156
00:42:16,720 --> 00:42:17,709
their own introspection.

1157
00:42:17,710 --> 00:42:18,710
The whole thing.

1158
00:42:18,730 --> 00:42:20,079
And people just you know, I'm

1159
00:42:21,060 --> 00:42:23,079
baffled by the way we do also

1160
00:42:23,080 --> 00:42:24,080
have an idea.

1161
00:42:27,420 --> 00:42:28,829
Good points all around.

1162
00:42:29,670 --> 00:42:31,649
So most

1163
00:42:31,650 --> 00:42:33,218
of the time programmers actually

1164
00:42:33,480 --> 00:42:35,610
spend maintaining versus

1165
00:42:35,670 --> 00:42:36,969
writing new code

1166
00:42:38,350 --> 00:42:39,690
out in the wild and.

1167
00:42:41,430 --> 00:42:43,189
What things or elements of a

1168
00:42:43,190 --> 00:42:45,020
programing language make it

1169
00:42:45,080 --> 00:42:46,880
easier to maintain?

1170
00:42:47,460 --> 00:42:48,850
And maybe we'll start with Fito.

1171
00:42:52,930 --> 00:42:54,730
Well, I found that.

1172
00:42:58,650 --> 00:42:59,650
TypeScript

1173
00:43:00,660 --> 00:43:02,630
is actually incredibly useful,

1174
00:43:02,640 --> 00:43:05,010
and so we're adding a very similar

1175
00:43:05,100 --> 00:43:06,949
idea to Python, adding it

1176
00:43:07,000 --> 00:43:09,039
in a slightly good way because.

1177
00:43:09,870 --> 00:43:11,460
We are taxed.

1178
00:43:28,950 --> 00:43:29,929
All right.

1179
00:43:29,930 --> 00:43:30,769
Go.

1180
00:43:30,770 --> 00:43:31,770
Anyway,

1181
00:43:33,050 --> 00:43:34,050
we have time.

1182
00:43:36,480 --> 00:43:37,699
I sort of

1183
00:43:38,480 --> 00:43:39,656
learned a painful lesson

1184
00:43:40,550 --> 00:43:42,859
of that before school programs.

1185
00:43:43,160 --> 00:43:45,409
Seafair dynamic typing

1186
00:43:45,410 --> 00:43:47,570
is great for large programs.

1187
00:43:47,600 --> 00:43:48,600
You have to

1188
00:43:49,520 --> 00:43:51,380
have a more disciplined

1189
00:43:51,410 --> 00:43:52,782
approach and it helps if the

1190
00:43:53,300 --> 00:43:54,559
language actually

1191
00:43:55,430 --> 00:43:56,209
gives you that.

1192
00:43:56,210 --> 00:43:58,159
It's rather than telling you,

1193
00:43:58,160 --> 00:43:59,469
well, you can do whatever you want.

1194
00:44:02,030 --> 00:44:02,979
Yeah, yeah.

1195
00:44:02,980 --> 00:44:04,440
That was part of our scale up

1196
00:44:04,990 --> 00:44:06,800
idea for Perl 6,

1197
00:44:07,170 --> 00:44:08,170
the typist would help

1198
00:44:09,100 --> 00:44:10,929
with programing in large because

1199
00:44:10,930 --> 00:44:12,550
we didn't notice those limitations.

1200
00:44:16,600 --> 00:44:18,119
One year ago, I started working

1201
00:44:18,130 --> 00:44:20,300
really heavily on Unbury

1202
00:44:20,310 --> 00:44:21,310
Factoring In.

1203
00:44:22,040 --> 00:44:23,040
And,

1204
00:44:24,140 --> 00:44:25,369
you know, being able to

1205
00:44:26,200 --> 00:44:27,909
do like large scale refactoring

1206
00:44:28,110 --> 00:44:30,219
is unlike millions of lines

1207
00:44:30,220 --> 00:44:31,540
of code at once,

1208
00:44:33,430 --> 00:44:34,749
that really changes the way you

1209
00:44:34,750 --> 00:44:36,309
think about maintaining code.

1210
00:44:37,200 --> 00:44:38,709
Because, you know, the world is

1211
00:44:38,710 --> 00:44:40,179
filled with libraries that have

1212
00:44:40,180 --> 00:44:41,503
things like stupid variable

1213
00:44:42,070 --> 00:44:43,929
names, stupid method

1214
00:44:43,930 --> 00:44:45,760
names, because,

1215
00:44:45,850 --> 00:44:47,049
you know, when you started out, it

1216
00:44:47,050 --> 00:44:48,050
meant one thing.

1217
00:44:48,520 --> 00:44:50,559
But now you realize that

1218
00:44:50,680 --> 00:44:52,269
you really didn't really understand

1219
00:44:52,270 --> 00:44:53,109
what you were doing in the

1220
00:44:53,110 --> 00:44:54,433
beginning. And so this this

1221
00:44:54,940 --> 00:44:57,010
method name really should be

1222
00:44:57,070 --> 00:44:58,070
different.

1223
00:44:58,180 --> 00:45:00,480
But nobody ever Reding's

1224
00:45:00,490 --> 00:45:01,813
methods because it's really

1225
00:45:02,440 --> 00:45:04,659
hard to go over a piece

1226
00:45:05,070 --> 00:45:07,030
of code and

1227
00:45:07,120 --> 00:45:08,780
rename. Exactly.

1228
00:45:08,960 --> 00:45:11,270
That's the right variable.

1229
00:45:11,370 --> 00:45:12,819
But if you've got a good refactoring

1230
00:45:12,820 --> 00:45:14,769
engine, you just you

1231
00:45:14,770 --> 00:45:16,436
just type, you know, medic control

1232
00:45:16,570 --> 00:45:18,340
are typing the new name,

1233
00:45:19,060 --> 00:45:20,829
you know, hundreds and hundreds of

1234
00:45:20,830 --> 00:45:22,660
files later, which takes about,

1235
00:45:23,080 --> 00:45:24,910
you know, maybe 30 seconds

1236
00:45:24,920 --> 00:45:26,289
if it's a lot of files.

1237
00:45:29,100 --> 00:45:30,100
You're done.

1238
00:45:30,710 --> 00:45:32,869
And so you basically be right.

1239
00:45:34,410 --> 00:45:35,909
Seven-part series,

1240
00:45:36,980 --> 00:45:37,980
I

1241
00:45:39,090 --> 00:45:41,070
hope this is this is actually like

1242
00:45:41,220 --> 00:45:42,919
you're describing, that's what

1243
00:45:43,080 --> 00:45:43,599
we what.

1244
00:45:43,600 --> 00:45:44,730
What was the genesis

1245
00:45:45,770 --> 00:45:47,790
of the typescript project, which was

1246
00:45:48,540 --> 00:45:50,590
these enormous JavaScript code basis

1247
00:45:50,640 --> 00:45:52,019
that we were seeing customers.

1248
00:45:52,020 --> 00:45:53,999
Right. And then in-house projects

1249
00:45:54,010 --> 00:45:55,259
that they were getting bigger and

1250
00:45:55,260 --> 00:45:56,789
bigger because JavaScript engines

1251
00:45:56,790 --> 00:45:58,139
were getting good enough for you to

1252
00:45:58,140 --> 00:45:59,669
write really big programs.

1253
00:46:00,200 --> 00:46:02,160
But maintaining them was impossible

1254
00:46:02,790 --> 00:46:04,819
because it turned into write

1255
00:46:04,860 --> 00:46:06,639
only code. You know, you do dare

1256
00:46:06,700 --> 00:46:08,159
not touch it. Once you've written it

1257
00:46:08,190 --> 00:46:10,020
because it worked then and then.

1258
00:46:10,350 --> 00:46:12,059
But, you know, there's this property

1259
00:46:12,060 --> 00:46:13,060
called text.

1260
00:46:13,440 --> 00:46:15,840
And I really want to change it to be

1261
00:46:16,400 --> 00:46:18,329
like some

1262
00:46:18,330 --> 00:46:20,159
other name, except there

1263
00:46:20,160 --> 00:46:21,479
is like a million other properties

1264
00:46:21,480 --> 00:46:23,244
called text. And if I'd just like to

1265
00:46:23,310 --> 00:46:24,639
do a global search and replace for

1266
00:46:24,640 --> 00:46:26,257
tax, then oh my God, it's like at

1267
00:46:26,460 --> 00:46:28,339
all. So I can't change anything.

1268
00:46:28,920 --> 00:46:30,599
But if you have semantic

1269
00:46:30,600 --> 00:46:32,315
understanding up like this property

1270
00:46:32,400 --> 00:46:34,019
call text is different from that

1271
00:46:34,050 --> 00:46:35,099
other property over here.

1272
00:46:35,100 --> 00:46:36,689
Call text. And if you want to rename

1273
00:46:36,690 --> 00:46:37,829
this one, that doesn't mean you want

1274
00:46:37,830 --> 00:46:38,830
to rename that one.

1275
00:46:39,080 --> 00:46:40,080
But but that takes

1276
00:46:41,010 --> 00:46:42,869
for something like a

1277
00:46:42,870 --> 00:46:44,639
type system to be in place.

