1 00:00:00,280 --> 00:00:02,400 Hi, and welcome to the Neil Ashton Podcast. 2 00:00:03,080 --> 00:00:05,920 In each episode, we explained some of the fascinating ways 3 00:00:05,920 --> 00:00:09,080 that science and engineering are changing the world around us. 4 00:00:09,800 --> 00:00:12,760 We talked to leading engineers from elite level sports like 5 00:00:12,840 --> 00:00:16,920 cycling and Formula One to some of the world's top academics to 6 00:00:16,920 --> 00:00:21,000 understand how fluid dynamics, machine learning, supercomputing 7 00:00:21,360 --> 00:00:23,000 are bringing in a new era of discovery. 8 00:00:23,960 --> 00:00:27,080 We also hear some of their life stories, their career advice, 9 00:00:27,640 --> 00:00:30,080 the lessons they've learned on the way that I hope will be 10 00:00:30,080 --> 00:00:33,840 helpful to you too. So sit back and enjoy this 11 00:00:33,840 --> 00:00:41,120 episode. Hi, and welcome back to the new 12 00:00:41,480 --> 00:00:44,520 Ashton podcast today. I think it's a pretty special 13 00:00:44,520 --> 00:00:46,040 episode. It's with really one of the 14 00:00:46,040 --> 00:00:49,720 icons of CFD, Professor Ansi Jamieson. 15 00:00:50,560 --> 00:00:54,760 It really is somebody who I'd always known about, you know, 16 00:00:54,760 --> 00:00:58,360 from the early days of me studying CFD and speaking to 17 00:00:58,360 --> 00:01:01,640 various people, it was clear that he was somebody who was a 18 00:01:01,640 --> 00:01:04,200 real pioneer. But I have to completely confess 19 00:01:04,200 --> 00:01:12,280 that I was I was learning a lot through the conversation I had 20 00:01:12,360 --> 00:01:17,320 with him. And as I was going through it, I 21 00:01:17,320 --> 00:01:25,200 was blown away by what he has done and what what it has led to 22 00:01:26,680 --> 00:01:29,080 it. It's pretty incredible that he 23 00:01:29,080 --> 00:01:34,560 was there at the beginning of both the use of computational 24 00:01:34,560 --> 00:01:40,000 fluid dynamics, but also the early days of computing, the 25 00:01:40,040 --> 00:01:44,640 early days of really the aeronautics, aeronautics, the 26 00:01:44,640 --> 00:01:47,960 move to transonic, the sort of evolution in aircraft design. 27 00:01:49,120 --> 00:01:53,040 And IT it's amazing. And I purposely have released it 28 00:01:53,600 --> 00:01:56,640 today, which is his 90th birthday. 29 00:01:57,320 --> 00:02:01,480 And so I consider this an episode to really celebrate the 30 00:02:01,480 --> 00:02:04,280 life and career of Professor Anthony Jameson. 31 00:02:05,480 --> 00:02:13,760 Just very briefly to summarize, he had incredible back story 32 00:02:15,080 --> 00:02:17,160 going into the British Army, being an officer in the British 33 00:02:17,160 --> 00:02:21,160 Army, living around the world in India, Malaya, studying at 34 00:02:21,160 --> 00:02:26,880 Cambridge, then moving over to to the US, working with some of 35 00:02:26,880 --> 00:02:30,120 the defense companies and aerospace companies, writing 36 00:02:30,120 --> 00:02:34,400 code, then moving to become a professor in New York, then 37 00:02:34,400 --> 00:02:39,400 eventually going to Princeton. But really the overall and later 38 00:02:39,400 --> 00:02:42,080 then he went to Stanford and he's now at Texas A&M. 39 00:02:42,920 --> 00:02:48,560 But what blows my mind is one, his intellect, he's just the way 40 00:02:48,560 --> 00:02:50,960 you can hear and speak about things. 41 00:02:51,560 --> 00:02:54,600 But also the fact that when you appreciate what he was writing 42 00:02:54,600 --> 00:02:59,600 these codes decades before any of the codes that we today take 43 00:02:59,600 --> 00:03:02,120 for granted, the commercial codes, the government codes. 44 00:03:03,480 --> 00:03:07,560 And you really can see that it, it really, I, I don't think it's 45 00:03:07,560 --> 00:03:11,320 too far of a stretch of imagination to say he was one of 46 00:03:11,320 --> 00:03:15,760 the key people that is responsible for modern day CFD. 47 00:03:15,760 --> 00:03:19,920 There's not many people who can say that it's, it's incredible. 48 00:03:19,920 --> 00:03:25,480 And yeah, he, he's known for certain certain things, you 49 00:03:25,480 --> 00:03:28,440 know, like adjoint methods, maybe for aircraft design or 50 00:03:28,440 --> 00:03:31,200 the, or some of the, the JST scheme shock capturing. 51 00:03:32,800 --> 00:03:37,080 But I think that that culture of writing the code, getting in, in 52 00:03:37,080 --> 00:03:41,440 the very early days, working with industry closely, he was 53 00:03:41,440 --> 00:03:44,880 not a pure academic nor a pure industry. 54 00:03:44,880 --> 00:03:49,840 It was clear that he was always moving between and you know, as 55 00:03:49,840 --> 00:03:52,520 we discussed a little bit, but probably didn't focus on too 56 00:03:52,520 --> 00:03:55,160 much. Just the sheer number of people 57 00:03:55,160 --> 00:03:58,840 that he has supervised and mentored who have gone on to do 58 00:03:58,840 --> 00:04:03,360 amazing things as top professors and and engineers throughout the 59 00:04:03,360 --> 00:04:06,640 world is a real testament to what he has done. 60 00:04:06,640 --> 00:04:12,000 Now this is a completely unedited, this is a more than 61 00:04:12,000 --> 00:04:16,399 two hour long conversation and I purposely kept it this way to 62 00:04:16,399 --> 00:04:20,040 not take anything away from what his his life story is. 63 00:04:20,240 --> 00:04:23,520 There was far many more things we could have spoken about but I 64 00:04:23,520 --> 00:04:26,240 wanted to be respectful of of his time. 65 00:04:27,880 --> 00:04:30,280 I will put some links in into the chat. 66 00:04:30,280 --> 00:04:34,280 If you're watching this on, on YouTube, it's probably one that 67 00:04:34,280 --> 00:04:36,680 maybe will take you a few times to go through because, you know, 68 00:04:36,800 --> 00:04:38,400 like I said, it's more than two hours long. 69 00:04:38,400 --> 00:04:43,200 But yeah, I, I just have incredible honour of being able 70 00:04:43,200 --> 00:04:46,240 to, to speak to him. And I want to thank him. 71 00:04:46,720 --> 00:04:49,960 And hopefully I can say that on behalf of everybody else to 72 00:04:49,960 --> 00:04:53,360 thank him for his amazing contribution to the field of 73 00:04:53,920 --> 00:04:58,120 computational fluid dynamics, computational aerodynamics and 74 00:04:58,120 --> 00:05:01,560 scientific computing. So sit back and enjoy this 75 00:05:01,560 --> 00:05:04,240 episode with Professor Anthony Jameson. 76 00:05:04,800 --> 00:05:09,680 So, so where did it, where did it all all begin? 77 00:05:09,680 --> 00:05:14,800 So did you have, were you one of these children who were always 78 00:05:14,800 --> 00:05:19,160 into aircraft from a very young age, or was this something that 79 00:05:19,240 --> 00:05:21,880 that took age? How did this love of aircraft 80 00:05:21,880 --> 00:05:24,120 and and later fluid dynamics begin? 81 00:05:24,920 --> 00:05:30,040 Well, so actually there's more boys. 82 00:05:30,040 --> 00:05:33,240 So I can tell you the following story, and of course it's just 83 00:05:33,800 --> 00:05:39,880 my memory, but my father was a British Army officer, so I think 84 00:05:40,760 --> 00:05:48,760 he was sent to India in 1938 when I was three years old, and 85 00:05:48,760 --> 00:05:52,520 he spent six months in Bombay. But then he was attached to the 86 00:05:53,360 --> 00:05:56,000 headquarters of the British Army. 87 00:05:56,800 --> 00:05:59,000 Remember, there was an Indian army, but there was also 88 00:05:59,000 --> 00:06:01,520 contingents of the British Army in India. 89 00:06:03,160 --> 00:06:08,840 So the headquarters is in Delhi. But because Delhi is very hot 90 00:06:09,400 --> 00:06:12,200 during the summer, they moved to this summer. 91 00:06:12,200 --> 00:06:15,760 Capital similar, which is at foothills of the similarities at 92 00:06:15,760 --> 00:06:21,720 7000 feet. So somewhere around when I was 93 00:06:22,000 --> 00:06:28,720 perhaps I've also had a nanny at this point and she took my 94 00:06:28,720 --> 00:06:31,880 brother and me. I have a brother who was one 95 00:06:31,880 --> 00:06:35,040 year younger than me, but born on the same day as me. 96 00:06:35,680 --> 00:06:40,320 On a walk in Delhi. We went past a small airport, 97 00:06:40,320 --> 00:06:47,200 which I'm sure doesn't exist anymore, and there was a, a 98 00:06:47,200 --> 00:06:51,240 biplane there, a twin engined biplane, something like a 99 00:06:51,240 --> 00:06:54,680 Vicar's Valletta if you want, but I don't know exactly what it 100 00:06:54,680 --> 00:06:57,320 was. But it had by playing with 101 00:06:57,320 --> 00:07:01,400 sticks and strings all over the place and the engines in the 102 00:07:01,400 --> 00:07:03,440 middle of each, between each pair of wings. 103 00:07:05,720 --> 00:07:10,560 I looked at that and you know, I had been looking at birds quite 104 00:07:10,560 --> 00:07:13,280 a bit and I thought this is all wrong. 105 00:07:14,720 --> 00:07:19,440 You know, birds don't have all these wires and sticks and 106 00:07:19,440 --> 00:07:21,840 things. But then it took another look 107 00:07:21,840 --> 00:07:26,200 and I was looking at, I saw it had these movable hinge surfaces 108 00:07:26,200 --> 00:07:32,640 on the wings and on the tail. And I looked at the tail and saw 109 00:07:32,680 --> 00:07:36,800 that if you, I saw the wind bearing over the plane. 110 00:07:36,800 --> 00:07:41,600 I can't tell you why, but I saw that if you push the elevator 111 00:07:41,600 --> 00:07:44,960 up, that the wind would hit the top of the elevator and push the 112 00:07:44,960 --> 00:07:51,040 tail down. And that was my first time I'd 113 00:07:51,040 --> 00:07:53,640 ever really had a close view of an aircraft. 114 00:07:54,960 --> 00:07:57,800 Well, I can't have been more than about six years old because 115 00:08:00,080 --> 00:08:08,000 soon after that my father was sent to Iraq and my mother spent 116 00:08:08,520 --> 00:08:11,320 moved to similar on a full time basis. 117 00:08:11,320 --> 00:08:15,440 And we lived in similar for the next between 1940 and 43. 118 00:08:18,120 --> 00:08:25,160 So my mother taught me to read and write by copying over her 119 00:08:25,160 --> 00:08:27,240 handwriting. She started teaching me the 120 00:08:27,240 --> 00:08:31,000 tables when I was about 7 and I told her that you don't need to 121 00:08:31,000 --> 00:08:33,000 teach me this because I can work it out. 122 00:08:35,159 --> 00:08:40,960 It's fairly obvious to me that you can, you know, 7 + 7 to 14 123 00:08:40,960 --> 00:08:43,600 and add 7 again, it's 21. I can do that. 124 00:08:45,120 --> 00:08:50,760 So at that point, my mother used to read to us. 125 00:08:50,760 --> 00:08:53,880 I haven't yet gone to school. I went to school when I was 7 126 00:08:53,880 --> 00:08:57,880 when my mother got pregnant at this Roman Catholic mission 127 00:08:57,880 --> 00:09:07,120 school in similar, but I found an observer's sort of aircraft 128 00:09:07,120 --> 00:09:11,200 spotting work with three view silhouettes of things like 129 00:09:11,200 --> 00:09:14,000 Spitfires and Measure Schmidt 109 and so on. 130 00:09:15,280 --> 00:09:19,400 And I persuaded my mother to buy this book for me and I forced 131 00:09:19,400 --> 00:09:22,960 her to read out of this book more or less everyday sort of 132 00:09:23,080 --> 00:09:25,400 wingspan, 36 feet and things like that. 133 00:09:27,520 --> 00:09:30,480 So that was that. All right. 134 00:09:31,960 --> 00:09:38,000 But we did. My father was spread first Iraq 135 00:09:38,000 --> 00:09:43,600 and then Iran and then he was moved back to fund against 136 00:09:43,600 --> 00:09:47,280 Japan, which by that time was in the not in Burma but in the 137 00:09:47,280 --> 00:09:54,040 North East of India. So when I was eight we moved to 138 00:09:54,040 --> 00:09:57,960 the Shalom NE of India and I went to another of these Roman 139 00:09:57,960 --> 00:09:59,600 Catholic mission schools for a year. 140 00:10:01,080 --> 00:10:07,080 My dad wanted to have something to do with the invasion of 141 00:10:07,080 --> 00:10:09,200 Normandy and he applied to come back to England. 142 00:10:09,200 --> 00:10:15,120 So we came back to England in March 1944, coming on a convoy 143 00:10:15,120 --> 00:10:17,280 that went through the Mediterranean and then went way 144 00:10:17,280 --> 00:10:22,400 out into the Atlantic and finally landed in Liverpool on a 145 00:10:22,400 --> 00:10:27,960 horrible cold day and they managed to get into luggage vans 146 00:10:27,960 --> 00:10:31,640 or some train we travelled to some. 147 00:10:32,720 --> 00:10:36,280 My father's sister was living in Sherborne at the time when we 148 00:10:36,280 --> 00:10:38,440 spent a few weeks there. And then I was dumped in a 149 00:10:38,440 --> 00:10:40,240 boarding school, a boarding prep school. 150 00:10:42,760 --> 00:10:45,880 And the reason for that particular school was that my 151 00:10:45,880 --> 00:10:52,520 father actually had been a Wimbledon tennis player and also 152 00:10:52,520 --> 00:10:54,400 one of the world's best squash players. 153 00:10:54,400 --> 00:11:00,360 And the owner of the school was somebody else who had been on 154 00:11:00,360 --> 00:11:05,800 the British national squash team that toured America in 1934. 155 00:11:07,000 --> 00:11:10,040 As far as I can make out, I was actually conceived in America 156 00:11:10,040 --> 00:11:17,200 during that squash tour. Anyway, I went to this boarding 157 00:11:17,720 --> 00:11:20,240 prep school that would normally be in Brighton, but it was 158 00:11:20,240 --> 00:11:28,800 evacuated to the Midlands during the war and well, we had I was 159 00:11:28,800 --> 00:11:32,800 put in the third form. It had 6 forms, so I was nine. 160 00:11:32,800 --> 00:11:36,840 I was put in the third form. But what happened was that you 161 00:11:36,840 --> 00:11:41,680 could be promoted each term. And each term I came in the top 162 00:11:41,680 --> 00:11:46,760 five and got promoted. So by the time I was 10, I was 163 00:11:46,760 --> 00:11:48,840 actually sitting in the top form of this school. 164 00:11:51,120 --> 00:11:56,840 And I sat there for the next couple of years kind of without 165 00:11:56,840 --> 00:12:01,600 much to do because the teachers there, I mean, these are during 166 00:12:01,600 --> 00:12:04,280 the war and they were not the only people available to teach, 167 00:12:04,280 --> 00:12:09,680 were sort of retirees and so on. I think that somebody, for 168 00:12:09,680 --> 00:12:12,440 example, could have built me calculus at this point, but 169 00:12:12,760 --> 00:12:15,240 there was nobody at the school who probably knew calculus and 170 00:12:17,200 --> 00:12:27,520 then wound up in 1940. Seven, my father applied to a 171 00:12:27,520 --> 00:12:31,880 job in New York for the what was going to be the United Nations 172 00:12:31,880 --> 00:12:34,040 Army, which never happened because of the Cold War. 173 00:12:35,000 --> 00:12:39,240 The reason it did is that my mother was actually born in 174 00:12:39,240 --> 00:12:42,400 America and we had American families. 175 00:12:42,400 --> 00:12:48,800 So they went to America and they left my brother and me going to 176 00:12:48,880 --> 00:12:50,440 these boarding schools in England. 177 00:12:50,440 --> 00:12:54,120 And we spent our school holidays with an aunt, actually. 