1
00:00:04,160 --> 00:00:07,160
Speaker 1: Get in touch with technology with tech Stuff from how

2
00:00:07,240 --> 00:00:13,800
Speaker 1: Stuff works dot com. Hey guys, welcome to tech Stuff.

3
00:00:13,880 --> 00:00:18,400
Speaker 1: I'm your host Jonathan strickland Um with super producer Dylan

4
00:00:18,880 --> 00:00:22,239
Speaker 1: in the teeny Tiny Stuff You Should Know Studio. Pretty soon,

5
00:00:22,280 --> 00:00:24,239
Speaker 1: I'm going to demand it be called the Stuff You

6
00:00:24,280 --> 00:00:28,319
Speaker 1: Should Know in Tech Stuff Studio, because who's gonna stop me.

7
00:00:29,000 --> 00:00:30,960
Speaker 1: Today I wanted to take a look at some of

8
00:00:31,000 --> 00:00:35,760
Speaker 1: the most spectacular computer bugs ever made, or at least

9
00:00:35,840 --> 00:00:37,839
Speaker 1: some of the more notable ones, and I got the

10
00:00:38,000 --> 00:00:41,400
Speaker 1: inspiration for this episode after researching a bug affecting the

11
00:00:41,440 --> 00:00:44,640
Speaker 1: Spotify desktop app. More on that in just a second,

12
00:00:45,880 --> 00:00:50,480
Speaker 1: but first I have to address a bit of apocryphal history,

13
00:00:50,800 --> 00:00:55,200
Speaker 1: and regrettably it's a story that we've repeated on tech Stuff.

14
00:00:55,240 --> 00:00:59,840
Speaker 1: So I'm sad to admit that I was complicit, although unknowingly,

15
00:01:00,600 --> 00:01:04,960
Speaker 1: in the spread of misinformation and that all has to

16
00:01:05,000 --> 00:01:07,960
Speaker 1: do with the origin of the term bug to describe

17
00:01:07,959 --> 00:01:12,280
Speaker 1: a flaw in programming. So here's the popular story, the

18
00:01:12,319 --> 00:01:19,039
Speaker 1: one that we have accidentally, uh promoted on tech Stuff

19
00:01:19,040 --> 00:01:22,280
Speaker 1: without knowing that we were in the wrong. It goes

20
00:01:22,319 --> 00:01:25,560
Speaker 1: that Grace Hopper, who was an early computer scientist who

21
00:01:25,640 --> 00:01:28,000
Speaker 1: rose to the rank of Rear admiral in the U. S.

22
00:01:28,080 --> 00:01:32,759
Speaker 1: Navy coined the phrase bug after discovering a moth coming

23
00:01:32,840 --> 00:01:38,200
Speaker 1: up Harvard's Mark two calculator, a literal bug. Generally speaking,

24
00:01:38,240 --> 00:01:43,000
Speaker 1: the story tends to be set in nineteen and there

25
00:01:43,080 --> 00:01:45,400
Speaker 1: is even a note in the log book that reads

26
00:01:45,560 --> 00:01:49,480
Speaker 1: first actual case of bug being found that's attributed to

27
00:01:49,520 --> 00:01:53,200
Speaker 1: Grace Hopper. But there are several points that are wrong

28
00:01:53,320 --> 00:01:57,160
Speaker 1: in this story. First, the year it didn't happen in ninet.

29
00:01:58,480 --> 00:02:02,000
Speaker 1: It happened on September. N We know because there's a

30
00:02:02,080 --> 00:02:05,560
Speaker 1: log book. At the log book that marks the incident

31
00:02:05,760 --> 00:02:08,720
Speaker 1: not only has the notes, it actually has the moth

32
00:02:08,880 --> 00:02:13,880
Speaker 1: taped into the book itself. It's taped onto the page. Second,

33
00:02:13,880 --> 00:02:17,280
Speaker 1: Grace Hopper wasn't the person to discover the moth or

34
00:02:17,480 --> 00:02:21,720
Speaker 1: make that log entry. She did tell the story about

35
00:02:21,720 --> 00:02:25,080
Speaker 1: the moth several times, but it wasn't in the context

36
00:02:25,200 --> 00:02:28,919
Speaker 1: of finding it or logging it. She just told the

37
00:02:28,960 --> 00:02:31,080
Speaker 1: story that, yeah, we really did have a bug in

38
00:02:31,120 --> 00:02:34,960
Speaker 1: the system. And most importantly, the word bug had already

39
00:02:35,000 --> 00:02:39,280
Speaker 1: been used to describe design flaws for decades before the

40
00:02:39,280 --> 00:02:42,720
Speaker 1: Mark two was even designed. In fact, if you look

41
00:02:42,760 --> 00:02:45,359
Speaker 1: at the log book. This makes sense. It says first

42
00:02:45,400 --> 00:02:49,519
Speaker 1: actual case of bug being found. That sentence doesn't make

43
00:02:49,680 --> 00:02:53,440
Speaker 1: sense unless you've already used the word bug to describe

44
00:02:53,520 --> 00:02:57,920
Speaker 1: a flaw, because you wouldn't say first actual case of

45
00:02:57,919 --> 00:03:00,880
Speaker 1: bug being found. That the wording doesn't make any sense.

46
00:03:00,880 --> 00:03:05,400
Speaker 1: The context makes no sense. Sadly, there are documented quotes

47
00:03:05,760 --> 00:03:09,040
Speaker 1: dating back to the nineteenth century using the word bug

48
00:03:09,080 --> 00:03:12,239
Speaker 1: to mean a design fault, and it could go back

49
00:03:12,240 --> 00:03:16,239
Speaker 1: even further than that. So is with much regret that

50
00:03:16,280 --> 00:03:18,840
Speaker 1: I admit I have unwittingly contributed to a bit of

51
00:03:18,880 --> 00:03:21,799
Speaker 1: misleading folklore making the rounds. But I'm glad I can

52
00:03:21,800 --> 00:03:24,880
Speaker 1: take this opportunity to address it. All Right, So let's

53
00:03:24,919 --> 00:03:29,799
Speaker 1: talk about design bugs, and I'll be covering several goofs, mistakes, flubs, flaws,

54
00:03:29,840 --> 00:03:34,040
Speaker 1: and outright catastrophes in this episode. But one thing I'm

55
00:03:34,080 --> 00:03:38,240
Speaker 1: not necessarily going to cover our software vulnerabilities that were

56
00:03:38,320 --> 00:03:42,680
Speaker 1: later exploited, either by opportunistic hackers or white hats who

57
00:03:42,680 --> 00:03:46,880
Speaker 1: are just trying to improve system security. Those vulnerabilities are

58
00:03:46,960 --> 00:03:50,000
Speaker 1: common in many types of software and arise not just

59
00:03:50,080 --> 00:03:55,240
Speaker 1: through mistakes, but sometimes simple oversights, and I think it

60
00:03:55,320 --> 00:03:57,440
Speaker 1: might be more fun to look at some real bugs,

61
00:03:57,440 --> 00:04:00,760
Speaker 1: like stuff that made things go wrong, stuff that may

62
00:04:00,800 --> 00:04:04,800
Speaker 1: have rendered a program defunct or otherwise caused headaches. Now

63
00:04:04,800 --> 00:04:06,840
Speaker 1: I'm gonna make an exception to this. I'm going to

64
00:04:06,960 --> 00:04:10,560
Speaker 1: start off with the ping of death, and I only

65
00:04:10,600 --> 00:04:14,240
Speaker 1: mention it because it has an awesome name. Now, this

66
00:04:14,280 --> 00:04:18,000
Speaker 1: flaw caused headaches back in ninety six. It was a

67
00:04:18,000 --> 00:04:23,160
Speaker 1: flawed i P fragmentation reassembly code, and it became possible

68
00:04:23,160 --> 00:04:25,919
Speaker 1: to crash lots of different types of computers using different

69
00:04:25,960 --> 00:04:30,000
Speaker 1: operating systems, although Windows machines were particularly vulnerable, and this

70
00:04:30,080 --> 00:04:32,880
Speaker 1: particular flaw would make a Windows machine revert to the

71
00:04:33,000 --> 00:04:36,400
Speaker 1: dreaded blue screen of death. And it all happened by

72
00:04:36,400 --> 00:04:40,039
Speaker 1: sending a special ping packet over the internet. So, for

73
00:04:40,160 --> 00:04:42,120
Speaker 1: those of you who aren't familiar with what that is,

74
00:04:42,279 --> 00:04:45,800
Speaker 1: a ping is essentially a simple message that checks for

75
00:04:45,880 --> 00:04:49,400
Speaker 1: a connection between two computers. You send one ping from

76
00:04:49,400 --> 00:04:52,200
Speaker 1: a computer to another one and look for a response,

77
00:04:52,480 --> 00:04:55,080
Speaker 1: so that way you verify there is in fact a connection.

78
00:04:55,120 --> 00:04:57,840
Speaker 1: You can also tell other things like how fast is

79
00:04:57,880 --> 00:05:01,400
Speaker 1: that connection between those two computers. Now, in this case,

80
00:05:01,720 --> 00:05:04,960
Speaker 1: you would have to actually design a malformed ping request

81
00:05:05,040 --> 00:05:07,840
Speaker 1: and send that to a target, and it would bring

82
00:05:07,880 --> 00:05:12,000
Speaker 1: that target down. That's the only security vulnerability story I

83
00:05:12,040 --> 00:05:15,040
Speaker 1: really wanted to focus on. The others are all just

84
00:05:15,760 --> 00:05:18,640
Speaker 1: design flaws. And let's begin with the bug that inspired

85
00:05:18,680 --> 00:05:20,320
Speaker 1: me to do this episode in the first place, That

86
00:05:20,400 --> 00:05:24,080
Speaker 1: Spotify bug I mentioned earlier. Ours Technica wrote a piece

87
00:05:24,120 --> 00:05:26,240
Speaker 1: on it in November two thousand and sixteen, but the

88
00:05:26,240 --> 00:05:29,120
Speaker 1: problem seems to date back at least as far as

89
00:05:29,240 --> 00:05:32,760
Speaker 1: June two thousand sixteen, and that's when a few savvy

90
00:05:32,839 --> 00:05:37,440
Speaker 1: Spotify users noticed some unusual activities on their computers. And

91
00:05:37,480 --> 00:05:39,240
Speaker 1: it took a little bit of detective work, but they

92
00:05:39,240 --> 00:05:43,120
Speaker 1: discovered that Spotify was apparently generating a huge amount of

93
00:05:43,200 --> 00:05:48,560
Speaker 1: data on a daily basis, like gigabytes of data per day.