1278
00:46:44,690 --> 00:46:46,454
And, you know, once you start adding

1279
00:46:46,800 --> 00:46:48,629
that, you add documentation to the

1280
00:46:48,630 --> 00:46:50,420
code and you add these capabilities

1281
00:46:50,430 --> 00:46:52,110
where we're now seeing people

1282
00:46:53,020 --> 00:46:55,139
like with regularity, refactor

1283
00:46:55,140 --> 00:46:56,699
one hundreds of thousands of lines

1284
00:46:56,700 --> 00:46:58,109
of JavaScript code, you know, we'd

1285
00:46:58,110 --> 00:46:59,730
like ten minutes

1286
00:47:00,280 --> 00:47:01,652
and end it and it just works

1287
00:47:02,280 --> 00:47:03,719
afterwards. And people are like

1288
00:47:03,720 --> 00:47:04,980
amazed that it's possible.

1289
00:47:06,130 --> 00:47:07,130
But but but

1290
00:47:08,040 --> 00:47:09,755
it's true. Said it takes a bit more

1291
00:47:10,050 --> 00:47:11,039
work to begin with.

1292
00:47:11,040 --> 00:47:12,420
And it's not maybe not right.

1293
00:47:12,450 --> 00:47:13,919
If you're writing five lines of code

1294
00:47:14,340 --> 00:47:15,989
that maybe more bother than you

1295
00:47:15,990 --> 00:47:16,979
really care to.

1296
00:47:16,980 --> 00:47:18,548
But but, you know, it started to

1297
00:47:18,840 --> 00:47:19,920
pay off pretty quickly.

1298
00:47:21,290 --> 00:47:22,710
There are different types

1299
00:47:23,280 --> 00:47:25,380
of really good lexical scoping

1300
00:47:25,710 --> 00:47:26,829
helps with refactoring.

1301
00:47:27,500 --> 00:47:29,400
It gets

1302
00:47:29,430 --> 00:47:30,430
back into this.

1303
00:47:31,410 --> 00:47:32,449
What's the right pick to hang the

1304
00:47:32,450 --> 00:47:33,820
sign up and off and it's slope.

1305
00:47:34,110 --> 00:47:35,820
It's a lexically skulked or

1306
00:47:35,840 --> 00:47:36,840
dynamically scale.

1307
00:47:36,850 --> 00:47:38,909
So now we're doing those right

1308
00:47:38,940 --> 00:47:40,369
and not interfering with threading.

1309
00:47:43,200 --> 00:47:45,269
So I have one more question

1310
00:47:45,270 --> 00:47:47,149
now before the break in

1311
00:47:47,150 --> 00:47:48,979
three minutes and we made

1312
00:47:49,050 --> 00:47:50,989
sort of a reference to Donald New,

1313
00:47:51,570 --> 00:47:52,920
one of the turning winners

1314
00:47:54,360 --> 00:47:55,739
in this last

1315
00:47:56,310 --> 00:47:57,840
section of conversation.

1316
00:47:57,930 --> 00:47:59,939
And he wrote, Premature

1317
00:47:59,970 --> 00:48:01,489
optimization is the root of all

1318
00:48:01,980 --> 00:48:02,980
evil.

1319
00:48:03,870 --> 00:48:04,997
So what's your approach

1320
00:48:05,880 --> 00:48:07,979
to optimizing code and gaining

1321
00:48:07,980 --> 00:48:09,860
efficiency, saying Python?

1322
00:48:12,520 --> 00:48:14,440
Well, I I'm terrible at that.

1323
00:48:14,470 --> 00:48:15,939
So I leave it to others.

1324
00:48:17,110 --> 00:48:18,110
That is

1325
00:48:19,410 --> 00:48:21,249
a very wise language creator

1326
00:48:21,910 --> 00:48:23,037
and they've lived up to

1327
00:48:23,860 --> 00:48:24,860
the task.

1328
00:48:26,200 --> 00:48:28,820
Piceance hash table implementation

1329
00:48:28,850 --> 00:48:30,980
has been rewritten so many times.

1330
00:48:32,590 --> 00:48:34,239
I don't understand any part of it,

1331
00:48:36,380 --> 00:48:38,229
but it is so much better than the

1332
00:48:38,230 --> 00:48:40,409
thing I actually copied out of truth

1333
00:48:41,230 --> 00:48:42,230
30 years ago.

1334
00:48:44,560 --> 00:48:45,729
Happy to have a clue.

1335
00:48:46,830 --> 00:48:47,870
Okey doke.

1336
00:48:52,790 --> 00:48:54,859
I kind of 50 percent believe that

1337
00:48:56,030 --> 00:48:57,889
because they're, you

1338
00:48:58,020 --> 00:48:59,809
know, the premature optimization in

1339
00:48:59,810 --> 00:49:01,369
terms of algorithms and code.

1340
00:49:01,910 --> 00:49:03,229
That part of it, I believe.

1341
00:49:04,280 --> 00:49:05,960
But people

1342
00:49:06,410 --> 00:49:07,939
often don't really think about the

1343
00:49:07,940 --> 00:49:09,349
role of data structures

1344
00:49:09,970 --> 00:49:11,030
in optimization

1345
00:49:11,900 --> 00:49:13,869
and interpret data structures

1346
00:49:13,940 --> 00:49:14,940
exposed.

1347
00:49:16,730 --> 00:49:17,780
You have to

1348
00:49:18,870 --> 00:49:20,420
know and you decide that, you know,

1349
00:49:20,990 --> 00:49:22,789
this algorithm is a problem.

1350
00:49:23,630 --> 00:49:25,369
But, you know, the reason that it's

1351
00:49:25,370 --> 00:49:27,169
a problem is because,

1352
00:49:27,800 --> 00:49:29,449
you know, this data structure was

1353
00:49:29,450 --> 00:49:30,920
just slapped together and they

1354
00:49:31,310 --> 00:49:32,719
thought, oh, they'd only ever be

1355
00:49:32,720 --> 00:49:34,610
like 10 items in the list

1356
00:49:35,090 --> 00:49:36,550
or 10 items in the set.

1357
00:49:37,220 --> 00:49:38,869
And then and then you start,

1358
00:49:39,130 --> 00:49:41,329
you know, having, you know, million

1359
00:49:41,330 --> 00:49:43,159
eight million items sets all the

1360
00:49:43,160 --> 00:49:44,160
time.

1361
00:49:44,300 --> 00:49:45,770
And all of a sudden, you know,

1362
00:49:45,800 --> 00:49:47,170
you're a little linkedlist

1363
00:49:47,660 --> 00:49:49,520
is. And that is not gonna happen.

1364
00:49:50,270 --> 00:49:50,819
Right.

1365
00:49:50,820 --> 00:49:52,600
Ex-directory or a unit

1366
00:49:52,640 --> 00:49:53,640
or. Yeah.

1367
00:49:53,910 --> 00:49:55,989
And and

1368
00:49:56,030 --> 00:49:57,980
so and so if you can you know,

1369
00:49:58,190 --> 00:50:00,320
it's like you know hyd

1370
00:50:00,350 --> 00:50:01,219
early hide.

1371
00:50:01,220 --> 00:50:02,220
Often.

1372
00:50:02,400 --> 00:50:03,400
Right.

1373
00:50:03,590 --> 00:50:05,449
Never tell the truth about what your

1374
00:50:05,450 --> 00:50:06,649
data structures are.

1375
00:50:09,320 --> 00:50:11,179
And on that note, we're

1376
00:50:11,180 --> 00:50:12,180
going to take

1377
00:50:13,100 --> 00:50:14,300
a quick look.

1378
00:50:14,330 --> 00:50:16,579
I just want to say that I

1379
00:50:16,580 --> 00:50:18,230
think like

1380
00:50:18,460 --> 00:50:20,479
with anything, the important thing

1381
00:50:20,480 --> 00:50:22,097
is to make sure that you actually

1382
00:50:22,580 --> 00:50:24,429
have the right data before

1383
00:50:24,430 --> 00:50:26,200
you you start optimizing.

1384
00:50:26,210 --> 00:50:28,099
Right. By too often you have

1385
00:50:28,100 --> 00:50:29,900
a hunch that, oh,

1386
00:50:29,930 --> 00:50:31,789
I think finally I need to do

1387
00:50:31,790 --> 00:50:32,219
this.

1388
00:50:32,220 --> 00:50:33,469
And I need to change this data

1389
00:50:33,470 --> 00:50:34,969
structure from being a linkedlist to

1390
00:50:34,970 --> 00:50:36,349
being overweight or whatever.

1391
00:50:36,980 --> 00:50:38,899
But but if you

1392
00:50:38,900 --> 00:50:40,129
measure, then you discover it

1393
00:50:40,130 --> 00:50:41,119
doesn't matter.

1394
00:50:41,120 --> 00:50:42,949
And but really, there's this other

1395
00:50:42,950 --> 00:50:44,269
thing that you haven't even thought

1396
00:50:44,270 --> 00:50:46,099
about that matters enormously.