178 00:12:56,080 --> 00:12:59,080 And I wound up getting a scholarship to Winchester in 179 00:12:59,080 --> 00:13:07,320 1948. And so that kind of had a big 180 00:13:07,320 --> 00:13:10,640 impact on me. I spent in Winchester courses 181 00:13:12,000 --> 00:13:15,600 probably about that time considered to be the best 182 00:13:15,600 --> 00:13:17,600 academic school in England probably. 183 00:13:19,360 --> 00:13:26,200 And at Winchester, I wound up focusing more on physics and 184 00:13:26,200 --> 00:13:34,400 chemistry, but I wound up eventually getting a scholarship 185 00:13:34,400 --> 00:13:38,280 to Cambridge. Well, we had state scholarships 186 00:13:38,280 --> 00:13:41,200 and county scholarships, but I also took this open scholarship 187 00:13:41,200 --> 00:13:46,680 exam in Cambridge. I went to Trinity Hall because 188 00:13:46,680 --> 00:13:51,680 my father had actually been sent to Cambridge for two years as an 189 00:13:51,680 --> 00:13:56,640 Army officer, but asked him what college he was, he said Trinity. 190 00:13:56,640 --> 00:13:59,080 And somebody at the War Office didn't know that Trinity Hall 191 00:13:59,080 --> 00:14:00,760 and Trinity were two different colleges. 192 00:14:00,760 --> 00:14:03,000 Said he he went to Chinterley Hall. 193 00:14:03,000 --> 00:14:06,320 Sorry, decided I might as well go to the same place. 194 00:14:07,240 --> 00:14:11,240 But I didn't go straight to Cambridge. 195 00:14:11,520 --> 00:14:16,440 See, we had national service in those days, so out of Winchester 196 00:14:16,440 --> 00:14:23,080 I went straight from the Army for two years when I was 18. 197 00:14:24,960 --> 00:14:30,000 And so now there's something else that's a little bit 198 00:14:30,000 --> 00:14:35,240 relevant to this. While I was at Winchester, one 199 00:14:35,240 --> 00:14:39,840 of the maths teachers was Doctor Thwaites. 200 00:14:41,080 --> 00:14:49,120 Thwaites actually was a quite well known expert in theoretical 201 00:14:49,120 --> 00:14:52,720 aerodynamics and he he went on from Winchester to become the 202 00:14:52,880 --> 00:14:56,200 chair of aeronautics at Southampton University. 203 00:14:57,760 --> 00:15:00,600 Now I was never in a class taught by him, but he organised 204 00:15:00,600 --> 00:15:07,880 some trip to visit Super Marine in South outside Southampton, 205 00:15:07,880 --> 00:15:15,480 which is pretty near Winchester. And I told him I was really 206 00:15:15,480 --> 00:15:20,920 interested in aircraft. And he told me that I ought to 207 00:15:20,920 --> 00:15:26,480 read the book by Claret Airfoil and Air School Theory, which I 208 00:15:26,480 --> 00:15:28,640 don't know whether you're familiar with that book or not. 209 00:15:30,960 --> 00:15:35,080 So I managed to to get a copy, I'm not quite sure how. 210 00:15:36,680 --> 00:15:41,800 So now when I'm in the Army, I spent several months in basic 211 00:15:41,800 --> 00:15:47,920 training and then I passed this War Office selection board for 212 00:15:48,200 --> 00:15:51,880 officer training. So another six months on officer 213 00:15:51,880 --> 00:15:56,920 training, then you could apply to go where you want. 214 00:15:56,920 --> 00:16:02,040 And I applied to go to, to Malaya was then called now 215 00:16:02,040 --> 00:16:07,160 called Malaysia, right? So we went out to Malaysia, 216 00:16:07,640 --> 00:16:13,000 Malaya by ship. And I had this book by glass. 217 00:16:14,400 --> 00:16:21,840 I read it on the ship and I've still got that copy of the book 218 00:16:21,840 --> 00:16:25,240 in very tattered form. And you can see that I've put 219 00:16:25,240 --> 00:16:29,120 handwritten notes all over the place when I tried to verify the 220 00:16:29,120 --> 00:16:33,280 formulas. But this, this book actually 221 00:16:34,400 --> 00:16:37,440 introduces complex variables and conformal mapping. 222 00:16:37,440 --> 00:16:42,400 So I actually learnt complex variables from this book and 223 00:16:42,400 --> 00:16:46,320 also conformal mapping. I've also learnt 3D lifting line 224 00:16:46,320 --> 00:16:52,960 wing theory of Prantle. So now I spend the year as a 225 00:16:52,960 --> 00:17:01,400 Lieutenant in Malaya and I came back and my father by this time 226 00:17:02,040 --> 00:17:06,960 was back in England and he knew somebody rather who had a 227 00:17:06,960 --> 00:17:09,800 contact with Bristol Aero Engines. 228 00:17:09,800 --> 00:17:17,040 So I spent six months from March to September working at Bristol 229 00:17:17,040 --> 00:17:21,319 Aero Engines before going to Cambridge. 230 00:17:23,400 --> 00:17:29,120 So I was in a compressor designed section, about 50 231 00:17:29,120 --> 00:17:35,600 people and everybody else in there had a university degree. 232 00:17:37,160 --> 00:17:41,920 But anyway, the first day or two somebody showed me like it was a 233 00:17:41,920 --> 00:17:43,520 side rule. We didn't. 234 00:17:44,120 --> 00:17:46,960 This is before electronic computing had really come in and 235 00:17:47,800 --> 00:17:51,160 I went to a book. I bought a book on 236 00:17:51,480 --> 00:17:55,280 thermodynamics and another one on jet engines and read them in 237 00:17:55,280 --> 00:17:58,480 a hurry. As it turned out, I could do the 238 00:17:58,480 --> 00:18:00,520 job. I didn't didn't need a 239 00:18:00,520 --> 00:18:06,320 university degree. Then the next thing that 240 00:18:06,320 --> 00:18:08,920 happened was that we were supposed to do these stage by 241 00:18:08,920 --> 00:18:14,920 stage calculations as slide rules looking at data from this 242 00:18:15,200 --> 00:18:18,560 Tess station which had like 25,000 horsepower electric 243 00:18:18,560 --> 00:18:24,200 motors driving this and the test station blew up. 244 00:18:24,640 --> 00:18:30,280 Well, the suddenly got out of balance and everything just torn 245 00:18:30,280 --> 00:18:32,240 apart. So there was no data. 246 00:18:33,760 --> 00:18:36,080 So it was the sort of stroke of that for me because then they 247 00:18:36,080 --> 00:18:40,680 had to think of something else for me to do and I looked at 248 00:18:40,680 --> 00:18:43,960 something called radial equilibrium that flows in jet 249 00:18:43,960 --> 00:18:47,400 engine compressors. And then finally I actually 250 00:18:49,120 --> 00:18:55,400 calculated the they were testing or had the 1st 2 spool jet 251 00:18:55,400 --> 00:19:00,360 engine to superstore the Olympus that went in the Valcon and 252 00:19:00,360 --> 00:19:02,320 later was adapted to the Concorde. 253 00:19:04,040 --> 00:19:06,560 And they had flight test data and they didn't know what the 254 00:19:06,560 --> 00:19:11,680 relative speed of the two high speed, high pressure and low 255 00:19:11,680 --> 00:19:15,040 pressure compressors would be. So based on the flight 256 00:19:15,160 --> 00:19:19,200 measurements of the RPM, I calculated the combined 257 00:19:19,600 --> 00:19:22,760 compressor characteristic for the whole 2 spools. 258 00:19:24,240 --> 00:19:27,040 So that's before I got to Cambridge. 259 00:19:29,680 --> 00:19:31,880 Before, yeah, yeah, before you went to university. 260 00:19:32,760 --> 00:19:35,240 All right. But by the time I got to 261 00:19:35,240 --> 00:19:39,600 Cambridge, of course I knew quite a bit of basic 262 00:19:39,600 --> 00:19:42,800 aerodynamics, like I also knew quite a bit about 263 00:19:42,800 --> 00:19:46,640 thermodynamics. I know at Cambridge I wasn't 264 00:19:46,640 --> 00:19:50,160 sure whether to try maths or physics or engineering, but 265 00:19:50,680 --> 00:19:56,400 seemed like the path of minimum effort was to do what we called 266 00:19:56,920 --> 00:19:58,760 mechanical sciences. At Cambridge. 267 00:20:00,840 --> 00:20:04,480 When you take, if you have a scholarship, you could take 268 00:20:04,480 --> 00:20:07,440 what's called a fast course admit you take the secondary 269 00:20:07,440 --> 00:20:11,400 exams the end of your first year and basically your degree exam 270 00:20:11,920 --> 00:20:16,400 after two years and then. The first third year you do 271 00:20:16,400 --> 00:20:21,280 something wants to be a kind of post graduation, sort of like 272 00:20:22,760 --> 00:20:28,560 approximately a master's degree. But but at Cambridge I was 273 00:20:28,560 --> 00:20:32,840 really, I thought they weren't teaching the maths that we 274 00:20:32,840 --> 00:20:37,680 needed, so they didn't teach factor analysis, for example. 275 00:20:39,440 --> 00:20:44,840 And so I started buying, they were various quite inexpensive 276 00:20:44,840 --> 00:20:48,440 books and I started buying these books on things like vector 277 00:20:48,440 --> 00:20:53,840 analysis and tensor analysis and so on and trying to teach myself 278 00:20:57,480 --> 00:21:03,880 anyway, I. Were you always a more than 279 00:21:03,880 --> 00:21:06,040 maths person than an engineering? 280 00:21:06,040 --> 00:21:10,080 Did you? Did you always have a a keenness 281 00:21:10,080 --> 00:21:12,520 for maths? You know, in your earlier years 282 00:21:12,520 --> 00:21:14,600 you said then that they didn't teach enough maths. 283 00:21:14,760 --> 00:21:17,040 Do you, do you have a very mathematical brain? 284 00:21:17,440 --> 00:21:20,040 Do you find maths comes very naturally? 285 00:21:20,120 --> 00:21:24,440 Well, yes, I think it does, but I didn't read maths at Cambridge 286 00:21:24,440 --> 00:21:27,800 and maybe I should have it actually I have a younger 287 00:21:27,800 --> 00:21:33,440 brother who but actually eight years younger than me and he did 288 00:21:33,680 --> 00:21:39,520 read maths at Cambridge. And so anyway, I but yeah, I'm 289 00:21:40,400 --> 00:21:50,240 really interested in, well, for example, I mean for the Greeks, 290 00:21:50,720 --> 00:21:53,160 you know, people think about Pythagoras theorem. 291 00:21:53,960 --> 00:21:56,200 But the thing that's actually interesting to me is a 292 00:21:56,880 --> 00:22:00,160 consequence of Pythagoras theorem is that if you have a 293 00:22:00,160 --> 00:22:07,360 triangle with two sides of one, I mean right triangle, then the 294 00:22:07,360 --> 00:22:15,320 length of third side of sqrt 2 and now there's no fraction that 295 00:22:15,560 --> 00:22:17,520 when you square it can produce 2. 296 00:22:18,400 --> 00:22:21,800 So you can prove that however large the fraction is. 297 00:22:21,960 --> 00:22:28,440 And the Greeks already knew this and it really wasn't sorted out 298 00:22:28,440 --> 00:22:32,560 until the 19th century when it had a theory for real numbers 299 00:22:32,560 --> 00:22:36,480 developed. Another kind of interesting 300 00:22:36,480 --> 00:22:41,280 thing is that Euclid, he's known for his geometry, but he 301 00:22:41,280 --> 00:22:44,400 actually had a proof that there's no largest prime number. 302 00:22:47,160 --> 00:22:49,720 So they knew there was an infinite number of primes 303 00:22:49,720 --> 00:22:53,760 already, the Greeks. So this is not the kind of 304 00:22:53,760 --> 00:22:58,400 question that people designing aeroplanes really usually 305 00:22:58,400 --> 00:23:00,680 interested in. But I am interested in this kind 306 00:23:00,680 --> 00:23:07,840 of thing, and there's certainly, which basically the theory of 307 00:23:08,000 --> 00:23:13,280 what's progressively larger infinities. 308 00:23:13,440 --> 00:23:16,240 There's countable numbers and then there's the real numbers 309 00:23:16,240 --> 00:23:18,920 and so on. There's a theorem that the set 310 00:23:18,920 --> 00:23:26,120 of all subsets is larger than the original set, which means it 311 00:23:26,120 --> 00:23:30,240 could get a whole hierarchy of these transfinite numbers. 312 00:23:30,720 --> 00:23:37,560 So yeah, this kind of thing continuous to interest me, but I 313 00:23:37,560 --> 00:23:46,440 was also still interested in aircraft and I was very lucky in 314 00:23:46,440 --> 00:23:52,360 the following sense that in England in those days, well 315 00:23:52,360 --> 00:23:55,080 under the Labour government, they tried to make 316 00:23:56,640 --> 00:23:59,600 opportunities, firstly for people to get into universities. 317 00:23:59,600 --> 00:24:04,400 But secondly, if you got first class honours, your degree 318 00:24:04,400 --> 00:24:10,800 exams, you automatically got support from the Ministry of 319 00:24:10,800 --> 00:24:19,000 Education to do doctoral studies so that she would pay the full 320 00:24:19,000 --> 00:24:25,680 tuition and lodging to any university who was willing to 321 00:24:25,680 --> 00:24:27,920 accept you as a graduate student. 322 00:24:29,360 --> 00:24:32,920 So of course I was able to take advantage of that, but I didn't 323 00:24:32,920 --> 00:24:36,240 really want to to leave Cambridge. 324 00:24:36,240 --> 00:24:44,160 So I had to try to find somebody inside Cambridge who doing 325 00:24:44,160 --> 00:24:49,120 something that interested me. And I wound up finding Arthur 326 00:24:49,120 --> 00:24:55,520 Shurcliffe, who later became well known kind of the world, 327 00:24:55,520 --> 00:24:58,320 society and so on. But at the time I knew he was 328 00:24:58,320 --> 00:25:03,680 only 7 years older than me and had just been made in a lecture. 329 00:25:04,560 --> 00:25:07,600 He was doing Magneto hydrodynamics. 330 00:25:08,720 --> 00:25:12,480 So that's what I wound up doing in my thesis. 331 00:25:14,040 --> 00:25:24,120 But and then about one year into the thesis, I was lucky that I 332 00:25:24,120 --> 00:25:28,280 was in kind of invited to apply for what's called a research 333 00:25:28,280 --> 00:25:35,160 fellowship at Trinity Hall. So research fellows at Oxford or 334 00:25:35,160 --> 00:25:38,160 Cambridge are people who are actually are on the governing 335 00:25:38,160 --> 00:25:41,760 bodies of their colleagues. I mean, it's a sort of pseudo 336 00:25:41,760 --> 00:25:48,200 faculty kind of position. So I had three fellowship at 337 00:25:48,200 --> 00:25:53,880 Trinity Hall. Actually I had a slightly 338 00:25:53,880 --> 00:26:03,360 difficult start because I had been getting my grant from the 339 00:26:03,360 --> 00:26:09,200 Ministry of Education every three months and that was paid 340 00:26:09,200 --> 00:26:11,800 in advance. And then as a researcher it 341 00:26:11,800 --> 00:26:14,520 turned out that I was going to get paid quarterly, but in 342 00:26:14,520 --> 00:26:18,360 arrears. So I had like six months and I 343 00:26:18,360 --> 00:26:22,920 had to ask the advisor of the college to to give me a loan 344 00:26:22,920 --> 00:26:31,240 because I, I. So the other thing is that my 345 00:26:31,240 --> 00:26:39,920 advisor, Arthur Shurcliffe, so he went on an exchange to MIT my 346 00:26:39,920 --> 00:26:42,080 third year as a doctoral student. 347 00:26:43,440 --> 00:26:46,920 So at this point a student who was one year ahead of me became 348 00:26:46,920 --> 00:26:56,800 my nominal thesis advisor and I had actually did experimental 349 00:26:57,440 --> 00:27:01,520 thesis and tried to generate alphane waves in a liquid 350 00:27:01,520 --> 00:27:05,000 sodium. My thesis by the way you can 351 00:27:05,000 --> 00:27:09,760 find on my Stanford homepage of I wound up building A10 tonne 352 00:27:09,760 --> 00:27:15,840 magnet and things like that and we needed 300 amperes of 353 00:27:16,120 --> 00:27:20,320 current, 200 volts. And there was a motor generator 354 00:27:20,320 --> 00:27:25,280 set from a World War and submergine in the lobby of the 355 00:27:25,280 --> 00:27:27,960 engineering department. And I managed to have that moved 356 00:27:27,960 --> 00:27:31,120 into the lab and use that to supply the current for this 357 00:27:31,120 --> 00:27:39,040 magnet. I sort of after all this I 358 00:27:39,040 --> 00:27:42,120 realised firstly I did not think I should be doing experiments. 