94
00:05:48,720 --> 00:05:51,960
Speaker 1: And the culprit turned out to be a vacuum process

95
00:05:52,040 --> 00:05:56,440
Speaker 1: for a database file containing the string mercury dot dB. Now,

96
00:05:56,480 --> 00:06:00,279
Speaker 1: the vacuum process is the digital equivalent of vacuum c link.

97
00:06:00,320 --> 00:06:03,000
Speaker 1: It's meant to repack data so that it takes up

98
00:06:03,120 --> 00:06:06,280
Speaker 1: less space on a drive. Now, this involves building a

99
00:06:06,279 --> 00:06:09,719
Speaker 1: new file to maximize efficiency, which is a good thing

100
00:06:09,839 --> 00:06:16,080
Speaker 1: generally speaking. The problem was that Spotify's version was making

101
00:06:16,120 --> 00:06:18,440
Speaker 1: it happen way too frequently, like on the order of

102
00:06:18,520 --> 00:06:22,640
Speaker 1: once every few minutes, so that's not generally necessary. You

103
00:06:22,640 --> 00:06:25,440
Speaker 1: don't need to rebuild a database file every few minutes

104
00:06:25,480 --> 00:06:28,440
Speaker 1: to make sure it's the most efficient size it can be.

105
00:06:29,720 --> 00:06:33,000
Speaker 1: So each rebuild represented a relatively small amount of data,

106
00:06:33,080 --> 00:06:36,039
Speaker 1: but over time it added up, which meant that if

107
00:06:36,080 --> 00:06:38,880
Speaker 1: you had Spotify on on your computer, even if it

108
00:06:38,960 --> 00:06:41,680
Speaker 1: was just running in the background, it would be generating

109
00:06:41,720 --> 00:06:45,400
Speaker 1: gigabytes worth of information rewriting this file over and over.

110
00:06:45,720 --> 00:06:48,479
Speaker 1: Now it wasn't filling up a hard drive. It was

111
00:06:48,520 --> 00:06:53,920
Speaker 1: just overwriting the same file. Now, if it had been

112
00:06:53,920 --> 00:06:56,320
Speaker 1: filling up a hard drive, people would have noticed much earlier,

113
00:06:56,320 --> 00:06:59,360
Speaker 1: and it wouldn't have just been savvy Spotify users, because

114
00:06:59,360 --> 00:07:01,800
Speaker 1: you would suddenly notice, hey, I don't I can't save

115
00:07:01,800 --> 00:07:05,800
Speaker 1: anything to my hard drive because everything is filling up. Instead. Again,

116
00:07:05,839 --> 00:07:08,480
Speaker 1: it was just sort of writing and deleting and writing

117
00:07:08,480 --> 00:07:11,280
Speaker 1: and deleting the same file over and over again. And

118
00:07:11,320 --> 00:07:13,240
Speaker 1: that probably doesn't sound like a big deal, but it

119
00:07:13,440 --> 00:07:15,800
Speaker 1: is a problem if you're using a solid state drive

120
00:07:16,040 --> 00:07:19,240
Speaker 1: or s s D. So one of the drawbacks of

121
00:07:19,240 --> 00:07:23,600
Speaker 1: an ss D is that over time it loses storage capacity,

122
00:07:23,720 --> 00:07:27,960
Speaker 1: like you can store less data on an SSD over time. Now,

123
00:07:28,400 --> 00:07:30,960
Speaker 1: by overtime I generally mean over a great deal of

124
00:07:31,040 --> 00:07:33,160
Speaker 1: time and a lot of different data being written to

125
00:07:33,200 --> 00:07:37,360
Speaker 1: it and overwritten. Uh. Generally speaking, most of us end

126
00:07:37,440 --> 00:07:40,320
Speaker 1: up replacing our drives before we get to a point

127
00:07:40,320 --> 00:07:44,320
Speaker 1: where the loss of capacity is a real issue. But

128
00:07:44,520 --> 00:07:46,560
Speaker 1: similar in a way to how a battery can lose

129
00:07:46,600 --> 00:07:49,400
Speaker 1: its ability to hold a full charge after you've gone

130
00:07:49,440 --> 00:07:53,560
Speaker 1: through lots of charging and discharging cycles, you know how

131
00:07:53,600 --> 00:07:55,960
Speaker 1: a battery won't be able to to hold as much

132
00:07:56,000 --> 00:07:58,720
Speaker 1: even if it says it's up to a but a

133
00:07:58,760 --> 00:08:01,120
Speaker 1: hundred percent doesn't last you as long as it used to.

134
00:08:01,680 --> 00:08:04,400
Speaker 1: That's because its capacity to hold a full charge has

135
00:08:04,440 --> 00:08:08,560
Speaker 1: decreased over time. But let's say you've got a program

136
00:08:08,640 --> 00:08:12,680
Speaker 1: that's just constantly overwriting data to your drive, you might

137
00:08:12,720 --> 00:08:15,560
Speaker 1: discover that your ss D s useful lifespan has been

138
00:08:15,640 --> 00:08:20,560
Speaker 1: drastically reduced. So as I record this episode, Spotify has

139
00:08:20,600 --> 00:08:24,360
Speaker 1: already rolled out an updated version of its desktop application,

140
00:08:24,400 --> 00:08:26,280
Speaker 1: and that, by the way, is the only version of

141
00:08:26,320 --> 00:08:29,520
Speaker 1: Spotify that was affected. If you use web based Spotify

142
00:08:29,800 --> 00:08:32,719
Speaker 1: or mobile Spotify, you're in the clear already if you

143
00:08:32,800 --> 00:08:35,480
Speaker 1: use a desktop version, as long as you have version

144
00:08:35,600 --> 00:08:40,160
Speaker 1: one point zero point four two or later, you are fine.

145
00:08:41,280 --> 00:08:43,520
Speaker 1: But if you did have that earlier version and you

146
00:08:43,640 --> 00:08:46,000
Speaker 1: just had Spotify running on in the background, chances are

147
00:08:46,160 --> 00:08:50,160
Speaker 1: it was writing to your hard drive like crazy. So

148
00:08:50,200 --> 00:08:53,400
Speaker 1: what about some of the other big bugs in computer history. Well,

149
00:08:53,440 --> 00:08:56,120
Speaker 1: some of the real doozies involve our attempts to explore

150
00:08:56,200 --> 00:08:59,839
Speaker 1: the final frontier. So we'll be talking about space a

151
00:09:00,000 --> 00:09:02,200
Speaker 1: few times in this episode, and we'll start with an

152
00:09:02,200 --> 00:09:06,640
Speaker 1: early US satellite. So first up is a nineteen sixty

153
00:09:06,679 --> 00:09:10,400
Speaker 1: two blunder involving the Mariner one. So some backstory on

154
00:09:10,440 --> 00:09:12,880
Speaker 1: this one. Uh, We're gonna talk a lot about the

155
00:09:12,880 --> 00:09:16,040
Speaker 1: Soviet Union in this episode two. It takes a couple

156
00:09:16,080 --> 00:09:18,880
Speaker 1: of roles as we go on. But in this case,

157
00:09:18,960 --> 00:09:22,880
Speaker 1: the then USSR had launched Sputnik into orbit in nineteen

158
00:09:22,920 --> 00:09:25,280
Speaker 1: fifty seven, which really kicked off the space race and

159
00:09:25,360 --> 00:09:28,640
Speaker 1: also was a big shot in the Cold War because

160
00:09:28,679 --> 00:09:31,640
Speaker 1: of so Union was essentially saying, hey, we can launch

161
00:09:31,720 --> 00:09:34,320
Speaker 1: this into space, we could also launch something at you.

162
00:09:35,520 --> 00:09:37,760
Speaker 1: In response, the US and done sort of the same thing.

163
00:09:37,760 --> 00:09:40,840
Speaker 1: They had launched some satellites into space and the Mariner

164
00:09:40,920 --> 00:09:43,880
Speaker 1: one was going to be a big, big feather in

165
00:09:43,920 --> 00:09:45,599
Speaker 1: the cap of the US. The whole idea was to

166
00:09:45,679 --> 00:09:48,440
Speaker 1: launch a probe that would be a fly by probe

167
00:09:48,559 --> 00:09:53,640
Speaker 1: and it would go by Venus. So uh NASA, which

168
00:09:53,679 --> 00:09:57,400
Speaker 1: was newly formed in nineteen sixty two, was taking control

169
00:09:57,440 --> 00:09:59,720
Speaker 1: of this and the budget for this particular project was

170
00:09:59,800 --> 00:10:02,600
Speaker 1: eight team point five million dollars, which if you were

171
00:10:02,640 --> 00:10:06,440
Speaker 1: to adjust for inflation, would be almost a hundred fifty

172
00:10:06,520 --> 00:10:10,520
Speaker 1: million dollars today, So a hundred fifty million dollar project

173
00:10:10,559 --> 00:10:14,200
Speaker 1: to launch the Mariner one and have it fly by Venus.

174
00:10:14,240 --> 00:10:16,760
Speaker 1: But as I'm sure you guys have figured out by

175
00:10:16,760 --> 00:10:20,520
Speaker 1: now based upon the topic of this podcast, not all

176
00:10:20,600 --> 00:10:24,760
Speaker 1: went according to plan. Not long at all. After the

177
00:10:24,880 --> 00:10:28,800
Speaker 1: rocket launched from the launch pad, it began to veer

178
00:10:28,880 --> 00:10:32,440
Speaker 1: off course, and neither the computer controls on the rocket

179
00:10:32,559 --> 00:10:37,000
Speaker 1: or manual controls back at HQ could correct for the problem.

180
00:10:37,080 --> 00:10:39,760
Speaker 1: The rockets course was such that it was going to

181
00:10:39,840 --> 00:10:42,640
Speaker 1: take it over shipping lanes, which meant there could be

182
00:10:42,640 --> 00:10:47,280
Speaker 1: a potential catastrophe, and so Arrange safety officer made the

183
00:10:47,320 --> 00:10:50,280
Speaker 1: difficult call and issued the command to blow the whole

184
00:10:50,360 --> 00:10:54,480
Speaker 1: thing up, just shy of three hundred seconds after it launched.