1397
00:50:46,100 --> 00:50:47,509
And maybe you should go look over

1398
00:50:47,510 --> 00:50:49,489
there. So get perf.,

1399
00:50:49,550 --> 00:50:51,479
get it, use a profile or

1400
00:50:51,480 --> 00:50:53,149
figure out where it is, the time

1401
00:50:53,150 --> 00:50:54,079
actually spent.

1402
00:50:54,080 --> 00:50:55,909
I am continually surprised and

1403
00:50:55,910 --> 00:50:57,439
I write compilers for a living and

1404
00:50:57,440 --> 00:50:59,239
I'm always surprised by where

1405
00:50:59,330 --> 00:51:00,330
time is spent.

1406
00:51:00,370 --> 00:51:01,039
All right.

1407
00:51:01,040 --> 00:51:02,960
And that is that

1408
00:51:03,110 --> 00:51:04,670
we're out of time for this section.

1409
00:51:05,030 --> 00:51:06,739
We're going to take a quick 15

1410
00:51:06,740 --> 00:51:07,740
minute break.

1411
00:51:09,080 --> 00:51:10,999
And when we come

1412
00:51:11,000 --> 00:51:12,029
back, the really hard

1413
00:51:12,920 --> 00:51:14,210
questions are going to come out

1414
00:51:14,270 --> 00:51:16,280
because the audience questions

1415
00:51:16,890 --> 00:51:18,709
that have been submitted are

1416
00:51:18,710 --> 00:51:20,260
going to be the ones that believe in

1417
00:51:20,380 --> 00:51:21,380
you.

1418
00:51:22,020 --> 00:51:23,020
So thank you.

1419
00:51:55,280 --> 00:51:56,899
And teach in every one of you has a

1420
00:51:56,900 --> 00:51:58,419
community that is formed around

1421
00:51:58,760 --> 00:52:00,079
the languages that you've created.

1422
00:52:00,110 --> 00:52:02,389
I'm curious if you were talk two

1423
00:52:02,390 --> 00:52:03,979
ways in which you've seen the design

1424
00:52:03,980 --> 00:52:05,059
choices you've made for your

1425
00:52:05,060 --> 00:52:07,109
language, so shaping the communities

1426
00:52:07,110 --> 00:52:08,110
that are formed around.

1427
00:52:11,440 --> 00:52:12,449
It really helps.

1428
00:52:12,580 --> 00:52:14,379
Well, good.

1429
00:52:14,590 --> 00:52:15,610
It really helps

1430
00:52:16,870 --> 00:52:18,040
if you're trying to

1431
00:52:19,270 --> 00:52:21,190
make everybody look

1432
00:52:21,240 --> 00:52:22,690
what's here with your slogan here.

1433
00:52:24,380 --> 00:52:25,970
Everybody comfortable?

1434
00:52:28,070 --> 00:52:29,860
We're doing that right now.

1435
00:52:30,600 --> 00:52:32,349
They think they can make

1436
00:52:32,840 --> 00:52:34,769
all of all Americans, I heard.

1437
00:52:37,920 --> 00:52:39,299
You know, feel like comfortable

1438
00:52:39,360 --> 00:52:40,360
programing.

1439
00:52:40,870 --> 00:52:42,050
You see it for all.

1440
00:52:42,150 --> 00:52:43,969
Yes. To make sense of

1441
00:52:43,980 --> 00:52:45,059
it. I guess you know

1442
00:52:46,120 --> 00:52:47,120
what?

1443
00:52:48,190 --> 00:52:49,600
The sign was a clue.

1444
00:52:51,250 --> 00:52:52,769
It's not see us for all of us, just

1445
00:52:52,770 --> 00:52:53,770
Americans.

1446
00:52:54,140 --> 00:52:55,590
No. There are many underserved

1447
00:52:56,520 --> 00:52:57,639
populations in the world.

1448
00:52:57,960 --> 00:52:59,879
And what we've really discovered

1449
00:52:59,910 --> 00:53:01,979
that it really helps

1450
00:53:01,980 --> 00:53:03,390
build international

1451
00:53:06,540 --> 00:53:08,070
camaraderie and community

1452
00:53:08,700 --> 00:53:10,499
to have world class

1453
00:53:11,010 --> 00:53:12,650
literally support for unico.

1454
00:53:13,440 --> 00:53:15,155
Yes. I would encourage any language

1455
00:53:15,330 --> 00:53:16,349
designers out there.

1456
00:53:16,920 --> 00:53:18,579
Don't do this halfway stuff with,

1457
00:53:18,830 --> 00:53:20,760
you know, code points.

1458
00:53:21,200 --> 00:53:22,200
Go all the graphics.

1459
00:53:23,160 --> 00:53:25,469
I completely agree.

1460
00:53:26,850 --> 00:53:28,109
Kita, you had something you were

1461
00:53:28,110 --> 00:53:29,110
going to say.

1462
00:53:29,940 --> 00:53:31,950
I was trying to say.

1463
00:53:33,280 --> 00:53:34,590
I how my

1464
00:53:35,160 --> 00:53:37,360
design choices necessary

1465
00:53:38,390 --> 00:53:39,390
to make

1466
00:53:40,570 --> 00:53:42,520
choices about how to use

1467
00:53:43,000 --> 00:53:44,200
energy certainly have.

1468
00:53:46,260 --> 00:53:47,260
Did

1469
00:53:48,250 --> 00:53:49,250
I?

1470
00:53:49,390 --> 00:53:51,309
I was somehow a dude

1471
00:53:51,510 --> 00:53:53,269
with the importance of

1472
00:53:54,260 --> 00:53:56,119
user centered based something

1473
00:53:56,630 --> 00:53:58,219
because by comparison.

1474
00:53:59,000 --> 00:54:01,189
That was ABC long

1475
00:54:01,370 --> 00:54:02,370
lost,

1476
00:54:04,260 --> 00:54:06,409
its authors had a of things

1477
00:54:06,410 --> 00:54:08,209
right, and one of those was

1478
00:54:08,780 --> 00:54:10,219
listen to the users

1479
00:54:10,850 --> 00:54:12,560
asking users what they

1480
00:54:12,710 --> 00:54:14,559
think. And then again, you making

1481
00:54:14,680 --> 00:54:16,052
a new one of their suggested

1482
00:54:16,580 --> 00:54:17,359
solutions.

1483
00:54:17,360 --> 00:54:18,360
And you first figure

1484
00:54:19,190 --> 00:54:20,699
out what nobody spends.

1485
00:54:21,530 --> 00:54:23,330
You gave users a voice.

1486
00:54:23,450 --> 00:54:24,699
And I said,

1487
00:54:25,610 --> 00:54:26,884
take that back and let the

1488
00:54:27,440 --> 00:54:28,879
users help.

1489
00:54:30,680 --> 00:54:32,829
And now they're raising of.

1490
00:54:36,790 --> 00:54:38,379
I was going to say I always felt

1491
00:54:38,380 --> 00:54:39,380
like

1492
00:54:40,630 --> 00:54:42,140
it's important as a language society

1493
00:54:42,170 --> 00:54:43,787
that you don't create unnecessary

1494
00:54:44,140 --> 00:54:45,904
work for your community because it's

1495
00:54:46,030 --> 00:54:47,451
it's very easy to in the name

1496
00:54:47,860 --> 00:54:49,149
of purity or whatever.

1497
00:54:49,180 --> 00:54:51,309
Go change something

1498
00:54:51,310 --> 00:54:53,169
or or strive

1499
00:54:53,170 --> 00:54:54,500
for the perfect language.

1500
00:54:54,560 --> 00:54:55,870
And I've always said, like, I mean,

1501
00:54:57,770 --> 00:54:59,109
show me the perfect language and

1502
00:54:59,110 --> 00:55:00,249
I'll show you your language with no

1503
00:55:00,250 --> 00:55:01,250
use or shoot over.

1504
00:55:01,300 --> 00:55:03,189
Every every language

1505
00:55:03,190 --> 00:55:04,500
has its functions.

1506
00:55:05,280 --> 00:55:06,620
Has them because.

1507
00:55:07,260 --> 00:55:09,150
Because it has evolved.

1508
00:55:09,240 --> 00:55:10,955
And the hard part of language aside

1509
00:55:11,040 --> 00:55:12,689
is actually not version one.

1510
00:55:12,960 --> 00:55:14,219
It's like the older versions that

1511
00:55:14,220 --> 00:55:16,230
come after you

1512
00:55:16,580 --> 00:55:18,719
try to sneak things

1513
00:55:18,720 --> 00:55:20,969
in without actually

1514
00:55:21,210 --> 00:55:23,249
causing extra burdens on

1515
00:55:23,250 --> 00:55:24,689
people who don't necessarily care

1516
00:55:24,690 --> 00:55:25,829
about the things that you're trying

1517
00:55:25,830 --> 00:55:26,699
to sneak in. Right.

1518
00:55:26,700 --> 00:55:27,629
And so.

1519
00:55:27,630 --> 00:55:29,550
So that's the hard part

1520
00:55:29,910 --> 00:55:31,590
of language design, I think

1521
00:55:31,770 --> 00:55:33,719
is is the evolution

1522
00:55:33,810 --> 00:55:34,810
of the language and

1523
00:55:35,730 --> 00:55:37,396
keeping your community with you as

1524
00:55:37,500 --> 00:55:38,500
you have all.