359 00:27:43,520 --> 00:27:47,160 I don't feel very uncomfortable trying to interpret sort of 360 00:27:47,720 --> 00:27:54,120 pictures and the sue escapes. I didn't actually think Magneto 361 00:27:54,120 --> 00:27:59,040 hydrodynamics is going anywhere in engineering, so I didn't 362 00:27:59,040 --> 00:28:09,760 really know what I wanted to do. And then I had some sort of 363 00:28:09,760 --> 00:28:17,680 interest in politics. I was quite upset when Macmillan 364 00:28:18,960 --> 00:28:23,680 stepped down as Prime Minister and engineered this Lord Hume 365 00:28:23,680 --> 00:28:29,240 becoming Prime Minister and I thought, well, maybe I should 366 00:28:29,240 --> 00:28:33,520 get involved. The trade student in Congress 367 00:28:33,520 --> 00:28:36,440 advertised at the time for people to work in the economics 368 00:28:36,480 --> 00:28:43,880 department, and I sent you the an application and they decided 369 00:28:43,880 --> 00:28:48,160 they would interview me. So I had about 3 weeks to read 370 00:28:48,160 --> 00:28:50,560 as much as I could about economics in a hurry. 371 00:28:50,560 --> 00:28:57,200 And then I had an interview with Len Murray, who at that time was 372 00:28:57,680 --> 00:29:00,400 head of the economics department but later became general 373 00:29:00,400 --> 00:29:04,560 secretary. He wound up hurrying me as an 374 00:29:04,560 --> 00:29:12,520 economist working for the Trade Union Congress, and it was quite 375 00:29:12,520 --> 00:29:16,320 interesting actually. They had what they called 376 00:29:16,600 --> 00:29:22,520 National Economic Development Council Nettie's, which were 1/3 377 00:29:22,520 --> 00:29:28,320 government, 1/3 management from industry and 1/3 union reps was 378 00:29:28,320 --> 00:29:31,880 supposed to be discussing how to improve things, different 379 00:29:31,880 --> 00:29:34,920 industries. So I was actually preparing the 380 00:29:34,920 --> 00:29:39,480 position for the trades unions on the aircraft industry and the 381 00:29:39,480 --> 00:29:48,880 electrical industry, but I knew that I wasn't going to be long 382 00:29:48,880 --> 00:29:57,320 term a trade union person. I found it, it was interesting 383 00:29:57,320 --> 00:30:03,160 and actually I did a kind of study of what happened to rates 384 00:30:03,160 --> 00:30:08,840 of inflation versus unemployment and going back to about 1900 and 385 00:30:09,640 --> 00:30:16,280 I found that interesting enough the fastest rate of inflation in 386 00:30:16,280 --> 00:30:19,560 England was usually on the downside of each economic cycle. 387 00:30:19,560 --> 00:30:28,280 And that's because as far as I can make out costs went up on 388 00:30:28,280 --> 00:30:31,240 the downside because there was less utilisation of the 389 00:30:31,560 --> 00:30:35,480 factories and so on and companies has passed the higher 390 00:30:35,480 --> 00:30:41,960 cost straight on high prices. I've certainly concluded that 391 00:30:41,960 --> 00:30:44,760 this idea that there's a sort of non inflationary rate of 392 00:30:44,760 --> 00:30:51,600 unemployment is is wrong and actually morally wrong as well 393 00:30:51,600 --> 00:30:53,240 as practically wrong in my opinion. 394 00:30:53,280 --> 00:30:59,360 I mean you deliberately a high rate of unemployment in order to 395 00:30:59,360 --> 00:31:02,920 prevent inflation is actually sort of a bit like human 396 00:31:02,920 --> 00:31:08,320 sacrifice used to be a sort of be historic good. 397 00:31:08,320 --> 00:31:11,840 You know you sacrifice somebody for the benefit of everybody 398 00:31:11,840 --> 00:31:17,360 else right. But anyway, quite by chance 399 00:31:19,760 --> 00:31:22,760 there was an advancement in one of the London Evening newspapers 400 00:31:22,760 --> 00:31:27,920 saying that Orcas Sidley was dynamics was trying to looking 401 00:31:27,920 --> 00:31:32,520 for engineers and they would have some representatives of the 402 00:31:32,520 --> 00:31:37,080 company in a hotel. Happened to me about 5 minutes 403 00:31:37,080 --> 00:31:38,800 walk from the Trades Union Congress. 404 00:31:41,120 --> 00:31:46,480 So I decided I might as well walk in there around about 5 in 405 00:31:46,480 --> 00:31:50,960 the evening at the end of the day and had a very weird 406 00:31:50,960 --> 00:31:54,640 interview. So there were two people in the 407 00:31:56,080 --> 00:31:57,920 room. Marathon said what do you think 408 00:31:57,920 --> 00:31:59,480 I can do? I said I think I can do 409 00:31:59,480 --> 00:32:02,800 aerodynamics. The other guy said well what are 410 00:32:02,800 --> 00:32:04,520 you doing now? I said, well, I'm working as an 411 00:32:04,520 --> 00:32:07,720 economist. So then there was complete 412 00:32:07,720 --> 00:32:11,880 silence for about two minutes. Then one of them said, well, 413 00:32:11,880 --> 00:32:13,840 what makes you think you can do aerodynamics? 414 00:32:13,840 --> 00:32:17,080 I said, well, I have a PhD in Magneto hydrodynamics and I 415 00:32:17,080 --> 00:32:20,160 think it would be relatively easy for me to do aerodynamics. 416 00:32:21,440 --> 00:32:23,080 So they thought about that for a bit. 417 00:32:24,800 --> 00:32:28,000 They finally just said, well, would am I interested in coming 418 00:32:28,000 --> 00:32:35,440 to the factory, which is in Warwick basically exactly what 419 00:32:35,440 --> 00:32:37,440 used to be the Armstrong Whitworth company during the 420 00:32:37,440 --> 00:32:39,960 Second World War made some very bad airplanes. 421 00:32:41,080 --> 00:32:44,320 There was a bomber called the Whitney sort of who knows down. 422 00:32:45,360 --> 00:32:49,360 But anyway, I said, well, it would depend on the salary. 423 00:32:49,360 --> 00:32:51,200 They said they didn't know what the salary would be. 424 00:32:51,240 --> 00:32:53,920 I said, in that case I don't, I don't know whether I'm 425 00:32:54,600 --> 00:32:56,880 interested in coming for an interview and I, I didn't 426 00:32:56,880 --> 00:33:02,200 certainly hear from them again, but but suddenly about a month 427 00:33:02,200 --> 00:33:04,960 later they, they wrote to me and asked me if they would go up 428 00:33:04,960 --> 00:33:09,560 there. So I found you travelled up for 429 00:33:09,560 --> 00:33:15,680 about a couple more interviews and the outcome was that they 430 00:33:15,680 --> 00:33:21,280 decided to offer me a position as chief mathematician of this 431 00:33:21,280 --> 00:33:23,280 Hawkeye City dynamics in Coventry. 432 00:33:25,560 --> 00:33:29,040 So actually not reporting to the head of aerodynamics, but on a 433 00:33:29,040 --> 00:33:36,000 level or even above him. So I thought, well, I got to try 434 00:33:36,000 --> 00:33:40,360 this. So I wound up in Hawkins City 435 00:33:40,360 --> 00:33:40,960 Dynamics. And. 436 00:33:41,080 --> 00:33:49,480 And how old were you then? Well, that was 1964, I guess. 437 00:33:49,480 --> 00:33:52,880 So I was just coming up to SO. You were 3030. 438 00:33:53,400 --> 00:33:56,120 Just approaching that, yeah. Yeah. 439 00:33:57,960 --> 00:34:03,000 So actually incredible. When I got to Orchestilla 440 00:34:03,000 --> 00:34:05,240 Dynamics, the first thing that happened was that the guy who 441 00:34:05,240 --> 00:34:11,360 had kind of initiated hiring me, head of aerodynamics, a guy 442 00:34:11,360 --> 00:34:14,719 called Malcolm James, told me that he decided to leave and was 443 00:34:14,719 --> 00:34:19,600 going to work for Douglas Aircraft in California. 444 00:34:20,920 --> 00:34:24,159 So I had expected that I would collaborate with him, and that 445 00:34:24,239 --> 00:34:26,320 didn't happen. You see it only overlapped a 446 00:34:26,320 --> 00:34:33,480 couple of weeks, but I did, I did quite a lot of antenna 447 00:34:33,480 --> 00:34:36,960 calculations for example. And again, if you go on my 448 00:34:36,960 --> 00:34:40,520 homepage, you can find some Hawker Sidley reports, which 449 00:34:40,880 --> 00:34:42,880 kind of thing there was during at the time. 450 00:34:45,239 --> 00:34:55,239 And well, what I concluded was after I'd been there about a 451 00:34:55,239 --> 00:34:57,800 year was that they were never actually going to build another 452 00:34:57,800 --> 00:34:59,880 missile. They were making anti aircraft 453 00:34:59,880 --> 00:35:03,720 missiles and they had us, it's called the C dot. 454 00:35:05,560 --> 00:35:08,920 And I didn't think we were ever going to build anything. 455 00:35:09,480 --> 00:35:14,080 So I finally decided, well I've either got to get out of the 456 00:35:14,080 --> 00:35:17,240 aircraft industry or I have to come to America, one or the 457 00:35:17,240 --> 00:35:21,160 other because British aircraft industry didn't seem to be going 458 00:35:21,160 --> 00:35:24,600 to build anything anyway. Every project got cancelled by 459 00:35:24,600 --> 00:35:27,320 the government, sometimes even when they got to the point of 460 00:35:27,320 --> 00:35:35,920 flying. So finally I had some interview 461 00:35:35,920 --> 00:35:38,840 actually, I think with the central that's just a generating 462 00:35:38,840 --> 00:35:42,280 board and for some teaching positions somewhere. 463 00:35:42,280 --> 00:35:46,040 I'm so lucky I was turned down by both of them. 464 00:35:47,360 --> 00:35:55,640 But, and then I went on this, I can't quite remember, but at 465 00:35:56,960 --> 00:36:01,640 that time Boeing, Douglas and so on were recruiting quite heavily 466 00:36:01,640 --> 00:36:08,560 in in England. Well, I knew that they had teams 467 00:36:08,880 --> 00:36:12,120 in England and I wasn't interviewed by Helberg or 468 00:36:12,240 --> 00:36:15,360 Douglas. And I probably thought I would 469 00:36:15,360 --> 00:36:19,800 probably rule out because they wouldn't like my trade union 470 00:36:19,800 --> 00:36:29,320 background. But only Drummond, who I didn't 471 00:36:29,320 --> 00:36:33,480 know a great deal about, but they made these sort of fatter 472 00:36:33,480 --> 00:36:35,080 plants during the Second World War. 473 00:36:37,120 --> 00:36:40,080 They said that they were interviewing in London and they 474 00:36:40,080 --> 00:36:43,640 in fact decided to invite me to an interview. 475 00:36:46,040 --> 00:36:51,440 So I wound up hitting it off rather well with the guy who 476 00:36:51,440 --> 00:36:57,160 interviewed me, and he said, well, would I like to be in what 477 00:36:57,160 --> 00:36:58,840 department? I said, well, the research 478 00:36:58,840 --> 00:37:00,960 department at Grumman. He said, well, he didn't 479 00:37:00,960 --> 00:37:03,040 represent the research department, but he could 480 00:37:03,040 --> 00:37:09,720 recommend me to in one of the aerodynamics or thermodynamics, 481 00:37:10,080 --> 00:37:13,840 this kind of thing. So I finally accepted an offer 482 00:37:15,640 --> 00:37:19,440 to work in the aerodynamics department of Grumman or Long 483 00:37:19,440 --> 00:37:24,880 Island. That didn't start very smoothly 484 00:37:26,120 --> 00:37:31,160 the day I got there. Well, firstly they had a kind of 485 00:37:33,240 --> 00:37:37,120 outside the main Grumman sort of campus. 486 00:37:37,160 --> 00:37:41,280 They had this plant where people who hadn't got security 487 00:37:41,280 --> 00:37:48,360 clearances were parked. So I went, I reported to this 488 00:37:49,200 --> 00:37:50,720 department. There was an Englishman in 489 00:37:50,720 --> 00:37:53,480 charge who said he was very sorry, but they didn't have any 490 00:37:54,360 --> 00:37:58,920 free desk for me to sit at. I mean, they had about 150 491 00:37:58,920 --> 00:38:02,200 people in there sitting side by side, but there was no free 492 00:38:02,200 --> 00:38:06,240 desk. So then they said, well, the 493 00:38:06,240 --> 00:38:11,120 best thing to do is we'll send you to to be interviewed by the 494 00:38:11,120 --> 00:38:14,920 head of the aero department. I mean, I had to be escorted by 495 00:38:15,560 --> 00:38:20,040 a copy. So I get to see this head of the 496 00:38:20,040 --> 00:38:23,320 aero department and it was Bill Murphy. 497 00:38:23,360 --> 00:38:27,240 And his opening words were that it was been a mistake to harm 498 00:38:27,240 --> 00:38:31,440 him, that he didn't have any need for anybody with a PhD, 499 00:38:31,440 --> 00:38:34,840 that he had a summoned with a PhD there before who was 500 00:38:34,840 --> 00:38:39,680 arrogant and lazy, had no need for me. 501 00:38:40,960 --> 00:38:44,120 So this didn't seem like a very good start. 502 00:38:45,960 --> 00:38:50,320 But then he handed me off with one of his group leaders, a guy 503 00:38:50,320 --> 00:38:56,320 called Rudy Mayo, who So they had control systems inside the 504 00:38:56,320 --> 00:39:00,440 aerodynamics department, because you have to be like a fighter. 505 00:39:00,440 --> 00:39:05,200 Then the aerodynamic stability depends a lot on things like 506 00:39:05,200 --> 00:39:09,840 dihedral and so on. So it's not just black boxes. 507 00:39:11,160 --> 00:39:18,040 So Rudy Meyer, I met him and he turned out to be an extremely 508 00:39:18,040 --> 00:39:20,880 nice guy who was actually a veteran of the Korean War. 509 00:39:22,720 --> 00:39:26,160 And he said he thought we have seen to look at Horowitz's 510 00:39:26,160 --> 00:39:31,480 method of discerning control systems or stability 511 00:39:31,480 --> 00:39:34,800 augmentation systems that I, I never heard of Horowitz. 512 00:39:35,480 --> 00:39:37,120 So I sort of thought about it a bit. 513 00:39:37,120 --> 00:39:41,360 I counted, I said, look, I think it's very interesting to look at 514 00:39:42,400 --> 00:39:48,600 issue of control systems, but why don't you let me go away and 515 00:39:48,840 --> 00:39:52,080 do a little bit of research for a few days and I'll come back 516 00:39:52,080 --> 00:39:55,840 and make a recommendation. So he he agreed to that. 517 00:39:57,520 --> 00:40:02,200 Now I'm not quite sure how I managed to find the literature, 518 00:40:02,200 --> 00:40:04,040 but I mean, we haven't had quite a good library. 519 00:40:04,040 --> 00:40:05,720 I suppose they may have allowed me into it. 520 00:40:05,920 --> 00:40:10,800 But anyway, when I met Rudy the following week, I suggested that 521 00:40:10,800 --> 00:40:13,960 we try the new optimal control theory. 522 00:40:13,960 --> 00:40:19,440 At that time was quite new. Pioneered by Kalman in 523 00:40:19,440 --> 00:40:25,160 particular with favours from the Kalman filter and really went 524 00:40:25,160 --> 00:40:31,880 along with that. So I started looking at how you 525 00:40:31,880 --> 00:40:35,600 would apply linear optimal control theory to designing 526 00:40:35,760 --> 00:40:42,240 control systems for fighter aircraft and such like or actual 527 00:40:42,240 --> 00:40:46,120 fact. Also, Graumman was building this 528 00:40:46,400 --> 00:40:50,280 electronic warfare plane, the E2 area, the big radome on top of 529 00:40:50,280 --> 00:40:54,640 it. And so that had really awkward 530 00:40:56,280 --> 00:40:58,200 stability characteristics you could have had. 531 00:40:58,400 --> 00:41:02,600 So the question was, could you design something that would make 532 00:41:02,600 --> 00:41:08,160 it flyable? And what I actually got into was 533 00:41:08,160 --> 00:41:12,440 that you don't necessarily optimal control requires you to 534 00:41:12,440 --> 00:41:15,600 measure everything. I mean, it's fairly obvious 535 00:41:16,480 --> 00:41:21,080 because you can't really have an optimal control if you can't 536 00:41:21,080 --> 00:41:25,320 predict the trajectory. So if you express the system of 537 00:41:25,320 --> 00:41:28,520 the 1st order system of differential regulations, you 538 00:41:28,520 --> 00:41:31,840 need to know the initial conditions for each variable. 