1 00:00:00,000 --> 00:00:01,120 2 00:00:01,120 --> 00:00:01,840 All right. 3 00:00:01,840 --> 00:00:09,270 Well, I became interested in artificial intelligence 4 00:00:09,270 --> 00:00:10,990 before I became interested in computers. 5 00:00:10,990 --> 00:00:14,430 This was in 1949. 6 00:00:14,430 --> 00:00:20,430 There was this symposium at Caltech on cerebral mechanisms 7 00:00:20,430 --> 00:00:23,430 and behavior, the so-called Hixon Symposium. 8 00:00:23,430 --> 00:00:26,170 And there were various bigshots there 9 00:00:26,170 --> 00:00:32,280 von Neumann and McCulloch, and psychologists, 10 00:00:32,280 --> 00:00:34,290 and physiologists. 11 00:00:34,290 --> 00:00:37,620 And at that time, I began to think 12 00:00:37,620 --> 00:00:41,370 about artificial intelligence from the point of view 13 00:00:41,370 --> 00:00:45,750 of special purpose computers. 14 00:00:45,750 --> 00:00:48,180 However, this was by no means my main activity, 15 00:00:48,180 --> 00:00:50,330 since I was a graduate student in mathematics 16 00:00:50,330 --> 00:00:52,170 and went to Princeton as a graduate student 17 00:00:52,170 --> 00:00:55,200 in mathematics, and felt that these ideas were 18 00:00:55,200 --> 00:00:59,640 too indefinite, although I continued 19 00:00:59,640 --> 00:01:03,810 to think about them for a serious thesis 20 00:01:03,810 --> 00:01:06,420 work at that time. 21 00:01:06,420 --> 00:01:08,310 I showed some ideas to von Neumann 22 00:01:08,310 --> 00:01:10,050 who said, write them up. 23 00:01:10,050 --> 00:01:11,642 Write them up. 24 00:01:11,642 --> 00:01:12,850 But they weren't good enough. 25 00:01:12,850 --> 00:01:14,310 And I didn't write it up. 26 00:01:14,310 --> 00:01:21,840 And I had listened to some lectures on computers 27 00:01:21,840 --> 00:01:25,830 at Caltech that time, which were actually-- 28 00:01:25,830 --> 00:01:29,160 I remember the exercise was to program divide 29 00:01:29,160 --> 00:01:31,530 for the SWAC, which was a computer which 30 00:01:31,530 --> 00:01:35,220 was going to appear. 31 00:01:35,220 --> 00:01:37,110 I don't think I understood it very well. 32 00:01:37,110 --> 00:01:40,950 Or if I did, I forgot, because I don't think-- well, 33 00:01:40,950 --> 00:01:43,672 there wasn't any SWAC to try it out on anyway. 34 00:01:43,672 --> 00:01:45,130 There was eventually, wasn't there? 35 00:01:45,130 --> 00:01:45,750 What? 36 00:01:45,750 --> 00:01:46,758 There was eventually. 37 00:01:46,758 --> 00:01:48,300 Sure, there was eventually, but there 38 00:01:48,300 --> 00:01:51,630 wasn't at the time these lectures were given. 39 00:01:51,630 --> 00:01:55,710 At least, we certainly weren't offered that opportunity. 40 00:01:55,710 --> 00:01:58,733 My first actual contact with computers-- 41 00:01:58,733 --> 00:01:59,400 well, let's see. 42 00:01:59,400 --> 00:02:01,970 Then I thought some more. 43 00:02:01,970 --> 00:02:08,160 The first time I spent full time on artificial intelligence 44 00:02:08,160 --> 00:02:12,000 was in the summer of '52 at Bell Labs, 45 00:02:12,000 --> 00:02:16,500 where Minsky and I were hired by Shannon. 46 00:02:16,500 --> 00:02:20,770 And again, I was interested in artificial intelligence 47 00:02:20,770 --> 00:02:25,050 and machines were doing it, but paid no attention to computers. 48 00:02:25,050 --> 00:02:29,940 Bell Labs at that time had IBM 603 or four, 49 00:02:29,940 --> 00:02:32,175 which were calculating punches. 50 00:02:32,175 --> 00:02:35,280 And again, no one thought that computers had much 51 00:02:35,280 --> 00:02:36,570 to do with it. 52 00:02:36,570 --> 00:02:40,080 Now, the one person who already had expressed the view 53 00:02:40,080 --> 00:02:42,660 that computers were the key to artificial intelligence 54 00:02:42,660 --> 00:02:45,960 was Turing who had written this article, "Computing Machinery 55 00:02:45,960 --> 00:02:49,140 and Intelligence" that was published in 1950. 56 00:02:49,140 --> 00:02:53,280 But I can't find anyone interested 57 00:02:53,280 --> 00:02:54,900 in artificial intelligence who had 58 00:02:54,900 --> 00:02:59,170 read that article, with the exception of Selfridge, who 59 00:02:59,170 --> 00:03:01,620 claimed he'd read it, but didn't believe it. 60 00:03:01,620 --> 00:03:04,530 At least, his approach towards artificial intelligence 61 00:03:04,530 --> 00:03:08,070 was with computers. 62 00:03:08,070 --> 00:03:14,220 Now, I became convinced first, that computers 63 00:03:14,220 --> 00:03:18,030 were the key to artificial intelligence in '55. 64 00:03:18,030 --> 00:03:22,560 And at that point, the idea of time sharing 65 00:03:22,560 --> 00:03:27,990 was immediately apparent to me, because being 66 00:03:27,990 --> 00:03:30,660 as I was thinking about artificial intelligence 67 00:03:30,660 --> 00:03:34,200 already, the question was, well, how are you going to interact 68 00:03:34,200 --> 00:03:38,700 with it, was the key question. 69 00:03:38,700 --> 00:03:41,730 And therefore, the idea that one should be sitting comfortably 70 00:03:41,730 --> 00:03:47,250 in one's office and interacting with a computer 71 00:03:47,250 --> 00:03:51,780 rather than thinking of it as a point-- 72 00:03:51,780 --> 00:03:53,950 as people did at that time-- 73 00:03:53,950 --> 00:03:57,030 of computation, is that you have this computation that you 74 00:03:57,030 --> 00:03:59,280 wanted to do, and you got a program 75 00:03:59,280 --> 00:04:02,400 to run this computation, and then you ran the computer. 76 00:04:02,400 --> 00:04:08,940 So the interactive use which was the immediate notion to me. 77 00:04:08,940 --> 00:04:11,520 And so, then I began to look around and talk to people 78 00:04:11,520 --> 00:04:15,300 and was very surprised that wasn't what everybody thought. 79 00:04:15,300 --> 00:04:18,360 80 00:04:18,360 --> 00:04:19,950 And now-- 81 00:04:19,950 --> 00:04:21,480 When was that John? 82 00:04:21,480 --> 00:04:24,600 '55 or thereabouts. 83 00:04:24,600 --> 00:04:27,150 But I was not in a position to do anything about it. 84 00:04:27,150 --> 00:04:30,300 I was Dartmouth at the time-- went back to Dartmouth. 85 00:04:30,300 --> 00:04:33,870 And I didn't actually do anything 86 00:04:33,870 --> 00:04:41,670 until I came to MIT as a Sloan Foundation fellow in '57. 87 00:04:41,670 --> 00:04:46,960 And at that point, I became interested in trying it out. 88 00:04:46,960 --> 00:04:50,460 And because I was quite unself-confident 89 00:04:50,460 --> 00:04:52,560 about hardware, I wanted to propose 90 00:04:52,560 --> 00:04:54,630 the minimal modification. 91 00:04:54,630 --> 00:04:59,220 And I proposed to D. Norden, who was, I believe, 92 00:04:59,220 --> 00:05:02,215 Director of Research or whatever it was for the computation 93 00:05:02,215 --> 00:05:03,730 center at that time. 94 00:05:03,730 --> 00:05:04,480 Yes, that's right. 95 00:05:04,480 --> 00:05:07,750 Since I was a visitor at MIT also, mind, 96 00:05:07,750 --> 00:05:11,620 I was very diffident about things-- 97 00:05:11,620 --> 00:05:18,610 that we installed a switch on the IBM 704 98 00:05:18,610 --> 00:05:21,820 that would allow it to be put into trapping mode 99 00:05:21,820 --> 00:05:28,990 from the outside, and also install equipment 100 00:05:28,990 --> 00:05:31,690 to allow six lines from flexor writer 101 00:05:31,690 --> 00:05:34,570 to take the place of the sense switches-- 102 00:05:34,570 --> 00:05:37,640 to put it parallel with the sense switches. 103 00:05:37,640 --> 00:05:42,250 So I think I'll stop at this point, which sort of gets me up 104 00:05:42,250 --> 00:05:46,394 to the beginning of the idea of my interaction with computers. 105 00:05:46,394 --> 00:05:47,756 Yes, John. 106 00:05:47,756 --> 00:05:50,458 It does John, but let me push you 107 00:05:50,458 --> 00:05:52,750 right back to the beginning, because you didn't tell us 108 00:05:52,750 --> 00:05:54,833 how you got interested in artificial intelligence, 109 00:05:54,833 --> 00:05:55,960 and what-- 110 00:05:55,960 --> 00:05:57,950 even before, you were in college-- 111 00:05:57,950 --> 00:05:59,870 Oh, the Hixon Symposium [INAUDIBLE] 112 00:05:59,870 --> 00:06:07,480 it before the ideas of computing and of machine intelligence 113 00:06:07,480 --> 00:06:11,440 were discussed by the various speakers at the Hixon Symposium 114 00:06:11,440 --> 00:06:13,240 in 1949. 115 00:06:13,240 --> 00:06:14,740 Where were you at the time? 116 00:06:14,740 --> 00:06:16,360 I was a graduate student at Caltech. 117 00:06:16,360 --> 00:06:17,110 Well, that's the-- 118 00:06:17,110 --> 00:06:18,040 Uh-huh. 119 00:06:18,040 --> 00:06:19,810 And what was your subject? 120 00:06:19,810 --> 00:06:22,520 Mathematics. 121 00:06:22,520 --> 00:06:26,670 What particular aspects of mathematics? 122 00:06:26,670 --> 00:06:28,380 Functional analysis, I guess. 123 00:06:28,380 --> 00:06:30,690 But I was just beginning. 124 00:06:30,690 --> 00:06:31,260 Did you-- 125 00:06:31,260 --> 00:06:32,530 And where did you come from? 126 00:06:32,530 --> 00:06:34,020 Where were you born, where did you grow up? 127 00:06:34,020 --> 00:06:34,380 In California? 128 00:06:34,380 --> 00:06:35,020 Oh, all right. 129 00:06:35,020 --> 00:06:36,990 I was born in Boston. 130 00:06:36,990 --> 00:06:40,950 And we started moving across the country when 131 00:06:40,950 --> 00:06:43,140 I was about eight or nine. 132 00:06:43,140 --> 00:06:45,540 I think we got to California when I was 10. 133 00:06:45,540 --> 00:06:50,190 So I went to the later grades of school from the sixth grade 134 00:06:50,190 --> 00:06:53,100 on in California, and went to Caltech as an undergraduate, 135 00:06:53,100 --> 00:06:56,050 and stayed there for one year of graduate work. 136 00:06:56,050 --> 00:07:00,120 And then went to Princeton, where I got my PhD, 137 00:07:00,120 --> 00:07:02,730 then I might as well finish it off, so to speak. 