1525
00:55:40,880 --> 00:55:42,630
Yeah. And so and so with with

1526
00:55:42,930 --> 00:55:43,930
Jarbawi like always

1527
00:55:45,150 --> 00:55:47,099
followed that we've

1528
00:55:47,100 --> 00:55:48,717
maybe been a little bit excessive

1529
00:55:49,590 --> 00:55:50,590
about it.

1530
00:55:51,420 --> 00:55:52,890
I mean, I've got binaries that

1531
00:55:53,250 --> 00:55:54,570
I compiled like

1532
00:55:55,680 --> 00:55:56,820
20 years ago,

1533
00:55:57,550 --> 00:55:58,650
but still run.

1534
00:55:59,490 --> 00:56:01,349
And and it's it's, you

1535
00:56:01,350 --> 00:56:02,967
know, on machines that were never

1536
00:56:03,120 --> 00:56:05,159
invented when the code was compiled.

1537
00:56:07,200 --> 00:56:09,380
And you only get there with

1538
00:56:09,390 --> 00:56:11,070
with crazy discipline

1539
00:56:11,640 --> 00:56:12,669
and and a a tolerance

1540
00:56:14,010 --> 00:56:15,750
for living with your mistakes.

1541
00:56:17,420 --> 00:56:19,260
But but it but it makes the

1542
00:56:19,420 --> 00:56:20,420
you.

1543
00:56:20,460 --> 00:56:21,329
It also help.

1544
00:56:21,330 --> 00:56:23,159
It also ends up selecting who

1545
00:56:23,160 --> 00:56:24,659
your users are

1546
00:56:25,210 --> 00:56:26,789
because in the in the Java universe,

1547
00:56:26,820 --> 00:56:28,143
pretty much everybody is is

1548
00:56:28,680 --> 00:56:30,599
really disciplined because,

1549
00:56:30,630 --> 00:56:31,829
you know, it's kind of like like

1550
00:56:31,920 --> 00:56:32,920
mountain climbing.

1551
00:56:32,950 --> 00:56:34,077
You know, you you don't

1552
00:56:34,770 --> 00:56:36,599
get there, get

1553
00:56:36,600 --> 00:56:38,217
sloppy with your gear when you're

1554
00:56:38,340 --> 00:56:39,340
mountain climbing

1555
00:56:40,700 --> 00:56:42,599
because it has a clear price.

1556
00:56:45,030 --> 00:56:47,039
And, you

1557
00:56:47,040 --> 00:56:48,167
know, when you you said

1558
00:56:48,930 --> 00:56:49,930
something about

1559
00:56:51,380 --> 00:56:52,619
Larry, you said it said something

1560
00:56:52,620 --> 00:56:53,796
about Unicode Java's had

1561
00:56:54,570 --> 00:56:56,099
Unicode since ninety two.

1562
00:56:56,700 --> 00:56:58,500
So, you know, it's,

1563
00:56:58,580 --> 00:56:59,990
you know, trying to be inclusive.

1564
00:57:01,110 --> 00:57:03,030
Yeah. And and that

1565
00:57:03,090 --> 00:57:05,070
really made a difference.

1566
00:57:05,850 --> 00:57:07,565
I mean lots of folks thought it was

1567
00:57:07,590 --> 00:57:09,419
like the lamest thing ever.

1568
00:57:10,200 --> 00:57:11,200
But

1569
00:57:12,240 --> 00:57:14,099
that was that that was during

1570
00:57:14,100 --> 00:57:15,330
my my, you know,

1571
00:57:16,230 --> 00:57:17,230
Lord of Java

1572
00:57:18,060 --> 00:57:19,060
phase which.

1573
00:57:20,190 --> 00:57:21,190
Yeah.

1574
00:57:21,460 --> 00:57:23,224
Yeah. But that that phase of my life

1575
00:57:23,400 --> 00:57:24,829
ended in about ninety five.

1576
00:57:25,800 --> 00:57:26,619
I'd like to say something

1577
00:57:26,620 --> 00:57:27,790
possibilitythat.

1578
00:57:30,150 --> 00:57:32,159
And that is we did this

1579
00:57:32,900 --> 00:57:34,769
perfect try to perfect

1580
00:57:35,310 --> 00:57:36,310
compatibility.

1581
00:57:36,540 --> 00:57:37,540
All the way through.

1582
00:57:37,560 --> 00:57:38,630
I didn't really

1583
00:57:40,200 --> 00:57:41,200
think.

1584
00:57:45,360 --> 00:57:47,099
That eventually runs into

1585
00:57:47,850 --> 00:57:49,569
the problem. But, you know, you

1586
00:57:49,620 --> 00:57:51,300
can't I can't fix your mistakes.

1587
00:57:51,930 --> 00:57:53,909
So one of the mistakes

1588
00:57:54,120 --> 00:57:55,640
that we made was was kind of a

1589
00:57:55,740 --> 00:57:57,389
mistake and that was assuming

1590
00:57:58,200 --> 00:58:00,179
that you had to always have

1591
00:58:00,180 --> 00:58:02,219
that kind of stability

1592
00:58:02,460 --> 00:58:03,630
and not change anything.

1593
00:58:04,890 --> 00:58:07,020
How would it be to design a language

1594
00:58:07,500 --> 00:58:09,089
where the language itself

1595
00:58:10,180 --> 00:58:12,120
could be evolved

1596
00:58:12,270 --> 00:58:13,270
from within

1597
00:58:14,290 --> 00:58:16,140
by the community

1598
00:58:16,620 --> 00:58:17,699
and the language we want?

1599
00:58:17,930 --> 00:58:19,639
Twenty five or 50 years from now,

1600
00:58:19,650 --> 00:58:21,449
this ties into what Polygram was

1601
00:58:21,450 --> 00:58:22,559
talking about, a hundred year

1602
00:58:22,590 --> 00:58:24,209
language. We kind of took that

1603
00:58:24,210 --> 00:58:25,210
seriously.

1604
00:58:26,040 --> 00:58:27,040
What is it that

1605
00:58:27,990 --> 00:58:29,879
prevents languages from evolving,

1606
00:58:30,490 --> 00:58:32,589
not being not having things still

1607
00:58:32,610 --> 00:58:33,610
thrive?

1608
00:58:34,710 --> 00:58:35,710
But

1609
00:58:36,630 --> 00:58:38,296
fundamentally, the Perl 6 compiler

1610
00:58:38,760 --> 00:58:40,050
is written in Perl 6.

1611
00:58:40,080 --> 00:58:41,789
Most the runtime system is written

1612
00:58:41,790 --> 00:58:42,769
in Perl 6.

1613
00:58:42,770 --> 00:58:44,099
The users can extend it.

1614
00:58:45,150 --> 00:58:47,219
Perl 6 itself does not care

1615
00:58:47,580 --> 00:58:49,260
whether things are built in or not.

1616
00:58:49,890 --> 00:58:51,099
It all works the same.

1617
00:58:51,450 --> 00:58:53,189
So most of the buildings are written

1618
00:58:53,190 --> 00:58:54,300
as if they were used or code.

1619
00:58:55,020 --> 00:58:57,269
So users, even if we say

1620
00:58:57,460 --> 00:58:58,460
no, we don't

1621
00:58:59,280 --> 00:59:00,420
want that language right now.

1622
00:59:00,450 --> 00:59:02,069
But put it in a module

1623
00:59:03,090 --> 00:59:04,090
you can do.

1624
00:59:04,140 --> 00:59:05,904
You can turn classes in two monitors

1625
00:59:06,090 --> 00:59:07,419
with about that much code.

1626
00:59:07,770 --> 00:59:09,499
You can put into implement actor

1627
00:59:09,520 --> 00:59:10,990
model classes with about about

1628
00:59:11,550 --> 00:59:12,550
how much code

1629
00:59:14,070 --> 00:59:15,269
you can you take.

1630
00:59:15,510 --> 00:59:17,399
Adding operators is just trivial.

1631
00:59:18,690 --> 00:59:20,579
When it does, it

1632
00:59:20,820 --> 00:59:22,979
does everything perfectly one pass

1633
00:59:22,980 --> 00:59:24,530
and drops into the sub compiler.

1634
00:59:24,960 --> 00:59:26,590
And with the end, it comes back out.

1635
00:59:26,880 --> 00:59:28,869
So there's never, ever

1636
00:59:28,900 --> 00:59:30,517
any ambiguity as to what language

1637
00:59:31,020 --> 00:59:32,020
you are in.

1638
00:59:32,820 --> 00:59:34,469
And so

1639
00:59:35,430 --> 00:59:37,210
because this scoping

1640
00:59:37,350 --> 00:59:38,879
of the exact language you're in is

1641
00:59:38,880 --> 00:59:40,889
always like if we scoped

1642
00:59:41,010 --> 00:59:42,510
any lexical scope can say

1643
00:59:43,290 --> 00:59:45,119
use whatever language I want a

1644
00:59:45,120 --> 00:59:46,760
feature version of Perl 6.

1645
00:59:46,860 --> 00:59:49,079
You could say use COBOL, use Python.