539 00:41:33,720 --> 00:41:36,160 In other words, if it's a feedback system, you need to 540 00:41:36,160 --> 00:41:39,240 measure everything the full state. 541 00:41:41,080 --> 00:41:46,080 But actually to make a workable control system, you don't 542 00:41:46,080 --> 00:41:50,240 necessarily need that. So you can try to make a what 543 00:41:50,240 --> 00:41:54,120 would be in some sense a suboptimal system nowadays, a 544 00:41:54,120 --> 00:41:58,400 quarterback so-called sparse control, where you restrict the 545 00:41:58,400 --> 00:42:04,360 number of permitted feedbacks. And actually for lateral control 546 00:42:04,360 --> 00:42:09,000 of an airplane, for example, you don't really want to measure the 547 00:42:09,000 --> 00:42:13,640 rail angle because otherwise if you rail 360°, the control 548 00:42:13,640 --> 00:42:15,480 system will try to the whole way back. 549 00:42:15,480 --> 00:42:23,360 It doesn't recognise it. All right, so nowadays, you 550 00:42:23,360 --> 00:42:26,360 know, quite a mathematical theory for sparse control, but I 551 00:42:26,360 --> 00:42:30,320 wound up actually publishing about 15 journal articles and 552 00:42:30,320 --> 00:42:33,720 control theory over the next few years. 553 00:42:37,520 --> 00:42:44,360 Some other strikes of luck at this point because so there was 554 00:42:44,360 --> 00:42:50,360 a a scientist called John Dee Young who was an ex NASA guy who 555 00:42:50,520 --> 00:42:59,280 was developing a theory for predicting very high lift 556 00:42:59,280 --> 00:43:01,800 aerodynamics, what we call V store aircraft. 557 00:43:01,800 --> 00:43:06,480 It was a thing like the Brigade 941 that has his big propellers 558 00:43:06,720 --> 00:43:10,080 and then it has flaps to try to deflect the propeller slipstream 559 00:43:10,080 --> 00:43:12,560 to take off in a very short distance. 560 00:43:13,920 --> 00:43:18,040 So John Deyoung had got a contract from Nasser Ames to 561 00:43:18,040 --> 00:43:23,880 develop this theory, and then he was married to a Brazilian woman 562 00:43:23,880 --> 00:43:25,680 who wasn't very happy, apparently long. 563 00:43:25,680 --> 00:43:29,640 He asked for a leave of absence for three months and then he 564 00:43:29,640 --> 00:43:35,680 didn't come back. So Brahman got out message from 565 00:43:35,680 --> 00:43:38,080 Nasr Ames saying what's happened to the contract? 566 00:43:38,080 --> 00:43:41,680 The Pi appears to be now working for Ling Temco Fort, which the 567 00:43:41,680 --> 00:43:46,040 government didn't know. So now there's a crisis. 568 00:43:47,880 --> 00:43:51,080 The head of aerodynamics said, well, we should just give up on 569 00:43:51,080 --> 00:43:53,280 the contract. And he was overruled by the vice 570 00:43:53,280 --> 00:43:56,880 president of engineering. We said we can't tell NASA that 571 00:43:56,880 --> 00:43:58,880 there's nobody else is able to do it here. 572 00:44:00,640 --> 00:44:02,880 At this point. It turned out that having 573 00:44:02,880 --> 00:44:05,880 someone with a PhD was a little bit of an advantage that they 574 00:44:05,880 --> 00:44:08,200 asked me if I was willing to take over the contract. 575 00:44:11,360 --> 00:44:14,960 I said, well, I don't really want to, but we need to, as long 576 00:44:14,960 --> 00:44:18,000 as I'm also allowed to continue doing some work on this control 577 00:44:18,000 --> 00:44:21,400 system theory. So they agreed to that. 578 00:44:21,400 --> 00:44:27,200 But now when I looked at the youngster, I thought it was 579 00:44:27,200 --> 00:44:31,440 actually not quite right. But I came up with fixes and 580 00:44:31,440 --> 00:44:38,760 kind of convinced the, I said, well, we have to write a 581 00:44:38,760 --> 00:44:42,520 computer, a curve also. And the first reaction was that 582 00:44:42,520 --> 00:44:45,560 that'll cost you much. And I went out to NASA and said, 583 00:44:45,560 --> 00:44:48,760 no, I can write this in within the cost of this contract. 584 00:44:50,160 --> 00:44:56,240 So that meant that at this point I felt that I was being 585 00:44:56,240 --> 00:45:01,000 underpaid by government because I knew that people with my level 586 00:45:01,000 --> 00:45:03,880 of experience were getting quite a bit more by just by reading 587 00:45:03,880 --> 00:45:06,720 the job advertisements that used to come out in the Sunday 588 00:45:06,720 --> 00:45:13,160 newspapers. I wound up having an interview 589 00:45:13,160 --> 00:45:16,880 with an outfit called General Research based in Santa Barbara, 590 00:45:16,880 --> 00:45:20,560 who were trying to develop a anti ballistic missile system. 591 00:45:22,040 --> 00:45:25,360 They offered me a job and my first reaction was to say yes. 592 00:45:26,680 --> 00:45:34,000 So I gave government notice on a Friday and on Monday I was 593 00:45:34,000 --> 00:45:37,200 called in before the VP of engineering who said that he was 594 00:45:37,200 --> 00:45:41,400 retrospectively raising my salary by 30% from the beginning 595 00:45:41,400 --> 00:45:42,840 of the year. Would I reconsider? 596 00:45:47,440 --> 00:45:52,080 So I said yes, but I don't really want to dust your 597 00:45:52,080 --> 00:45:53,920 dynamics. I mean, I want to look at a 598 00:45:53,920 --> 00:45:57,920 broader picture. And finally, they created a 599 00:45:57,920 --> 00:46:01,280 staff position for me where I reported to the manager of 600 00:46:01,360 --> 00:46:03,640 Flight Sciences, a man called Stu Harvey. 601 00:46:08,440 --> 00:46:12,760 Stu first asked me if I could try to design AA wind tunnel 602 00:46:12,760 --> 00:46:17,080 that Common might build. I looked at this for a few 603 00:46:17,080 --> 00:46:19,040 months. I don't really believe that 604 00:46:19,040 --> 00:46:20,960 Common is ever going to go anywhere with it. 605 00:46:21,840 --> 00:46:30,000 Well, then all of a sudden, and this is in I guess like November 606 00:46:30,000 --> 00:46:34,400 1969, Stu Harvey called in and says that we've got to do 607 00:46:34,440 --> 00:46:36,280 something about supercritical wings. 608 00:46:37,440 --> 00:46:39,760 And I said, well, what do you mean by supercritical wing? 609 00:46:39,760 --> 00:46:44,240 And he said, well, the wing that works well in transonic flow. 610 00:46:46,320 --> 00:46:49,680 And Wickham, at this point, Archie Wickham and NASA Langley 611 00:46:49,680 --> 00:46:53,520 had made a big splash with what he called supercritical airfoils 612 00:46:53,520 --> 00:46:58,560 that managed to reduce the shock drag transonic. 613 00:47:00,640 --> 00:47:05,960 And Wickham is experimentalist who just sort of used a file to 614 00:47:05,960 --> 00:47:08,880 change the shapes and so on. And Stu Howie says, well, we 615 00:47:08,880 --> 00:47:14,520 can't depend on Wickham's file. So I said to Stu, well look, 616 00:47:14,520 --> 00:47:17,920 this is actually very interesting. 617 00:47:21,520 --> 00:47:26,560 I said there's a aerospace sciences meeting being held in 618 00:47:26,560 --> 00:47:29,920 New York in January, and maybe I should. 619 00:47:30,520 --> 00:47:34,920 After all, Grumman's 40 miles from New York City. 620 00:47:34,920 --> 00:47:37,720 And I said maybe I should go to this and try to find out what's 621 00:47:37,720 --> 00:47:41,560 being done in transonics. And I went to this meeting and 622 00:47:43,200 --> 00:47:49,040 the whole moment presented what became quite a famous paper on 623 00:47:49,160 --> 00:47:52,560 calculating a solution to the transonic small disturbance 624 00:47:52,560 --> 00:47:59,280 equation. And The thing is I knew very 625 00:47:59,280 --> 00:48:06,240 well that transonic flow cannot be treated by linearising 626 00:48:06,240 --> 00:48:10,800 anything. So you have linearised subsonic 627 00:48:10,800 --> 00:48:13,680 equations called Prandtl Kraut equation or you have linearised 628 00:48:13,680 --> 00:48:17,600 supersonic flow. But those what happens in 629 00:48:17,600 --> 00:48:21,240 transonic flow is it it's actually an elliptic equation in 630 00:48:21,240 --> 00:48:24,520 the subsonic part of the Fairfield and hyperbolic and 631 00:48:24,560 --> 00:48:28,560 supersonic pattern. You can indicate that change in 632 00:48:28,560 --> 00:48:31,080 type and cheap hold on inner equations. 633 00:48:32,920 --> 00:48:36,960 And there were books, for example by Otsochich and Ben 634 00:48:36,960 --> 00:48:42,000 Bersen, attempts at analytic solutions of transonic flow 635 00:48:42,000 --> 00:48:47,280 basically gone no way. So I knew that you had to do 636 00:48:47,280 --> 00:48:53,760 this by some kind of numerical calculation, and I also saw that 637 00:48:53,800 --> 00:48:56,960 computers are getting just fast enough that you actually 638 00:48:56,960 --> 00:49:01,240 probably could do some useful calculations. 639 00:49:03,040 --> 00:49:11,040 So that was kind of my chance to try to solve the problem. 640 00:49:11,040 --> 00:49:15,320 This was a pivotal moment. Just at the right time, down in 641 00:49:15,320 --> 00:49:20,240 the right it. Seems like you had, but you had 642 00:49:20,240 --> 00:49:23,400 a build up. I, I, I assumed that from an 643 00:49:23,440 --> 00:49:26,880 early stage of your career that you were working on transonic 644 00:49:26,880 --> 00:49:33,400 flow, but you almost had 36 years of build up to this point 645 00:49:34,000 --> 00:49:38,880 to then be given the opportunity to, to focus on this. 646 00:49:39,640 --> 00:49:42,560 Yeah, well, that's is it. No, I hadn't actually worked in 647 00:49:42,920 --> 00:49:47,320 Sonic at all, and I knew very little about it, and I didn't 648 00:49:48,760 --> 00:49:51,320 nothing much about numerical methods. 649 00:49:51,320 --> 00:49:56,840 I'd never taken a course in numerical methods, but I had 650 00:49:56,920 --> 00:50:00,040 learned how to solve partial differential equations as part 651 00:50:00,040 --> 00:50:03,920 of my thesis, because when I did this semi magical hydrodynamics, 652 00:50:03,920 --> 00:50:08,800 I had to solve this. I solved the equations for these 653 00:50:08,960 --> 00:50:14,320 alphane waves by separation of variables and expansions and 654 00:50:14,320 --> 00:50:15,960 Bessel functions and things like that. 655 00:50:15,960 --> 00:50:18,440 So I had some background in Pdes. 656 00:50:19,680 --> 00:50:22,400 I've never taken a course in Pdes either. 657 00:50:24,200 --> 00:50:29,560 But now that I realised that you could do something numerically, 658 00:50:31,440 --> 00:50:40,440 turned out that nobody, the head of the aero department, grandma, 659 00:50:40,440 --> 00:50:42,560 wasn't willing to put anybody on to this. 660 00:50:42,560 --> 00:50:47,880 I finally decided I'd better try to do it myself, so I started 661 00:50:47,880 --> 00:50:56,120 trying to write computer programmes for fluid flow and I 662 00:50:56,120 --> 00:50:59,120 didn't. I knew I had to learn more about 663 00:50:59,120 --> 00:51:01,400 numerical methods. So I managed to get hold of a 664 00:51:01,400 --> 00:51:06,480 book by Rick Meyer and Morton called Find out Different 665 00:51:06,480 --> 00:51:11,240 Methods for initial value Problems and that was very hard 666 00:51:11,240 --> 00:51:14,480 for me. Talked things like Banach spaces 667 00:51:14,480 --> 00:51:17,920 and so on. I never heard of, but I fought 668 00:51:17,920 --> 00:51:24,760 my way through this book. And then what we needed to do 669 00:51:24,760 --> 00:51:27,280 was not to solve the small disturbance equation because 670 00:51:27,280 --> 00:51:31,440 they're not small disturbances. When you have a supercritical 671 00:51:31,440 --> 00:51:37,120 airfoil will win. So the best model was to use the 672 00:51:37,120 --> 00:51:40,520 what's called the full transonic potential flare equation. 673 00:51:42,640 --> 00:51:46,320 I ruled out trying to solve the Euler equations because you 674 00:51:46,320 --> 00:51:50,000 didn't have enough memory on the computer at the time. 675 00:51:51,280 --> 00:51:55,120 Do a 3D, see the potential flow solution. 676 00:51:56,400 --> 00:52:00,200 You just got to store 1 variable instead of like 5 variables if 677 00:52:00,200 --> 00:52:01,640 you're solving the Euler equations. 678 00:52:01,640 --> 00:52:09,680 And so I could see that we could actually write a repair game 679 00:52:09,720 --> 00:52:14,720 that would work. And in the meanwhile also I had 680 00:52:14,720 --> 00:52:18,840 discovered that there was a group at the Quran Institute led 681 00:52:18,840 --> 00:52:22,240 by Paul Garabedian and I went in to meet them. 682 00:52:24,840 --> 00:52:30,560 Paul Garabedian had a student called David Corn who they had 683 00:52:30,560 --> 00:52:34,320 developed a method of calculating so-called shock free 684 00:52:34,320 --> 00:52:40,280 airfoils in transonic flow by transformation to the photograph 685 00:52:40,280 --> 00:52:44,160 plane and then a method of solution so-called complex 686 00:52:44,160 --> 00:52:48,360 characteristics. It is new to Bulgaria, but I got 687 00:52:48,360 --> 00:52:54,840 to know them and I thought that we could solve the full 688 00:52:54,840 --> 00:52:58,760 transformation flow equation, but you would need to use some 689 00:52:58,760 --> 00:53:03,920 kind of conformal mapping so as to map the flow outside the 690 00:53:03,920 --> 00:53:06,520 airfoil to the flow inside a unit disk. 691 00:53:08,520 --> 00:53:11,160 Well, that's I had the background and conformal mapping 692 00:53:11,280 --> 00:53:16,320 way back to the Clark book. So I wrote a book Codes to solve 693 00:53:18,080 --> 00:53:22,800 absolutely incompressible Invisible Flow by conformal 694 00:53:22,800 --> 00:53:27,000 mapping with. My first step was flow one, but 695 00:53:27,000 --> 00:53:32,320 I envisaged a sequence of codes that would become progressively 696 00:53:33,880 --> 00:53:37,040 more complicated, more complicated geometries go to 3D 697 00:53:37,040 --> 00:53:40,640 and so on, and eventually go to solving the Euler equation. 698 00:53:40,640 --> 00:53:46,360 So I in Fortune at the time you were not allowed more than six 699 00:53:46,360 --> 00:53:49,520 characters in a variable name or a programme name. 700 00:53:49,520 --> 00:53:54,560 So I truncated FLOW to Flo to allow for three digits. 701 00:53:59,040 --> 00:54:07,040 And after, about, about sometime during 1971 or after, I managed 702 00:54:07,040 --> 00:54:12,680 to get a flow solver for transonic potential flow 703 00:54:12,680 --> 00:54:20,320 working. And at this point, Paul Garabin, 704 00:54:20,800 --> 00:54:24,720 I had persuaded grandma to hire Corn as a consultant. 705 00:54:24,720 --> 00:54:28,120 So we had some contacts and they'd also been writing an 706 00:54:28,840 --> 00:54:31,920 analysis code. And when they found out that I'd 707 00:54:31,920 --> 00:54:33,680 read mine, that appeared to work as well. 708 00:54:35,080 --> 00:54:41,320 Paul asked me if I would spend 3 months at the Quran Institute as 709 00:54:41,360 --> 00:54:46,640 a visitor and I thought I said yes. 710 00:54:46,640 --> 00:54:51,560 I got, I, I realised that I'm not going to be able to do this 711 00:54:51,560 --> 00:54:56,880 in the long run at Grameen. So I had to be considered moving 712 00:54:56,880 --> 00:55:01,680 into academia somehow. And so I went there for three 713 00:55:01,680 --> 00:55:05,440 months and then Paul kind of tested me in various ways during 714 00:55:05,440 --> 00:55:08,040 that time. And then he said, well, why 715 00:55:08,040 --> 00:55:10,040 don't you stay here? And he said, well, I can't just 716 00:55:10,040 --> 00:55:12,680 disappear from grammar and I have to go back and give them 717 00:55:12,680 --> 00:55:18,040 notice. But I agreed and he had me as a 718 00:55:18,920 --> 00:55:24,360 senior research scientist. That's in 1972. 