138 00:07:02,730 --> 00:07:04,800 I stayed at Princeton for two more years 139 00:07:04,800 --> 00:07:09,500 as an instructor then went to Stanford, 140 00:07:09,500 --> 00:07:16,800 and then went to Dartmouth, and then went to MIT-- 141 00:07:16,800 --> 00:07:19,080 first as a Sloan Foundation Fellow 142 00:07:19,080 --> 00:07:23,460 and then as an assistant professor of communication 143 00:07:23,460 --> 00:07:25,830 sciences in the EE Department. 144 00:07:25,830 --> 00:07:27,477 [LAUGHTER] 145 00:07:27,477 --> 00:07:28,310 I can remember that. 146 00:07:28,310 --> 00:07:30,360 Thank you. 147 00:07:30,360 --> 00:07:32,730 That was essentially because the math department 148 00:07:32,730 --> 00:07:36,690 couldn't take more than one computer specialist. 149 00:07:36,690 --> 00:07:38,580 So Marvin Minsky was there. 150 00:07:38,580 --> 00:07:39,693 Right. 151 00:07:39,693 --> 00:07:42,235 Because Marvin eventually moved to the electrical engineering 152 00:07:42,235 --> 00:07:42,735 department. 153 00:07:42,735 --> 00:07:46,050 [INAUDIBLE] Now, this has sort of the aura of this 154 00:07:46,050 --> 00:07:47,922 is your life. 155 00:07:47,922 --> 00:07:49,380 Which I'm going to try to telescope 156 00:07:49,380 --> 00:07:50,550 to the relevant parts. 157 00:07:50,550 --> 00:07:51,750 [LAUGHTER] 158 00:07:51,750 --> 00:07:57,210 But I think the thing that I didn't realize it 159 00:07:57,210 --> 00:07:59,040 at the time, but probably the most-- 160 00:07:59,040 --> 00:08:02,460 one of the key episodes of my origins 161 00:08:02,460 --> 00:08:08,190 was during World War II, when I was still in high school, 162 00:08:08,190 --> 00:08:12,120 and facing the inevitability of being drafted. 163 00:08:12,120 --> 00:08:17,310 And I had actually rushed through high school 164 00:08:17,310 --> 00:08:18,840 and gotten into UCLA for a while. 165 00:08:18,840 --> 00:08:24,610 And maybe I've somehow got approached 166 00:08:24,610 --> 00:08:27,430 by the Navy recruiting office. 167 00:08:27,430 --> 00:08:29,500 And they offered me this tempting opportunity 168 00:08:29,500 --> 00:08:31,500 to do something interesting while in the service 169 00:08:31,500 --> 00:08:35,440 rather than just become kind of an ordinary soldier. 170 00:08:35,440 --> 00:08:38,200 And then, the thing they offered was the Eddy Program, 171 00:08:38,200 --> 00:08:42,789 which turned in for electronic technicians. 172 00:08:42,789 --> 00:08:47,110 And so I enlisted, thinking that was at least more interesting. 173 00:08:47,110 --> 00:08:49,780 Indeed, after a year of training, 174 00:08:49,780 --> 00:08:51,250 became an electronic technician. 175 00:08:51,250 --> 00:08:53,090 And in the process-- 176 00:08:53,090 --> 00:08:57,280 so I basically had a high school background-- 177 00:08:57,280 --> 00:09:00,790 and in the process got exposed to working 178 00:09:00,790 --> 00:09:03,760 with all the large electronic systems which 179 00:09:03,760 --> 00:09:07,510 were developed World War II, which were radar, 180 00:09:07,510 --> 00:09:12,550 sonar, Lorand, and on and on. 181 00:09:12,550 --> 00:09:14,380 Most of them I didn't realize at the time 182 00:09:14,380 --> 00:09:17,110 were developed here at the Rad Lab at MIT. 183 00:09:17,110 --> 00:09:20,150 But that was a later connection. 184 00:09:20,150 --> 00:09:23,070 But what I got out of that was something 185 00:09:23,070 --> 00:09:24,740 that I realized afterwards, was working 186 00:09:24,740 --> 00:09:26,560 with large systems, which one didn't fully 187 00:09:26,560 --> 00:09:28,650 comprehend every module's purpose, 188 00:09:28,650 --> 00:09:31,420 but one had a very strong sense of the interfaces 189 00:09:31,420 --> 00:09:34,240 and modularity and functionalities 190 00:09:34,240 --> 00:09:37,450 that were expected out of all the boxes. 191 00:09:37,450 --> 00:09:41,590 So that turned out to be an important piece of background 192 00:09:41,590 --> 00:09:43,690 for me. 193 00:09:43,690 --> 00:09:47,110 After the war, I was able to use the GI Bill 194 00:09:47,110 --> 00:09:49,880 and that helped finance going to Caltech 195 00:09:49,880 --> 00:09:53,410 as an undergraduate, where I, along with most 196 00:09:53,410 --> 00:09:56,590 of the other hotshots in the class, 197 00:09:56,590 --> 00:09:58,160 thought physics was a thing to do. 198 00:09:58,160 --> 00:10:00,070 And so, I got into physics. 199 00:10:00,070 --> 00:10:02,710 And from there, I went on to graduate school 200 00:10:02,710 --> 00:10:07,150 here at MIT, starting as a physicist. 201 00:10:07,150 --> 00:10:11,020 And I can still remember Phil Morris one day 202 00:10:11,020 --> 00:10:12,832 calling me into his office and saying, 203 00:10:12,832 --> 00:10:14,290 how would you like to be a research 204 00:10:14,290 --> 00:10:18,410 assistant on a new kind of project we have, 205 00:10:18,410 --> 00:10:21,070 which is a [INAUDIBLE] I forget the exact title. 206 00:10:21,070 --> 00:10:26,020 But it was basically a computational assistantship, 207 00:10:26,020 --> 00:10:27,130 multidisciplinary. 208 00:10:27,130 --> 00:10:31,300 So there were people from many departments. 209 00:10:31,300 --> 00:10:33,080 But I was one of the early ones there. 210 00:10:33,080 --> 00:10:35,240 And so I signed up for that. 211 00:10:35,240 --> 00:10:39,430 And in the process, in the summer of '51, 212 00:10:39,430 --> 00:10:42,160 began to learn about computing equipment, 213 00:10:42,160 --> 00:10:46,468 ranging from punch cards to a Whirlwind. 214 00:10:46,468 --> 00:10:48,010 And Whirlwind, we were just beginning 215 00:10:48,010 --> 00:10:49,593 to roll with in those days. 216 00:10:49,593 --> 00:10:51,010 And in fact, I think when we first 217 00:10:51,010 --> 00:10:56,260 began to take lessons, or instruction on it, 218 00:10:56,260 --> 00:11:01,480 they had maybe 256 18-bit words. 219 00:11:01,480 --> 00:11:02,530 And it was still-- 220 00:11:02,530 --> 00:11:05,650 every month it was expanding by a factor of 2, 221 00:11:05,650 --> 00:11:07,690 which more or less-- 222 00:11:07,690 --> 00:11:09,790 With a parity check every two minutes, right? 223 00:11:09,790 --> 00:11:10,540 Yeah. 224 00:11:10,540 --> 00:11:13,730 Maybe two, maybe six, but not more than 20. 225 00:11:13,730 --> 00:11:18,315 [LAUGHTER] These were using those-- 226 00:11:18,315 --> 00:11:18,940 Williams tubes. 227 00:11:18,940 --> 00:11:20,620 No, they weren't Williams tubes. 228 00:11:20,620 --> 00:11:23,485 They were special designed. 229 00:11:23,485 --> 00:11:25,360 In fact, that's what distinguished Whirlwind. 230 00:11:25,360 --> 00:11:27,627 They had their own design of tubes, 231 00:11:27,627 --> 00:11:29,210 which made them slightly more reliable 232 00:11:29,210 --> 00:11:30,210 than the Williams tubes. 233 00:11:30,210 --> 00:11:33,760 234 00:11:33,760 --> 00:11:37,210 But that was a very pivotal exposure, 235 00:11:37,210 --> 00:11:40,000 because in fact, I ended up doing my doctoral work 236 00:11:40,000 --> 00:11:43,960 on molecular physics, primarily because it 237 00:11:43,960 --> 00:11:46,840 required a lot of extensive calculations. 238 00:11:46,840 --> 00:11:51,040 And in fact, somewhere in that period, in the mid '50s, 239 00:11:51,040 --> 00:11:54,320 I began to get more interested in the computational aspects 240 00:11:54,320 --> 00:11:56,140 than I was in the physics. 241 00:11:56,140 --> 00:12:02,530 And so, after I got my degree, Phil Morris 242 00:12:02,530 --> 00:12:04,092 approached me again. 243 00:12:04,092 --> 00:12:05,800 This time, because he had been organizing 244 00:12:05,800 --> 00:12:10,330 the start of the computation center at MIT 245 00:12:10,330 --> 00:12:17,860 in '56, which was to acquire a gift from IBM of an IBM 704. 246 00:12:17,860 --> 00:12:23,340 So I signed on as a staff member, an assistant director, 247 00:12:23,340 --> 00:12:27,950 who's responsibility for the research assistantship program 248 00:12:27,950 --> 00:12:32,860 that was one of the things that also went with the center. 249 00:12:32,860 --> 00:12:34,780 And in the process, continued to try 250 00:12:34,780 --> 00:12:37,120 to do some physics research involving computation 251 00:12:37,120 --> 00:12:40,030 on molecular diatomic molecules. 252 00:12:40,030 --> 00:12:46,660 But I experienced in that period, an observation 253 00:12:46,660 --> 00:12:48,850 that computers were getting tougher and tougher 254 00:12:48,850 --> 00:12:52,520 to get at, as people became to realize they were 255 00:12:52,520 --> 00:12:54,085 very useful and interesting. 256 00:12:54,085 --> 00:12:55,960 So the queues kept getting longer and longer. 257 00:12:55,960 --> 00:12:58,660 And it was harder to debug programs in the old batch 258 00:12:58,660 --> 00:13:00,100 processing style. 259 00:13:00,100 --> 00:13:01,810 Computation centers were getting more 260 00:13:01,810 --> 00:13:04,870 and more obsessed with trying to be more efficient. 261 00:13:04,870 --> 00:13:06,910 And the users were getting squeezed out. 262 00:13:06,910 --> 00:13:10,660 And debugging was getting harder and harder. 263 00:13:10,660 --> 00:13:12,993 And that was another important piece of background, 264 00:13:12,993 --> 00:13:14,410 because it was quite clear that it 265 00:13:14,410 --> 00:13:16,090 was getting harder and harder to write programs 266 00:13:16,090 --> 00:13:17,170 and to get things done. 267 00:13:17,170 --> 00:13:20,440 268 00:13:20,440 --> 00:13:22,950 And I guess it was about that time, 269 00:13:22,950 --> 00:13:27,240 shortly after the center began, John showed up on the scene. 270 00:13:27,240 --> 00:13:33,420 And the recollection was that John was probably 271 00:13:33,420 --> 00:13:41,370 the most articulate in agitating for this new way of dealing 272 00:13:41,370 --> 00:13:42,930 with computers. 273 00:13:42,930 --> 00:13:45,060 Namely, trying to get more interactive. 274 00:13:45,060 --> 00:13:47,640 My exposure to Whirlwind had been always 275 00:13:47,640 --> 00:13:48,960 on the unclassified side. 276 00:13:48,960 --> 00:13:52,340 So I didn't know about to say stuff, though I had some hints. 277 00:13:52,340 --> 00:13:54,680 But certainly, the notion of being 278 00:13:54,680 --> 00:13:56,180 able to get your hands on a computer 279 00:13:56,180 --> 00:13:57,920 was part of my background. 