185
00:10:55,000 --> 00:10:57,400
Speaker 1: So what happened? What why did it go off course

186
00:10:57,400 --> 00:10:59,520
Speaker 1: in the first place? Well, there was a flaw in

187
00:10:59,559 --> 00:11:02,880
Speaker 1: the space crafts guidance software which diverted the rocket, and

188
00:11:02,960 --> 00:11:05,840
Speaker 1: no amount of commands from ground control could correct for it.

189
00:11:06,280 --> 00:11:10,800
Speaker 1: After a lengthy investigation, NASA discovered the era error was

190
00:11:11,080 --> 00:11:16,560
Speaker 1: the result of a mistake transcribing handwritten notes into computer code.

191
00:11:17,360 --> 00:11:21,880
Speaker 1: So someone just took some handwritten notes and misinterpreted one

192
00:11:21,880 --> 00:11:28,000
Speaker 1: of them, and that one mistake was enough to crash

193
00:11:28,280 --> 00:11:33,680
Speaker 1: the rocket, or to to necessitate it being destroyed. The

194
00:11:33,760 --> 00:11:38,400
Speaker 1: great science fiction author Arthur C. Clark wrote that the

195
00:11:38,480 --> 00:11:43,800
Speaker 1: Mariner one was wrecked by the most expensive hyphen in history,

196
00:11:43,960 --> 00:11:47,120
Speaker 1: which isn't quite right, but it's pretty funny. I mean,

197
00:11:47,280 --> 00:11:51,480
Speaker 1: come on, it's humorous phrase. So the actual punctuation mark

198
00:11:51,520 --> 00:11:54,840
Speaker 1: that caused the problem was not technically a hyphen. It

199
00:11:54,920 --> 00:11:59,120
Speaker 1: was a superscript bar. Superscript bars, by the way, not

200
00:11:59,200 --> 00:12:01,319
Speaker 1: a place where player rights hang out to get tore up.

201
00:12:02,120 --> 00:12:05,440
Speaker 1: A superscript bar just means it's a horizontal bar that

202
00:12:05,600 --> 00:12:08,360
Speaker 1: is above some other symbol. In this case, it was

203
00:12:08,400 --> 00:12:12,920
Speaker 1: a radius symbol, and that was a symbol along with

204
00:12:12,960 --> 00:12:17,280
Speaker 1: the superscript bar to describe a smoothing function, which means

205
00:12:17,320 --> 00:12:20,920
Speaker 1: the formula was meant to calculate smoothed values over the

206
00:12:20,960 --> 00:12:25,760
Speaker 1: time derivative of a radius. Now, without the smoothing function,

207
00:12:26,320 --> 00:12:30,439
Speaker 1: tiny deviations in course sent commands to the rockets thrusters

208
00:12:30,440 --> 00:12:34,000
Speaker 1: to kick in big time to overcorrect for that problem.

209
00:12:34,040 --> 00:12:36,760
Speaker 1: As an analogy, imagine you're driving a vehicle and you

210
00:12:36,800 --> 00:12:39,600
Speaker 1: see a pothole in the road and you're approaching it,

211
00:12:39,720 --> 00:12:44,160
Speaker 1: and instead of gently steering out of the way, you

212
00:12:44,400 --> 00:12:47,160
Speaker 1: wrenched the wheel really hard to the left or to

213
00:12:47,200 --> 00:12:49,440
Speaker 1: the right in order to try and get around this pothole.

214
00:12:49,880 --> 00:12:51,920
Speaker 1: That's kind of what was happening with the rocket. It

215
00:12:51,960 --> 00:12:55,120
Speaker 1: didn't have the smoothing function and so as a result,

216
00:12:55,200 --> 00:12:59,240
Speaker 1: it was having these wild deviations and course. So it

217
00:12:59,360 --> 00:13:02,000
Speaker 1: wasn't a high fin that caused the problem, but is

218
00:13:02,040 --> 00:13:05,800
Speaker 1: close enough. Our next space story takes place in nineteen

219
00:13:06,400 --> 00:13:10,640
Speaker 1: with the European Space Agencies RAN five Flight five O

220
00:13:10,880 --> 00:13:14,200
Speaker 1: one rocket. Now, this rocket was to launch into space

221
00:13:14,240 --> 00:13:19,400
Speaker 1: on June four nine, and instead the rocket disintegrated forty

222
00:13:19,559 --> 00:13:23,640
Speaker 1: seconds after taking off. So what the heck happened? Well,

223
00:13:23,679 --> 00:13:25,640
Speaker 1: it largely had to do with the e s A

224
00:13:26,080 --> 00:13:30,239
Speaker 1: reusing old work. This actually becomes a theme in this episode.

225
00:13:30,640 --> 00:13:34,480
Speaker 1: One of the morals of of this entire podcast is

226
00:13:34,600 --> 00:13:38,440
Speaker 1: if you're designing something a successor to an earlier product,

227
00:13:39,720 --> 00:13:42,800
Speaker 1: and you'd want to reuse some of the features that

228
00:13:42,880 --> 00:13:47,320
Speaker 1: you created in your previous product, test the heck out

229
00:13:47,360 --> 00:13:50,760
Speaker 1: of it in its new form factor, because it could

230
00:13:50,840 --> 00:13:53,760
Speaker 1: be that things that worked perfectly fine in the earlier

231
00:13:53,800 --> 00:13:58,240
Speaker 1: model will go awry in the new one. That's what

232
00:13:58,360 --> 00:14:01,760
Speaker 1: happened here. So as you might guess from the name,

233
00:14:02,000 --> 00:14:06,360
Speaker 1: the Aeryan five marked the fifth generation of launch vehicles

234
00:14:06,520 --> 00:14:11,160
Speaker 1: under that name. The Arian four's inertial reference system would

235
00:14:11,160 --> 00:14:14,920
Speaker 1: convert sixty four bit floating point numbers into a sixteen

236
00:14:14,960 --> 00:14:20,680
Speaker 1: bits signed integer, and it worked just fine. But the

237
00:14:20,760 --> 00:14:26,400
Speaker 1: Arian five stats were beefier than its predecessor with faster engines,

238
00:14:26,440 --> 00:14:29,640
Speaker 1: and that was where the problem really started. The engine

239
00:14:29,640 --> 00:14:33,480
Speaker 1: output meant those sixty four bit floating point numbers were

240
00:14:33,560 --> 00:14:37,520
Speaker 1: significantly larger than the ones generated by the engines on

241
00:14:37,560 --> 00:14:41,680
Speaker 1: the Arian four. They didn't anticipate this, so during the

242
00:14:41,680 --> 00:14:46,400
Speaker 1: conversion process there was actually data overflow, and that overflow

243
00:14:46,520 --> 00:14:50,360
Speaker 1: caused both the backup computer and the primary computer aboard

244
00:14:50,360 --> 00:14:53,440
Speaker 1: the Arian five to crash, and they crashed in that order.

245
00:14:53,480 --> 00:14:57,040
Speaker 1: The backup computer crashed first, followed by the primary computer

246
00:14:57,080 --> 00:14:59,960
Speaker 1: a couple of seconds later. The whole thing took less

247
00:15:00,000 --> 00:15:04,120
Speaker 1: than a minute to go from launch to disintegration. Oops,

248
00:15:05,600 --> 00:15:07,920
Speaker 1: now we're gonna stick with space. But jumped forward to

249
00:15:10,240 --> 00:15:16,200
Speaker 1: and the Mars Climate Orbiter. This was an unfortunate problem.

250
00:15:16,560 --> 00:15:19,960
Speaker 1: So this particular spacecraft was meant to study Mars's climate,

251
00:15:20,080 --> 00:15:22,920
Speaker 1: atmosphere and surface changes, and it was also supposed to

252
00:15:22,960 --> 00:15:26,200
Speaker 1: be a kind of relay station for landers that would

253
00:15:26,240 --> 00:15:30,320
Speaker 1: explore Mars's surface, but none of that would last because

254
00:15:30,440 --> 00:15:36,040
Speaker 1: of some pretty significant goofs. So on September, the orbiter

255
00:15:36,160 --> 00:15:40,680
Speaker 1: passed into the upper atmosphere of Mars and did so

256
00:15:40,880 --> 00:15:44,320
Speaker 1: at a pretty low altitude. And this is what folks

257
00:15:44,320 --> 00:15:47,920
Speaker 1: in the space industry called a bad thing. The Dragon,

258
00:15:48,000 --> 00:15:51,800
Speaker 1: the spacecraft was significant. It began to fall apart and

259
00:15:51,960 --> 00:15:57,680
Speaker 1: it was destroyed upon entering Mars's atmosphere. That's what happened.

260
00:15:58,840 --> 00:16:02,280
Speaker 1: So the software guide the orbiter was to blame, and

261
00:16:03,160 --> 00:16:06,360
Speaker 1: it's a dumb, dumb mistake. It was supposed to make

262
00:16:06,360 --> 00:16:10,600
Speaker 1: adjustments to the orbiter's flight in SI units, specifically in

263
00:16:10,640 --> 00:16:15,560
Speaker 1: Newton seconds. That's what the contract with Lockheed and NASA said,

264
00:16:16,280 --> 00:16:20,040
Speaker 1: Newton seconds, use SI units for all of your all

265
00:16:20,080 --> 00:16:24,400
Speaker 1: of your calculations, but the software instead made calculations in

266
00:16:24,520 --> 00:16:30,840
Speaker 1: non SI units, namely pounds seconds. So Lockheeds software gave

267
00:16:30,920 --> 00:16:34,960
Speaker 1: information to NASA's systems using the wrong units of measure.

268
00:16:36,040 --> 00:16:39,240
Speaker 1: NASA systems then took that information, assuming it was with

269
00:16:39,360 --> 00:16:43,800
Speaker 1: the right units of measure, and executed commands based upon that.