719 00:55:25,440 --> 00:55:30,400 And he told me that there's no faculty positions going to be 720 00:55:30,400 --> 00:55:33,000 available at the Quran Institute, but if you make us 721 00:55:33,000 --> 00:55:35,200 flat, you might get an offer from somewhere else. 722 00:55:37,640 --> 00:55:40,320 So I said, OK, I'll settle for that. 723 00:55:41,840 --> 00:55:47,320 So when I got to the Quran Institute started working for 724 00:55:47,320 --> 00:55:51,800 Paul who he agreed that I would try to write a 3D solver for the 725 00:55:52,000 --> 00:55:55,920 transonic potential flow, but he wanted it to be for a yard 726 00:55:55,920 --> 00:55:59,600 flying wing because he had a contract with Archie Jones at 727 00:55:59,600 --> 00:56:03,920 Nassau Ames. Archie Jones is advocating that 728 00:56:05,160 --> 00:56:11,840 supersonic transport aircraft to adjust a flying wing which would 729 00:56:12,000 --> 00:56:14,800 fly at a skew angle, so-called yard flying wing. 730 00:56:16,560 --> 00:56:19,920 Of course, that never actually happened, but but the meanwhile 731 00:56:19,920 --> 00:56:28,520 also had a contract with NASA Langley because Whitcomb also 732 00:56:28,520 --> 00:56:31,440 understood that if you could calculate transonic flows, it 733 00:56:31,440 --> 00:56:34,120 might help him to obtain better supercritical airfoils. 734 00:56:35,120 --> 00:56:40,400 So I wound up being the point of contact of both Archie Jones and 735 00:56:40,400 --> 00:56:43,360 Nasser Ames and Archie Whitcomb and Langley. 736 00:56:45,640 --> 00:56:49,800 And both of them sort of encouraged me. 737 00:56:49,800 --> 00:56:51,960 They, they said move, Jameson move. 738 00:56:52,760 --> 00:56:55,280 But that's so that maybe I could take this further. 739 00:56:56,680 --> 00:57:04,960 So then another thing that happened was that I'd only been 740 00:57:04,960 --> 00:57:09,040 at the Quran Institute a week when they they had a they 741 00:57:09,040 --> 00:57:12,680 didn't, they didn't take teaching very seriously there. 742 00:57:12,680 --> 00:57:17,240 But what they did do is they teach a lot of courses in the 743 00:57:17,240 --> 00:57:20,600 evenings because they were accommodating people who were 744 00:57:20,600 --> 00:57:23,640 working in the city who wanted to study nights. 745 00:57:25,120 --> 00:57:30,280 Turned out that the graduate numerical analysis course was 746 00:57:30,280 --> 00:57:34,800 normally taught by Isaacson, who's a co-author of the book 747 00:57:34,800 --> 00:57:36,760 Isaacson and Keller, which is quite famous. 748 00:57:36,760 --> 00:57:40,320 And Isaacson was on leave and they hadn't got an instructor. 749 00:57:41,160 --> 00:57:44,320 So they asked me if I would like to try to teach it. 750 00:57:46,840 --> 00:57:56,560 So I said yes, I can try. That was that in some sense is a 751 00:57:56,680 --> 00:58:00,120 circle that for me, because I didn't know really numerical 752 00:58:00,120 --> 00:58:03,440 methods other than the book I mentioned I'd read by Morton 753 00:58:03,440 --> 00:58:07,160 Rickman, but I had to teach some Isaacs and in Keller it really 754 00:58:07,160 --> 00:58:09,360 mean I had to read the book ahead of the class. 755 00:58:12,200 --> 00:58:14,720 But the standby time I've done that, I mean, they're pretty 756 00:58:15,360 --> 00:58:17,560 thorough knowledge of numerical methods. 757 00:58:20,960 --> 00:58:29,440 So then after two years, the maths department at UCLA was 758 00:58:29,440 --> 00:58:31,880 interested. I got an approach and well, they 759 00:58:31,880 --> 00:58:34,480 turned out that they were really bored in trying to hire Paul 760 00:58:34,480 --> 00:58:37,480 Garabinia, but it was the kind of suggestion that they would 761 00:58:37,480 --> 00:58:45,120 hire both of us and they offered me full professorship at UCLA, 762 00:58:45,120 --> 00:58:47,960 but never in writing. But that's what was on the table 763 00:58:49,720 --> 00:58:55,680 at this point. Quran Institute decided to 764 00:58:55,680 --> 00:59:00,400 respond by eventually the first offering me an associate 765 00:59:00,400 --> 00:59:02,160 professor. I said no, I'll take the full 766 00:59:02,160 --> 00:59:04,920 professor at UCLA. So then they did offer me a full 767 00:59:05,000 --> 00:59:08,640 professor said turn it kind of just like that. 768 00:59:08,640 --> 00:59:11,440 So I've never been in a an assistant professor or an 769 00:59:11,800 --> 00:59:14,240 associate professor went straight from the yes. 770 00:59:14,280 --> 00:59:22,120 Straight to the top. So after that I was no longer 771 00:59:22,120 --> 00:59:24,760 working directly for Paul Garrabedian, but of course he 772 00:59:24,760 --> 00:59:30,160 remained in close contact and I went on to he'd. 773 00:59:30,160 --> 00:59:35,320 Actually another thing that happened was just before I got 774 00:59:35,320 --> 00:59:43,200 that offer, I had written this Flow 17 which did a 775 00:59:43,200 --> 00:59:46,200 three-dimensional flow plasters yard wing. 776 00:59:47,600 --> 00:59:50,560 But I knew perfectly well that Craft engine was never going to 777 00:59:50,560 --> 00:59:54,000 build a yard ring that they needed was a swept ring code. 778 00:59:56,080 --> 01:00:01,640 So Paul had another visitor from McDonald, Ludlow said we have a 779 01:00:01,640 --> 01:00:06,600 guy called John Darlene, John Raj who said we need a swept 780 01:00:06,600 --> 01:00:09,080 ring coated and our cell code. I said yes, I know that. 781 01:00:10,760 --> 01:00:13,880 I said, look, I can write the codes for you if you'll help me 782 01:00:13,880 --> 01:00:18,680 debug them. So I wrote the two codes. 783 01:00:19,280 --> 01:00:24,400 Flow 21 was for Nacelle and Flow 22 was for Swept Wing. 784 01:00:27,160 --> 01:00:30,920 In a few weeks, actually. Each was a box of cards. 785 01:00:31,960 --> 01:00:33,920 You know, we were still using card readers. 786 01:00:34,640 --> 01:00:38,960 We had this control letter 6600, which was kind of the world's 787 01:00:38,960 --> 01:00:41,760 fastest computer at the time. That's because Quran had all 788 01:00:41,760 --> 01:00:45,800 these contacts with people like Von Norman and it had the 789 01:00:45,800 --> 01:00:52,240 so-called computer centre in Quran Institute who would get 790 01:00:52,480 --> 01:01:01,560 the latest computers. But the 6600 was fast but stands 791 01:01:01,560 --> 01:01:04,320 to the day. But it only had 130,000 words in 792 01:01:04,320 --> 01:01:06,800 memory. That turns out to be 64 bit 793 01:01:06,800 --> 01:01:08,680 words, turns out to be about 2 megabytes. 794 01:01:10,800 --> 01:01:14,000 So you couldn't really do a 3D calculation and store it on the 795 01:01:14,800 --> 01:01:17,120 core memory. You had to go plane by plane 796 01:01:17,160 --> 01:01:22,360 reading off the disk. Anyway, I So I wrote the two 797 01:01:23,440 --> 01:01:28,280 codes and then John and myself went to see Borgara Billion and 798 01:01:30,600 --> 01:01:36,280 he was furiously angry. He said I was supposed to be 799 01:01:36,280 --> 01:01:39,320 helping him write his book on simply good wing sections and I 800 01:01:39,320 --> 01:01:41,760 was only interested in being famous and so on. 801 01:01:43,480 --> 01:01:46,720 And I said, well yes, I'd certainly like to be famous, but 802 01:01:46,720 --> 01:01:49,160 actually I really need a threat to encourage me. 803 01:01:51,920 --> 01:01:59,800 But John Darlene was not. After that, John Darlene felt he 804 01:01:59,800 --> 01:02:07,920 couldn't whisk confronting Paul and trying to debug his codes or 805 01:02:07,920 --> 01:02:11,920 help me. And so he I decided perhaps 806 01:02:11,920 --> 01:02:13,440 you'd better leave it alone for the moment. 807 01:02:13,440 --> 01:02:15,960 So the two boxes of cards remained on the floor of my 808 01:02:15,960 --> 01:02:18,720 office at the Grown Institute for the next few months. 809 01:02:19,920 --> 01:02:22,600 I never tried to run either of them. 810 01:02:24,720 --> 01:02:28,320 And then when John went back to Douglas, he told them that these 811 01:02:28,320 --> 01:02:32,080 two codes existed but hadn't been debugged. 812 01:02:34,040 --> 01:02:38,280 And then I got a call from a guy, David Curie, who was 813 01:02:38,280 --> 01:02:41,400 working for the McDonnell Douglas Research Centre in Saint 814 01:02:41,400 --> 01:02:45,480 Louis, saying that they had somebody coming for the summer 815 01:02:45,480 --> 01:02:49,360 and would it be possible for them to have this guy see if he 816 01:02:49,360 --> 01:02:54,680 can debug Flow 21. So I said that's fine. 817 01:02:54,680 --> 01:02:59,440 And I flew out to Saint Louis through the box of cards, flew 818 01:02:59,440 --> 01:03:03,480 to one and gave it to Lewin. Well, this guy did nothing that 819 01:03:03,480 --> 01:03:11,480 summer. But then in fall of 74, David 820 01:03:11,480 --> 01:03:15,080 Coe was suddenly offered a position as a visiting assistant 821 01:03:15,080 --> 01:03:21,320 professor at Cornell. They'd had a crisis because two 822 01:03:21,320 --> 01:03:25,760 of their top people had both left, actually a sea bass in 823 01:03:27,040 --> 01:03:31,840 particular. So David asked me if he could 824 01:03:31,840 --> 01:03:36,800 take Flow 21, and I said sure. So a few weeks later he called 825 01:03:36,800 --> 01:03:38,720 up to say that he'd fixed Flow 21. 826 01:03:38,720 --> 01:03:40,080 It was working. And I said, well, you're 827 01:03:40,120 --> 01:03:44,680 interested in Flow 22. And he said yes. 828 01:03:44,680 --> 01:03:49,000 And I said, well, thing about Flow 22 is it's got all these 829 01:03:49,000 --> 01:03:52,280 analytic transformations and the huge, very, very complicated 830 01:03:52,280 --> 01:03:58,360 formulas. We need to have some independent 831 01:03:58,360 --> 01:04:00,600 checking of this. I said, so are you willing to 832 01:04:00,600 --> 01:04:04,520 try to work out all the formulas separately from me? 833 01:04:04,800 --> 01:04:08,880 He agreed to do that. And then we had a meeting and 834 01:04:09,880 --> 01:04:13,000 came to New York and 1st his formulas looked different from 835 01:04:13,000 --> 01:04:14,920 mine, but then we found they were equivalent. 836 01:04:17,240 --> 01:04:22,040 So I gave him Flo 22 and took it back with him to Cornell. 837 01:04:23,320 --> 01:04:29,120 Called up a week later to say, well, he'd fixed that too, and 838 01:04:29,200 --> 01:04:33,680 turned out that I had got a miscount on reading plane by 839 01:04:33,680 --> 01:04:37,200 plane off the disk and writing updates back onto the disk. 840 01:04:37,200 --> 01:04:39,120 But he fixed that and the code worked. 841 01:04:41,360 --> 01:04:51,160 So Flow 22 was really the first computer programme that could 842 01:04:51,160 --> 01:04:55,560 calculate transonic flow solutions for wings with 843 01:04:55,560 --> 01:05:02,400 reasonable accuracy. Those days it took about several 844 01:05:02,400 --> 01:05:09,680 hours to run a calculation on a Contravenator 6600, but Douglas 845 01:05:09,680 --> 01:05:15,720 says attributed a cost of about $3000 for each run. 846 01:05:15,720 --> 01:05:18,560 But they started running it about five times a day when they 847 01:05:18,560 --> 01:05:22,280 were using it to trying to design the C-17 wing and things 848 01:05:22,280 --> 01:05:30,880 like that. And then Canada was trying to 849 01:05:30,880 --> 01:05:35,200 develop what became the Canada, what do you call it? 850 01:05:35,640 --> 01:05:41,720 Well, now it's bombarded, but the regional jet and they wanted 851 01:05:41,760 --> 01:05:46,680 to use Flow 22, but they found a consulting company in Los 852 01:05:46,680 --> 01:05:49,120 Angeles who was running Flow 22 for them. 853 01:05:50,880 --> 01:05:57,560 And this Canada regional jet was the first commercial aircraft 854 01:05:57,560 --> 01:06:02,240 that had a supercritical wing. They see the design of using 855 01:06:02,240 --> 01:06:11,680 Flow 22 and eventually NASA kind of released listing of the code 856 01:06:11,680 --> 01:06:15,240 on some NASA website and it wound up being picked up by the 857 01:06:15,240 --> 01:06:17,080 Russians and the Chinese and everybody. 858 01:06:22,320 --> 01:06:30,200 So the other thing that kind of unusual about my girls, right? 859 01:06:30,200 --> 01:06:34,560 I was a professor now normally of computer science, the Quran 860 01:06:34,560 --> 01:06:38,760 Institute, but I didn't have any students, you know, the people, 861 01:06:39,120 --> 01:06:42,120 the students all wanted to work with the famous professors like 862 01:06:42,120 --> 01:06:46,040 Peter Lacks or Huey Nehrenberg or Jack Schwartz. 863 01:06:46,600 --> 01:06:49,200 You know, we had ten members of the National Academy of Sciences 864 01:06:49,200 --> 01:06:52,880 at the Croyde Institute. So I had no students. 865 01:06:54,720 --> 01:06:58,240 So the only way to continue is to go on writing the programmes 866 01:06:58,240 --> 01:07:01,320 myself. And I had this collaboration 867 01:07:01,320 --> 01:07:04,320 that developed with David Coe and then we managed to get 868 01:07:04,640 --> 01:07:08,400 support from the Office of Naval Research to go pursue this 869 01:07:08,400 --> 01:07:16,120 further. So I realised that we couldn't 870 01:07:16,120 --> 01:07:18,720 really use analytical transformations to do anything 871 01:07:18,720 --> 01:07:20,800 more complicated than a winged fuse large. 872 01:07:20,800 --> 01:07:25,280 So we, David and I developed what we call the finite volume 873 01:07:25,280 --> 01:07:29,640 method for transonic potential for it was really a finite 874 01:07:29,640 --> 01:07:34,720 element method with try linear isoparametric elements, but I 875 01:07:34,720 --> 01:07:46,200 didn't know that at the time. And then I got into the going to 876 01:07:46,200 --> 01:07:52,480 the Euler equations kind of again by sort of circuitous 877 01:07:52,480 --> 01:07:59,440 rooted. I went to various international 878 01:07:59,440 --> 01:08:03,920 meetings those days, something called Symposium Transonicum 2 879 01:08:03,920 --> 01:08:09,920 that was held in Gerdingham at that time. 880 01:08:09,920 --> 01:08:14,000 There was a Wolfgang Schmidt who was a rather charismatic 881 01:08:14,000 --> 01:08:19,560 character who had a small group doing CFD and Dornier, well, 882 01:08:19,560 --> 01:08:24,800 Dornier was a sort of nothing much of a company that somehow 883 01:08:24,800 --> 01:08:28,600 survived after the Second World War, still learned by the 884 01:08:28,600 --> 01:08:33,040 Dornier family, but Afghan. We are very effective in getting 885 01:08:34,439 --> 01:08:37,399 support from various German government agencies. 886 01:08:39,960 --> 01:08:44,520 I met him in Girdingham. In fact, he persuaded me to try 887 01:08:44,520 --> 01:08:48,840 to swim in the town's swimming pool, which is very embarrassing 888 01:08:48,840 --> 01:08:55,560 to me when it turned out that I had to rent some kind of bikini 889 01:08:55,560 --> 01:08:59,800 like swimsuit and I was required to wear the cap on my head. 890 01:08:59,800 --> 01:09:02,520 And the other cap available was a sort of pink cap with feathers 891 01:09:02,520 --> 01:09:08,000 or something. Well, Gang turned out to be a A 892 01:09:08,000 --> 01:09:11,120 star Mortimer player and an expert swimmer and he was. 893 01:09:12,160 --> 01:09:13,800 He kind of forced me into the pool. 894 01:09:15,279 --> 01:09:19,439 But anyway, a couple of years later he persuaded me to come 895 01:09:19,439 --> 01:09:22,319 and spend several weeks at Dawnia. 