280 00:13:57,920 --> 00:14:02,720 So trying to make interactive computing accessible 281 00:14:02,720 --> 00:14:05,450 was certainly something that made sense. 282 00:14:05,450 --> 00:14:07,370 And when the actual thoughts began 283 00:14:07,370 --> 00:14:12,080 to gel in the form of some of these early memos 284 00:14:12,080 --> 00:14:17,090 that John wrote, and later Strachey, which I remember John 285 00:14:17,090 --> 00:14:18,140 called my attention to-- 286 00:14:18,140 --> 00:14:22,040 Strachey's memo, or paper actually. 287 00:14:22,040 --> 00:14:25,490 It seemed more and more obvious and to myself and everybody 288 00:14:25,490 --> 00:14:28,190 else in the circle at the time. 289 00:14:28,190 --> 00:14:36,140 And in fact, as we began to start a set of committees 290 00:14:36,140 --> 00:14:40,370 to try to do something in an organized way. 291 00:14:40,370 --> 00:14:42,830 And it was out of that set of committees 292 00:14:42,830 --> 00:14:49,090 that we each began to look for what we could do. 293 00:14:49,090 --> 00:14:53,680 John alluded to some of his attempts. 294 00:14:53,680 --> 00:14:56,420 I can still remember [INAUDIBLE] and had to plug. 295 00:14:56,420 --> 00:15:01,250 It was also to support John's [INAUDIBLE] connection. 296 00:15:01,250 --> 00:15:04,260 But that's as far as I've got, I think. 297 00:15:04,260 --> 00:15:07,260 I started on a quick and dirty version of time sharing, 298 00:15:07,260 --> 00:15:12,570 which was supposed to be more of a demo system, which it later 299 00:15:12,570 --> 00:15:16,350 got called a compatible time-sharing system, CKSS. 300 00:15:16,350 --> 00:15:19,560 We'll probably come back to that. 301 00:15:19,560 --> 00:15:22,050 And we were off and running. 302 00:15:22,050 --> 00:15:26,550 And I think that's probably a good place to start, 303 00:15:26,550 --> 00:15:31,904 because we'll pick it up from there I'm sure. 304 00:15:31,904 --> 00:15:33,690 Go ahead bob. 305 00:15:33,690 --> 00:15:35,670 Well, my story is a little bit longer, 306 00:15:35,670 --> 00:15:37,710 simply because I come from further away. 307 00:15:37,710 --> 00:15:38,910 [LAUGHTER] 308 00:15:38,910 --> 00:15:46,740 And I was born in Torino, Italy, and I did four years 309 00:15:46,740 --> 00:15:51,840 of engineering school there. 310 00:15:51,840 --> 00:15:54,750 And I started-- 311 00:15:54,750 --> 00:15:58,600 I was going into electrical engineering. 312 00:15:58,600 --> 00:16:06,030 But in Italy, the engineering is a five year program. 313 00:16:06,030 --> 00:16:09,970 And the specialization comes only in the fifth year. 314 00:16:09,970 --> 00:16:15,960 So when I came over here in 1939, 315 00:16:15,960 --> 00:16:19,950 I still really had a full year of work 316 00:16:19,950 --> 00:16:24,300 to do before I could get a bachelor's degree 317 00:16:24,300 --> 00:16:25,560 in electrical engineering. 318 00:16:25,560 --> 00:16:30,210 And I started at MIT in February 1940, 319 00:16:30,210 --> 00:16:35,670 and finished, got my bachelors in 1941. 320 00:16:35,670 --> 00:16:40,110 The question is, why did I go into electrical engineering? 321 00:16:40,110 --> 00:16:41,810 I really don't know. 322 00:16:41,810 --> 00:16:44,200 As far as back as I can remember, 323 00:16:44,200 --> 00:16:46,680 I always wanted to be an electrical engineer. 324 00:16:46,680 --> 00:16:50,130 And I have no idea why. 325 00:16:50,130 --> 00:16:52,140 So I became an electrical engineer. 326 00:16:52,140 --> 00:16:54,960 I must also say that there was another thing that 327 00:16:54,960 --> 00:16:58,170 was very clear to me, it was that I 328 00:16:58,170 --> 00:17:00,695 didn't want to be a professor. 329 00:17:00,695 --> 00:17:02,550 [INAUDIBLE] I am a professor, and I 330 00:17:02,550 --> 00:17:07,770 have been a professor, because of strange reason 331 00:17:07,770 --> 00:17:09,339 that I won't get into it. 332 00:17:09,339 --> 00:17:17,069 When I graduated in February 1941, I look for a job. 333 00:17:17,069 --> 00:17:20,430 And it was practically impossible for me to get a job. 334 00:17:20,430 --> 00:17:25,349 And the reason why, is that I was still an Italian citizen. 335 00:17:25,349 --> 00:17:29,400 And essentially the whole electrical industry 336 00:17:29,400 --> 00:17:34,050 in the United States was already beginning to do military work. 337 00:17:34,050 --> 00:17:37,290 So a clearance was necessary, and which I could not get. 338 00:17:37,290 --> 00:17:41,850 So I ended up with to work for General Motors 339 00:17:41,850 --> 00:17:45,540 in the official body plant in Grand Rapids. 340 00:17:45,540 --> 00:17:48,840 And without spending time on that, 341 00:17:48,840 --> 00:17:51,360 I really didn't like that job. 342 00:17:51,360 --> 00:17:56,820 And then, I remember that Carlton Tucker, who 343 00:17:56,820 --> 00:18:00,000 was an executive officer of the department 344 00:18:00,000 --> 00:18:04,290 had offered me a teaching assistantship when I graduated. 345 00:18:04,290 --> 00:18:06,270 And I had said, no, because I was not 346 00:18:06,270 --> 00:18:09,250 interested in a teaching career. 347 00:18:09,250 --> 00:18:13,590 So I got in touch with him, and told him that at that point 348 00:18:13,590 --> 00:18:14,610 I was interested. 349 00:18:14,610 --> 00:18:17,940 To make a long story short, I started as a teaching assistant 350 00:18:17,940 --> 00:18:24,240 in September 1941, and I haven't moved from MIT since. 351 00:18:24,240 --> 00:18:27,960 Now, how did I get eventually into computers? 352 00:18:27,960 --> 00:18:34,560 Well, I'd been a fill hopper from the very beginning. 353 00:18:34,560 --> 00:18:38,010 First of all in teaching, because of the war, 354 00:18:38,010 --> 00:18:40,650 since most of the faculty of the electrical engineering 355 00:18:40,650 --> 00:18:45,390 department disappeared, I was asked to do, 356 00:18:45,390 --> 00:18:49,260 as teaching assistant, things that were unheard of. 357 00:18:49,260 --> 00:18:51,630 Like being in charge of a graduate course 358 00:18:51,630 --> 00:18:54,915 after two years that I had been a teaching assistant 359 00:18:54,915 --> 00:18:56,800 in microwaves. 360 00:18:56,800 --> 00:18:59,700 I had taught all sorts of things-- 361 00:18:59,700 --> 00:19:05,850 electronics, for communication, microwaves. 362 00:19:05,850 --> 00:19:09,900 And then, in 1944, finally they opened the door 363 00:19:09,900 --> 00:19:12,780 to the radiation laboratory for non-citizens. 364 00:19:12,780 --> 00:19:14,730 So I moved to the radiation laboratory, 365 00:19:14,730 --> 00:19:19,050 where I was what was being called a microwave plumber. 366 00:19:19,050 --> 00:19:22,330 And at the end of the radiation laboratory, 367 00:19:22,330 --> 00:19:27,210 I was one of the founding research associate and graduate 368 00:19:27,210 --> 00:19:29,970 student of [INAUDIBLE]. 369 00:19:29,970 --> 00:19:33,990 I finished my doctorate in June '47, 370 00:19:33,990 --> 00:19:37,110 and I went on into faculty. 371 00:19:37,110 --> 00:19:40,320 Now, after getting my doctorate, and my thesis 372 00:19:40,320 --> 00:19:42,730 was in network theory. 373 00:19:42,730 --> 00:19:45,390 So I had been in microwave and network theory. 374 00:19:45,390 --> 00:19:48,390 By then I got interested in information theory. 375 00:19:48,390 --> 00:19:52,245 There was a field for a little more than a decade. 376 00:19:52,245 --> 00:19:57,300 377 00:19:57,300 --> 00:20:00,540 Now that got me in the direction of computers 378 00:20:00,540 --> 00:20:01,920 in a certain sense. 379 00:20:01,920 --> 00:20:09,540 And I remember that around 1960, I 380 00:20:09,540 --> 00:20:13,260 got involved into computer affairs, simply 381 00:20:13,260 --> 00:20:19,200 because there was a great scarcity of senior faculty-- 382 00:20:19,200 --> 00:20:22,050 I was a full professor by then-- 383 00:20:22,050 --> 00:20:25,210 that had any connection whatsoever with computers. 384 00:20:25,210 --> 00:20:27,060 So for instance, when a committee 385 00:20:27,060 --> 00:20:31,110 was set up to look at the future of computation, 386 00:20:31,110 --> 00:20:33,360 chaired by Philip, or was Al Hill? 387 00:20:33,360 --> 00:20:34,510 No, I started it. 388 00:20:34,510 --> 00:20:35,850 Yeah. 389 00:20:35,850 --> 00:20:38,400 I was asked to be a member of that committee. 390 00:20:38,400 --> 00:20:42,990 I knew practically nothing about computers. 391 00:20:42,990 --> 00:20:45,600 But another thing that I recall very well, 392 00:20:45,600 --> 00:20:48,300 that in my work in communication theory, 393 00:20:48,300 --> 00:20:55,200 I became very evident in the late '50s 394 00:20:55,200 --> 00:21:01,470 that the process of signal structuring 395 00:21:01,470 --> 00:21:04,290 and particularly the process of the texting signal 396 00:21:04,290 --> 00:21:07,060 in the presence of noise and decoding 397 00:21:07,060 --> 00:21:10,300 involved very complicated operation. 398 00:21:10,300 --> 00:21:12,880 And at that time, communication engineers 399 00:21:12,880 --> 00:21:17,560 were thinking primarily of decoding or detecting 400 00:21:17,560 --> 00:21:21,460 in terms of primitive hardware. 401 00:21:21,460 --> 00:21:23,500 Hardware primitive, I'm sorry. 402 00:21:23,500 --> 00:21:28,160 Like an amplifier, an oscillator, a detector, 403 00:21:28,160 --> 00:21:29,720 and so forth and so on. 404 00:21:29,720 --> 00:21:34,120 And it occurred to me that since computers existed, 405 00:21:34,120 --> 00:21:37,810 we ought to be thinking instead in functional terms. 406 00:21:37,810 --> 00:21:41,910 And then the operation that was logically possible 407 00:21:41,910 --> 00:21:44,310 would eventually be possible to implement 408 00:21:44,310 --> 00:21:45,450 by means of a computer. 409 00:21:45,450 --> 00:21:47,610 This is what got me started. 410 00:21:47,610 --> 00:21:55,390 It so happens that in 1961, I completed various things 411 00:21:55,390 --> 00:21:56,850 that I was working at. 412 00:21:56,850 --> 00:22:00,360 I completed in my book an information theory. 