1646
00:59:49,380 --> 00:59:50,909
If someone actually people have

1647
00:59:50,910 --> 00:59:52,739
written Python parsers in

1648
00:59:52,870 --> 00:59:54,480
April 6, I don't think anybody's

1649
00:59:54,510 --> 00:59:55,619
going to call for sure yet,

1650
00:59:57,390 --> 00:59:58,899
but we'd like to think that, you

1651
00:59:58,900 --> 01:00:00,820
know, a lot of language

1652
01:00:00,890 --> 01:00:02,279
stuff like languages and Perl 6.

1653
01:00:03,030 --> 01:00:04,050
So I think that

1654
01:00:06,190 --> 01:00:08,149
I think the third module

1655
01:00:08,160 --> 01:00:09,440
of semantics would

1656
01:00:10,350 --> 01:00:11,350
likely.

1657
01:00:11,970 --> 01:00:12,970
But I think

1658
01:00:13,800 --> 01:00:14,800
there is a way us.

1659
01:00:15,590 --> 01:00:16,889
I think there is a way forward to

1660
01:00:16,890 --> 01:00:18,030
get the stability and

1661
01:00:18,040 --> 01:00:19,040
rehabilitation.

1662
01:00:19,310 --> 01:00:20,310
And I think.

1663
01:00:23,200 --> 01:00:24,200
Awesome.

1664
01:00:24,400 --> 01:00:26,115
Do we want to take another audience

1665
01:00:26,170 --> 01:00:27,170
question?

1666
01:00:28,960 --> 01:00:29,960
All right.

1667
01:00:31,650 --> 01:00:32,650
Thanks.

1668
01:00:36,650 --> 01:00:38,089
Can you talk into the mike like

1669
01:00:38,090 --> 01:00:38,899
you're eating the mike?

1670
01:00:38,900 --> 01:00:39,900
Hi. Yeah.

1671
01:00:39,980 --> 01:00:40,980
Can you hear me?

1672
01:00:41,190 --> 01:00:42,190
Yeah.

1673
01:00:42,380 --> 01:00:43,879
There's been a lot of talk tonight

1674
01:00:43,880 --> 01:00:45,559
about the evolve ability of

1675
01:00:45,560 --> 01:00:46,999
languages and

1676
01:00:47,510 --> 01:00:48,931
the trouble with implementing

1677
01:00:49,820 --> 01:00:51,169
things that might not be backwards

1678
01:00:51,170 --> 01:00:52,170
compatible.

1679
01:00:52,220 --> 01:00:53,899
What's something that you wish you

1680
01:00:53,900 --> 01:00:55,009
could have implemented with your

1681
01:00:55,010 --> 01:00:56,869
language? But for that, or

1682
01:00:56,870 --> 01:00:58,039
maybe another reason that wasn't

1683
01:00:58,040 --> 01:00:59,040
possible.

1684
01:01:02,160 --> 01:01:03,160
A

1685
01:01:06,300 --> 01:01:07,449
word,

1686
01:01:11,090 --> 01:01:12,090
right?

1687
01:01:13,370 --> 01:01:14,719
As all the

1688
01:01:15,820 --> 01:01:16,820
words

1689
01:01:19,070 --> 01:01:20,829
or these

1690
01:01:22,540 --> 01:01:23,689
or other forms of

1691
01:01:28,440 --> 01:01:29,440
naming is hard.

1692
01:01:31,690 --> 01:01:33,039
Well, I'll I'll I'll go.

1693
01:01:33,040 --> 01:01:34,989
I've. I mean I my favorite is always

1694
01:01:35,200 --> 01:01:37,059
like d that the two billion dollar

1695
01:01:37,060 --> 01:01:38,060
mistake of the

1696
01:01:38,920 --> 01:01:40,269
billion dollar mistake of having

1697
01:01:40,270 --> 01:01:41,379
Nall in the language.

1698
01:01:41,400 --> 01:01:42,400
And since JavaScript

1699
01:01:43,540 --> 01:01:45,009
has both null and undefined, it's

1700
01:01:45,010 --> 01:01:46,630
the two billion dollar mistake.

1701
01:01:48,730 --> 01:01:49,730
But

1702
01:01:50,770 --> 01:01:51,820
I mean if it did.

1703
01:01:52,320 --> 01:01:54,219
And I mean what's done is done

1704
01:01:54,220 --> 01:01:56,169
right. And and now now,

1705
01:01:56,230 --> 01:01:58,539
you know, I spend a significant

1706
01:01:58,540 --> 01:02:00,699
portion of my time actually working

1707
01:02:00,700 --> 01:02:02,910
on ways to

1708
01:02:04,110 --> 01:02:06,310
to make code null safe

1709
01:02:06,880 --> 01:02:07,880
to and who's

1710
01:02:08,770 --> 01:02:10,749
never who in here has ever had

1711
01:02:10,750 --> 01:02:11,949
a null pointer exception.

1712
01:02:12,010 --> 01:02:13,179
Right. There you go.

1713
01:02:14,530 --> 01:02:16,780
It's it is by far the most

1714
01:02:16,960 --> 01:02:17,960
problematic part of

1715
01:02:18,820 --> 01:02:20,439
language design. It it's a single

1716
01:02:20,440 --> 01:02:21,440
value that

1717
01:02:22,560 --> 01:02:24,280
that if only that wasn't there.

1718
01:02:25,060 --> 01:02:26,679
Imagine all the problems we wouldn't

1719
01:02:26,680 --> 01:02:28,199
have. Right. If if type systems

1720
01:02:28,570 --> 01:02:30,139
were designed that way and some type

1721
01:02:30,140 --> 01:02:31,757
systems are and some type systems

1722
01:02:32,020 --> 01:02:33,279
are getting there. But boy, trying

1723
01:02:33,280 --> 01:02:35,139
to retrofit that on top of a

1724
01:02:35,140 --> 01:02:36,806
type system that has normal in the

1725
01:02:36,820 --> 01:02:38,739
first place is

1726
01:02:39,400 --> 01:02:40,400
quite an undertaking.

1727
01:02:41,740 --> 01:02:43,357
The features I wanted to add were

1728
01:02:43,360 --> 01:02:44,360
negative features.

1729
01:02:46,660 --> 01:02:48,081
I think all of us as language

1730
01:02:48,580 --> 01:02:50,739
designers have borrowed

1731
01:02:50,740 --> 01:02:51,729
things.

1732
01:02:51,730 --> 01:02:53,259
You know, we all steal from each

1733
01:02:53,260 --> 01:02:54,550
other's languages all the time.

1734
01:02:55,300 --> 01:02:57,230
And and

1735
01:02:57,920 --> 01:02:59,096
it's often we steal good

1736
01:02:59,920 --> 01:03:00,279
things.

1737
01:03:00,280 --> 01:03:02,229
But for some reason, we also steal

1738
01:03:02,230 --> 01:03:03,230
bad things.

1739
01:03:05,580 --> 01:03:06,580
You know, like what?

1740
01:03:07,160 --> 01:03:08,750
Like regular expression syntax.

1741
01:03:08,850 --> 01:03:10,570
Oh, yeah.

1742
01:03:11,480 --> 01:03:12,379
I did that one.

1743
01:03:12,380 --> 01:03:14,150
Like the C precedence table.

1744
01:03:14,750 --> 01:03:15,750
OK. Another.

1745
01:03:15,870 --> 01:03:16,870
OK.

1746
01:03:16,880 --> 01:03:18,049
These are things that I could not

1747
01:03:18,050 --> 01:03:19,090
fix in four or five.

1748
01:03:20,450 --> 01:03:21,289
And we did fix it.

1749
01:03:21,290 --> 01:03:22,290
Burleson's.

1750
01:03:24,620 --> 01:03:25,620
Awesome.

1751
01:03:25,940 --> 01:03:26,940
All right.

1752
01:03:27,140 --> 01:03:28,806
We do tend to be doing pretty well

1753
01:03:29,060 --> 01:03:29,989
with the audience. So how would

1754
01:03:29,990 --> 01:03:31,249
another audience question?

1755
01:03:36,680 --> 01:03:37,680
Hi. Hello there.

1756
01:03:38,280 --> 01:03:40,399
So, well, what I wanted to ask

1757
01:03:40,400 --> 01:03:42,319
is you can sort of know what is

1758
01:03:42,770 --> 01:03:44,690
these sort of pendulum effect

1759
01:03:45,200 --> 01:03:47,269
in regards to tech and programing,

1760
01:03:47,500 --> 01:03:48,530
throw it out. Time

1761
01:03:49,430 --> 01:03:50,929
to Vinick's several different

1762
01:03:50,930 --> 01:03:51,859
paradigms.

1763
01:03:51,860 --> 01:03:53,219
There's a certain time where people

1764
01:03:53,230 --> 01:03:55,039
were like, are we going to do things

1765
01:03:55,040 --> 01:03:55,959
imperatively?

1766
01:03:55,960 --> 01:03:57,439
Are we going to go object oriented

1767
01:03:57,440 --> 01:03:58,440
or functional?