896 01:09:23,680 --> 01:09:31,800 This is 1979. And then in while 1979 there was 897 01:09:31,800 --> 01:09:35,479 a workshop in Stockholm which tried to compare if it's to 898 01:09:35,479 --> 01:09:38,800 solve the Euler equations with potential flow solutions. 899 01:09:41,040 --> 01:09:47,160 I participated in that. I saw that nobody had a solver 900 01:09:47,160 --> 01:09:50,000 for the Euler equations that reached a steady state, and I 901 01:09:50,000 --> 01:09:52,920 felt that it was wrong. You should be able to get a 902 01:09:52,920 --> 01:09:58,200 steady state. So in 1980, Wolfgang asked you 903 01:09:58,200 --> 01:10:03,120 to come back and he had a onola solvent that had been developed 904 01:10:03,120 --> 01:10:10,720 jointly with ART, received at Dornea, and he wasn't working 905 01:10:10,760 --> 01:10:12,720 all that well. It was one of the ones that 906 01:10:13,080 --> 01:10:16,280 wouldn't get a steady state that had been in his Stockholm 907 01:10:16,280 --> 01:10:19,800 workshop. He asked me if I'd like to look 908 01:10:19,800 --> 01:10:23,240 at it and I said sure. But I said, look, did she really 909 01:10:23,240 --> 01:10:26,560 want me to do this? We're going to have to spend 910 01:10:26,560 --> 01:10:32,000 several hours a day trying to do computer runs and really to pay 911 01:10:32,000 --> 01:10:33,160 for this. He said don't worry. 912 01:10:34,680 --> 01:10:40,840 So Donia had this IBM 4341 and they were charging like one 913 01:10:40,840 --> 01:10:43,200 Deutsche mark per seat to be used second or something like 914 01:10:43,200 --> 01:10:48,040 that. And Donna was kind of on you. 915 01:10:48,080 --> 01:10:51,560 You sat, They had two men in an office. 916 01:10:51,640 --> 01:10:58,800 They would sit facing each other inside each you would sit one. 917 01:10:58,800 --> 01:11:01,680 Each had a desk, but the desks were pushed against each other 918 01:11:01,680 --> 01:11:03,040 and you were looking at each other. 919 01:11:03,040 --> 01:11:06,040 And they all call themselves her, like him and her Fritz. 920 01:11:06,040 --> 01:11:12,600 They didn't use first names And I showed up and I told Wolfgang 921 01:11:12,600 --> 01:11:14,800 by his first name and they were all shocked. 922 01:11:14,800 --> 01:11:17,880 But anyway Wolfgang said don't worry about it. 923 01:11:18,920 --> 01:11:23,600 So I actually looked at their solver and decided that they got 924 01:11:23,600 --> 01:11:29,720 some series, we could call them errors. 925 01:11:30,520 --> 01:11:34,200 They had a kind of, it was implemented what was called the 926 01:11:34,200 --> 01:11:39,160 McCormack scheme, but they had an additional artificial 927 01:11:39,160 --> 01:11:41,400 viscosity term. And the way that was 928 01:11:41,400 --> 01:11:45,800 constructed, turned out it was going to take differences of the 929 01:11:45,800 --> 01:11:50,760 cell volumes if the flow was uniform in the far field. 930 01:11:50,760 --> 01:11:53,360 And that's just generating big error term. 931 01:11:53,560 --> 01:11:59,840 And so rewrote the code with the first step to what became the 932 01:11:59,840 --> 01:12:02,200 James of Spit Turkel scheme in the next year. 933 01:12:02,200 --> 01:12:09,720 And at the end of the three weeks, Hofgang got a bill for 934 01:12:09,720 --> 01:12:13,880 250,000 Deutsche Marks for the computer cost of my visit there. 935 01:12:15,960 --> 01:12:17,880 They see that his problem was not mine. 936 01:12:22,880 --> 01:12:29,600 This is the point when I had been approached by both Stanford 937 01:12:29,600 --> 01:12:35,440 and Princeton with Office of Faculty Positions. 938 01:12:35,440 --> 01:12:44,760 And what had actually happened was that in the academic world 939 01:12:47,840 --> 01:12:50,960 they valued applied mathematicians basically and 940 01:12:50,960 --> 01:12:57,440 people like Milton Van Dyken mortal land on at MIT, they did 941 01:12:57,440 --> 01:13:01,680 all these asymptotic expansions and things like that, but they 942 01:13:01,680 --> 01:13:04,400 none of them had ever written a computer programme or had any 943 01:13:04,400 --> 01:13:09,120 interest in that. And I suddenly discovered that 944 01:13:09,480 --> 01:13:14,520 CFD actually worked quite well. But the thing was that there was 945 01:13:14,520 --> 01:13:18,200 nobody, not many people around who knew how to write these 946 01:13:18,200 --> 01:13:23,960 programmes, mostly from NASA, but none of them had any 947 01:13:23,960 --> 01:13:28,400 teaching experience was I had been teaching at the Groan 948 01:13:28,400 --> 01:13:32,440 Institute. So I wound up being offered jobs 949 01:13:32,440 --> 01:13:39,680 by both Stanford and Princeton and I went out to Stanford and 950 01:13:39,680 --> 01:13:45,400 founded. They were offering me 42,000 951 01:13:45,400 --> 01:13:52,200 years salary and you couldn't buy a house there for less than 952 01:13:52,200 --> 01:13:55,840 about $400,000. And at that time the interest 953 01:13:55,840 --> 01:13:58,360 rate for mortgages is around 13%. 954 01:13:59,480 --> 01:14:04,080 So it turned out that the $42,000 would just about pay the 955 01:14:04,080 --> 01:14:13,720 interest on the mortgage house. So I then I got approached by a 956 01:14:13,720 --> 01:14:17,280 direct phone call from the president of Princeton saying, 957 01:14:17,280 --> 01:14:19,360 well, I'd like to come down and play squash with him. 958 01:14:20,600 --> 01:14:23,360 I didn't know he was the New Jersey State squash champion, 959 01:14:23,360 --> 01:14:28,400 but I said OK, and eventually I took the job in Princeton. 960 01:14:28,400 --> 01:14:30,880 As a result, I went down and lost to him at squash. 961 01:14:30,880 --> 01:14:37,320 It didn't matter very much, but but this 1980 was in transition. 962 01:14:37,320 --> 01:14:48,120 So I had got a first cut at this Euler Savo Donia, but I hadn't 963 01:14:48,120 --> 01:14:50,520 got it yet to converge to a steady state. 964 01:14:51,160 --> 01:14:55,760 And when I first got to Princeton, we didn't have a 965 01:14:55,760 --> 01:14:59,560 computer when I convinced that we had to buy a mini computer 966 01:14:59,560 --> 01:15:07,520 when I would have gone for digital equipment, but Princeton 967 01:15:07,520 --> 01:15:10,200 was kind of IBM territory. So we wound up getting an IBM 968 01:15:10,200 --> 01:15:14,840 4341, which is the same machine as it had it. 969 01:15:15,200 --> 01:15:21,200 Don't have my own. That had filled a huge room with 970 01:15:21,200 --> 01:15:24,640 cabinets all over the place, you know, the disk and one cabinet 971 01:15:24,640 --> 01:15:28,920 and so on. It actually had two megabytes of 972 01:15:28,920 --> 01:15:35,840 memory. I had access to that from 973 01:15:35,840 --> 01:15:44,000 roughly November, and I couldn't get this thing to converge to 974 01:15:44,000 --> 01:15:46,960 state. It would go down to some level 975 01:15:46,960 --> 01:15:49,880 of error around himmatri and start oscillating. 976 01:15:52,160 --> 01:15:56,160 I thought it was maybe reflection from the far field 977 01:15:56,160 --> 01:15:59,400 boundary conditions. That's when I consulted Eli 978 01:15:59,400 --> 01:16:01,960 Tokyo about non reflecting boundary conditions. 979 01:16:02,680 --> 01:16:04,400 That actually didn't fix the problem. 980 01:16:06,000 --> 01:16:11,600 Then I tried running it 1st order accurate and it converged 981 01:16:11,600 --> 01:16:15,880 very well. So then I saw that I needed to 982 01:16:15,880 --> 01:16:22,320 have but that was not accurate enough so that I came up with 983 01:16:22,320 --> 01:16:26,920 this idea of you blend force differences and 2nd differences 984 01:16:26,920 --> 01:16:32,240 of artificial viscosity and use the force differences as 985 01:16:32,240 --> 01:16:34,600 background. But in the nearer shock that 986 01:16:34,600 --> 01:16:37,480 causes an oscillation. So you turn them off and drop to 987 01:16:37,520 --> 01:16:39,760 a locally 1st order accurate through the shock. 988 01:16:39,760 --> 01:16:46,240 And this is eventually we became this so-called JST scheme that 989 01:16:46,520 --> 01:16:50,520 got away with a lot. Yeah. 990 01:16:52,120 --> 01:16:57,480 But actually for me something else happened around that time. 991 01:16:57,480 --> 01:17:05,000 So I I presented this JST scheme and by that time I had got a 3D 992 01:17:05,000 --> 01:17:13,440 implementation with Flow 57 at the AI AA meeting in 1981 in the 993 01:17:13,440 --> 01:17:19,600 summer in Palo Alto I think it was and nobody paid much 994 01:17:19,600 --> 01:17:23,760 attention. But then of course the company 995 01:17:23,760 --> 01:17:27,000 started getting my code Flow 57 and they found it worked. 996 01:17:28,840 --> 01:17:34,040 But just about that time, Phil Rowe came out with his famous 997 01:17:34,040 --> 01:17:42,600 paper with his so-called row matrix of basically the 998 01:17:42,600 --> 01:17:50,280 character it's based and it's the work of genius that paper I 999 01:17:51,240 --> 01:17:56,320 so that kind of got the academic land. 1000 01:17:56,320 --> 01:18:02,720 I don't deserve it yourself. But when I tried coding the Phil 1001 01:18:02,720 --> 01:18:05,880 Roe's scheme, I found that it wasn't really doing any better 1002 01:18:05,880 --> 01:18:13,280 than my scheme, but it wasn't really helpful to to push that. 1003 01:18:13,280 --> 01:18:19,640 But what I do find out about thorough scheme especially you 1004 01:18:19,640 --> 01:18:23,720 had this mean value of matrix, it's called a row matrix and you 1005 01:18:23,720 --> 01:18:25,840 found an analytic way of calculating and then. 1006 01:18:25,840 --> 01:18:30,200 So it's a stroke of genius and turns out that that enables 1007 01:18:32,240 --> 01:18:35,680 analysis of other schemes, not just a row scheme. 1008 01:18:35,680 --> 01:18:40,680 If you use this row matrix, you can actually prove things about 1009 01:18:40,680 --> 01:18:45,720 the numerical shock structure, shock waves, being in a shock 1010 01:18:45,720 --> 01:18:49,720 capturing scheme and so on. So it really opened the way to a 1011 01:18:50,000 --> 01:18:53,440 much broader way of looking at sort of theoretical side of 1012 01:18:53,440 --> 01:19:03,440 these schemes that well, I don't know where you want to go from 1013 01:19:03,440 --> 01:19:07,480 there, but so we're in 1980. So what happened? 1014 01:19:10,760 --> 01:19:19,480 Yes, so you would then at that time at Princeton, one question 1015 01:19:19,480 --> 01:19:24,080 I had was the code. So you mentioned that people 1016 01:19:24,080 --> 01:19:26,880 were using the code. You were, you know that NASA 1017 01:19:26,880 --> 01:19:29,880 were using it different companies and we Fast forward to 1018 01:19:29,880 --> 01:19:34,400 now and there are these large commercial companies, you know, 1019 01:19:35,320 --> 01:19:38,480 and SIS Fluent and Siemens and all these other. 1020 01:19:38,480 --> 01:19:43,000 Did you did you think about or, or maybe this is what's coming 1021 01:19:43,000 --> 01:19:46,600 next, the commercialization of this or was it very much at that 1022 01:19:46,600 --> 01:19:50,680 time this idea of the code was for everybody to use. 1023 01:19:50,800 --> 01:19:54,120 It was sort of for the good of the community. 1024 01:19:55,400 --> 01:20:03,240 Well, there are several things that I think you need to 1025 01:20:03,240 --> 01:20:06,240 realize, you see. We didn't have copyright 1026 01:20:06,240 --> 01:20:08,520 protection of computer codes at that time. 1027 01:20:11,160 --> 01:20:14,440 Actually, that was only established by Bill Gates around 1028 01:20:14,440 --> 01:20:21,560 about early 198283. So there was only no way to 1029 01:20:21,720 --> 01:20:23,520 prevent codes being passed around. 1030 01:20:27,000 --> 01:20:35,160 But in fact, of course, if you wrote a code that worked and 1031 01:20:35,320 --> 01:20:39,520 Flow 57 went up, being used by just about everybody in the 1032 01:20:39,520 --> 01:20:46,240 aircraft industry in America, and then it got, well, Dawnia, I 1033 01:20:46,240 --> 01:20:49,200 had it because of course I'd been working with Forfgang at 1034 01:20:49,440 --> 01:20:54,080 Dawnia. But then I had a consulting 1035 01:20:54,080 --> 01:20:57,960 arrangement with Cray Research and I didn't mention this, but 1036 01:20:58,960 --> 01:21:08,480 actually when I wrote 3D code Flow 57 that was in Kandra 1037 01:21:09,320 --> 01:21:15,880 around February, March of 1981, I could only run a measure of 1038 01:21:15,880 --> 01:21:20,160 like 48 by 8 by 8 on this, I bet 4341. 1039 01:21:20,160 --> 01:21:26,360 But I had been approached by Cray Research because we're no 1040 01:21:26,360 --> 01:21:29,080 trying to sell their first Cray One computers. 1041 01:21:30,160 --> 01:21:33,600 They got benchmarked typically by floor 22. 1042 01:21:35,440 --> 01:21:38,000 So they decided it might be better to have me on board. 1043 01:21:40,560 --> 01:21:46,400 And I was interested too, because I thought that this JST 1044 01:21:46,400 --> 01:21:48,680 scheme should run Vector without a problem. 1045 01:21:50,680 --> 01:21:58,920 So they gave me access to Cream One when I first tried to run, 1046 01:22:00,440 --> 01:22:02,280 but his code that didn't run vector. 1047 01:22:02,280 --> 01:22:04,960 And I went out there and discovered that there were one 1048 01:22:04,960 --> 01:22:08,640 or two minor changes I still had to make to the code, but it 1049 01:22:08,640 --> 01:22:13,080 would run vector. But when I tried to write Flow 1050 01:22:13,080 --> 01:22:18,720 57, I had a terminal with Princeton connected to Cray 1051 01:22:18,720 --> 01:22:21,920 Research. So after I got it to run on 1052 01:22:22,360 --> 01:22:29,800 4341, I was able to submit the same code to Cray through this 1053 01:22:29,800 --> 01:22:36,320 terminal, and I could run something like 96 by 16 by 16. 1054 01:22:36,320 --> 01:22:39,480 It still wasn't enough memory on the Cray one at the time, but at 1055 01:22:39,480 --> 01:22:42,400 least that was enough to kind of meaningful calculation. 1056 01:22:43,120 --> 01:22:44,680 But now it meant Cray had the code. 1057 01:22:46,640 --> 01:22:52,720 So they were trying to sell Crays in Europe and they sold 1058 01:22:52,720 --> 01:22:55,920 one to the British Royal Aircraft Establishment and 1059 01:22:55,920 --> 01:23:02,440 another to the Dutch in Alaro. They they threw in Flow 57 as 1060 01:23:02,440 --> 01:23:05,400 part of the deal without my knowledge. 1061 01:23:05,480 --> 01:23:15,600 So this kind of got to Holland and England and well, I didn't 1062 01:23:15,600 --> 01:23:19,000 get it money out of it, but I did get a good deal of 1063 01:23:19,000 --> 01:23:26,600 recognition. And so now my sort of attitude 1064 01:23:26,640 --> 01:23:31,400 started to change around there because the Air Force is pseudo 1065 01:23:33,320 --> 01:23:39,880 contract with a request for proposals to kind of solve more 1066 01:23:39,880 --> 01:23:43,720 general configurations in the obvious thing to do, we go with 1067 01:23:43,720 --> 01:23:54,840 some kind of multiple extension of Flo 57 and bunches of 1068 01:23:54,840 --> 01:23:59,840 companies like Lockheed and Northrop and Sohn all submitted 1069 01:23:59,840 --> 01:24:04,400 proposals to the Air Force, which amounted to further 1070 01:24:04,400 --> 01:24:07,800 developing Flow 57. And seemed to me that the Air 1071 01:24:07,800 --> 01:24:10,160 Force could have asked me whether I wanted to permanently 1072 01:24:10,400 --> 01:24:13,480 the code rather than funding locking what they wound up 1073 01:24:13,480 --> 01:24:23,000 doing. So I was sort of started to 1074 01:24:23,000 --> 01:24:27,560 wonder, well, how are we going to get more control? 