413 00:22:00,360 --> 00:22:03,270 I completed the book on electromagnetic field, 414 00:22:03,270 --> 00:22:06,110 that later was the textbook for undergraduate. 415 00:22:06,110 --> 00:22:10,650 It so happens also my doctoral student finished that thesis 416 00:22:10,650 --> 00:22:11,890 at that time. 417 00:22:11,890 --> 00:22:13,530 So I was a free bird. 418 00:22:13,530 --> 00:22:18,330 And I went on sabbatical in Lincoln Lab 419 00:22:18,330 --> 00:22:24,630 in the academic year '61, '62. 420 00:22:24,630 --> 00:22:26,760 And I decided at that time, that I 421 00:22:26,760 --> 00:22:29,090 had enough with information theory and communication 422 00:22:29,090 --> 00:22:33,120 theory, that I wanted to get in the computer field. 423 00:22:33,120 --> 00:22:37,620 So when I came back in fall, I had already arranged 424 00:22:37,620 --> 00:22:40,500 with Bob Gallagher to take over my graduate course 425 00:22:40,500 --> 00:22:41,760 in information theory. 426 00:22:41,760 --> 00:22:44,980 And I was ready to go. 427 00:22:44,980 --> 00:22:47,290 When project Mac happened. 428 00:22:47,290 --> 00:22:49,870 And probably I should stop at that point, 429 00:22:49,870 --> 00:22:54,070 because we will talk about it, essentially 430 00:22:54,070 --> 00:22:58,580 how I got into the computer field not as a worker, 431 00:22:58,580 --> 00:23:00,800 but as an administrator. 432 00:23:00,800 --> 00:23:02,900 Probably I should stop at that point. 433 00:23:02,900 --> 00:23:04,204 Project Mac didn't just happen. 434 00:23:04,204 --> 00:23:04,704 Hmm? 435 00:23:04,704 --> 00:23:06,440 Project Mac just didn't happen. 436 00:23:06,440 --> 00:23:07,590 [LAUGHTER] 437 00:23:07,590 --> 00:23:11,270 No, I mean, the opportunity to start Project Mac came up. 438 00:23:11,270 --> 00:23:11,870 OK. 439 00:23:11,870 --> 00:23:16,880 Perhaps I should say just a couple of words about it. 440 00:23:16,880 --> 00:23:21,950 That fall, [INAUDIBLE] started coming around MIT, 441 00:23:21,950 --> 00:23:25,170 was at the [INAUDIBLE]. 442 00:23:25,170 --> 00:23:27,367 Yeah, he couldn't make it today. 443 00:23:27,367 --> 00:23:30,500 [INTERPOSING VOICES] 444 00:23:30,500 --> 00:23:34,910 [INAUDIBLE] was at MIT in the immediate post-war period. 445 00:23:34,910 --> 00:23:38,480 And I knew very well that he went to [INAUDIBLE] Neumann. 446 00:23:38,480 --> 00:23:41,315 And then eventually, he got interested in computing. 447 00:23:41,315 --> 00:23:45,770 He wrote the classic paper about men, machines, and [INAUDIBLE].. 448 00:23:45,770 --> 00:23:48,380 He went to the Pentagon, essentially, 449 00:23:48,380 --> 00:23:54,080 to get that sort of movement started. 450 00:23:54,080 --> 00:23:56,600 And he got funded by IARPA. 451 00:23:56,600 --> 00:24:02,470 And he started going around the country building up enthusiasm 452 00:24:02,470 --> 00:24:04,820 for man machine interaction and time-sharing. 453 00:24:04,820 --> 00:24:07,370 Of course, he came to MIT also. 454 00:24:07,370 --> 00:24:14,285 And he was quite anxious to get a big operation started at MIT. 455 00:24:14,285 --> 00:24:18,560 I recall he wanted to get a few centers of excellence-- 456 00:24:18,560 --> 00:24:20,120 that's what he was calling them-- 457 00:24:20,120 --> 00:24:23,780 set up in universities. 458 00:24:23,780 --> 00:24:27,200 And he wanted in particular one at MIT. 459 00:24:27,200 --> 00:24:36,260 And it was quite clear that the problem of getting such a thing 460 00:24:36,260 --> 00:24:38,450 started there at MIT was serious. 461 00:24:38,450 --> 00:24:44,090 And the reason was that all the computer faculty 462 00:24:44,090 --> 00:24:47,000 were young at that time. 463 00:24:47,000 --> 00:24:49,100 John McCarthy was an assistant professor. 464 00:24:49,100 --> 00:24:52,500 You had just joined the faculty, perhaps not even yet. 465 00:24:52,500 --> 00:24:54,020 Marvin, I don't know why he did-- 466 00:24:54,020 --> 00:24:56,390 I'm sure that Marvin didn't have tenure yet. 467 00:24:56,390 --> 00:24:57,290 That's right. 468 00:24:57,290 --> 00:25:00,230 The only senior person involved with computing 469 00:25:00,230 --> 00:25:03,950 was Phil, who was director of the computation center 470 00:25:03,950 --> 00:25:06,620 and had his hands full, because he also 471 00:25:06,620 --> 00:25:08,760 was doing lots of other things. 472 00:25:08,760 --> 00:25:11,690 So it became immediately clear that something 473 00:25:11,690 --> 00:25:12,690 was going to happen. 474 00:25:12,690 --> 00:25:18,180 That MIT had to be pushed by somebody who was not a computer 475 00:25:18,180 --> 00:25:22,020 person, but came from a field sufficiently related 476 00:25:22,020 --> 00:25:26,260 to make some sort of a sense out of that. 477 00:25:26,260 --> 00:25:30,550 Well, I became quite interested in the matter, largely because 478 00:25:30,550 --> 00:25:35,330 of the point of view that John McCarthy had introduced. 479 00:25:35,330 --> 00:25:39,370 This was in the lecture that he gave at the what was it-- 480 00:25:39,370 --> 00:25:42,070 50 years ago? 481 00:25:42,070 --> 00:25:45,060 That is in the book edited by Greenberg. 482 00:25:45,060 --> 00:25:45,560 Yeah. 483 00:25:45,560 --> 00:25:46,360 [INTERPOSING VOICES] 484 00:25:46,360 --> 00:25:46,860 Computer-- 485 00:25:46,860 --> 00:25:47,700 [INTERPOSING VOICES] 486 00:25:47,700 --> 00:25:49,408 Management in the computer of the future. 487 00:25:49,408 --> 00:25:52,420 That is what really appealed to me was the notion of a computer 488 00:25:52,420 --> 00:25:54,510 utility, and all the implications that 489 00:25:54,510 --> 00:25:56,480 had for a lot of things. 490 00:25:56,480 --> 00:26:04,160 So essentially, the whole thing was decided in a week. 491 00:26:04,160 --> 00:26:06,380 And I think I want to mention the story, 492 00:26:06,380 --> 00:26:09,920 because that is typical of the sort of things 493 00:26:09,920 --> 00:26:13,320 that were happening at MIT. 494 00:26:13,320 --> 00:26:16,400 [INAUDIBLE] kept talking with me, 495 00:26:16,400 --> 00:26:19,210 and trying to get me hooked. 496 00:26:19,210 --> 00:26:22,030 And we were together at a meeting 497 00:26:22,030 --> 00:26:25,490 organized by [INAUDIBLE],, down in Virginia, in a homestead. 498 00:26:25,490 --> 00:26:26,770 And we talked a lot. 499 00:26:26,770 --> 00:26:29,740 We talked a lot on the train coming back. 500 00:26:29,740 --> 00:26:32,760 And it was the day before Thanksgiving. 501 00:26:32,760 --> 00:26:34,320 And I kept thinking about it. 502 00:26:34,320 --> 00:26:37,730 Then on Thanksgiving Day, I decided, I'm going to do it. 503 00:26:37,730 --> 00:26:38,800 What the heck. 504 00:26:38,800 --> 00:26:40,320 This is in what year, '62? 505 00:26:40,320 --> 00:26:44,610 That was '62, on Thanksgiving Day. 506 00:26:44,610 --> 00:26:47,850 On Friday, in those days, MIT was open the day 507 00:26:47,850 --> 00:26:48,840 after Thanksgiving. 508 00:26:48,840 --> 00:26:52,270 I had an appointment with the provost, who 509 00:26:52,270 --> 00:26:55,830 was Charlie Towns at that time, because I was also 510 00:26:55,830 --> 00:27:00,210 chairman of the Computer Science Committee. 511 00:27:00,210 --> 00:27:02,470 And I had to talk with him about it. 512 00:27:02,470 --> 00:27:04,590 And I told him about the idea. 513 00:27:04,590 --> 00:27:07,650 And I said, well, why don't you think about it? 514 00:27:07,650 --> 00:27:09,030 Let me know what you think. 515 00:27:09,030 --> 00:27:10,500 That's something we should do. 516 00:27:10,500 --> 00:27:11,460 And he said, oh yeah. 517 00:27:11,460 --> 00:27:12,270 That's a good idea. 518 00:27:12,270 --> 00:27:13,830 Go ahead. 519 00:27:13,830 --> 00:27:17,670 So over the weekend, I wrote a two-page memorandum 520 00:27:17,670 --> 00:27:20,970 outlining what I had in mind. 521 00:27:20,970 --> 00:27:24,120 On Tuesday I met with Jay Stratton, 522 00:27:24,120 --> 00:27:27,310 who was President of MIT at that time. 523 00:27:27,310 --> 00:27:29,910 And the main question that he asked me was, where are you 524 00:27:29,910 --> 00:27:32,860 going to do it? 525 00:27:32,860 --> 00:27:36,520 Because as Phil Morris said at that time, 526 00:27:36,520 --> 00:27:40,630 MIT was caught with [INAUDIBLE]. 527 00:27:40,630 --> 00:27:46,480 And by then, I had heard that the eighth and ninth floor 528 00:27:46,480 --> 00:27:51,970 of the new technology square building that was not up yet 529 00:27:51,970 --> 00:27:54,910 had been leased by CEIR. 530 00:27:54,910 --> 00:27:57,580 The CEIR wanted to get rid of that lease, 531 00:27:57,580 --> 00:28:01,250 because they weren't doing so well. 532 00:28:01,250 --> 00:28:03,940 So I told Jay Stratton, I heard this, 533 00:28:03,940 --> 00:28:08,060 and then maybe we can get that space, and he said, fine. 534 00:28:08,060 --> 00:28:10,180 And then, on Thursday, [INAUDIBLE] 535 00:28:10,180 --> 00:28:14,430 had already plans to come to MIT from Washington. 536 00:28:14,430 --> 00:28:18,250 And we all met in Jay Stratton's office and kind of formally 537 00:28:18,250 --> 00:28:22,130 agreed that was going to happen. 538 00:28:22,130 --> 00:28:25,850 So in a week span, essentially the decisionary principle 539 00:28:25,850 --> 00:28:26,570 was made. 540 00:28:26,570 --> 00:28:28,220 And that's-- 541 00:28:28,220 --> 00:28:29,415 Let me ask you one question. 542 00:28:29,415 --> 00:28:30,110 Yeah? 543 00:28:30,110 --> 00:28:32,120 Somewhere along that line, and I'm 544 00:28:32,120 --> 00:28:37,310 interested to know where, you met in the hall. 545 00:28:37,310 --> 00:28:37,880 Yeah. 546 00:28:37,880 --> 00:28:42,030 I think it was at the corner of Building 6. 547 00:28:42,030 --> 00:28:45,530 And you said, look. 548 00:28:45,530 --> 00:28:49,910 This project is going to get started. 549 00:28:49,910 --> 00:28:51,590 Don't you want to be the director? 550 00:28:51,590 --> 00:28:52,430 That's right. 551 00:28:52,430 --> 00:28:53,748 And I said, no. 552 00:28:53,748 --> 00:28:54,290 That's right. 