270
00:16:45,000 --> 00:16:47,760
Speaker 1: Uh So, this is why if you're ever in a

271
00:16:47,840 --> 00:16:52,600
Speaker 1: math course and the teacher makes you stop in the

272
00:16:52,640 --> 00:16:54,760
Speaker 1: middle of writing a problem on the board and says,

273
00:16:54,800 --> 00:16:57,920
Speaker 1: where are your units? This is why you have to

274
00:16:58,000 --> 00:17:01,520
Speaker 1: make sure you're using the right units, because if you're

275
00:17:01,560 --> 00:17:04,800
Speaker 1: saying a number and you don't associate a unit with it,

276
00:17:05,080 --> 00:17:08,040
Speaker 1: someone could make an incorrect decision based on that, and

277
00:17:08,080 --> 00:17:10,639
Speaker 1: it could be disastrous, as it was with the case

278
00:17:10,760 --> 00:17:14,360
Speaker 1: of this orbiter. The thrusters fired at four point four

279
00:17:14,440 --> 00:17:17,960
Speaker 1: or five times the power they were supposed to, and

280
00:17:18,000 --> 00:17:20,760
Speaker 1: the orbiter didn't stand a chance. And this was a

281
00:17:20,760 --> 00:17:24,080
Speaker 1: pretty expensive mistake. That mission's cost came in at three

282
00:17:24,800 --> 00:17:28,639
Speaker 1: seven point six million dollars, but on the bright side

283
00:17:28,680 --> 00:17:31,359
Speaker 1: with all of these stories, at least no human lives

284
00:17:31,440 --> 00:17:36,000
Speaker 1: were ever in real danger as a result of the mistake. Now,

285
00:17:36,080 --> 00:17:39,840
Speaker 1: I've got a lot more to say about bugs, but

286
00:17:40,040 --> 00:17:42,240
Speaker 1: before I get into that, let's take a quick break

287
00:17:42,400 --> 00:17:53,000
Speaker 1: to thank our sponsor. All right, Now, let's make a

288
00:17:53,040 --> 00:17:55,479
Speaker 1: switch to A T and T, which is a company

289
00:17:55,480 --> 00:17:58,160
Speaker 1: that had a pretty big problem with switches once upon

290
00:17:58,200 --> 00:18:00,920
Speaker 1: a time. I'm talking about issue that popped up on

291
00:18:01,000 --> 00:18:04,560
Speaker 1: January nine nine. That's when A T and T long

292
00:18:04,600 --> 00:18:07,200
Speaker 1: distance customers discovered they were unable to make any long

293
00:18:07,240 --> 00:18:11,879
Speaker 1: distance calls. Why why could they no longer reach anybody? Well,

294
00:18:12,440 --> 00:18:17,119
Speaker 1: A T and T s long distance switches, which control

295
00:18:17,240 --> 00:18:20,639
Speaker 1: that and allow for the actual connections to be made,

296
00:18:20,960 --> 00:18:23,840
Speaker 1: were on the fritz. They were trying to reboot over

297
00:18:23,880 --> 00:18:27,560
Speaker 1: and over again. They were just stuck in a reboot cycle. Now,

298
00:18:27,640 --> 00:18:31,520
Speaker 1: initially the company thought it was being hacked, But like

299
00:18:31,560 --> 00:18:33,000
Speaker 1: I said at the top of the show, I'm not

300
00:18:33,119 --> 00:18:36,680
Speaker 1: covering stories about hackers here. I'm talking about big design

301
00:18:36,720 --> 00:18:41,680
Speaker 1: flaws that caused problems. So they weren't getting hacked. That's

302
00:18:41,920 --> 00:18:44,320
Speaker 1: not what was going on with those D fourteen long

303
00:18:44,320 --> 00:18:47,600
Speaker 1: distance switches. No, there was a design problem at fault.

304
00:18:48,440 --> 00:18:49,959
Speaker 1: So what had happened was a T and D had

305
00:18:50,040 --> 00:18:53,480
Speaker 1: rolled out an update to the code that managed the switches,

306
00:18:53,880 --> 00:18:56,080
Speaker 1: and it was meant to increase the efficiency. It was

307
00:18:56,119 --> 00:18:58,879
Speaker 1: meant to speed things up, but the problem was it

308
00:18:59,040 --> 00:19:01,680
Speaker 1: sped things up so much that the system got caught

309
00:19:01,800 --> 00:19:04,679
Speaker 1: up in itself. It gets pretty technical, but I can

310
00:19:04,680 --> 00:19:08,439
Speaker 1: give you kind of a overview of what the problem was. Alright,

311
00:19:08,480 --> 00:19:12,560
Speaker 1: so each switch had a function that allowed it to

312
00:19:12,760 --> 00:19:16,800
Speaker 1: alert the next switch down the line if things were

313
00:19:16,800 --> 00:19:20,560
Speaker 1: starting to get hairy. So imagine that switch number one

314
00:19:20,960 --> 00:19:24,200
Speaker 1: is handling traffic, but it's getting really close to capacity.

315
00:19:24,240 --> 00:19:26,320
Speaker 1: So it sends a message over to switch number two

316
00:19:26,359 --> 00:19:29,680
Speaker 1: and says, I can't take on any more work because

317
00:19:29,720 --> 00:19:33,200
Speaker 1: if I do, I'll be overloaded. Switch to then says,

318
00:19:33,400 --> 00:19:36,679
Speaker 1: no problem, I'll take on the any oncoming work for

319
00:19:36,720 --> 00:19:40,159
Speaker 1: you and we'll handle it from there. And if switched

320
00:19:40,200 --> 00:19:43,000
Speaker 1: number two order to get into the same source situation,

321
00:19:43,359 --> 00:19:45,240
Speaker 1: it would say the same thing to switch number three,

322
00:19:45,320 --> 00:19:48,840
Speaker 1: and so on and so forth. Now, eventually each switch

323
00:19:48,920 --> 00:19:51,679
Speaker 1: will contact the one below it and say, hey, how

324
00:19:51,680 --> 00:19:54,440
Speaker 1: are you doing there? And if the answer is okay,

325
00:19:54,520 --> 00:19:58,240
Speaker 1: then everything switches back and you go back to normal operation.

326
00:19:59,200 --> 00:20:02,679
Speaker 1: That's how it's suppo was to work up. But A

327
00:20:02,800 --> 00:20:05,560
Speaker 1: T and t s updated code sped things up so

328
00:20:05,640 --> 00:20:09,360
Speaker 1: much it caused some real issues, and there was some

329
00:20:09,560 --> 00:20:14,080
Speaker 1: poor timing, just coincidental timing that made things worse. So

330
00:20:14,440 --> 00:20:17,159
Speaker 1: switch number one starts to get overwhelmed and sends a

331
00:20:17,200 --> 00:20:19,679
Speaker 1: message over to switch number two, but switched number two

332
00:20:19,920 --> 00:20:23,840
Speaker 1: was just in the middle of resetting itself. So switch

333
00:20:23,920 --> 00:20:26,680
Speaker 1: number two goes into reset mode, which says do not disturb.

334
00:20:26,760 --> 00:20:29,680
Speaker 1: Sends a message over to switch number three. That prompted

335
00:20:29,720 --> 00:20:32,239
Speaker 1: switch number three to overload and put up I do

336
00:20:32,280 --> 00:20:34,760
Speaker 1: not disturb sign. Move that down to switch number four.

337
00:20:35,119 --> 00:20:39,120
Speaker 1: This whole thing goes down the entire line of on switches.

338
00:20:39,200 --> 00:20:42,040
Speaker 1: They all end up getting overloaded as a result of this,

339
00:20:42,440 --> 00:20:45,600
Speaker 1: and I'll go into reset mode and they get stuck there.

340
00:20:47,160 --> 00:20:51,359
Speaker 1: That problem lasted for nine hours before A T and

341
00:20:51,359 --> 00:20:54,080
Speaker 1: T was finally able to address the message load on

342
00:20:54,119 --> 00:20:56,520
Speaker 1: the entire system and get the switches back to normal.

343
00:20:57,040 --> 00:21:00,560
Speaker 1: The estimated cost of lost revenue for that was about

344
00:21:00,560 --> 00:21:04,600
Speaker 1: sixty million dollars in long distance calls, and there were

345
00:21:04,640 --> 00:21:07,560
Speaker 1: a lot of angry customers to boot, so to placate them,

346
00:21:07,600 --> 00:21:10,719
Speaker 1: A T and T offered reduced long distance rates on

347
00:21:10,840 --> 00:21:15,440
Speaker 1: Valentine's Day pretty ugly. By the A T and T

348
00:21:15,440 --> 00:21:17,280
Speaker 1: tried to handle it, at least in a way that

349
00:21:17,440 --> 00:21:21,040
Speaker 1: didn't turn it into a pr nightmare. Not so with Intel.

350
00:21:21,760 --> 00:21:24,320
Speaker 1: That's what it brings us to the Pendium problem. I

351
00:21:24,359 --> 00:21:27,040
Speaker 1: don't know if you guys remember when Pentium processors first

352
00:21:27,080 --> 00:21:29,720
Speaker 1: came out, but they were a big deal. It was

353
00:21:30,680 --> 00:21:33,800
Speaker 1: a redesign of the architecture of the microprocessor and it

354
00:21:33,840 --> 00:21:37,239
Speaker 1: was meant to really speed things up. Well, Intel had

355
00:21:37,280 --> 00:21:41,840
Speaker 1: a massive nightmare in nine thanks to a flaw in

356
00:21:41,880 --> 00:21:47,080
Speaker 1: the entire first generation of Pentium processors. Now, when you

357
00:21:47,119 --> 00:21:50,400
Speaker 1: break it all down, a CPU is all about performing

358
00:21:50,440 --> 00:21:53,399
Speaker 1: mathematical operations on data, so it's kind of important that

359
00:21:53,440 --> 00:21:57,920
Speaker 1: it does this correctly. Unfortunately, the flaw and the Pentium

360
00:21:57,920 --> 00:22:00,840
Speaker 1: processors kind of messed that up. And the issue has

361
00:22:00,880 --> 00:22:04,280
Speaker 1: to do with floating point operations. So the predecessor to

362
00:22:04,280 --> 00:22:07,879
Speaker 1: the Pentium, the four six, used a shift and subtract

363
00:22:07,960 --> 00:22:12,800
Speaker 1: algorithm for floating point operations, which was effective but relatively

364
00:22:12,880 --> 00:22:17,000
Speaker 1: slow compared to what Intel thought they could do by

365
00:22:17,280 --> 00:22:22,960
Speaker 1: totally redesigning that structure and using a look up table approach. Now,

366
00:22:22,960 --> 00:22:26,159
Speaker 1: the table was supposed to have one thousand sixty six

367
00:22:26,440 --> 00:22:30,800
Speaker 1: entries programmed directly onto the logic array of the Pentium processor,

368
00:22:31,920 --> 00:22:36,320
Speaker 1: but for some reason only one thousand sixty one entries

369
00:22:36,400 --> 00:22:40,800
Speaker 1: made it. Five entries went missing and essentially returned an

370
00:22:40,800 --> 00:22:45,400
Speaker 1: answer of zero instead of what they were supposed to say,

371
00:22:45,480 --> 00:22:47,919
Speaker 1: So if a calculation accessed one of those missing cells,

372
00:22:47,960 --> 00:22:50,280
Speaker 1: it got zero, even though that's not the correct answer.