1768
01:03:59,770 --> 01:04:01,729
And for example, right now there's

1769
01:04:01,730 --> 01:04:03,130
a whole bunch of language used are

1770
01:04:03,140 --> 01:04:04,610
sort of like very aggressively

1771
01:04:05,480 --> 01:04:07,280
taking that standpoint of

1772
01:04:07,820 --> 01:04:09,589
being very pro,

1773
01:04:09,890 --> 01:04:12,409
being friendly with concurrency

1774
01:04:12,410 --> 01:04:14,270
or being bery

1775
01:04:14,720 --> 01:04:16,429
regardless of what immutability with

1776
01:04:16,430 --> 01:04:17,430
memory.

1777
01:04:17,600 --> 01:04:19,070
And I think that that kind of

1778
01:04:19,160 --> 01:04:20,287
pendulum effect happens

1779
01:04:21,080 --> 01:04:22,999
because technology has actually

1780
01:04:23,000 --> 01:04:24,339
been evolving throughout time.

1781
01:04:24,920 --> 01:04:26,749
Our machines are like beefier and

1782
01:04:26,750 --> 01:04:28,010
we have more memory now.

1783
01:04:28,520 --> 01:04:29,520
So that waste on we

1784
01:04:30,380 --> 01:04:32,239
design programing languages now

1785
01:04:32,790 --> 01:04:34,099
are probably a lot different than

1786
01:04:34,100 --> 01:04:36,079
there were 20 or 30 years

1787
01:04:36,080 --> 01:04:37,429
ago. So

1788
01:04:37,970 --> 01:04:39,919
considering that, where do you think

1789
01:04:39,980 --> 01:04:42,019
or how do you think the programing

1790
01:04:42,020 --> 01:04:43,579
languages of 10 years into the

1791
01:04:43,580 --> 01:04:44,630
future are going to look like?

1792
01:04:45,290 --> 01:04:46,999
Or in your opinion,

1793
01:04:48,070 --> 01:04:50,239
where where is language you design

1794
01:04:50,510 --> 01:04:52,489
headed to in the future?

1795
01:04:53,220 --> 01:04:54,079
So. So.

1796
01:04:54,080 --> 01:04:56,210
So my favorite candidate for that

1797
01:04:57,020 --> 01:04:58,909
is that, you

1798
01:04:58,910 --> 01:05:00,919
know, these, you know, major breaks

1799
01:05:00,920 --> 01:05:02,299
and things like that always happen

1800
01:05:02,300 --> 01:05:03,672
because, you know, something

1801
01:05:04,280 --> 01:05:05,799
major happens in the underlying

1802
01:05:05,870 --> 01:05:06,870
technology.

1803
01:05:07,550 --> 01:05:09,469
And, you know, one

1804
01:05:09,470 --> 01:05:10,579
of the areas

1805
01:05:11,540 --> 01:05:12,540
of underlying

1806
01:05:13,550 --> 01:05:15,069
technology that is kind of like

1807
01:05:15,650 --> 01:05:17,750
a computer that

1808
01:05:17,930 --> 01:05:20,420
I think is really underserved is

1809
01:05:20,900 --> 01:05:22,880
is writing code for GP use.

1810
01:05:24,000 --> 01:05:25,520
Right. I mean, right now there are

1811
01:05:26,270 --> 01:05:28,190
there are no languages

1812
01:05:28,550 --> 01:05:29,550
worth that crap.

1813
01:05:30,500 --> 01:05:32,719
And that's a technical for trademark

1814
01:05:35,010 --> 01:05:37,070
that work well

1815
01:05:37,160 --> 01:05:38,389
on GP use.

1816
01:05:38,960 --> 01:05:41,020
And, you know, maybe because there's

1817
01:05:41,030 --> 01:05:42,030
like a limited

1818
01:05:43,160 --> 01:05:44,929
number of algorithms that people are

1819
01:05:44,930 --> 01:05:46,498
willing to are really interested

1820
01:05:46,820 --> 01:05:48,049
in, but they're really important

1821
01:05:48,050 --> 01:05:48,889
ones.

1822
01:05:48,890 --> 01:05:50,479
Or maybe it's because they're

1823
01:05:50,540 --> 01:05:52,640
they're inherently all mathematical.

1824
01:05:53,330 --> 01:05:55,039
And the number of people who care

1825
01:05:55,040 --> 01:05:56,929
about writing numerical code

1826
01:05:56,960 --> 01:05:58,699
is small.

1827
01:06:00,080 --> 01:06:01,670
I'm one of those, though.

1828
01:06:02,160 --> 01:06:03,826
And you know, and so every now and

1829
01:06:03,980 --> 01:06:05,749
then, I spend time thinking about,

1830
01:06:06,350 --> 01:06:07,569
you know, what would you do to make

1831
01:06:07,570 --> 01:06:08,900
GPE you programing

1832
01:06:09,530 --> 01:06:10,530
really easy.

1833
01:06:17,550 --> 01:06:19,429
Yeah, but but but it's more than

1834
01:06:19,430 --> 01:06:20,749
just vector operators.

1835
01:06:20,800 --> 01:06:22,639
I mean that that's kind of my my

1836
01:06:22,640 --> 01:06:24,109
problem with the whole thing is that

1837
01:06:24,110 --> 01:06:26,239
most of these things are just

1838
01:06:26,360 --> 01:06:28,309
libraries of hand coded

1839
01:06:28,310 --> 01:06:30,840
assembly that do vector operators.

1840
01:06:31,850 --> 01:06:33,889
And and surely there's

1841
01:06:33,890 --> 01:06:35,510
something more clever than that.

1842
01:06:38,960 --> 01:06:40,439
Maybe I'll just add one with

1843
01:06:40,460 --> 01:06:41,509
language design. You know, one of

1844
01:06:41,510 --> 01:06:42,409
one of the things that's interesting,

1845
01:06:42,410 --> 01:06:43,879
you look at like all of us old

1846
01:06:43,880 --> 01:06:45,209
geezers sitting up here and there

1847
01:06:45,230 --> 01:06:46,969
they were, we're proof positive that

1848
01:06:46,970 --> 01:06:48,219
languages move slowly.

1849
01:06:49,860 --> 01:06:51,800
And I mean, it's not

1850
01:06:51,890 --> 01:06:52,890
you know, like

1851
01:06:53,930 --> 01:06:55,339
a lot of people make the mistake of

1852
01:06:55,340 --> 01:06:56,839
thinking that languages move at the

1853
01:06:56,840 --> 01:06:58,159
same speed as like

1854
01:06:58,730 --> 01:07:00,619
hardware or like all of the other

1855
01:07:00,620 --> 01:07:02,480
technologies that that that

1856
01:07:02,600 --> 01:07:04,190
we live with. But but

1857
01:07:04,670 --> 01:07:06,379
languages are much more like math

1858
01:07:06,410 --> 01:07:08,329
and much more like the human brain.

1859
01:07:08,360 --> 01:07:10,049
And, you know, and they all evolve

1860
01:07:10,250 --> 01:07:11,879
slowly. And we're still programing

1861
01:07:11,880 --> 01:07:13,595
and languages that were invented 50

1862
01:07:13,880 --> 01:07:15,019
years ago, like functional.

1863
01:07:15,020 --> 01:07:16,159
But all of the principles of

1864
01:07:16,160 --> 01:07:17,599
functional programing were thought

1865
01:07:17,600 --> 01:07:19,070
of more than 50 years ago.

1866
01:07:20,420 --> 01:07:22,369
I do think one of the things that

1867
01:07:22,370 --> 01:07:24,469
that is luckily happening is

1868
01:07:24,470 --> 01:07:26,140
that like as Larry said, everyone's

1869
01:07:26,150 --> 01:07:27,739
boring from everyone and languages

1870
01:07:27,740 --> 01:07:29,760
are becoming more multi paradigm.

1871
01:07:29,810 --> 01:07:31,699
Yeah, I think it's wrong to talk

1872
01:07:31,700 --> 01:07:33,559
about oh, I only

1873
01:07:33,560 --> 01:07:35,079
like object oriented programing

1874
01:07:35,150 --> 01:07:37,099
languages or I only like imperative

1875
01:07:37,100 --> 01:07:38,960
programing or functional programing.

1876
01:07:38,990 --> 01:07:40,520
I mean, it's important to

1877
01:07:41,300 --> 01:07:42,829
look at where's the research or

1878
01:07:42,850 --> 01:07:44,149
where the new where's the new

1879
01:07:44,150 --> 01:07:45,709
thinking and where where are new

1880
01:07:45,710 --> 01:07:47,419
paradigms that are interesting and

1881
01:07:47,420 --> 01:07:49,400
then try to try to incorporate them.

1882
01:07:49,430 --> 01:07:51,320
But do so tastefully

1883
01:07:51,680 --> 01:07:53,809
in a sense and and work them in to

1884
01:07:53,900 --> 01:07:55,615
work them into whatever is is there

1885
01:07:56,060 --> 01:07:56,669
already.

1886
01:07:56,670 --> 01:07:58,336
And and I think we're all learning

1887
01:07:58,730 --> 01:08:00,109
a lot from functional programing

1888
01:08:00,110 --> 01:08:00,979
languages these days.