553 00:28:54,290 --> 00:28:55,680 And you said, why? 554 00:28:55,680 --> 00:28:56,840 And I told you. 555 00:28:56,840 --> 00:28:59,690 And you started giggling, and said, thanks, and went off. 556 00:28:59,690 --> 00:29:00,440 Yeah. 557 00:29:00,440 --> 00:29:02,720 Well I think this was probably before-- 558 00:29:02,720 --> 00:29:04,040 Yeah. 559 00:29:04,040 --> 00:29:07,430 --was at the time when I felt that something ought to happen, 560 00:29:07,430 --> 00:29:10,170 and obviously you were the right person to do it. 561 00:29:10,170 --> 00:29:11,500 So I asked you. 562 00:29:11,500 --> 00:29:13,030 Let me add another slight footnote. 563 00:29:13,030 --> 00:29:13,550 Yeah. 564 00:29:13,550 --> 00:29:15,830 [INAUDIBLE] came up-- there were several meetings when 565 00:29:15,830 --> 00:29:18,590 Lech first approached us. 566 00:29:18,590 --> 00:29:23,045 He came up to MIT, and there was about a dozen people in a room, 567 00:29:23,045 --> 00:29:26,180 it was in the old Busch room. 568 00:29:26,180 --> 00:29:30,890 And I can still recall the meeting, where everyone sort of 569 00:29:30,890 --> 00:29:33,920 said what was on their minds in response to [INAUDIBLE].. 570 00:29:33,920 --> 00:29:35,560 And it was kind of a cat and dog fight. 571 00:29:35,560 --> 00:29:36,890 And it sounded like a miss. 572 00:29:36,890 --> 00:29:39,880 And the result of it was, that you came out of that meeting 573 00:29:39,880 --> 00:29:41,360 Bob, saying, we can't-- 574 00:29:41,360 --> 00:29:43,430 we've got to do something about this. 575 00:29:43,430 --> 00:29:45,200 You were very appalled at the fact 576 00:29:45,200 --> 00:29:47,210 that we had such a disorganized front. 577 00:29:47,210 --> 00:29:50,870 And I think that was the actual point where you first 578 00:29:50,870 --> 00:29:53,970 began to seriously think about doing something. 579 00:29:53,970 --> 00:29:58,200 Well, really, the critical time was when 580 00:29:58,200 --> 00:30:00,775 I was down at this meeting in-- 581 00:30:00,775 --> 00:30:01,275 Yeah. 582 00:30:01,275 --> 00:30:02,580 That's when it gelled. 583 00:30:02,580 --> 00:30:07,240 Because I became clear on what a mess 584 00:30:07,240 --> 00:30:11,790 the information processing field was at the time, 585 00:30:11,790 --> 00:30:13,800 and that something had to be done 586 00:30:13,800 --> 00:30:15,750 to try to think through the issue 587 00:30:15,750 --> 00:30:17,130 and do something about it. 588 00:30:17,130 --> 00:30:20,030 Of course, I was already familiar by then 589 00:30:20,030 --> 00:30:23,565 with all the work on time sharing, 590 00:30:23,565 --> 00:30:27,720 and I had been influenced by John, and so forth and so on. 591 00:30:27,720 --> 00:30:30,330 592 00:30:30,330 --> 00:30:30,830 Yeah. 593 00:30:30,830 --> 00:30:35,010 And it seems to me that it was very 594 00:30:35,010 --> 00:30:38,100 hard to recreate the incredible arrogance 595 00:30:38,100 --> 00:30:40,920 and rigidity that manufacturers had already taken on. 596 00:30:40,920 --> 00:30:41,420 [LAUGHTER] 597 00:30:41,420 --> 00:30:42,420 Yeah. 598 00:30:42,420 --> 00:30:45,300 And that was one of the reasons why we had to get organized. 599 00:30:45,300 --> 00:30:47,970 Only matched by our own. 600 00:30:47,970 --> 00:30:48,600 Maybe so. 601 00:30:48,600 --> 00:30:51,803 But why don't we keep going around, because we'll 602 00:30:51,803 --> 00:30:52,970 get into a long distraction. 603 00:30:52,970 --> 00:30:55,816 Yeah, I'm going to touch on that too. 604 00:30:55,816 --> 00:30:57,280 Well, you finished? 605 00:30:57,280 --> 00:30:59,170 Yeah. 606 00:30:59,170 --> 00:31:03,770 So I'm a little bit of a place here, 607 00:31:03,770 --> 00:31:07,070 because I never have been a computer specialist neither 608 00:31:07,070 --> 00:31:09,830 for hardware nor software. 609 00:31:09,830 --> 00:31:16,110 But as a physicist and later as someone 610 00:31:16,110 --> 00:31:18,540 interested in operations research, 611 00:31:18,540 --> 00:31:21,150 I've always been interested in the use of computers 612 00:31:21,150 --> 00:31:29,700 and got more and more convinced that somehow 613 00:31:29,700 --> 00:31:37,020 or other every person going through a technical school, 614 00:31:37,020 --> 00:31:40,830 and eventually any college, ought to know something 615 00:31:40,830 --> 00:31:43,230 about computer. 616 00:31:43,230 --> 00:31:46,215 I started way back pre-war. 617 00:31:46,215 --> 00:31:49,070 618 00:31:49,070 --> 00:31:54,740 Got my bachelor's degree at Case under Dayton Miller 619 00:31:54,740 --> 00:31:57,110 in the '20s. 620 00:31:57,110 --> 00:31:59,870 Went on to Princeton, got my doctors degree there 621 00:31:59,870 --> 00:32:00,440 in physics. 622 00:32:00,440 --> 00:32:03,660 623 00:32:03,660 --> 00:32:06,680 Finished my thesis in a couple of years. 624 00:32:06,680 --> 00:32:09,710 And then, spent my last year-- 625 00:32:09,710 --> 00:32:11,180 you couldn't get a doctor's degree 626 00:32:11,180 --> 00:32:13,700 in two years at Princeton then. 627 00:32:13,700 --> 00:32:18,000 Spent my last year helping Ed [INAUDIBLE] write 628 00:32:18,000 --> 00:32:21,960 a text on quantum mechanics. 629 00:32:21,960 --> 00:32:27,260 And I came up to MIT at the time that Compton 630 00:32:27,260 --> 00:32:30,680 moved from Princeton up to MIT. 631 00:32:30,680 --> 00:32:37,365 And been in the physics department ever since until-- 632 00:32:37,365 --> 00:32:42,710 well, I still got an office down in the physics department. 633 00:32:42,710 --> 00:32:49,140 But very early, I got interested in computers, 634 00:32:49,140 --> 00:32:54,810 because the work we were all doing in quantum mechanics 635 00:32:54,810 --> 00:32:58,230 got more and more complicated and needed 636 00:32:58,230 --> 00:33:00,690 more and more computing. 637 00:33:00,690 --> 00:33:06,660 So two of my papers that were published during the '30s 638 00:33:06,660 --> 00:33:09,870 were based on some work that I did with the Bush differential 639 00:33:09,870 --> 00:33:13,170 analyzer. 640 00:33:13,170 --> 00:33:16,770 Then during the war, when I was down in Washington 641 00:33:16,770 --> 00:33:21,480 working with the Navy, I think my group 642 00:33:21,480 --> 00:33:25,810 introduced the first computing equipment to the Navy. 643 00:33:25,810 --> 00:33:30,100 It was a straight punch card equipment, with a sorter, 644 00:33:30,100 --> 00:33:30,810 and so forth. 645 00:33:30,810 --> 00:33:34,150 We did a lot of statistical work on it. 646 00:33:34,150 --> 00:33:36,820 Kept a great deal of data on it. 647 00:33:36,820 --> 00:33:41,590 Reported on all of the contacts with submarines-- 648 00:33:41,590 --> 00:33:45,060 the German submarines in the Atlantic on it, 649 00:33:45,060 --> 00:33:51,680 and reported every morning to Admiral King's staff. 650 00:33:51,680 --> 00:33:52,220 Question. 651 00:33:52,220 --> 00:33:52,800 Yes? 652 00:33:52,800 --> 00:33:55,140 Did you know anything about the electronic computing 653 00:33:55,140 --> 00:33:56,610 that was going on for cryptograph data systems? 654 00:33:56,610 --> 00:33:57,540 I knew nothing. 655 00:33:57,540 --> 00:34:02,460 I was too very much busy at that time. 656 00:34:02,460 --> 00:34:06,750 Matter of fact, I didn't get in contact with anybody working 657 00:34:06,750 --> 00:34:11,250 in the digital computer line until I came back to MIT in, 658 00:34:11,250 --> 00:34:14,280 I think it was '51 or '50, when I started 659 00:34:14,280 --> 00:34:16,850 talking to Jay Forrester. 660 00:34:16,850 --> 00:34:21,610 No, I was I was convinced that we needed computing, 661 00:34:21,610 --> 00:34:24,110 but I knew very little about what 662 00:34:24,110 --> 00:34:28,870 the developments were going on. 663 00:34:28,870 --> 00:34:35,050 When I got back to MIT, I soon began talking to Jay Forrester. 664 00:34:35,050 --> 00:34:38,380 He was, of course, very busy designing and building 665 00:34:38,380 --> 00:34:41,810 [INAUDIBLE]. 666 00:34:41,810 --> 00:34:46,520 But my interest was to get students, 667 00:34:46,520 --> 00:34:49,969 to start with graduate students, all 668 00:34:49,969 --> 00:34:58,340 around the Institute acquainted and familiar with computing. 669 00:34:58,340 --> 00:35:01,280 And the more I talked to Jay-- 670 00:35:01,280 --> 00:35:04,020 this was of course second hand, because I 671 00:35:04,020 --> 00:35:06,840 wasn't doing any work along that line-- the more I talked 672 00:35:06,840 --> 00:35:11,340 to Jay, the more I got convinced that the digital way was 673 00:35:11,340 --> 00:35:13,570 the way to go. 674 00:35:13,570 --> 00:35:17,550 I can remember some very strong arguments I had with van Bush. 675 00:35:17,550 --> 00:35:18,840 [LAUGHTER] 676 00:35:18,840 --> 00:35:22,710 Van was quite convinced that digital computers 677 00:35:22,710 --> 00:35:25,140 were the wrong way to go. 678 00:35:25,140 --> 00:35:30,720 I remember very well, that in [INAUDIBLE] of Project Mac, 679 00:35:30,720 --> 00:35:34,080 you suggested that in the fall that we should 680 00:35:34,080 --> 00:35:36,000 have launched with van Bush. 681 00:35:36,000 --> 00:35:40,020 And we had lunch, the two of us with them, at the faculty club. 682 00:35:40,020 --> 00:35:40,560 Yeah. 683 00:35:40,560 --> 00:35:44,070 And I was completely flabbergasted 684 00:35:44,070 --> 00:35:47,760 by the powerful emotion that there were in Van 685 00:35:47,760 --> 00:35:49,590 against digital computers. 686 00:35:49,590 --> 00:35:51,840 And then it became clear, that what 687 00:35:51,840 --> 00:35:55,410 he meant by digital computer was binary computers. 688 00:35:55,410 --> 00:35:57,480 He didn't mean discrete. 689 00:35:57,480 --> 00:35:59,910 And he was perfectly willing to buy-- 690 00:35:59,910 --> 00:36:04,740 and in fact, he said there was a lot of need for computation. 691 00:36:04,740 --> 00:36:08,100 But he had in mind something other than binary. 692 00:36:08,100 --> 00:36:12,510 I could never understand why in particular, he 693 00:36:12,510 --> 00:36:15,450 was against binary computers, unless it 694 00:36:15,450 --> 00:36:19,572 had to do with some early history or argument with-- 695 00:36:19,572 --> 00:36:22,590 Yeah, I think so. 696 00:36:22,590 --> 00:36:23,490 --Norbert Wiener? 