373
00:22:51,200 --> 00:22:53,720
Speaker 1: All the first generation pentiums went out with this error

374
00:22:53,800 --> 00:22:56,200
Speaker 1: because it was so minor that it wasn't even picked

375
00:22:56,320 --> 00:23:01,680
Speaker 1: up by Intel's quality control at the time. Now, processes

376
00:23:01,680 --> 00:23:05,159
Speaker 1: work just fine up to the eighth decimal point. Beyond

377
00:23:05,240 --> 00:23:08,360
Speaker 1: that things got messy. But for most folks that wasn't

378
00:23:08,359 --> 00:23:12,959
Speaker 1: a problem because they weren't doing mathematical calculations that needed

379
00:23:12,960 --> 00:23:16,400
Speaker 1: that level of precision. It just wasn't a thing. In fact,

380
00:23:16,440 --> 00:23:19,399
Speaker 1: there was only a one in three sixty billion chance

381
00:23:19,880 --> 00:23:23,320
Speaker 1: that this error would cause a a big enough problem

382
00:23:23,560 --> 00:23:27,480
Speaker 1: to reach up to the fourth decimal place. So most

383
00:23:27,480 --> 00:23:31,000
Speaker 1: calculations that were simple were bulletproof. You you're fine, But

384
00:23:31,080 --> 00:23:34,439
Speaker 1: if you needed that precision, if you needed that really

385
00:23:34,560 --> 00:23:38,399
Speaker 1: fine degree, that's when you would encounter the flaw, and

386
00:23:38,560 --> 00:23:42,439
Speaker 1: that happened because there are math professors in this world,

387
00:23:42,920 --> 00:23:48,040
Speaker 1: and one of those, Thomas nicely discovered in October that

388
00:23:48,119 --> 00:23:51,200
Speaker 1: he was getting errors because of this issue. He needed

389
00:23:51,240 --> 00:23:54,840
Speaker 1: the processor to work correctly, and so he contacted Intel

390
00:23:54,880 --> 00:23:58,800
Speaker 1: about the problem. And this is where we take a

391
00:23:58,880 --> 00:24:01,639
Speaker 1: moment to acknowledge there's a right way and a wrong

392
00:24:01,680 --> 00:24:06,479
Speaker 1: way to handle an issue. That's your fault until decided

393
00:24:06,520 --> 00:24:09,199
Speaker 1: to go the wrong way. My opinion is, if you

394
00:24:09,240 --> 00:24:11,520
Speaker 1: make a mistake, it's usually a good idea to just

395
00:24:11,720 --> 00:24:13,920
Speaker 1: own up to it and try to make it better.

396
00:24:14,600 --> 00:24:16,960
Speaker 1: But Intel's response was more along the lines of yeah,

397
00:24:17,040 --> 00:24:19,200
Speaker 1: we didn't think it was a big deal. And then

398
00:24:19,240 --> 00:24:23,800
Speaker 1: Intel made other pr blenders. But because people began to hear, hey,

399
00:24:23,880 --> 00:24:27,119
Speaker 1: that Pentium processor in your computer that you just bought,

400
00:24:27,880 --> 00:24:31,480
Speaker 1: it doesn't work properly. So people wanted to get replacements.

401
00:24:32,160 --> 00:24:34,520
Speaker 1: But Intel said, oh, we're only going to replace the

402
00:24:34,560 --> 00:24:37,880
Speaker 1: ones if you can prove that the mistakes that it

403
00:24:37,960 --> 00:24:42,119
Speaker 1: makes affect you in some meaningful way. So they weren't

404
00:24:42,600 --> 00:24:45,320
Speaker 1: They weren't denying that there was a problem. They were

405
00:24:45,359 --> 00:24:47,840
Speaker 1: just saying, hey, unless you can prove the problem affects

406
00:24:47,840 --> 00:24:51,880
Speaker 1: you where we don't care. That didn't go well. If

407
00:24:51,880 --> 00:24:53,919
Speaker 1: you create a product and you market it as the

408
00:24:53,960 --> 00:24:56,680
Speaker 1: future of computing and then it's discovered there's a flaw

409
00:24:56,720 --> 00:24:59,280
Speaker 1: on the design, and then you say we'll replace it,

410
00:24:59,320 --> 00:25:02,880
Speaker 1: but only if you proof you deserve it, it doesn't

411
00:25:03,000 --> 00:25:06,919
Speaker 1: tend to make your customer base very happy. So ultimately,

412
00:25:07,040 --> 00:25:11,080
Speaker 1: until reverse that decision and offered to replace the processor

413
00:25:11,160 --> 00:25:15,240
Speaker 1: for anyone who wanted it who had a first generation Pentium,

414
00:25:15,400 --> 00:25:22,760
Speaker 1: and that mistake ended up costing the company four million dollars. Yikes.

415
00:25:23,560 --> 00:25:27,280
Speaker 1: All right, now we're gonna switch Gears over to Microsoft. First.

416
00:25:27,320 --> 00:25:30,399
Speaker 1: I think you could claim that all of Microsoft Bob

417
00:25:31,920 --> 00:25:34,600
Speaker 1: product that was supposed to be an easy, accessible computer

418
00:25:34,640 --> 00:25:38,720
Speaker 1: interface was really just a massive software bug. I mean,

419
00:25:39,080 --> 00:25:44,400
Speaker 1: it introduced comic sands for goodness sakes, the cluttered organization system,

420
00:25:44,520 --> 00:25:49,000
Speaker 1: the lack of meaningful security, and other numerous issues plagued

421
00:25:49,200 --> 00:25:52,560
Speaker 1: that software. But we did an entire episode of Tech

422
00:25:52,560 --> 00:25:55,239
Speaker 1: Stuff about Microsoft Bob a couple of years ago, so

423
00:25:55,240 --> 00:25:57,920
Speaker 1: I'm not gonna dwell on it anymore, but if you

424
00:25:57,960 --> 00:26:00,720
Speaker 1: want to hear more about it, go in that episode.

425
00:26:00,800 --> 00:26:03,760
Speaker 1: It was a fun one. Now in two thousand seven,

426
00:26:04,720 --> 00:26:08,280
Speaker 1: Microsoft experienced a massive headache when a bug on their

427
00:26:08,320 --> 00:26:11,960
Speaker 1: servers notified thousands of Windows customers that they were filthy,

428
00:26:12,040 --> 00:26:15,280
Speaker 1: dirty software pirates and they should be punished. These include

429
00:26:15,320 --> 00:26:20,120
Speaker 1: people who actually had legitimate, legal purchase copies of Windows

430
00:26:20,240 --> 00:26:24,679
Speaker 1: XP or Vista. So the problem here was Microsoft had

431
00:26:24,720 --> 00:26:29,280
Speaker 1: an initiative called Windows Genuine Advantage, and it was a

432
00:26:29,359 --> 00:26:35,040
Speaker 1: nice name for a strategy meant to curtail operating system piracy. Essentially,

433
00:26:35,080 --> 00:26:37,960
Speaker 1: it was a component in Windows that would allow Microsoft

434
00:26:38,040 --> 00:26:40,680
Speaker 1: to figure out if the copy of Windows on any

435
00:26:40,720 --> 00:26:45,199
Speaker 1: given computer was legit. In other words, it was a

436
00:26:45,280 --> 00:26:48,080
Speaker 1: d r M strategy. But in two thousand seven, a

437
00:26:48,119 --> 00:26:54,000
Speaker 1: buggy install of software on a server misidentified thousands of legitimate,

438
00:26:54,520 --> 00:26:59,400
Speaker 1: law abiding customers as pirates for nineteen hours. The software

439
00:26:59,440 --> 00:27:01,679
Speaker 1: just laid on the law, and so people began to

440
00:27:01,720 --> 00:27:05,000
Speaker 1: receive sternly written warnings about their choice to indulge in

441
00:27:05,119 --> 00:27:08,440
Speaker 1: bad behavior. And if you were a Windows Vista customer,

442
00:27:08,920 --> 00:27:11,639
Speaker 1: you had it the worst, not just because you were

443
00:27:11,720 --> 00:27:15,239
Speaker 1: using Windows Vista, which I think we all agree was

444
00:27:15,400 --> 00:27:20,200
Speaker 1: not one of the bright points and Microsoft's operating system history,

445
00:27:20,520 --> 00:27:23,919
Speaker 1: but also because Microsoft had built in the ability for

446
00:27:24,000 --> 00:27:28,760
Speaker 1: Windows genuine advantage to switch off certain operating system features

447
00:27:29,000 --> 00:27:32,680
Speaker 1: in Windows Vista if it determined that the copy someone

448
00:27:32,760 --> 00:27:36,560
Speaker 1: was using was a pirated version, so it was misidentifying

449
00:27:36,640 --> 00:27:40,280
Speaker 1: real versions as pirated ones turning off features, and these

450
00:27:40,280 --> 00:27:43,119
Speaker 1: are for people who have bought legitimate copies. This, by

451
00:27:43,119 --> 00:27:45,639
Speaker 1: the way, is one of the big arguments people have

452
00:27:45,880 --> 00:27:51,240
Speaker 1: against DRM. It has the tendency to punish legitimate customers.