1075 01:24:29,360 --> 01:24:34,680 But then I also knew that I wanted to do a complete aircraft 1076 01:24:34,680 --> 01:24:39,920 calculation. And my first idea was that you 1077 01:24:39,920 --> 01:24:43,480 could use a multi block mesh. But then when I looked at how 1078 01:24:43,480 --> 01:24:46,120 hard it would be to generate these meshes, I saw that we 1079 01:24:46,120 --> 01:24:49,200 didn't have anybody at Princeton who was going to do that. 1080 01:24:51,440 --> 01:24:59,480 And I had this time had brought in Tim Baker from England, as 1081 01:24:59,480 --> 01:25:03,600 was research scientist working for me, and he had his friend 1082 01:25:03,600 --> 01:25:08,160 Nigel Wetherill who was working at the Aircraft Research 1083 01:25:08,160 --> 01:25:14,320 Establishment in Bedford. Nigel was interested in spending 1084 01:25:15,640 --> 01:25:19,520 six months sort of sabbatical at Princeton. 1085 01:25:21,280 --> 01:25:24,640 The ARA were very worried that they had developed a multi block 1086 01:25:24,640 --> 01:25:29,880 extension of Flow 57 and they were very worried that Nigel 1087 01:25:29,880 --> 01:25:31,360 would bring it with him to Britson. 1088 01:25:31,360 --> 01:25:36,400 Right. I said no, I I've concluded that 1089 01:25:36,400 --> 01:25:39,280 we have a better chance if we go to an unstructured mesh. 1090 01:25:42,080 --> 01:25:51,200 So I told Nigel to forget about the multi block code. 1091 01:25:51,200 --> 01:25:54,800 We were going to see work from scratch, the triangulated mesh. 1092 01:25:56,600 --> 01:26:00,200 When Nigel arrived I asked him if he could By this time I'd 1093 01:26:00,200 --> 01:26:03,200 become aware of this process called Deloni triangulation and 1094 01:26:03,200 --> 01:26:06,080 I asked him if he would try decoding it. 1095 01:26:06,080 --> 01:26:08,680 And he got it to run in two DA week or two. 1096 01:26:10,720 --> 01:26:15,280 So we could do sort of in principle multiolu and airfoils 1097 01:26:15,280 --> 01:26:21,160 and things like that. But then I had a meeting with 1098 01:26:21,160 --> 01:26:23,240 Tim and Nigel. I said, well, I think we should 1099 01:26:23,240 --> 01:26:31,560 try to do a complete aircraft. And the idea was that Tim could 1100 01:26:31,560 --> 01:26:37,280 generate meshes around the different components, rather 1101 01:26:37,280 --> 01:26:41,000 like an overset mesh. But then we would take the 1102 01:26:41,000 --> 01:26:45,440 points from all these meshes, just a cloud of points, and 1103 01:26:45,880 --> 01:26:49,360 triangulate them with only triangulation, and that would be 1104 01:26:49,360 --> 01:26:51,280 our grade. And meanwhile, I wrote a flow 1105 01:26:51,280 --> 01:27:01,760 solver to work on tetrahedral meshes, and we had an agreement 1106 01:27:01,760 --> 01:27:06,200 from Cramer Search that they would support this project. 1107 01:27:09,760 --> 01:27:13,880 But then we finally got a first version of the code together. 1108 01:27:15,480 --> 01:27:18,520 We called it the airplane code. Turned out that it used the 1109 01:27:18,520 --> 01:27:22,200 whole memory of the Quay. So they put it to the back of 1110 01:27:22,200 --> 01:27:25,640 the queue. And then at the end of the day 1111 01:27:26,520 --> 01:27:30,400 they would turn the shut the machine down and start running 1112 01:27:30,480 --> 01:27:32,080 experiments with operating systems. 1113 01:27:32,080 --> 01:27:37,720 So it just didn't run. So at that point I tried to 1114 01:27:38,240 --> 01:27:43,680 contact Millwagon who was the president of the CEO of Craig 1115 01:27:43,680 --> 01:27:47,480 had agreed to support this and he was out of time, but I got 1116 01:27:47,480 --> 01:27:55,880 his number 2. They agreed to let us come out 1117 01:27:55,880 --> 01:28:01,160 and use the machine during weekends, overnight. 1118 01:28:03,360 --> 01:28:10,280 And I also, I had a, because I was a consultant, I had some 1119 01:28:10,280 --> 01:28:14,560 kind of immediate monitor Cray and he was very angry that I'd 1120 01:28:14,560 --> 01:28:17,480 bypassed him and he dismissed me as a consultant to Cray. 1121 01:28:18,640 --> 01:28:20,520 I said, well, we still have accidental machine. 1122 01:28:20,520 --> 01:28:22,640 It turned out it is, I didn't care about that. 1123 01:28:24,640 --> 01:28:30,760 But we went out in actually in July, we were given the machine 1124 01:28:30,880 --> 01:28:33,960 nights during the weekend. So we spent sort of two nights, 1125 01:28:34,000 --> 01:28:37,680 we were up all night. Two nights in a row came clear 1126 01:28:37,680 --> 01:28:45,600 that Nigel's Delaunay triangulator wasn't working, and 1127 01:28:45,600 --> 01:28:48,040 we couldn't really disentangle exactly why not. 1128 01:28:48,840 --> 01:28:58,960 But then Nigel was due to go back to England and so we 1129 01:28:58,960 --> 01:29:03,640 decided to basically to put everything on hold for a bit. 1130 01:29:04,960 --> 01:29:09,960 Tim and I had summer vacations when I was spending time no near 1131 01:29:09,960 --> 01:29:17,840 again, but when we got back in fall, we concluded that we 1132 01:29:17,840 --> 01:29:21,880 couldn't debug Natural's triangulation because he was 1133 01:29:22,840 --> 01:29:27,560 doing some things that are not well. 1134 01:29:27,600 --> 01:29:31,320 He kind of assumed that when you called a subroutine that some 1135 01:29:31,400 --> 01:29:34,920 data values would still be there when you recall the subroutine. 1136 01:29:34,920 --> 01:29:37,080 It's not actually the case in Fortune. 1137 01:29:37,080 --> 01:29:40,360 It depends on somebody's compiler whether it is or, which 1138 01:29:40,360 --> 01:29:43,520 isn't still there. It was just almost impossible to 1139 01:29:43,520 --> 01:29:47,000 figure out exactly what had gone wrong. 1140 01:29:47,000 --> 01:29:51,280 So Tim started overwriting a Delaunay triangulator. 1141 01:29:53,840 --> 01:29:59,640 And finally in November we used to go out to Cray because also 1142 01:29:59,640 --> 01:30:03,560 we couldn't visualize anything. But Cray had these evidence of 1143 01:30:03,560 --> 01:30:06,960 Sutherland terminals where you could actually take a look at 1144 01:30:06,960 --> 01:30:14,800 what the mesh. Finally in N1 weekend round 1145 01:30:14,800 --> 01:30:19,560 about 9:00 in the morning after you've been up all night, we saw 1146 01:30:19,640 --> 01:30:21,640 that we've got a mesh that seemed to work. 1147 01:30:24,920 --> 01:30:30,960 Then Jim and I had to catch a flight back to Princeton and we 1148 01:30:30,960 --> 01:30:36,040 barely caught the flight, but we were upgraded to 1st class, 1149 01:30:36,040 --> 01:30:38,880 having showed up late and got free drinks. 1150 01:30:39,680 --> 01:30:41,920 That wasn't a really good idea. Then we got to Newark. 1151 01:30:41,920 --> 01:30:46,040 I had this junk car that I looked at the airport and Tim 1152 01:30:46,040 --> 01:30:49,400 asked you if I could drive, and I said I thought so, but he got 1153 01:30:49,400 --> 01:30:51,920 to the car and he just passed out. 1154 01:30:53,720 --> 01:30:57,000 I started driving and I discovered that I just had 1155 01:30:57,000 --> 01:30:59,800 double vision, which wasn't really good to me. 1156 01:30:59,920 --> 01:31:04,080 Two New Jersey Turnpike, basically I could get rid of the 1157 01:31:04,080 --> 01:31:06,000 double vision if I took my glasses off. 1158 01:31:06,000 --> 01:31:10,600 But then I decided it was kind of lucky that we made it back to 1159 01:31:10,600 --> 01:31:16,040 Princeton without a crash. But anyway, we we finally did 1160 01:31:16,040 --> 01:31:22,240 get a successful calculation for a Boeing 747 which we presented 1161 01:31:22,240 --> 01:31:27,760 in the Aerospace at Sands Meeting Arena in January to 6th 1162 01:31:29,680 --> 01:31:36,480 that was well at this point A. Big moment. 1163 01:31:37,440 --> 01:31:44,880 Yeah. But I thought now I have to tell 1164 01:31:44,880 --> 01:31:51,240 Princeton that we've got the code because Tomb has been paid 1165 01:31:51,240 --> 01:31:55,640 full time as a research scientist by of our groundsmith 1166 01:31:56,160 --> 01:31:59,600 by Princeton. And normally that meant that any 1167 01:31:59,600 --> 01:32:02,040 code we wrote that probably belongs to Princeton. 1168 01:32:05,200 --> 01:32:08,800 So I approached my department head at the time and said, well, 1169 01:32:08,800 --> 01:32:13,960 I think this code is actually has potential commercial value 1170 01:32:16,720 --> 01:32:20,640 and it's also possible that you could patent the numerical 1171 01:32:20,640 --> 01:32:24,160 scheme. Well, it turns out that 1172 01:32:24,160 --> 01:32:30,640 Princeton had signed some deal with an outfit, some of the 1173 01:32:30,640 --> 01:32:37,480 small companies, to at first write a refusal for them. 1174 01:32:39,320 --> 01:32:43,240 Property rights are really software developed in Princeton. 1175 01:32:43,240 --> 01:32:48,080 And I got a call from somebody. I thought I was talking to a 1176 01:32:48,440 --> 01:32:51,000 somebody from the legal department inside Princeton, and 1177 01:32:51,000 --> 01:32:54,400 I was talking to these people and I told them those things 1178 01:32:54,400 --> 01:32:55,880 that I probably would not have told them. 1179 01:32:55,880 --> 01:33:04,960 I don't know who they were. OK Anyway, these people said a 1180 01:33:09,600 --> 01:33:15,480 form to Tim and me which we agreed to give them all rights 1181 01:33:15,480 --> 01:33:19,320 and perpetuities to the code and needed derivatives for $1.00 and 1182 01:33:19,560 --> 01:33:22,120 two, and I refused to sign it. So we're not going to agree with 1183 01:33:22,120 --> 01:33:25,360 that. And there really nothing 1184 01:33:25,360 --> 01:33:27,920 happened because these people, we said, look, we've written the 1185 01:33:27,920 --> 01:33:31,680 code, but you ought to patent it. 1186 01:33:32,640 --> 01:33:37,800 So that's, and they didn't. But the meanwhile the code was 1187 01:33:38,840 --> 01:33:41,040 kind of in limbo for three years. 1188 01:33:42,440 --> 01:33:44,640 And then the Wrights reverted Princeton. 1189 01:33:44,640 --> 01:33:48,520 And at this point, Tim and I went back to Princeton and said, 1190 01:33:48,520 --> 01:33:53,400 well, look, the only people who can really commercialised the 1191 01:33:53,400 --> 01:33:56,760 code or actually asked because we can, we can make it work. 1192 01:33:57,880 --> 01:34:00,960 So if I had agreed to let us form a company that we called 1193 01:34:00,960 --> 01:34:10,280 ESOP, Aerodynamic Simulation and Optimization, we would pay 15% 1194 01:34:10,280 --> 01:34:19,040 of the proceeds to Princeton. So then we tried to market the 1195 01:34:19,040 --> 01:34:22,040 airplane code, but the thing was during those three years, that's 1196 01:34:22,040 --> 01:34:29,520 when Florence started to get going with the mesh code. 1197 01:34:30,440 --> 01:34:43,440 So really we lost the window and I, I didn't really want to run a 1198 01:34:43,440 --> 01:34:48,880 company where you, I mean, actually commercial software has 1199 01:34:48,880 --> 01:34:51,480 a great deal of an issue of marketing, isn't it? 1200 01:34:51,480 --> 01:34:57,120 And such like I really care about doing calculations on some 1201 01:34:57,120 --> 01:34:58,720 of the schooling fan or something. 1202 01:35:00,680 --> 01:35:06,600 So, but we did, we did make, we did sell about 5 copies of this 1203 01:35:07,120 --> 01:35:13,240 code 1 to Douglas where they were used it very seriously. 1204 01:35:13,240 --> 01:35:17,960 And John Vassman became very involved in actually he managed 1205 01:35:17,960 --> 01:35:26,760 to fix some issues with the code because it, it was by that time 1206 01:35:26,760 --> 01:35:32,040 we had versions of Cray with four processes and things like 1207 01:35:32,040 --> 01:35:33,560 that. And for some reason this 1208 01:35:33,560 --> 01:35:36,280 Delaunay triangulation should have been sequential when it was 1209 01:35:36,280 --> 01:35:39,440 running faster with the multiprocesses. 1210 01:35:39,440 --> 01:35:43,440 And John figured out that what was happening was that there was 1211 01:35:43,440 --> 01:35:47,960 one particular process going on which was consuming all the 1212 01:35:47,960 --> 01:35:49,280 time. It was. 1213 01:35:52,640 --> 01:35:55,480 So what this Delaunay triangulation, what you do is 1214 01:35:55,480 --> 01:35:58,320 you have a triangulation, you insert one more point and then 1215 01:35:58,320 --> 01:36:03,680 you re triangulate in the vicinity of that point you add. 1216 01:36:05,600 --> 01:36:11,200 But Jim had some flags and at the end of each additional point 1217 01:36:11,200 --> 01:36:15,120 he was setting his flags back to 0 through the whole mash and 1218 01:36:15,480 --> 01:36:21,240 instead of just in the region where it had been re 1219 01:36:21,240 --> 01:36:28,200 triangulated. It turned out that that was 90% 1220 01:36:28,200 --> 01:36:31,040 of the time the triangulation was spent resetting these flags 1221 01:36:31,040 --> 01:36:35,320 to 0. And when you removed that, the 1222 01:36:35,320 --> 01:36:41,320 code ran in like 40 minutes instead of 30 hours 1223 01:36:42,160 --> 01:36:45,480 triangulated. But John Rasberg was kind of 1224 01:36:45,480 --> 01:36:47,880 responsible when it's coming. So they he was secured 1225 01:36:47,960 --> 01:36:53,720 successfully. Donald Louglas to to fix the 1226 01:36:54,280 --> 01:36:58,520 some of the problems that the MD 11 on the nacelle wing pylon 1227 01:36:58,520 --> 01:37:04,600 interaction, things like that. But at this time, I also came 1228 01:37:04,600 --> 01:37:10,680 back to my original first love. You see, I never thought it's 1229 01:37:10,680 --> 01:37:12,560 just a matter of calculating the flow. 1230 01:37:14,640 --> 01:37:20,440 I thought we're trying to find the best design on actually 1231 01:37:20,440 --> 01:37:25,920 going all the way back to 1970. I wrote a quote called SYN one 1232 01:37:26,640 --> 01:37:31,320 same time as flow 1 and SIN one actually solved the inverse 1233 01:37:31,320 --> 01:37:33,120 problem. We input the crusher 1234 01:37:33,120 --> 01:37:37,760 distribution and it finds the shape, but we've conform a 1235 01:37:37,760 --> 01:37:41,320 mapping. But I'd never given up on this 1236 01:37:41,320 --> 01:37:47,800 thought. But this is where I I went to a 1237 01:37:47,800 --> 01:37:57,640 meeting in NASA Langley on flow control probably in maybe 1238 01:37:58,000 --> 01:38:02,880 February 1988, something like around that time, kind of a sort 1239 01:38:02,880 --> 01:38:05,880 of light bulb went up in my head. 1240 01:38:05,880 --> 01:38:10,160 I suddenly realised, oh look, these people are trying to do 1241 01:38:10,160 --> 01:38:16,400 flow control, but actually you could use the control theory to 1242 01:38:16,880 --> 01:38:20,280 do shape vision. In other words, the control 1243 01:38:20,280 --> 01:38:22,000 would be changing the shape of the boundary. 1244 01:38:24,600 --> 01:38:27,320 And I should have known this because I had all the background 1245 01:38:27,320 --> 01:38:29,120 and control theory. But I think, see when I did 1246 01:38:29,120 --> 01:38:32,840 control theory, I always did it for finite dimensional systems. 1247 01:38:33,880 --> 01:38:37,120 This would be, I've said something by mistake. 