697 00:36:23,490 --> 00:36:24,592 Was that it? 698 00:36:24,592 --> 00:36:26,925 Well, I don't know whether it was Norbert Wiener or not, 699 00:36:26,925 --> 00:36:28,570 but it certainly was-- 700 00:36:28,570 --> 00:36:29,070 No. 701 00:36:29,070 --> 00:36:30,750 It was, he didn't like to-- 702 00:36:30,750 --> 00:36:34,840 I talked to him before Project Mac as part of the study. 703 00:36:34,840 --> 00:36:35,460 Yeah. 704 00:36:35,460 --> 00:36:40,080 And his hatred of binary was as an engineer, 705 00:36:40,080 --> 00:36:43,950 he didn't like the idea of using 10, ones, and zeros 706 00:36:43,950 --> 00:36:47,010 to represent something he could do with three decimal number. 707 00:36:47,010 --> 00:36:47,890 That was it? 708 00:36:47,890 --> 00:36:48,757 That was it. 709 00:36:48,757 --> 00:36:49,590 [INTERPOSING VOICES] 710 00:36:49,590 --> 00:36:51,240 But there was really powerful emotion-- 711 00:36:51,240 --> 00:36:52,990 I don't think-- I tried to explain to him, 712 00:36:52,990 --> 00:36:55,500 that we can make a converter for you. 713 00:36:55,500 --> 00:36:56,430 There were-- 714 00:36:56,430 --> 00:36:57,450 Did you know his work? 715 00:36:57,450 --> 00:37:00,010 --terribly powerful emotion behind it. 716 00:37:00,010 --> 00:37:00,510 Yes. 717 00:37:00,510 --> 00:37:05,500 And really, I could not justify on the basis of-- 718 00:37:05,500 --> 00:37:06,420 [INTERPOSING VOICES] 719 00:37:06,420 --> 00:37:07,447 Historically, it was-- 720 00:37:07,447 --> 00:37:08,280 [INTERPOSING VOICES] 721 00:37:08,280 --> 00:37:10,650 --by Norman Snowdon at one point. 722 00:37:10,650 --> 00:37:11,483 [INTERPOSING VOICES] 723 00:37:11,483 --> 00:37:13,025 In other words, there was something-- 724 00:37:13,025 --> 00:37:14,190 there was something about-- 725 00:37:14,190 --> 00:37:15,010 Bizarre. 726 00:37:15,010 --> 00:37:17,220 He felt he lost out during the war. 727 00:37:17,220 --> 00:37:21,510 I mean, he built up this very complicated differential 728 00:37:21,510 --> 00:37:23,650 analyzer with vacuum tubes and so on, 729 00:37:23,650 --> 00:37:25,230 which was in the middle of MIT. 730 00:37:25,230 --> 00:37:27,610 And it never did anything. 731 00:37:27,610 --> 00:37:31,170 And he felt that people just passed it by and didn't 732 00:37:31,170 --> 00:37:32,730 pay any attention to it. 733 00:37:32,730 --> 00:37:34,497 And that was as well part of it. 734 00:37:34,497 --> 00:37:35,330 [INTERPOSING VOICES] 735 00:37:35,330 --> 00:37:40,230 Did you know if his work in trying to build a digital base 736 00:37:40,230 --> 00:37:43,200 four computer in about 1938, to '39 737 00:37:43,200 --> 00:37:45,000 with Ratford and [INAUDIBLE]? 738 00:37:45,000 --> 00:37:46,000 This is the Rockefeller? 739 00:37:46,000 --> 00:37:49,070 No, before the Rockefeller analyzer. 740 00:37:49,070 --> 00:37:50,023 There was a-- 741 00:37:50,023 --> 00:37:50,856 [INTERPOSING VOICES] 742 00:37:50,856 --> 00:37:51,356 Yes. 743 00:37:51,356 --> 00:37:54,420 --computer project in the EE Department 744 00:37:54,420 --> 00:37:57,120 just before World War II. 745 00:37:57,120 --> 00:37:59,670 And Ratford among others was involved. 746 00:37:59,670 --> 00:38:01,267 Truman Gray also had some-- 747 00:38:01,267 --> 00:38:02,100 [INTERPOSING VOICES] 748 00:38:02,100 --> 00:38:03,060 Sam Caldwell 749 00:38:03,060 --> 00:38:04,920 Sam Caldwell. 750 00:38:04,920 --> 00:38:09,420 I found out about it from Carl [INAUDIBLE] right up. 751 00:38:09,420 --> 00:38:09,930 Yes. 752 00:38:09,930 --> 00:38:15,600 And he was focused primarily as I recall now 753 00:38:15,600 --> 00:38:19,710 in the development of a vacuum tube that 754 00:38:19,710 --> 00:38:21,180 would do the switching. 755 00:38:21,180 --> 00:38:22,020 Yeah. 756 00:38:22,020 --> 00:38:26,335 And of course, the project was stopped at the beginning 757 00:38:26,335 --> 00:38:27,900 of World War II. 758 00:38:27,900 --> 00:38:34,160 It was restarted, according to Carl, 759 00:38:34,160 --> 00:38:36,450 immediately after World War II. 760 00:38:36,450 --> 00:38:40,380 And then the [INAUDIBLE],, who was the [INAUDIBLE] 761 00:38:40,380 --> 00:38:44,220 at that time, came to the conclusion [INAUDIBLE] 762 00:38:44,220 --> 00:38:47,850 involved that there was no point in having a computer 763 00:38:47,850 --> 00:38:52,530 project in the department on an essentially shoestring support 764 00:38:52,530 --> 00:38:57,750 when there was Whirlwind next door going on 765 00:38:57,750 --> 00:39:00,460 with the millions of dollars of funding. 766 00:39:00,460 --> 00:39:01,980 So he was stopped. 767 00:39:01,980 --> 00:39:05,850 And [INAUDIBLE] returned to the sponsor, 768 00:39:05,850 --> 00:39:10,440 it was a foundation of some sort, a check, 769 00:39:10,440 --> 00:39:13,770 for the balance of the unexpended money. 770 00:39:13,770 --> 00:39:17,100 And probably that it might have been the last time 771 00:39:17,100 --> 00:39:21,570 that MIT returned any money to a sponsor. 772 00:39:21,570 --> 00:39:23,520 Anyway, when you said-- 773 00:39:23,520 --> 00:39:27,730 What was the year of that luncheon that you talked about? 774 00:39:27,730 --> 00:39:29,920 That was the before '62. 775 00:39:29,920 --> 00:39:30,420 In '62. 776 00:39:30,420 --> 00:39:30,962 That's right. 777 00:39:30,962 --> 00:39:33,880 The decision to go ahead with Project Mac had been made. 778 00:39:33,880 --> 00:39:34,380 Yeah. 779 00:39:34,380 --> 00:39:38,160 And I was in the planning stage, and I had a lot of contact 780 00:39:38,160 --> 00:39:40,380 with Phil at that time, and he made 781 00:39:40,380 --> 00:39:42,300 that suggestion, which was-- 782 00:39:42,300 --> 00:39:45,270 Well, I had talked to them much earlier 783 00:39:45,270 --> 00:39:48,330 and got some of that feeling. 784 00:39:48,330 --> 00:39:50,880 Of course, we were reasonably close back 785 00:39:50,880 --> 00:39:54,330 the '30s when I worked with his machine. 786 00:39:54,330 --> 00:39:58,430 Anyway, coming back to Whirlwind, after-- 787 00:39:58,430 --> 00:40:04,770 it wasn't much more than the six months that after I arrived, 788 00:40:04,770 --> 00:40:07,020 I became convinced that something 789 00:40:07,020 --> 00:40:11,880 had to be done to make most of MIT 790 00:40:11,880 --> 00:40:15,750 aware of digital computers-- of computers in general. 791 00:40:15,750 --> 00:40:21,112 And Whirlwind was the obvious. 792 00:40:21,112 --> 00:40:25,090 And Forester admitted, a little grudgingly, 793 00:40:25,090 --> 00:40:30,280 that there was some spare time, and that people 794 00:40:30,280 --> 00:40:35,080 could come and use the machine when they weren't busy. 795 00:40:35,080 --> 00:40:38,830 796 00:40:38,830 --> 00:40:48,380 So I went down to my friends in the Navy, ONR, and said, 797 00:40:48,380 --> 00:40:51,700 look, why don't you add a little bit to the Whirlwind 798 00:40:51,700 --> 00:41:01,980 funds to set up a set of research assistants 799 00:41:01,980 --> 00:41:03,930 to use Whirlwind. 800 00:41:03,930 --> 00:41:07,830 And I'll set up a committee at MIT. 801 00:41:07,830 --> 00:41:14,880 And we'll appoint research assistants 802 00:41:14,880 --> 00:41:20,330 in as many departments as are willing to take them. 803 00:41:20,330 --> 00:41:22,250 My feeling was that the only way you 804 00:41:22,250 --> 00:41:28,700 were going to get faculty interested and using computers 805 00:41:28,700 --> 00:41:35,570 was to get a graduate student under their direction using 806 00:41:35,570 --> 00:41:37,220 computers. 807 00:41:37,220 --> 00:41:40,970 So ONR agreed. 808 00:41:40,970 --> 00:41:44,750 And I forget when it was, '53 or '54, 809 00:41:44,750 --> 00:41:48,870 something like that, when the first assistantships were 810 00:41:48,870 --> 00:41:49,875 available. 811 00:41:49,875 --> 00:41:51,480 No, it was earlier. 812 00:41:51,480 --> 00:41:52,110 Was it earlier? 813 00:41:52,110 --> 00:41:52,440 Oh yeah. 814 00:41:52,440 --> 00:41:53,273 [INTERPOSING VOICES] 815 00:41:53,273 --> 00:41:54,435 I started the summer '51. 816 00:41:54,435 --> 00:41:54,935 Really? 817 00:41:54,935 --> 00:41:56,650 I was in the first crop. 818 00:41:56,650 --> 00:42:00,920 Well, then it went faster than I thought. 819 00:42:00,920 --> 00:42:06,620 Anyway, we had somewhere between six and eight assistantships 820 00:42:06,620 --> 00:42:08,482 all spread around. 821 00:42:08,482 --> 00:42:10,190 There were a couple in physics, and there 822 00:42:10,190 --> 00:42:11,930 were a number in math, and there were 823 00:42:11,930 --> 00:42:15,350 a number in the various electrical engineering 824 00:42:15,350 --> 00:42:17,120 departments. 825 00:42:17,120 --> 00:42:22,640 And one or two in economics even somewhere along the line. 826 00:42:22,640 --> 00:42:26,680 827 00:42:26,680 --> 00:42:29,160 So things began to build up. 828 00:42:29,160 --> 00:42:34,170 But it was obvious that very soon, 829 00:42:34,170 --> 00:42:39,630 the available time in Whirlwind was going to be saturated. 830 00:42:39,630 --> 00:42:43,240 And something would have had to be done. 831 00:42:43,240 --> 00:42:48,790 So I can't remember how I got acquainted 832 00:42:48,790 --> 00:42:52,770 with Cuthbert Hurd in IBM. 833 00:42:52,770 --> 00:42:58,800 834 00:42:58,800 --> 00:43:09,270 Somewhere along in the early '50s, he started to visit here. 835 00:43:09,270 --> 00:43:12,480 And I was telling him some of the complications. 836 00:43:12,480 --> 00:43:18,500 And one of the things that I kept pressing on him 837 00:43:18,500 --> 00:43:23,070 was the point that the first computer manufacturer that 838 00:43:23,070 --> 00:43:27,780 installs a machine in a place like MIT 839 00:43:27,780 --> 00:43:31,440 was going to have a great advantage in selling machines 840 00:43:31,440 --> 00:43:34,770 after these graduates got out. 841 00:43:34,770 --> 00:43:38,040 You'll notice that you're not acquainted with Cuthbert 842 00:43:38,040 --> 00:43:39,720 Hurd through Fortran? 843 00:43:39,720 --> 00:43:40,980 Is that correct? 