453
00:27:52,080 --> 00:27:56,280
Speaker 1: And you feel like you're stupid for buying a copy

454
00:27:56,520 --> 00:27:58,960
Speaker 1: of a piece of software rather than just stealing one

455
00:27:59,040 --> 00:28:02,119
Speaker 1: that has had those feet yours or those defenses removed,

456
00:28:02,640 --> 00:28:05,439
Speaker 1: Like why you're you're creating more incentives for people to

457
00:28:05,480 --> 00:28:11,000
Speaker 1: go outside and get a pirated copy. Alright, so imagine

458
00:28:11,040 --> 00:28:14,240
Speaker 1: you've purchased this legitimate copy of Windows Vista. First of all,

459
00:28:14,280 --> 00:28:17,200
Speaker 1: you you already feel bad. Then you're told you're a thief,

460
00:28:17,280 --> 00:28:20,320
Speaker 1: so you feel worse. Then someone remotely switches off several

461
00:28:20,359 --> 00:28:22,560
Speaker 1: features of your operating system. That was not a great

462
00:28:22,600 --> 00:28:26,880
Speaker 1: pr message, So that was a real issue. They did

463
00:28:26,920 --> 00:28:29,800
Speaker 1: eventually fix it after that nineteen hours, but by then

464
00:28:29,840 --> 00:28:32,720
Speaker 1: people were already very upset. Also, I don't want to

465
00:28:33,040 --> 00:28:38,080
Speaker 1: just you know, pile lots of abuse onto Microsoft I

466
00:28:38,080 --> 00:28:42,240
Speaker 1: gotta talk about Apple here too. So the company prides

467
00:28:42,280 --> 00:28:45,880
Speaker 1: itself on a high standard of quality and general it's

468
00:28:45,920 --> 00:28:48,920
Speaker 1: pretty good about living up to that standard of quality,

469
00:28:49,080 --> 00:28:51,600
Speaker 1: depending upon your point of view of their various products.

470
00:28:52,200 --> 00:28:55,880
Speaker 1: But that hasn't stopped a few clunkers getting through and

471
00:28:56,440 --> 00:28:59,400
Speaker 1: into the public hands. And that was the case in

472
00:28:59,400 --> 00:29:02,400
Speaker 1: two thousand twelve with Apple Maps. If you owned an

473
00:29:02,400 --> 00:29:04,960
Speaker 1: iPhone back in two thousand twelve when Apple Maps came out,

474
00:29:05,360 --> 00:29:08,960
Speaker 1: you may remember this problem. It's pretty well publicized. Maps

475
00:29:08,960 --> 00:29:12,120
Speaker 1: were inaccurate, sometimes leaving out important details like you know,

476
00:29:12,320 --> 00:29:15,440
Speaker 1: a river or a lake between you and your destination,

477
00:29:16,200 --> 00:29:18,160
Speaker 1: things that might be important if I don't know you

478
00:29:18,240 --> 00:29:21,680
Speaker 1: don't drive an amphibious vehicle, might not have a road

479
00:29:21,760 --> 00:29:25,800
Speaker 1: on there that's important. Might misidentify the location of a

480
00:29:25,920 --> 00:29:29,600
Speaker 1: historical landmark. For instance, that thought the Washington Monument was

481
00:29:29,640 --> 00:29:32,800
Speaker 1: across the street from where it is, But nope, it's

482
00:29:32,880 --> 00:29:36,600
Speaker 1: just where we left it. Despite all of Roland emericks

483
00:29:36,920 --> 00:29:40,880
Speaker 1: best attempts to move it or destroy it, it's still there.

484
00:29:43,280 --> 00:29:45,800
Speaker 1: The real problem here was that the Apple software just

485
00:29:45,880 --> 00:29:49,320
Speaker 1: wasn't ready for public unveiling. It was. It needed a

486
00:29:49,400 --> 00:29:52,400
Speaker 1: lot more testing. It was trying to play catch up

487
00:29:52,440 --> 00:29:55,479
Speaker 1: to Google Maps, but Google had the advantage of working

488
00:29:55,480 --> 00:29:58,880
Speaker 1: with companies that have been doing mapping software for years.

489
00:29:58,960 --> 00:30:02,000
Speaker 1: Google acquired the those companies and acquired the expertise of

490
00:30:02,040 --> 00:30:04,880
Speaker 1: people who have been working on that software. And Apple

491
00:30:05,040 --> 00:30:08,800
Speaker 1: was really just trying to create their own version and

492
00:30:08,840 --> 00:30:10,600
Speaker 1: get it out as fast as it could. But it

493
00:30:10,680 --> 00:30:13,400
Speaker 1: got out a little too early, and the company spent

494
00:30:13,480 --> 00:30:15,800
Speaker 1: the next several months tweaking maps and trying to keep

495
00:30:15,800 --> 00:30:17,840
Speaker 1: control of the situation. But by that time, many of

496
00:30:17,840 --> 00:30:20,840
Speaker 1: Apple's fans, even the most devoted ones, had kind of

497
00:30:20,880 --> 00:30:24,920
Speaker 1: given up and switched over to Google Maps instead. Well,

498
00:30:24,960 --> 00:30:27,440
Speaker 1: that's most of the fun stuff. I've got some really

499
00:30:27,680 --> 00:30:31,760
Speaker 1: serious bugs to cover. But before I do that, let's

500
00:30:31,760 --> 00:30:43,400
Speaker 1: take another quick break and thank our sponsor. Now I'm

501
00:30:43,440 --> 00:30:46,760
Speaker 1: going to transition into some serious bugs. These are ones

502
00:30:46,840 --> 00:30:51,400
Speaker 1: that either threatened the lives of people or they contributed

503
00:30:51,480 --> 00:30:56,680
Speaker 1: to people dying. The ones I've talked about now up

504
00:30:56,720 --> 00:30:59,760
Speaker 1: to now rather have cost companies millions of dollars, but

505
00:31:00,160 --> 00:31:04,480
Speaker 1: one's life was truly threatened. Unfortunately, that's not the case

506
00:31:04,520 --> 00:31:07,200
Speaker 1: with all software bugs. Now, a couple of bugs had

507
00:31:07,200 --> 00:31:11,960
Speaker 1: the potential to kill millions of people. One of those

508
00:31:12,000 --> 00:31:17,560
Speaker 1: happened in night a famous, famous bug, or at least

509
00:31:17,560 --> 00:31:20,360
Speaker 1: a faulty circuit, and that was a faulty circuit in

510
00:31:20,440 --> 00:31:23,680
Speaker 1: nora ADS computer system which caused it to mistakenly conclude

511
00:31:23,720 --> 00:31:26,720
Speaker 1: the US was under nuclear attack from the Soviet Union.

512
00:31:28,400 --> 00:31:32,600
Speaker 1: So displays on nora AD systems showed seemingly random attacks,

513
00:31:32,600 --> 00:31:35,440
Speaker 1: and they didn't correspond with each other. So the display

514
00:31:35,520 --> 00:31:38,240
Speaker 1: might show, Hey, there two missiles heading over from the

515
00:31:38,280 --> 00:31:41,600
Speaker 1: Soviet Union. No, they're two hundred. No they're fifty. No,

516
00:31:41,760 --> 00:31:45,880
Speaker 1: there's three, And it wasn't consistent, and command posts around

517
00:31:46,080 --> 00:31:49,840
Speaker 1: the US all had conflicting information, which led leaders to

518
00:31:49,880 --> 00:31:54,400
Speaker 1: conclude the whole thing was a regrettable computer error, and

519
00:31:54,440 --> 00:31:57,560
Speaker 1: they were right to do so. To be fair, they

520
00:31:57,560 --> 00:31:59,880
Speaker 1: were kind of prepared for this because there was another

521
00:32:00,040 --> 00:32:02,600
Speaker 1: incident that had actually happened in nineteen seventy nine that

522
00:32:02,720 --> 00:32:06,760
Speaker 1: was way scarier, and in that case, someone mistakenly inserted

523
00:32:06,760 --> 00:32:09,360
Speaker 1: a training scenario into the computer system that made it

524
00:32:09,400 --> 00:32:11,760
Speaker 1: seem like the Soviet Union had launched an all out

525
00:32:11,840 --> 00:32:15,200
Speaker 1: nuclear attack on the US. But that wasn't a bug.

526
00:32:15,360 --> 00:32:17,080
Speaker 1: That was a mistake on the part of a human

527
00:32:17,120 --> 00:32:20,960
Speaker 1: who had accidentally uploaded the wrong or rather executed the

528
00:32:20,960 --> 00:32:23,560
Speaker 1: wrong command. It didn't have something to do with a

529
00:32:23,600 --> 00:32:27,680
Speaker 1: flaw in the computer system itself, however, because that thing

530
00:32:27,760 --> 00:32:32,360
Speaker 1: happened and everybody was freaked out and then was able

531
00:32:32,360 --> 00:32:34,320
Speaker 1: to determine that, in fact it was a false alarm,

532
00:32:34,640 --> 00:32:37,520
Speaker 1: it meant that calmer heads could prevail in the nineteen

533
00:32:37,560 --> 00:32:42,600
Speaker 1: eighty incident, so the Soviets also had a close call

534
00:32:42,680 --> 00:32:44,720
Speaker 1: just a few years later. It was a bug in

535
00:32:44,760 --> 00:32:48,640
Speaker 1: the early warning detection software that the USSR was using

536
00:32:48,840 --> 00:32:51,600
Speaker 1: in the early eighties, and on September twenty three, nine

537
00:32:51,840 --> 00:32:55,560
Speaker 1: eight three and so Union received an alert that the

538
00:32:55,680 --> 00:32:58,640
Speaker 1: US had launched a nuclear attack in the form of

539
00:32:58,800 --> 00:33:03,200
Speaker 1: five nuclear war It's uh, technically two different attacks. The

540
00:33:03,280 --> 00:33:06,280
Speaker 1: first would have been a single nuclear warhead and the

541
00:33:06,360 --> 00:33:09,920
Speaker 1: second was four nuclear warheads. And this was during a

542
00:33:09,920 --> 00:33:13,600
Speaker 1: particularly stressful period in the history of both countries and

543
00:33:13,640 --> 00:33:16,880
Speaker 1: their relationship with each other, at the height of the

544
00:33:16,920 --> 00:33:23,800
Speaker 1: Cold War nine three now fortunately, UH Soviet Air Defense

545
00:33:23,840 --> 00:33:29,960
Speaker 1: Forces Lieutenant Colonel Stanislav Petrov suspected that this report was