1889
01:08:00,980 --> 01:08:03,139
I certainly feel like I am because

1890
01:08:03,140 --> 01:08:05,089
a lot of interesting research

1891
01:08:05,090 --> 01:08:06,090
has happened there.

1892
01:08:06,290 --> 01:08:07,440
But functional programing

1893
01:08:08,210 --> 01:08:09,337
is imperfect and no one

1894
01:08:10,070 --> 01:08:11,690
writes pure functional programs.

1895
01:08:11,750 --> 01:08:13,429
I mean because they they don't

1896
01:08:13,430 --> 01:08:14,606
exist. So it's all about

1897
01:08:15,260 --> 01:08:17,270
like how can you tastefully

1898
01:08:17,399 --> 01:08:18,399
sneak in mutation in

1899
01:08:19,340 --> 01:08:21,289
ways that you can better reason

1900
01:08:21,290 --> 01:08:23,119
about as opposed to

1901
01:08:23,750 --> 01:08:25,879
mutation and free threading for

1902
01:08:25,880 --> 01:08:26,818
everyone, you know?

1903
01:08:26,819 --> 01:08:28,579
And that's like just a

1904
01:08:28,970 --> 01:08:30,259
recipe for disaster.

1905
01:08:30,290 --> 01:08:30,659
Right.

1906
01:08:30,660 --> 01:08:31,660
So,

1907
01:08:32,740 --> 01:08:33,740
yes.

1908
01:08:40,560 --> 01:08:41,999
All languages are imperfect.

1909
01:08:42,029 --> 01:08:43,439
Let's just settle it right there.

1910
01:08:43,540 --> 01:08:44,540
You know.

1911
01:09:04,760 --> 01:09:06,619
Yes, we've been

1912
01:09:06,620 --> 01:09:07,959
thinking a lot about the

1913
01:09:08,470 --> 01:09:10,550
the the parallelism

1914
01:09:10,580 --> 01:09:11,959
concurrency issue.

1915
01:09:12,770 --> 01:09:14,599
This is one of those issues where we

1916
01:09:14,600 --> 01:09:15,739
were waiting to be smarter.

1917
01:09:17,229 --> 01:09:18,229
Uh, we we

1918
01:09:19,130 --> 01:09:20,130
did borrow.

1919
01:09:21,050 --> 01:09:22,399
I mean, steal a bunch of

1920
01:09:23,029 --> 01:09:24,529
ideas from Haskell

1921
01:09:24,920 --> 01:09:26,194
in terms of lazy lists and

1922
01:09:26,750 --> 01:09:28,579
things like that. But in

1923
01:09:28,580 --> 01:09:30,649
terms of how you actually

1924
01:09:30,890 --> 01:09:32,059
are going to deal with the end of

1925
01:09:32,060 --> 01:09:33,060
Moore's Law

1926
01:09:34,130 --> 01:09:35,389
or if not the end of Moore's Law, at

1927
01:09:35,390 --> 01:09:37,759
least, you know, multi, multi,

1928
01:09:38,990 --> 01:09:41,090
multiple C.P.S., many, many cores.

1929
01:09:41,540 --> 01:09:42,688
When you have a thousand cores, what

1930
01:09:42,689 --> 01:09:43,689
are you gonna do with them?

1931
01:09:44,180 --> 01:09:46,309
And so a lot of our

1932
01:09:46,760 --> 01:09:48,648
early design on Perl 6

1933
01:09:48,649 --> 01:09:50,449
was we didn't understand how

1934
01:09:50,510 --> 01:09:52,309
we wanted to do it yet, but

1935
01:09:52,550 --> 01:09:54,140
we knew very

1936
01:09:54,620 --> 01:09:56,539
we knew very deeply that we didn't

1937
01:09:56,540 --> 01:09:57,409
want to do anything. It would

1938
01:09:57,410 --> 01:09:58,410
prevent it.

1939
01:09:59,060 --> 01:10:00,899
So a lot of the early design

1940
01:10:00,980 --> 01:10:02,930
decisions were just saying, well,

1941
01:10:03,290 --> 01:10:04,290
we have to keep this open.

1942
01:10:05,210 --> 01:10:06,778
Then when the right smart people

1943
01:10:06,950 --> 01:10:08,600
came along, they said, look,

1944
01:10:09,230 --> 01:10:10,896
what you really have to look at is

1945
01:10:11,210 --> 01:10:12,209
things that are composed.

1946
01:10:12,210 --> 01:10:13,210
All threads are

1947
01:10:14,060 --> 01:10:15,169
not composed of, well, we can't take

1948
01:10:15,170 --> 01:10:16,489
two threads and turn it into a third

1949
01:10:16,490 --> 01:10:17,490
thread.

1950
01:10:18,440 --> 01:10:19,819
What you want is things like

1951
01:10:20,360 --> 01:10:21,529
we stole from go from.

1952
01:10:21,530 --> 01:10:23,689
We stole promises and channels

1953
01:10:24,290 --> 01:10:26,029
from C-sharp. We stole functional

1954
01:10:26,030 --> 01:10:27,039
reaction will programing.

1955
01:10:29,270 --> 01:10:31,279
And with these sort of primitives

1956
01:10:32,510 --> 01:10:34,078
which end up being duels of your

1957
01:10:34,340 --> 01:10:35,719
regular list processing kind of

1958
01:10:35,720 --> 01:10:37,459
primitives, they just happened to be

1959
01:10:37,460 --> 01:10:38,460
working on events.

1960
01:10:38,890 --> 01:10:40,939
And you're you can have loops, event

1961
01:10:40,940 --> 01:10:42,139
loops that were just like regular

1962
01:10:42,140 --> 01:10:43,979
loops that they're running on and

1963
01:10:44,510 --> 01:10:45,379
same control flow.

1964
01:10:45,380 --> 01:10:46,700
So that is easy to learn.

1965
01:10:47,370 --> 01:10:48,938
There are ways of sneaking these

1966
01:10:49,310 --> 01:10:51,710
things into a language,

1967
01:10:53,000 --> 01:10:54,109
at least in the ways that we

1968
01:10:54,110 --> 01:10:55,220
understand right now

1969
01:10:55,940 --> 01:10:58,100
that I heard.

1970
01:10:59,750 --> 01:11:00,750
Was it.

1971
01:11:01,250 --> 01:11:02,869
Who is talking about adding fibers

1972
01:11:02,870 --> 01:11:04,420
to their runtime?

1973
01:11:04,640 --> 01:11:06,109
I think you'd think it was a Java

1974
01:11:06,230 --> 01:11:07,230
news item,

1975
01:11:08,600 --> 01:11:10,099
but like lightweight threads that

1976
01:11:10,100 --> 01:11:11,178
can you, as Erlang has

1977
01:11:12,110 --> 01:11:13,689
done, has shown, you know, you can

1978
01:11:13,690 --> 01:11:14,839
run a hundred thousand threads in

1979
01:11:14,840 --> 01:11:15,840
your process.

1980
01:11:15,990 --> 01:11:17,089
That's another language that's been

1981
01:11:17,090 --> 01:11:18,049
around for a long.

1982
01:11:18,050 --> 01:11:19,420
Oh yeah. And we stole from them too.

1983
01:11:19,600 --> 01:11:20,600
Yeah.

1984
01:11:21,080 --> 01:11:22,219
Should we. We like their pattern

1985
01:11:22,220 --> 01:11:23,099
matching.

1986
01:11:23,100 --> 01:11:24,919
But one of the many therms is

1987
01:11:24,920 --> 01:11:26,389
that all of that language design is

1988
01:11:26,390 --> 01:11:27,589
theft of course.

1989
01:11:28,230 --> 01:11:29,209
Yeah. So yeah.

1990
01:11:29,210 --> 01:11:30,380
I mean we're thinking about it.

1991
01:11:30,410 --> 01:11:31,549
We don't know have all the answers

1992
01:11:31,550 --> 01:11:32,550
yet but

1993
01:11:34,370 --> 01:11:35,899
I think all these languages, you

1994
01:11:35,900 --> 01:11:37,468
know, long term tend to converge

1995
01:11:37,790 --> 01:11:38,899
on similar solutions.

1996
01:11:39,620 --> 01:11:40,747
So very sadly we are at

1997
01:11:41,540 --> 01:11:43,580
the last question.

1998
01:11:44,990 --> 01:11:46,910
I know you went so quickly

1999
01:11:46,940 --> 01:11:48,770
and unfortunately

2000
01:11:49,190 --> 01:11:51,319
all good things must end

2001
01:11:52,130 --> 01:11:53,899
except for maybe good languages.

2002
01:11:56,150 --> 01:11:58,520
So I'm going to start and

2003
01:11:59,180 --> 01:12:00,920
maybe quickly wrap up

2004
01:12:01,400 --> 01:12:03,259
as you look back over

2005
01:12:03,320 --> 01:12:04,839
the long lives of the languages

2006
01:12:05,390 --> 01:12:06,390
that you've written.

2007
01:12:06,830 --> 01:12:08,630
What have you found most rewarding?

2008
01:12:10,150 --> 01:12:11,029
The people?

2009
01:12:11,030 --> 01:12:12,590
The people by far

2010
01:12:12,880 --> 01:12:13,880
after the.