844 00:43:40,980 --> 00:43:42,840 Could be. 845 00:43:42,840 --> 00:43:43,800 Mrs. or Mr.? 846 00:43:43,800 --> 00:43:48,000 Fortran is another story. 847 00:43:48,000 --> 00:43:49,035 Because IBM persuaded-- 848 00:43:49,035 --> 00:43:49,868 [INTERPOSING VOICES] 849 00:43:49,868 --> 00:43:55,350 I vaguely remember that the Whirlwind people were very much 850 00:43:55,350 --> 00:43:58,980 against compilers-- 851 00:43:58,980 --> 00:44:02,850 Fortran in particular as it was being developed. 852 00:44:02,850 --> 00:44:06,980 Because it was going to use a lot more of computer time. 853 00:44:06,980 --> 00:44:10,700 They had to run the problem through twice. 854 00:44:10,700 --> 00:44:15,060 First to compile it, and then to run the problem. 855 00:44:15,060 --> 00:44:19,390 And the old timers didn't like this. 856 00:44:19,390 --> 00:44:20,360 [INTERPOSING VOICES] 857 00:44:20,360 --> 00:44:22,130 --some of them, but not all. 858 00:44:22,130 --> 00:44:23,900 No, that's true. 859 00:44:23,900 --> 00:44:26,056 There were a number of them took part of it. 860 00:44:26,056 --> 00:44:28,640 And considered amicably, they became old timers after that. 861 00:44:28,640 --> 00:44:30,350 [LAUGHTER] 862 00:44:30,350 --> 00:44:34,610 Did I recall, Corby, in '51, didn't Charlie Adams 863 00:44:34,610 --> 00:44:36,350 have a summer thing? 864 00:44:36,350 --> 00:44:39,230 And a couple of his people, Sheldon, something or other-- 865 00:44:39,230 --> 00:44:40,190 Sheldon Best. 866 00:44:40,190 --> 00:44:42,560 --put together an algebraic compiler. 867 00:44:42,560 --> 00:44:46,040 And then, by the time I came back to MIT '55, 868 00:44:46,040 --> 00:44:48,550 I heard rumors they'd gone to MIT to work on Fortran. 869 00:44:48,550 --> 00:44:49,050 Yeah. 870 00:44:49,050 --> 00:44:49,310 Yeah. 871 00:44:49,310 --> 00:44:50,148 Yeah, there was-- 872 00:44:50,148 --> 00:44:51,190 We had a session on that. 873 00:44:51,190 --> 00:44:51,710 Yeah. 874 00:44:51,710 --> 00:44:53,703 Maybe Sheldon did do something. 875 00:44:53,703 --> 00:44:55,120 Of course, Hal Lanning was the one 876 00:44:55,120 --> 00:44:56,825 who did it in the instrumentation lab, 877 00:44:56,825 --> 00:45:01,890 did the early algebraic interpreter. 878 00:45:01,890 --> 00:45:03,230 Was he at Fortran too? 879 00:45:03,230 --> 00:45:06,710 No, he never got related to the-- 880 00:45:06,710 --> 00:45:08,420 he just stayed within the Draper orbit 881 00:45:08,420 --> 00:45:11,150 and continued to work with compilers for them. 882 00:45:11,150 --> 00:45:14,600 Lanny was here, and sort of interesting story quickly, 883 00:45:14,600 --> 00:45:17,000 because it's a diversion really, but you'd be interested. 884 00:45:17,000 --> 00:45:19,040 He developed a compiler, and he's 885 00:45:19,040 --> 00:45:20,810 so good at [INAUDIBLE] which is still is 886 00:45:20,810 --> 00:45:23,100 that you just never trusted the compiler. 887 00:45:23,100 --> 00:45:23,600 Wow. 888 00:45:23,600 --> 00:45:26,540 And when he was sitting here, he said he still 889 00:45:26,540 --> 00:45:27,540 doesn't trust compilers. 890 00:45:27,540 --> 00:45:29,748 It's much easier to do it with the sounding language. 891 00:45:29,748 --> 00:45:30,350 He pushed it. 892 00:45:30,350 --> 00:45:33,270 And so, you get that kind of personal thing. 893 00:45:33,270 --> 00:45:35,620 But they did do it around '54 I think it was. 894 00:45:35,620 --> 00:45:36,440 Yeah. 895 00:45:36,440 --> 00:45:41,660 Well, they're always people that have clung to the details. 896 00:45:41,660 --> 00:45:47,290 But no, Phil basically right. 897 00:45:47,290 --> 00:45:49,940 I don't quite know how the liaison got started. 898 00:45:49,940 --> 00:45:53,115 But MIT had showed them-- and wasn't there 899 00:45:53,115 --> 00:45:54,240 another-- one other person? 900 00:45:54,240 --> 00:45:55,270 Yeah there was somebody. 901 00:45:55,270 --> 00:45:56,437 I can't remember their name. 902 00:45:56,437 --> 00:45:59,330 And Sheldon was part of that original-- 903 00:45:59,330 --> 00:46:00,110 [INAUDIBLE] 904 00:46:00,110 --> 00:46:02,210 Now, there was another MIT person. 905 00:46:02,210 --> 00:46:04,130 Sheldon was part of that original team, 906 00:46:04,130 --> 00:46:06,350 and it was the one that got involved 907 00:46:06,350 --> 00:46:09,710 in the index [INAUDIBLE] dropped optimization, which 908 00:46:09,710 --> 00:46:13,880 became a monument to obscurity, in the sense it was very 909 00:46:13,880 --> 00:46:16,310 effective, and people never understood it 910 00:46:16,310 --> 00:46:21,980 for about five or 10 years later how they did it. 911 00:46:21,980 --> 00:46:24,240 Because they didn't write it up very well. 912 00:46:24,240 --> 00:46:26,630 [INAUDIBLE] at all. 913 00:46:26,630 --> 00:46:28,250 But it was one of key. 914 00:46:28,250 --> 00:46:30,740 Anyway, Cuthbert, I guess, managed 915 00:46:30,740 --> 00:46:33,260 to persuade the top people at IBM 916 00:46:33,260 --> 00:46:35,000 that something had to be done. 917 00:46:35,000 --> 00:46:38,360 And the next thing I heard was that Jim Killian called in 918 00:46:38,360 --> 00:46:42,830 and said, they understood that Watson 919 00:46:42,830 --> 00:46:46,760 wanted to donate to a big computer to MIT. 920 00:46:46,760 --> 00:46:49,440 921 00:46:49,440 --> 00:46:51,765 Because he was mad at Harvard. 922 00:46:51,765 --> 00:46:52,265 [LAUGHTER] 923 00:46:52,265 --> 00:46:54,050 Hmm? 924 00:46:54,050 --> 00:46:59,920 So that's essentially the way the thing got started. 925 00:46:59,920 --> 00:47:04,370 MIT was building the Compton building, what is it 26? 926 00:47:04,370 --> 00:47:06,950 Building 26, right. 927 00:47:06,950 --> 00:47:11,210 And the ticker was that IBM would 928 00:47:11,210 --> 00:47:17,030 pay for some additional amount of the building, 929 00:47:17,030 --> 00:47:21,650 plus a hunk that stuck outside, that 930 00:47:21,650 --> 00:47:24,230 was going to be the housing-- 931 00:47:24,230 --> 00:47:26,990 the place for the computer. 932 00:47:26,990 --> 00:47:31,170 So we had a lot of arguments as to how the thing be installed, 933 00:47:31,170 --> 00:47:36,540 and how to cut down the noise on the air conditioning 934 00:47:36,540 --> 00:47:37,790 and all the rest of the stuff. 935 00:47:37,790 --> 00:47:41,510 But it got in. 936 00:47:41,510 --> 00:47:47,640 In the meantime, IBM was willing to donate additional money 937 00:47:47,640 --> 00:47:50,130 for research assistants. 938 00:47:50,130 --> 00:47:56,700 Both at MIT and at New England universities, a dozen 939 00:47:56,700 --> 00:47:58,110 each, I believe. 940 00:47:58,110 --> 00:48:04,320 So I had to set up committees, both inside MIT 941 00:48:04,320 --> 00:48:12,850 and with their related colleges. 942 00:48:12,850 --> 00:48:16,690 I don't know, was that the time that you got acquainted 943 00:48:16,690 --> 00:48:19,175 with the fact that we were going to get a computer, 944 00:48:19,175 --> 00:48:22,220 or had you heard about it earlier? 945 00:48:22,220 --> 00:48:26,020 I became acquainted with it when two IBMers-- 946 00:48:26,020 --> 00:48:28,990 all I remember is that they were two tall men in dark suits. 947 00:48:28,990 --> 00:48:29,796 [LAUGHTER] 948 00:48:29,796 --> 00:48:30,820 Wearing ties. 949 00:48:30,820 --> 00:48:32,560 And one of them was Truman Hunter, 950 00:48:32,560 --> 00:48:35,265 and the other one was either Cuthbert Hurd or Nat Rochester. 951 00:48:35,265 --> 00:48:36,640 I've forgotten which, but I think 952 00:48:36,640 --> 00:48:38,350 it was probably Rochester. 953 00:48:38,350 --> 00:48:41,470 Came up to MIT to-- sorry. 954 00:48:41,470 --> 00:48:44,620 Came up to Dartmouth to sell this concept. 955 00:48:44,620 --> 00:48:49,660 And their main contact with John Kemeny. 956 00:48:49,660 --> 00:48:52,480 And then I went to lunch with them 957 00:48:52,480 --> 00:48:55,450 and became acquainted with Rochester 958 00:48:55,450 --> 00:48:56,980 and harangued him about something. 959 00:48:56,980 --> 00:48:58,610 I suppose it was AI. 960 00:48:58,610 --> 00:49:01,150 And he offered me a summer job, which I took so. 961 00:49:01,150 --> 00:49:04,570 I was with IBM in the summer of '55. 962 00:49:04,570 --> 00:49:07,137 And then I became, I guess, Dartmouth's representative 963 00:49:07,137 --> 00:49:07,720 to the school. 964 00:49:07,720 --> 00:49:08,345 Right, you did. 965 00:49:08,345 --> 00:49:09,550 Right. 966 00:49:09,550 --> 00:49:12,680 That was my first acquaintance. 967 00:49:12,680 --> 00:49:17,390 At any rate, when we got going, sometime in '57 968 00:49:17,390 --> 00:49:20,270 I believe the machine got in. 969 00:49:20,270 --> 00:49:22,270 We began to roll. 970 00:49:22,270 --> 00:49:30,790 And the machine very soon got overloaded, of course. 971 00:49:30,790 --> 00:49:33,610 As more and more people began to use it, 972 00:49:33,610 --> 00:49:37,410 as more and more faculty began to learn what could be done. 973 00:49:37,410 --> 00:49:41,340 974 00:49:41,340 --> 00:49:47,410 It wasn't very long before these ideas that you, 975 00:49:47,410 --> 00:49:51,960 and you too, began to percolate. 976 00:49:51,960 --> 00:49:57,210 The idea of interactive work, or [INAUDIBLE] 977 00:49:57,210 --> 00:50:00,670 and some of his early work too. 978 00:50:00,670 --> 00:50:09,510 And I started talking about it to IBM. 979 00:50:09,510 --> 00:50:14,040 Well, they weren't interested. 980 00:50:14,040 --> 00:50:19,200 This was one of these fancy long-haired university 981 00:50:19,200 --> 00:50:22,980 sort of things that would never get off the ground. 982 00:50:22,980 --> 00:50:26,770 And they were perfectly willing for us to play around with it. 983 00:50:26,770 --> 00:50:30,210 But first, we're not at all interested in helping out 984 00:50:30,210 --> 00:50:31,890 with additional equipment. 985 00:50:31,890 --> 00:50:35,610 It soon became quite obvious that in order to get very far, 986 00:50:35,610 --> 00:50:39,100 we had to have more memory anyway. 987 00:50:39,100 --> 00:50:42,130 I would add a footnote, that there 988 00:50:42,130 --> 00:50:44,350 was very little recognition in those days 989 00:50:44,350 --> 00:50:49,240 of the rapid obsolescence that would characterize computing 990 00:50:49,240 --> 00:50:50,140 equipment. 