546
00:33:30,040 --> 00:33:32,920
Speaker 1: an error and that there was some sort of bug

547
00:33:32,920 --> 00:33:36,479
Speaker 1: in the software or a mistake in the reporting system

548
00:33:36,520 --> 00:33:39,960
Speaker 1: that caused this, he gave a command to hold off

549
00:33:40,000 --> 00:33:43,240
Speaker 1: on any sort of retaliatory strike, which would have initiated

550
00:33:43,240 --> 00:33:47,040
Speaker 1: a full scale nuclear war had it happened. Petrov was

551
00:33:47,360 --> 00:33:49,680
Speaker 1: the officer in charge of a bunker that served as

552
00:33:49,680 --> 00:33:52,840
Speaker 1: the command center for this early warning system, and he

553
00:33:52,840 --> 00:33:56,640
Speaker 1: he had said afterward that his reckoning was any real

554
00:33:56,640 --> 00:34:00,680
Speaker 1: attack would consist of hundreds of warheads, not five. No

555
00:34:00,760 --> 00:34:04,560
Speaker 1: one would start an attack with just five warheads, so

556
00:34:04,600 --> 00:34:06,320
Speaker 1: it was more likely to be an error than a

557
00:34:06,360 --> 00:34:09,440
Speaker 1: genuine attack. So he gave the command to wait until

558
00:34:09,440 --> 00:34:12,480
Speaker 1: the reported missiles would pass into the range of radar,

559
00:34:13,000 --> 00:34:16,720
Speaker 1: which only extended as far as the horizon, so if

560
00:34:16,880 --> 00:34:18,959
Speaker 1: it had in fact been a real attack, it would

561
00:34:18,960 --> 00:34:22,759
Speaker 1: have potentially limited the Soviet Union's ability to respond. But

562
00:34:23,120 --> 00:34:29,040
Speaker 1: no missile showed up, and he was vindicated in his decision. Now,

563
00:34:29,040 --> 00:34:31,080
Speaker 1: the cause of the false alarm in this case was

564
00:34:31,760 --> 00:34:36,840
Speaker 1: a combination of factors that the designers didn't anticipate, uh

565
00:34:36,920 --> 00:34:41,279
Speaker 1: which largely consisted of sunlight hitting high altitude clouds at

566
00:34:41,320 --> 00:34:45,359
Speaker 1: a particular angle from a particular perspective of the satellites,

567
00:34:45,760 --> 00:34:52,239
Speaker 1: so the satellites misidentified that reflection as a warhead. Now,

568
00:34:52,280 --> 00:34:54,480
Speaker 1: they were. The subvieys were able to address this error

569
00:34:54,520 --> 00:34:58,120
Speaker 1: in the future by adding another step in which these

570
00:34:58,160 --> 00:35:02,520
Speaker 1: satellites would cross referenced data from other geostationary satellites to

571
00:35:02,640 --> 00:35:07,680
Speaker 1: make certain that they are identifying actual rockets as opposed

572
00:35:07,719 --> 00:35:13,640
Speaker 1: to high altitude clouds. Now, there are several cases of

573
00:35:14,640 --> 00:35:19,360
Speaker 1: software bugs leading to actual deaths. For example, the the

574
00:35:19,560 --> 00:35:22,880
Speaker 1: RACK was such a case. Now that was a radiation

575
00:35:22,960 --> 00:35:26,799
Speaker 1: therapy machine that could deliver two different modes of radiation treatments.

576
00:35:27,400 --> 00:35:30,719
Speaker 1: The first was a low powered direct electron beam and

577
00:35:30,760 --> 00:35:33,960
Speaker 1: the second was a mega volt X ray beam. Now,

578
00:35:33,960 --> 00:35:36,440
Speaker 1: the x ray beam was far more intense and it

579
00:35:36,520 --> 00:35:40,839
Speaker 1: required physicians to provide shielding to patients to limit exposure

580
00:35:40,880 --> 00:35:44,880
Speaker 1: to the beam. But the therac had inherited its code

581
00:35:44,920 --> 00:35:49,000
Speaker 1: from its predecessor, which had different hardware constraints. Now the

582
00:35:49,040 --> 00:35:53,160
Speaker 1: new machine meant that these constraints weren't there, and it

583
00:35:53,239 --> 00:35:56,600
Speaker 1: created a deadly problem if operators changed the machines mode

584
00:35:56,719 --> 00:35:59,600
Speaker 1: too quickly from one to the other, it would actually

585
00:35:59,600 --> 00:36:03,000
Speaker 1: send sets of instructions to the processor, one for each

586
00:36:03,080 --> 00:36:06,160
Speaker 1: mode of operation, and whichever set of instructions reached the

587
00:36:06,160 --> 00:36:10,680
Speaker 1: processor first. That's what the machine would switch to So

588
00:36:10,760 --> 00:36:15,640
Speaker 1: let's say you've been operating the therac in the mega

589
00:36:15,719 --> 00:36:18,759
Speaker 1: volt X ray mode, but now you're going to have

590
00:36:18,880 --> 00:36:21,720
Speaker 1: a patient come in. You need to administer radiation therapy,

591
00:36:22,040 --> 00:36:24,560
Speaker 1: so you want to switch it to low electron. Being

592
00:36:25,080 --> 00:36:27,880
Speaker 1: you switch it too quickly, it sends two sets of

593
00:36:27,920 --> 00:36:30,560
Speaker 1: instructions to the processor, and the one that arises the

594
00:36:30,600 --> 00:36:33,759
Speaker 1: mega volt x ray instruction, so instead of switching it,

595
00:36:34,040 --> 00:36:39,000
Speaker 1: you confirm to stay on the more intense, deadlier radiation.

596
00:36:40,520 --> 00:36:45,799
Speaker 1: The tragic news is this did happen several times. Six

597
00:36:45,840 --> 00:36:49,840
Speaker 1: patients were documented as dying from complications due to radiation

598
00:36:49,880 --> 00:36:53,560
Speaker 1: poisoning from there at twenty five machines between night five

599
00:36:53,640 --> 00:36:56,840
Speaker 1: and nineteen eight six, And while the machine would send

600
00:36:56,960 --> 00:37:01,360
Speaker 1: error messages when these conditions were present, the documentation for

601
00:37:01,400 --> 00:37:04,640
Speaker 1: the machine didn't explain what the errors meant. It didn't say, hey,

602
00:37:04,640 --> 00:37:07,400
Speaker 1: if you get this error, it means that you've switched

603
00:37:07,480 --> 00:37:11,399
Speaker 1: modes too quickly and you need to address this. So,

604
00:37:11,440 --> 00:37:16,000
Speaker 1: since operators weren't told that this was necessarily a hazardous condition,

605
00:37:16,080 --> 00:37:19,440
Speaker 1: they would just clear the error and proceed, and there

606
00:37:19,440 --> 00:37:24,680
Speaker 1: were deadly results in a similar vein in Panama City, Panama,

607
00:37:25,520 --> 00:37:29,120
Speaker 1: there was an incident involving a Cobalt sixty system, actually

608
00:37:29,200 --> 00:37:32,840
Speaker 1: several incidents involving this Cobalt sixty system that was running

609
00:37:32,880 --> 00:37:35,960
Speaker 1: therapy planning software made by a company called Multi Data

610
00:37:36,000 --> 00:37:39,799
Speaker 1: Systems International. Now, the software's purpose was to calculate the

611
00:37:39,840 --> 00:37:44,040
Speaker 1: amount of radiation that cancer patients should receive in radiation

612
00:37:44,080 --> 00:37:49,600
Speaker 1: therapy sessions. During these radiation therapy sessions, the therapists were

613
00:37:49,600 --> 00:37:53,920
Speaker 1: meant to place metal shields on the patient to protect

614
00:37:54,000 --> 00:37:59,080
Speaker 1: healthy tissue from radiation damage. And the software would allow

615
00:37:59,160 --> 00:38:03,560
Speaker 1: therapists to use a methodology to show where those shields

616
00:38:03,600 --> 00:38:06,560
Speaker 1: were on the patient, to indicate where the shields are present.

617
00:38:07,640 --> 00:38:11,000
Speaker 1: But they could only draw up to four shields, and

618
00:38:11,040 --> 00:38:14,040
Speaker 1: the doctors in Panama wanted to use five shields for

619
00:38:14,160 --> 00:38:18,239
Speaker 1: particular therapy sessions. They were overloaded, they had a long

620
00:38:18,280 --> 00:38:20,440
Speaker 1: waiting list of patients, and they were trying to make

621
00:38:20,480 --> 00:38:24,439
Speaker 1: things more efficient, and they discovered that they could kind

622
00:38:24,480 --> 00:38:29,000
Speaker 1: of work around this limitation of four shields by drawing

623
00:38:29,040 --> 00:38:31,120
Speaker 1: a design on the computer screen as if they were

624
00:38:31,200 --> 00:38:34,319
Speaker 1: using just one large shield that has a hole in

625
00:38:34,360 --> 00:38:36,279
Speaker 1: the middle of it. And so what they would do

626
00:38:36,320 --> 00:38:39,080
Speaker 1: is they would arrange the five shields to essentially be

627
00:38:39,080 --> 00:38:41,880
Speaker 1: in the same sort of shape with the middle of

628
00:38:41,920 --> 00:38:44,480
Speaker 1: it being open so that they can have the radiation

629
00:38:44,520 --> 00:38:49,600
Speaker 1: therapy passed through it. Uh, But they didn't realize that

630
00:38:49,719 --> 00:38:52,000
Speaker 1: the software had a bug in it, and that bug

631
00:38:52,160 --> 00:38:55,400
Speaker 1: was if you drew the whole in one direction, you

632
00:38:55,440 --> 00:38:58,399
Speaker 1: get the correct dose of radiation, but if you drew

633
00:38:58,440 --> 00:39:02,600
Speaker 1: it in the other direction, so like clockwise versus counterclockwise,

634
00:39:03,120 --> 00:39:06,839
Speaker 1: the software would recommend a dosage twice as strong as

635
00:39:06,880 --> 00:39:10,480
Speaker 1: what was needed, and the result was devastating. Eight patients

636
00:39:10,640 --> 00:39:14,840
Speaker 1: died as a result of this, and another twenty received

637
00:39:14,880 --> 00:39:17,640
Speaker 1: doses high enough to potentially cause health problems. Later on,