2011
01:12:16,810 --> 01:12:17,810
Oh, yeah.

2012
01:12:18,630 --> 01:12:20,269
Yeah. I mean, that that's what keeps

2013
01:12:20,270 --> 01:12:21,800
me coming back, that to work

2014
01:12:21,830 --> 01:12:23,899
everyday is like the incredible

2015
01:12:23,900 --> 01:12:25,700
excitement that

2016
01:12:25,790 --> 01:12:27,409
the community shows.

2017
01:12:27,740 --> 01:12:29,839
You and I I am so grateful

2018
01:12:29,840 --> 01:12:31,550
for all the people who

2019
01:12:32,590 --> 01:12:33,600
who have used that.

2020
01:12:34,190 --> 01:12:35,509
I mean, it there's nothing more

2021
01:12:35,510 --> 01:12:36,784
rewarding than have people

2022
01:12:37,550 --> 01:12:39,430
just like on Twitter or

2023
01:12:39,440 --> 01:12:41,204
wherever it is, go, oh, my God, this

2024
01:12:41,270 --> 01:12:42,990
is so great. And it's like, yeah,

2025
01:12:43,430 --> 01:12:44,749
that makes you want to just keep

2026
01:12:44,750 --> 01:12:45,559
doing more of it.

2027
01:12:45,560 --> 01:12:46,049
I mean.

2028
01:12:46,050 --> 01:12:47,716
Oh, yeah. You know, somebody comes

2029
01:12:48,050 --> 01:12:49,618
up to me in the street somewhere

2030
01:12:50,120 --> 01:12:51,629
and says, you know,

2031
01:12:52,370 --> 01:12:54,080
thanks for giving me a career.

2032
01:12:54,500 --> 01:12:55,640
Can I have a selfie?

2033
01:12:57,370 --> 01:12:58,791
You know, that's like this is

2034
01:12:59,270 --> 01:13:01,189
kind of like, you know, light and

2035
01:13:01,190 --> 01:13:02,089
dark thing.

2036
01:13:02,090 --> 01:13:03,266
But being this is really

2037
01:13:03,950 --> 01:13:04,939
wonderful.

2038
01:13:04,940 --> 01:13:05,929
This is all great.

2039
01:13:05,930 --> 01:13:07,369
I mean, I have to be in Silicon

2040
01:13:07,370 --> 01:13:09,289
Valley at a movie theater, the

2041
01:13:09,290 --> 01:13:10,290
tire shop.

2042
01:13:14,350 --> 01:13:15,739
Yeah. Well, what's really weird is

2043
01:13:15,740 --> 01:13:17,269
when you're when your kids are with

2044
01:13:17,270 --> 01:13:18,410
you when that happens.

2045
01:13:20,430 --> 01:13:21,430
But but, yes, it's.

2046
01:13:21,440 --> 01:13:23,239
It's when the when people say, you

2047
01:13:23,240 --> 01:13:24,612
gave me a career, you get my

2048
01:13:24,990 --> 01:13:25,990
livelihood.

2049
01:13:26,480 --> 01:13:27,480
You fed my family.

2050
01:13:27,540 --> 01:13:28,540
Yeah.

2051
01:13:29,600 --> 01:13:30,739
That's the best.

2052
01:13:32,150 --> 01:13:33,150
It's.

2053
01:13:33,320 --> 01:13:35,060
And you know, and

2054
01:13:35,210 --> 01:13:36,289
and not just.

2055
01:13:36,990 --> 01:13:38,389
And we're not talking about a star

2056
01:13:38,390 --> 01:13:39,390
relationship here.

2057
01:13:39,920 --> 01:13:41,719
Each of us to those individuals,

2058
01:13:41,990 --> 01:13:43,369
we're talking about the second

2059
01:13:43,370 --> 01:13:45,469
order, Community 2.0

2060
01:13:46,010 --> 01:13:47,010
kind of effects.

2061
01:13:47,300 --> 01:13:48,300
It's it's one

2062
01:13:49,220 --> 01:13:50,800
thing to see them, you know,

2063
01:13:51,950 --> 01:13:52,950
getting help from you.

2064
01:13:53,030 --> 01:13:54,229
It's another thing to see them

2065
01:13:54,230 --> 01:13:55,230
helping each other.

2066
01:13:56,660 --> 01:13:58,100
That's a whole whole level above

2067
01:13:58,870 --> 01:14:00,980
even the individual relationships.

2068
01:14:01,280 --> 01:14:03,680
I feel to see a community

2069
01:14:04,070 --> 01:14:05,539
that is learning how to love,

2070
01:14:06,230 --> 01:14:07,339
you know, kind of building a little

2071
01:14:07,340 --> 01:14:08,340
bit of heaven on earth.

2072
01:14:10,460 --> 01:14:11,460
There's just nothing like it.

2073
01:14:13,850 --> 01:14:15,020
Completely agree.

2074
01:14:15,480 --> 01:14:16,849
Kita, you get the last word.

2075
01:14:22,250 --> 01:14:23,659
And we love learning from you as

2076
01:14:23,660 --> 01:14:24,199
well.

2077
01:14:24,200 --> 01:14:26,180
And with that,

2078
01:14:26,810 --> 01:14:28,699
I want to say we've all

2079
01:14:28,700 --> 01:14:29,876
learned so much from all

2080
01:14:30,560 --> 01:14:32,599
of our wonderful panelists tonight.

2081
01:14:33,230 --> 01:14:34,357
I want to thank you for

2082
01:14:35,120 --> 01:14:36,786
allowing me to share and celebrate

2083
01:14:37,370 --> 01:14:38,600
this moment with you.

2084
01:14:39,260 --> 01:14:41,030
The little fifth grade kid in me

2085
01:14:41,090 --> 01:14:42,829
that was plinking away on basic at

2086
01:14:42,830 --> 01:14:44,779
Bell Labs who later

2087
01:14:44,780 --> 01:14:45,889
watch the P.C.

2088
01:14:45,890 --> 01:14:47,929
cell phone, Internet, Web and cloud

2089
01:14:47,930 --> 01:14:49,489
computing and change our world.

2090
01:14:49,850 --> 01:14:51,709
Never dreamed that I would be

2091
01:14:51,710 --> 01:14:53,510
chatting with four individuals

2092
01:14:53,780 --> 01:14:55,760
who profoundly impacted our world

2093
01:14:55,790 --> 01:14:57,979
by doing what they love, creating

2094
01:14:58,100 --> 01:14:59,239
and building languages.

2095
01:14:59,540 --> 01:15:01,489
Please join me in

2096
01:15:01,490 --> 01:15:03,529
a huge round of applause to thank

2097
01:15:03,620 --> 01:15:04,620
Anders

2098
01:15:07,040 --> 01:15:09,180
for inspiring this evening.

2099
01:15:15,990 --> 01:15:16,990
Thank you, Carol.

2100
01:15:18,050 --> 01:15:19,050
OK.

2101
01:15:19,320 --> 01:15:19,629
OK.

2102
01:15:19,630 --> 01:15:20,720
Was that epic?

2103
01:15:22,200 --> 01:15:24,539
So, Marina, I have an idea.

2104
01:15:24,840 --> 01:15:26,457
If you guys could step down right

2105
01:15:26,820 --> 01:15:28,109
here, we're going to get a giant

2106
01:15:28,110 --> 01:15:29,110
picture of everybody.

2107
01:15:30,720 --> 01:15:32,130
You guys stand right there.

2108
01:15:32,770 --> 01:15:33,629
We're going to get up here with your

2109
01:15:33,630 --> 01:15:34,630
camera.

2110
01:15:38,610 --> 01:15:39,869
All right. We're getting a picture

2111
01:15:39,870 --> 01:15:40,920
of everybody in that room.

2112
01:15:42,300 --> 01:15:44,189
All right. I need some excitement.

2113
01:15:47,070 --> 01:15:48,600
Oh, my God. I met my idol.

2114
01:15:50,460 --> 01:15:51,359
She can stand on the chair.

2115
01:15:51,360 --> 01:15:52,360
You can use my shoulder.

2116
01:15:53,220 --> 01:15:54,059
OK.

2117
01:15:54,060 --> 01:15:55,628
All right. Show some excitement,

2118
01:15:55,680 --> 01:15:56,760
people. Let's see it.

2119
01:16:01,350 --> 01:16:02,350
Monahan's.

2120
01:16:05,480 --> 01:16:07,180
Three times it's kinda weird and

2121
01:16:07,190 --> 01:16:08,190
lumpy.

2122
01:16:11,630 --> 01:16:12,630
OK,

2123
01:16:14,480 --> 01:16:16,189
thank you, everyone, for coming.

2124
01:16:16,220 --> 01:16:18,199
Thank you to all of our speakers.

2125
01:16:18,230 --> 01:16:19,790
Thank you to our organizers.

2126
01:16:19,820 --> 01:16:22,039
Let's do something equally awesome

2127
01:16:22,040 --> 01:16:23,040
next year.

2128
01:16:23,780 --> 01:16:24,780
And

2129
01:16:25,640 --> 01:16:26,939
if you want a robot there, 100

2130
01:16:26,960 --> 01:16:28,070
bucks. So come see me.