991 00:50:50,140 --> 00:50:53,860 And I can still recall a kind of battleship. 992 00:50:53,860 --> 00:50:56,470 In those days, there were no false floors. 993 00:50:56,470 --> 00:51:01,050 And the original computer floor was steel, maybe a quarter 994 00:51:01,050 --> 00:51:04,830 1/2 inch, or maybe 1/2 inch steel plate, which 995 00:51:04,830 --> 00:51:07,170 a welder would come in with a cutting torch, 996 00:51:07,170 --> 00:51:10,230 make a hole for cables to go down underneath the floor. 997 00:51:10,230 --> 00:51:12,285 It was a gigantic construction project that 998 00:51:12,285 --> 00:51:13,200 would lay off the machine room. 999 00:51:13,200 --> 00:51:15,408 And every time you wanted to change the machine room, 1000 00:51:15,408 --> 00:51:18,940 you have to get new steel plate, or weld up the holes. 1001 00:51:18,940 --> 00:51:20,190 It was really wild. 1002 00:51:20,190 --> 00:51:23,670 And then, you had persuaded them over this noise issue, 1003 00:51:23,670 --> 00:51:28,887 that they should customize the airflow on each of the units, 1004 00:51:28,887 --> 00:51:30,720 so that in fact, they were inducted directly 1005 00:51:30,720 --> 00:51:32,745 from the plant and rather, all the internal fans 1006 00:51:32,745 --> 00:51:33,370 were taken out. 1007 00:51:33,370 --> 00:51:35,490 Much against IBM's-- 1008 00:51:35,490 --> 00:51:37,320 IBM objected pretty hard to it, because -- 1009 00:51:37,320 --> 00:51:38,250 Yeah. 1010 00:51:38,250 --> 00:51:41,530 --and we'd never do that ever again. 1011 00:51:41,530 --> 00:51:44,750 But it seemed like a good idea at the time. 1012 00:51:44,750 --> 00:51:45,570 Yeah. 1013 00:51:45,570 --> 00:51:47,010 Well, it quieted things. 1014 00:51:47,010 --> 00:51:49,170 Yeah. 1015 00:51:49,170 --> 00:51:52,890 It also led to silver migration, because our air conditioning 1016 00:51:52,890 --> 00:51:55,323 unit was-- 1017 00:51:55,323 --> 00:51:56,990 this is really a footnote on a footnote. 1018 00:51:56,990 --> 00:52:02,670 Our air conditioning unit was the usual crude affair. 1019 00:52:02,670 --> 00:52:06,360 And we got excessive moisture condensation 1020 00:52:06,360 --> 00:52:09,670 around the terminal strips of the 704. 1021 00:52:09,670 --> 00:52:10,170 [LAUGHTER] 1022 00:52:10,170 --> 00:52:11,160 And it turned out that-- 1023 00:52:11,160 --> 00:52:11,910 Forgot about that. 1024 00:52:11,910 --> 00:52:16,980 --it was a design, a production flaw back at IBM, 1025 00:52:16,980 --> 00:52:19,410 which they didn't really find out about until about two 1026 00:52:19,410 --> 00:52:20,970 or three years later. 1027 00:52:20,970 --> 00:52:28,350 That somebody had substituted silver as a contact material. 1028 00:52:28,350 --> 00:52:31,230 And that led to a phenomena which was well 1029 00:52:31,230 --> 00:52:32,800 known in the research literature, 1030 00:52:32,800 --> 00:52:35,160 but the production people didn't understand, 1031 00:52:35,160 --> 00:52:40,200 that the silver ions would migrate across the insulation 1032 00:52:40,200 --> 00:52:43,140 and proceed to form a high resistance bridge, which 1033 00:52:43,140 --> 00:52:46,140 would lead to shorting and signal failures, and so forth. 1034 00:52:46,140 --> 00:52:48,690 And about the time we shipped out that 704, 1035 00:52:48,690 --> 00:52:50,640 it was just about ready to be scrapped anyway, 1036 00:52:50,640 --> 00:52:54,620 because we were getting silver migration problems everywhere. 1037 00:52:54,620 --> 00:52:57,260 1038 00:52:57,260 --> 00:53:05,120 Well, my problem at the time was to keep up 1039 00:53:05,120 --> 00:53:08,300 with the demand for computing. 1040 00:53:08,300 --> 00:53:11,500 At the same time help out as much 1041 00:53:11,500 --> 00:53:18,340 as I could by persuading [INAUDIBLE] to add equipment as 1042 00:53:18,340 --> 00:53:19,010 needed. 1043 00:53:19,010 --> 00:53:24,610 And I felt myself a little bit between two hard rocks. 1044 00:53:24,610 --> 00:53:29,920 1045 00:53:29,920 --> 00:53:33,280 I guess I can jump ahead and let the rest of you 1046 00:53:33,280 --> 00:53:34,615 fill in in between. 1047 00:53:34,615 --> 00:53:37,150 1048 00:53:37,150 --> 00:53:39,910 About the time that ac got started, of course, 1049 00:53:39,910 --> 00:53:46,690 there was a general feeling that IBM was not the manufacturer 1050 00:53:46,690 --> 00:53:51,230 to help out with time-sharing, because they've been pretty 1051 00:53:51,230 --> 00:53:54,450 much against it right along. 1052 00:53:54,450 --> 00:53:56,255 [INAUDIBLE] complicate it. 1053 00:53:56,255 --> 00:53:58,940 Yeah, I know you will. 1054 00:53:58,940 --> 00:54:01,520 And so, when the final decision was 1055 00:54:01,520 --> 00:54:06,200 made to get a GE machine for the Mac, 1056 00:54:06,200 --> 00:54:10,970 I was the one that got it in the neck. 1057 00:54:10,970 --> 00:54:13,230 I got called in by-- 1058 00:54:13,230 --> 00:54:17,840 or no, I guess, I forget the guy's name. 1059 00:54:17,840 --> 00:54:23,270 Head of IBM for New England. 1060 00:54:23,270 --> 00:54:23,770 Bradley. 1061 00:54:23,770 --> 00:54:24,485 Phil Bradley. 1062 00:54:24,485 --> 00:54:27,920 Yeah Phil Bradley, came in and gave me the worst 1063 00:54:27,920 --> 00:54:29,930 dressing down I have ever had. 1064 00:54:29,930 --> 00:54:31,810 [LAUGHTER] 1065 00:54:31,810 --> 00:54:33,170 MIT was-- 1066 00:54:33,170 --> 00:54:35,454 [INTERPOSING VOICES] 1067 00:54:35,454 --> 00:54:38,360 1068 00:54:38,360 --> 00:54:44,540 --just didn't have any conscience or-- 1069 00:54:44,540 --> 00:54:47,590 1070 00:54:47,590 --> 00:54:48,965 Harvard had betrayed them before. 1071 00:54:48,965 --> 00:54:49,798 [INTERPOSING VOICES] 1072 00:54:49,798 --> 00:54:51,430 Any gratitude toward a company that 1073 00:54:51,430 --> 00:54:55,550 had befriended the institute for so long, 1074 00:54:55,550 --> 00:55:01,990 we went out of our way to rub IBM's nose in the dirt. 1075 00:55:01,990 --> 00:55:03,910 It took about a half hour before I 1076 00:55:03,910 --> 00:55:09,640 had a chance to even say thank you or goodbye. 1077 00:55:09,640 --> 00:55:14,170 Well, IBM in those days would punish people badly, 1078 00:55:14,170 --> 00:55:15,310 financially. 1079 00:55:15,310 --> 00:55:16,420 Oh yeah. 1080 00:55:16,420 --> 00:55:18,970 What happened to Bradley? 1081 00:55:18,970 --> 00:55:21,370 He survived a few more years. 1082 00:55:21,370 --> 00:55:23,270 With an account in [INAUDIBLE]? 1083 00:55:23,270 --> 00:55:24,770 But the person that really caught it 1084 00:55:24,770 --> 00:55:26,740 was Lauren Bullet, who worked in the IBM-- 1085 00:55:26,740 --> 00:55:27,573 [INTERPOSING VOICES] 1086 00:55:27,573 --> 00:55:29,320 --university representative, and who 1087 00:55:29,320 --> 00:55:33,640 had been one of our staunchest friends in trying to help us. 1088 00:55:33,640 --> 00:55:37,420 And they of course, in standard fashion, 1089 00:55:37,420 --> 00:55:39,940 punished him the worst. 1090 00:55:39,940 --> 00:55:40,900 Yeah. 1091 00:55:40,900 --> 00:55:43,760 Why don't we go on to Herb, because-- 1092 00:55:43,760 --> 00:55:44,650 Yeah. 1093 00:55:44,650 --> 00:55:47,220 I think so. 1094 00:55:47,220 --> 00:55:49,200 --there are many things in this conversation 1095 00:55:49,200 --> 00:55:54,180 which I would consider to be quite inaccurate. 1096 00:55:54,180 --> 00:55:55,140 OK. 1097 00:55:55,140 --> 00:55:59,580 But let's sort of go through the preliminary part first. 1098 00:55:59,580 --> 00:56:00,420 Herb, you're on. 1099 00:56:00,420 --> 00:56:02,630 OK. 1100 00:56:02,630 --> 00:56:07,290 One footnote or aside before I start. 1101 00:56:07,290 --> 00:56:09,180 This is 20 years ago. 1102 00:56:09,180 --> 00:56:12,270 And as Bob [INAUDIBLE] pointed out, 1103 00:56:12,270 --> 00:56:14,940 we were awfully God damn young. 1104 00:56:14,940 --> 00:56:17,550 And we were awfully insecure. 1105 00:56:17,550 --> 00:56:22,230 And we didn't know each other very well, truth be known. 1106 00:56:22,230 --> 00:56:26,670 I've heard John and Corby say things about their background 1107 00:56:26,670 --> 00:56:28,260 I never knew. 1108 00:56:28,260 --> 00:56:30,990 And I rather suspect there are things in my background 1109 00:56:30,990 --> 00:56:32,860 they never knew. 1110 00:56:32,860 --> 00:56:34,770 We never got to be friends. 1111 00:56:34,770 --> 00:56:37,200 And that's a pity, because I think 1112 00:56:37,200 --> 00:56:40,920 computing and time-sharing and the whole megillah 1113 00:56:40,920 --> 00:56:43,500 might have gone off in a different direction 1114 00:56:43,500 --> 00:56:46,400 if we hadn't been quite so pitted against each other. 1115 00:56:46,400 --> 00:56:50,110 If there had been a little more of a vector dot product. 1116 00:56:50,110 --> 00:56:53,160 Be as it may, I was born in Ohio. 1117 00:56:53,160 --> 00:56:56,820 I matriculated high school in Brooklyn, New York. 1118 00:56:56,820 --> 00:57:01,200 I was a real pain in the ass. 1119 00:57:01,200 --> 00:57:02,490 I was a hot shot. 1120 00:57:02,490 --> 00:57:06,870 I won a Westinghouse as a senior, 1121 00:57:06,870 --> 00:57:10,450 and figured I was one of the smartest kids in the country. 1122 00:57:10,450 --> 00:57:13,470 I came to MIT convince of one thing. 1123 00:57:13,470 --> 00:57:15,780 I did want to be a professor. 1124 00:57:15,780 --> 00:57:18,210 My dad had tried the industrial route, 1125 00:57:18,210 --> 00:57:21,000 and I watched him bleed a little bit too much. 1126 00:57:21,000 --> 00:57:24,540 And I thought there was such a thing as an ivory tower. 1127 00:57:24,540 --> 00:57:27,420 And I couldn't decide what I wanted to be. 1128 00:57:27,420 --> 00:57:30,420 So I solved the problem by tripling 1129 00:57:30,420 --> 00:57:34,440 in electrical engineering, physics, and chemical 1130 00:57:34,440 --> 00:57:35,610 engineering. 1131 00:57:35,610 --> 00:57:39,360 One of my dad's good friends was a chemical engineer. 1132 00:57:39,360 --> 00:57:40,610 [INAUDIBLE]