638
00:39:18,960 --> 00:39:21,959
Speaker 1: the physicians were actually arrested and brought up on murder

639
00:39:22,120 --> 00:39:25,760
Speaker 1: charges because they were supposed to double check all calculations

640
00:39:25,800 --> 00:39:28,440
Speaker 1: by hand to ensure that they were going to give

641
00:39:28,480 --> 00:39:31,960
Speaker 1: the proper dose of radiation treatment. So while the software

642
00:39:32,160 --> 00:39:37,040
Speaker 1: was calculating the incorrect dose, the physicians were responsible for

643
00:39:37,120 --> 00:39:40,640
Speaker 1: making sure that any dose that was calculated was in

644
00:39:40,680 --> 00:39:42,480
Speaker 1: fact the correct one, and they failed to do so,

645
00:39:42,840 --> 00:39:46,600
Speaker 1: or at least that was the charge. There are also

646
00:39:46,680 --> 00:39:49,640
Speaker 1: bugs that involved military applications that have resulted in the

647
00:39:49,719 --> 00:39:52,640
Speaker 1: loss of life. During the Persian Gulf War in Iraqi,

648
00:39:52,840 --> 00:39:55,800
Speaker 1: fired scud missile hit a US base in Saudi Arabia

649
00:39:56,080 --> 00:40:00,120
Speaker 1: and it killed twenty eight soldiers. Now the base had

650
00:40:00,160 --> 00:40:04,000
Speaker 1: detected the missile and had launched and fired a Patriot

651
00:40:04,000 --> 00:40:06,719
Speaker 1: missile in return. The purpose of the Patriot missile was

652
00:40:06,760 --> 00:40:09,880
Speaker 1: to intercept and destroy incoming missiles, and the way a

653
00:40:09,880 --> 00:40:12,840
Speaker 1: Patriot missile did this was to use radar pulses to

654
00:40:12,960 --> 00:40:17,120
Speaker 1: guide trajectory calculations so that it would end up getting

655
00:40:17,160 --> 00:40:19,880
Speaker 1: close to the incoming missile. This is harder than it

656
00:40:19,960 --> 00:40:23,520
Speaker 1: sounds because both missiles are moving very very quickly, so

657
00:40:23,560 --> 00:40:26,600
Speaker 1: we need a very precise information in order to adjust

658
00:40:26,680 --> 00:40:31,400
Speaker 1: its trajectory properly and make sure it was on target. Now,

659
00:40:31,440 --> 00:40:34,279
Speaker 1: once it gets within range, which is between five and

660
00:40:34,520 --> 00:40:38,399
Speaker 1: ten meters I think uh, it would then fire out

661
00:40:38,840 --> 00:40:42,440
Speaker 1: a thousand pellets from the Patriot missile at high velocity

662
00:40:42,480 --> 00:40:46,080
Speaker 1: with the goal of causing the incoming warhead to explode prematurely.

663
00:40:47,719 --> 00:40:50,560
Speaker 1: In this case, the Patriot missile missed and the military

664
00:40:50,600 --> 00:40:52,880
Speaker 1: investigated the issue in the wake of the loss of

665
00:40:52,920 --> 00:40:55,600
Speaker 1: life and found a problem with the software guiding the

666
00:40:55,600 --> 00:40:58,759
Speaker 1: Patriot missile, and it was a problem that actually the

667
00:40:58,760 --> 00:41:01,000
Speaker 1: military kind of knew about already. So one of the

668
00:41:01,040 --> 00:41:04,080
Speaker 1: processes in the Patriots programming was to convert time into

669
00:41:04,120 --> 00:41:10,000
Speaker 1: floating point operations for increased accuracy. But not all subroutines

670
00:41:10,560 --> 00:41:15,200
Speaker 1: that depended on tracking time did this. Some of them

671
00:41:15,200 --> 00:41:19,759
Speaker 1: remained UH clock units rather than floating point operations, which

672
00:41:19,800 --> 00:41:22,960
Speaker 1: meant that they would get out of sync after a while.

673
00:41:23,000 --> 00:41:26,160
Speaker 1: There'd be a disagreement in various subroutines as to what

674
00:41:26,600 --> 00:41:29,279
Speaker 1: how much time had actually passed. And like I said,

675
00:41:29,280 --> 00:41:31,560
Speaker 1: the military was aware of this issue and they had

676
00:41:31,600 --> 00:41:35,440
Speaker 1: a work around, which was not ideal. The workaround was

677
00:41:35,880 --> 00:41:38,840
Speaker 1: you would occasionally reboot the system, which would reset the

678
00:41:38,840 --> 00:41:41,480
Speaker 1: clocks and synchronize them, but over time they would fall

679
00:41:41,520 --> 00:41:44,320
Speaker 1: out of sync because they're not tracking time the same way.

680
00:41:44,960 --> 00:41:47,520
Speaker 1: And since there was no hard and fast rule as

681
00:41:47,520 --> 00:41:50,799
Speaker 1: to how frequently you'd reset the system, problems like this

682
00:41:50,840 --> 00:41:53,160
Speaker 1: one where possible, and in fact, in this case it

683
00:41:53,200 --> 00:41:57,040
Speaker 1: did happen. So prior to this particular incident, that specific

684
00:41:57,080 --> 00:42:00,080
Speaker 1: Patriot system had been running for one hours with how

685
00:42:00,120 --> 00:42:05,319
Speaker 1: to reboot, and the clock disagreement amounted to about one

686
00:42:05,400 --> 00:42:07,680
Speaker 1: third of a second. Now, that seems like it's no

687
00:42:07,800 --> 00:42:09,760
Speaker 1: time at all. One third of a second is so

688
00:42:09,760 --> 00:42:13,440
Speaker 1: so short, But a scutt missile's top speed is about

689
00:42:13,480 --> 00:42:17,200
Speaker 1: one point one miles per second or one point seven

690
00:42:17,280 --> 00:42:20,120
Speaker 1: kilometers per second, which means if you take a third

691
00:42:20,120 --> 00:42:22,399
Speaker 1: of a second, the missile could travel more than five

692
00:42:24,000 --> 00:42:26,160
Speaker 1: And since the patriot needs to be within ten ms

693
00:42:26,160 --> 00:42:28,279
Speaker 1: of a target to destroy it, that resulted in a

694
00:42:28,360 --> 00:42:32,960
Speaker 1: catastrophic failure. So software bugs can be a matter of

695
00:42:33,000 --> 00:42:36,239
Speaker 1: life or death. It's not all just Hey, this irritating

696
00:42:36,320 --> 00:42:40,480
Speaker 1: thing meant people couldn't make long distance phone calls or uh,

697
00:42:40,520 --> 00:42:44,680
Speaker 1: this issue caused my computer to start writing massive amounts

698
00:42:44,680 --> 00:42:47,799
Speaker 1: of data to its hard drive. And this is why

699
00:42:48,200 --> 00:42:53,080
Speaker 1: it's so important to have really qualified q A personnel

700
00:42:53,360 --> 00:42:55,879
Speaker 1: go through code and make sure it's doing what it's

701
00:42:55,880 --> 00:42:58,720
Speaker 1: supposed to do, because the problems that can arise can

702
00:42:58,760 --> 00:43:01,680
Speaker 1: be non trivial and in fact life or death situations

703
00:43:02,000 --> 00:43:05,959
Speaker 1: depending upon the application of technology. So technology is a

704
00:43:06,000 --> 00:43:09,799
Speaker 1: fascinating thing. It's a wonderful thing. It has benefited us

705
00:43:09,840 --> 00:43:13,399
Speaker 1: in ways that I can't even begin to describe. It's

706
00:43:13,440 --> 00:43:16,000
Speaker 1: just too broad a topic, and it's something I've been

707
00:43:16,040 --> 00:43:19,200
Speaker 1: tackling for, you know, eight years and I haven't haven't

708
00:43:19,239 --> 00:43:22,279
Speaker 1: even gotten close to getting toward the finishing point. So

709
00:43:23,400 --> 00:43:26,000
Speaker 1: I don't want to suggest that technology is bad, but

710
00:43:26,080 --> 00:43:30,360
Speaker 1: we definitely have the need to check, double check, and

711
00:43:30,400 --> 00:43:32,799
Speaker 1: triple check all this work to make certain things are

712
00:43:32,880 --> 00:43:35,840
Speaker 1: working properly before we release them out into the wild.

713
00:43:35,880 --> 00:43:40,640
Speaker 1: That particularly applies if, again, you are reusing old code

714
00:43:41,000 --> 00:43:45,680
Speaker 1: or old components in a new way, because you have

715
00:43:45,760 --> 00:43:47,959
Speaker 1: to make absolutely certain that there's not going to be

716
00:43:48,080 --> 00:43:53,200
Speaker 1: some unintended problem that results when a new form factor

717
00:43:53,440 --> 00:43:57,200
Speaker 1: is using old code. I hope you guys found this

718
00:43:57,280 --> 00:44:00,719
Speaker 1: episode interesting. I plan on doing a lot of other

719
00:44:00,960 --> 00:44:06,479
Speaker 1: kind of list sort of podcasts in the future, things

720
00:44:06,520 --> 00:44:11,680
Speaker 1: like some of the the most successful pieces of software

721
00:44:11,680 --> 00:44:15,439
Speaker 1: of all time, the most uh popular gadgets of all time,

722
00:44:15,480 --> 00:44:17,480
Speaker 1: that sort of stuff. But if you have any suggestions

723
00:44:17,480 --> 00:44:20,640
Speaker 1: for that kind of topic or anything else, really you

724
00:44:20,680 --> 00:44:23,839
Speaker 1: should write me. My email address is tech stuff at

725
00:44:23,920 --> 00:44:26,400
Speaker 1: how stuff works dot com, or you can drop me

726
00:44:26,440 --> 00:44:30,000
Speaker 1: a line on Facebook for Twitter to handle it. Both

727
00:44:30,000 --> 00:44:33,560
Speaker 1: of those is text stuff H s W. And I'll

728
00:44:33,600 --> 00:44:42,719
Speaker 1: talk to you guys again really soon for more on

729
00:44:42,719 --> 00:44:45,200
Speaker 1: this and thousands of other topics. Is it how stuff

730
00:44:45,200 --> 00:44:55,680
Speaker 1: Works dot Com