1248 01:38:37,120 --> 01:38:38,720 I'm not quite sure what's going on here. 1249 01:38:40,280 --> 01:38:43,600 Can you see the pictures still? I've got a second little picture 1250 01:38:43,600 --> 01:38:46,560 of myself. I don't know why that happened. 1251 01:38:46,560 --> 01:38:50,200 Oh. I can, yeah. 1252 01:38:50,200 --> 01:38:51,640 I'm not sure. I can still see you and hear 1253 01:38:51,640 --> 01:38:53,200 you. You can see me. 1254 01:38:55,640 --> 01:39:00,800 Yeah, all good. Well, anyway, I sort of hit me 1255 01:39:00,800 --> 01:39:05,840 that we need you could just use control theory and well, that's 1256 01:39:05,840 --> 01:39:09,200 control of infinite dimensional systems PDS. 1257 01:39:10,520 --> 01:39:15,640 So I knew that Jacques Louis Leons, this famous French 1258 01:39:15,640 --> 01:39:20,320 mathematician who written a bunch of books on controller 1259 01:39:20,320 --> 01:39:23,480 pedes. So I had to go off and try to 1260 01:39:23,480 --> 01:39:25,080 fight my way through some of this book. 1261 01:39:25,080 --> 01:39:28,560 Well, it's very, very difficult. It's all in civil of space, so 1262 01:39:28,560 --> 01:39:31,520 on. But I got to enough just to 1263 01:39:31,520 --> 01:39:34,480 understand that you need to solve these adjoint equations. 1264 01:39:36,640 --> 01:39:41,560 And so I came up with this aerodynamic design by a control 1265 01:39:41,560 --> 01:39:44,200 theory of paper that I wrote, and later that. 1266 01:39:44,200 --> 01:39:49,120 Yeah. And they did something that I'd 1267 01:39:49,120 --> 01:39:54,480 never done before. Aerodynamic design by control 1268 01:39:54,480 --> 01:39:58,240 theory doesn't actually have any computer programme in it at all. 1269 01:39:59,400 --> 01:40:03,000 It just develops a theory of saying if you did this, it 1270 01:40:03,000 --> 01:40:04,040 should work. All right. 1271 01:40:05,840 --> 01:40:09,720 And. I think I was partly reacting 1272 01:40:09,720 --> 01:40:13,920 to, I'd had feedback that people said I was a good computer 1273 01:40:13,920 --> 01:40:15,960 programmer. And I said, well, OK, I'm going 1274 01:40:15,960 --> 01:40:17,640 to show you that there's a theory here. 1275 01:40:17,640 --> 01:40:19,160 I'm going to publish the theory first. 1276 01:40:22,800 --> 01:40:26,800 And I gave it to Steve Orszag, who was a Princeton at the time, 1277 01:40:26,800 --> 01:40:29,560 and we were pretty good friends. And he was just trying to start 1278 01:40:29,560 --> 01:40:32,720 this new journal of scientific computing. 1279 01:40:34,560 --> 01:40:40,240 Because I knew that I felt it would help him starting the 1280 01:40:40,240 --> 01:40:42,600 journal, and I knew that he would publish it without my 1281 01:40:42,600 --> 01:40:45,760 having to fight by way of arguing with reviewers and so 1282 01:40:45,800 --> 01:40:52,480 on. So that came out, and now nobody 1283 01:40:52,480 --> 01:40:56,400 seemed to believe me. I read presentations in various 1284 01:40:56,400 --> 01:41:00,840 places like NASA Langley who got blank stares, who gave a seminar 1285 01:41:00,840 --> 01:41:03,640 at Stanford and just got blank stares. 1286 01:41:05,400 --> 01:41:10,120 So then I had to sit down and try to program it myself and I 1287 01:41:10,120 --> 01:41:13,880 managed to get to work for transonic potential flow 1288 01:41:13,880 --> 01:41:17,440 equation. I actually first made a 1289 01:41:17,440 --> 01:41:23,720 presentation at the Israel an Oracle conference in I think 1290 01:41:23,720 --> 01:41:30,160 1989. And once I'd written a computer 1291 01:41:30,160 --> 01:41:33,360 programme to take work then of course started to attract 1292 01:41:33,360 --> 01:41:43,080 attention and I so that was became my main focus after about 1293 01:41:43,080 --> 01:41:55,120 1989 nineteen 95 or thereabouts. I eventually. 1294 01:41:55,480 --> 01:42:01,160 I I guess that is no, no carry on, sorry. 1295 01:42:02,520 --> 01:42:08,280 Well, I think the basic thing is, you know, I was basically 1296 01:42:08,280 --> 01:42:10,520 trying to solve the problem of a transonic flow. 1297 01:42:11,000 --> 01:42:15,160 I wasn't really basically trying to develop numerical methods as 1298 01:42:15,160 --> 01:42:20,200 such for the real sake. If you look at a number of the 1299 01:42:20,200 --> 01:42:25,720 other people who've been big contributors in CFG, Philro, 1300 01:42:26,280 --> 01:42:35,000 Bran Vanneo and Martin or Stanos and so on, and none of them were 1301 01:42:35,000 --> 01:42:45,840 interested in aircraft actually they were interested in CFD as 1302 01:42:45,840 --> 01:42:49,360 such, whereas to me. Yes. 1303 01:42:50,040 --> 01:42:52,880 I was interested in designing better aircraft, and in 1304 01:42:52,880 --> 01:42:56,560 particular the problem of transonic flow was kind of 1305 01:42:56,560 --> 01:43:00,240 perfect for me because it was a combination of mathematical 1306 01:43:00,320 --> 01:43:09,800 challenge and everything else. And how much do you think that 1307 01:43:09,800 --> 01:43:13,320 your time in industry shaped that? 1308 01:43:13,880 --> 01:43:17,680 You know, you went from Cambridge then for more than 10 1309 01:43:17,680 --> 01:43:23,440 years in industry before you eventually moved to a pure or 1310 01:43:23,960 --> 01:43:26,960 more of a pure academic role in the in the 70s. 1311 01:43:26,960 --> 01:43:33,320 Do you think that industrial time made you more focus on the 1312 01:43:33,320 --> 01:43:38,400 practical challenges rather than a purely theoretical, you know, 1313 01:43:38,400 --> 01:43:43,560 research path? Well, I think yes, to some 1314 01:43:43,560 --> 01:43:48,680 extent. I mean, I was involved in, I 1315 01:43:48,680 --> 01:43:52,800 mean actually I didn't have much involvement in any particular 1316 01:43:53,720 --> 01:43:59,680 designer Grumman, but I was, you know, I was exposed to what was 1317 01:43:59,680 --> 01:44:01,680 going on. Of course, they were the main 1318 01:44:01,680 --> 01:44:07,560 project half drivers. There was this, the F-14 swing 1319 01:44:07,560 --> 01:44:11,600 wings fighting, which I didn't believe was the best solution 1320 01:44:11,600 --> 01:44:15,760 myself, but, but I used to walk around. 1321 01:44:17,440 --> 01:44:20,760 They'd still had draftsman and drawing boards. 1322 01:44:21,840 --> 01:44:25,880 I used to walk around sometimes where they had the draftsman and 1323 01:44:26,520 --> 01:44:31,480 chat with people. You know, it was a plan for some 1324 01:44:31,480 --> 01:44:38,680 kind of government, I think some kind of regional jet, and the 1325 01:44:38,680 --> 01:44:41,920 starting point was to trace the three of you drawings from the 1326 01:44:41,920 --> 01:44:44,160 Canada regional jet or something like that. 1327 01:44:44,200 --> 01:44:46,800 Watch this stuff and doing it. You. 1328 01:44:48,600 --> 01:44:54,920 Know, but I, yeah, I wanted to have a real Roman designing an 1329 01:44:55,240 --> 01:45:00,920 aircraft and odd enough it finally did happen in a very 1330 01:45:00,920 --> 01:45:08,040 direct way when we got to the Gulfstream G650 which was 1331 01:45:08,800 --> 01:45:17,840 designed around about 2006 so maybe I should mention that to 1332 01:45:17,840 --> 01:45:21,520 you so. Yes, please. 1333 01:45:22,120 --> 01:45:28,320 Dating back to, yes, when I was Douglas was using Flay 22, he 1334 01:45:28,320 --> 01:45:36,280 was Presheni was one of the really outstanding engineer who 1335 01:45:36,280 --> 01:45:39,480 kind of climbed the ladder at Douglas eventually was the 1336 01:45:39,720 --> 01:45:44,240 programme manager for MD-95 and things like that. 1337 01:45:44,240 --> 01:45:48,680 And I can't remember the MD 90 I think it was. 1338 01:45:48,680 --> 01:45:53,600 But then he became vice president of engineering at 1339 01:45:53,600 --> 01:46:00,240 Gulfstream and he was particularly interested. 1340 01:46:01,000 --> 01:46:10,760 Well, he developed the G5, but then early 2000s he was really 1341 01:46:10,760 --> 01:46:15,240 interested in developing a, a supersonic business jet. 1342 01:46:15,240 --> 01:46:22,400 He had, they had deals with the Russians, people like Sequoia 1343 01:46:22,400 --> 01:46:30,960 and so on. But so he actually asked me if I 1344 01:46:30,960 --> 01:46:37,560 would consult for Gulfstream and I said, yeah, but actually I 1345 01:46:37,560 --> 01:46:41,080 believe I can help you more with Transonics and his supersonic 1346 01:46:41,160 --> 01:46:48,400 business Jetman. And then around about 2006, they 1347 01:46:48,400 --> 01:46:55,120 decided to launch this G650, and Press wanted to have a plane 1348 01:46:55,120 --> 01:47:02,160 that could fly 7000 nautical miles at mark .85, but he also 1349 01:47:02,160 --> 01:47:06,160 wanted to fly 5000 miles at mark .9. 1350 01:47:08,800 --> 01:47:14,640 And they're having a problem because they're actually getting 1351 01:47:14,640 --> 01:47:18,880 about 3500 miles at mark .9 according to their first efforts 1352 01:47:18,880 --> 01:47:22,360 at designing it. The press sent me an e-mail 1353 01:47:22,360 --> 01:47:26,240 saying, well, would I like to try to to help that or I said 1354 01:47:26,320 --> 01:47:32,000 yes. So I started working with Bob 1355 01:47:32,000 --> 01:47:38,280 Mills, who was the aero lead for the G650 Australian tremendous 1356 01:47:38,680 --> 01:47:41,440 guy. And this interesting because it 1357 01:47:41,440 --> 01:47:44,880 really was an example of remote cooperation. 1358 01:47:46,800 --> 01:47:59,760 So I had this wing body shape optimization code SIM 107 and I 1359 01:47:59,760 --> 01:48:04,560 could run. I had a cluster in my garage in 1360 01:48:04,560 --> 01:48:06,560 this house actually. So I could run these 1361 01:48:07,000 --> 01:48:11,920 optimization calculations in a few hours. 1362 01:48:14,040 --> 01:48:24,760 And then I gave him an account on my cluster and he could pick 1363 01:48:24,760 --> 01:48:31,680 up the geometries from I drop with SIM 107 and he would then 1364 01:48:33,800 --> 01:48:37,240 put that geometry into a solver for the complete aircraft. 1365 01:48:37,240 --> 01:48:41,240 And they were using USM 3D so that, but I could actually come 1366 01:48:41,240 --> 01:48:47,280 up with optimised shapes faster than they could do the analysis 1367 01:48:47,360 --> 01:48:55,360 of Gulfstream. But eventually it took hundreds 1368 01:48:55,360 --> 01:48:57,680 of iterations because the question of course, exactly what 1369 01:48:57,680 --> 01:48:59,520 to optimise and what we were trying to. 1370 01:48:59,840 --> 01:49:04,520 Firstly, it was clear that the original green design was going 1371 01:49:04,520 --> 01:49:11,760 nowhere. So I told, I told him that 1372 01:49:11,840 --> 01:49:17,840 that's why don't I just put in my own wing sections. 1373 01:49:17,840 --> 01:49:21,040 I had sort of lots of wing sections that had developed in 1374 01:49:21,360 --> 01:49:24,200 different sort of studies over there and just start over. 1375 01:49:24,240 --> 01:49:28,840 And so he agreed to that. Then we found that we could get 1376 01:49:29,320 --> 01:49:34,960 a wing which basically didn't have drag rise to about mark .88 1377 01:49:37,320 --> 01:49:44,880 and could fly mark .9 if you came down to a lower lift 1378 01:49:44,880 --> 01:49:49,040 coefficient without a real drag rise or not much. 1379 01:49:51,120 --> 01:49:56,880 And essentially the wing was designed kind of semi 1380 01:49:56,880 --> 01:50:03,320 automatically by SIM 107. It's so there's a picture of the 1381 01:50:03,320 --> 01:50:07,440 two of us in I went to the first high speed wind tunnel test of 1382 01:50:07,440 --> 01:50:12,080 this plane. Yes, Sir, Lang is transonic 1383 01:50:12,520 --> 01:50:15,160 facility and the quadrenic tunnel. 1384 01:50:16,560 --> 01:50:20,080 There's a picture on my homepage that shows the two of us inside 1385 01:50:20,080 --> 01:50:21,960 the wind tunnel with the bottle wheel and plane. 1386 01:50:21,960 --> 01:50:25,720 And yes, when we actually did. That's amazing. 1387 01:50:26,560 --> 01:50:33,040 The actual high speed run, this quadrenic tunnel verified that 1388 01:50:33,240 --> 01:50:39,120 our predictions are corrected and it was a completely new 1389 01:50:39,120 --> 01:50:42,080 level of performance when we were long range business jet at 1390 01:50:42,080 --> 01:50:46,720 the time. So we left something on the 1391 01:50:46,720 --> 01:50:52,000 table because I wanted to have a highly slightly more sweep back 1392 01:50:52,000 --> 01:50:57,480 in the wing impression. He vetoed that because it was 1393 01:50:57,480 --> 01:51:01,200 the first plane the Gulfstream design was going to go fly by 1394 01:51:01,200 --> 01:51:05,120 well and he wasn't sure that they could handle the lateral 1395 01:51:05,120 --> 01:51:07,680 stability problems. If you had more feedback, and I 1396 01:51:07,680 --> 01:51:12,240 wouldn't have the same feedback as you see on a Cessna Citation 1397 01:51:12,240 --> 01:51:16,360 10. And I think we could have taken 1398 01:51:16,360 --> 01:51:25,800 it just a little bit further. That's the one time that I. 1399 01:51:25,800 --> 01:51:27,080 That's amazing. That was. 1400 01:51:27,320 --> 01:51:29,040 Designing a plane, basically, yeah, yeah. 1401 01:51:32,120 --> 01:51:36,480 But during that time, do you think back, you know, when you 1402 01:51:36,480 --> 01:51:42,040 were writing the code, you know, 2 megabytes of memory, you could 1403 01:51:42,040 --> 01:51:47,400 barely fit something on to then still within your career, 1404 01:51:48,120 --> 01:51:52,880 designing a whole aircraft and designing 10 in your garage. 1405 01:51:53,520 --> 01:51:57,400 Do you appreciate, does it still amaze you, the development of 1406 01:51:57,400 --> 01:51:59,960 code and computers over your lifetime? 1407 01:52:01,880 --> 01:52:10,240 Yeah, well, I think the thing was I always thought that you 1408 01:52:10,240 --> 01:52:13,000 had to look at what's feasible at any given time, given 1409 01:52:13,000 --> 01:52:14,960 computer resources. In fact, I'm going to tell you 1410 01:52:15,920 --> 01:52:20,120 one thing I learned in the British Army doing opposite 1411 01:52:20,120 --> 01:52:27,920 training. So they told us that in kind of 1412 01:52:27,920 --> 01:52:29,880 situation you have to make a plan, right? 1413 01:52:29,880 --> 01:52:32,760 So you said the first thing to do is to make an appreciation of 1414 01:52:32,760 --> 01:52:38,560 the situation. You know, the enemy has 10,000 1415 01:52:38,560 --> 01:52:41,520 troops and you have 100. So a frontal assault, for 1416 01:52:41,520 --> 01:52:44,000 example, is not going to work whatever, right? 1417 01:52:45,440 --> 01:52:52,320 And then you have to, to find a goal, let's say defeat the 1418 01:52:52,320 --> 01:52:56,160 enemy. But then you do a based on the 1419 01:52:56,160 --> 01:52:58,880 appreciation, you might have to refine the girl for something 1420 01:52:58,880 --> 01:53:04,760 that could conceivably succeed. So you iterate that a few times 1421 01:53:05,720 --> 01:53:09,800 and when you arrive at a girl that's potentially feasible, 1422 01:53:09,800 --> 01:53:13,080 then you draw up a plan and you carry it out. 1423 01:53:15,800 --> 01:53:21,560 So I would say that I kind of applied that kind of thinking. 1424 01:53:23,200 --> 01:53:33,560 It was feasible to do 2D airfoils in 197071 and then by 1425 01:53:33,560 --> 01:53:37,480 the time we got to the 1980s, you could do even 3D Euler 1426 01:53:37,480 --> 01:53:41,760 equations. And of course eventually, but I 1427 01:53:41,760 --> 01:53:45,600 had. So it took about 15 years to get 1428 01:53:45,600 --> 01:53:54,080 from 2D airfoil to 3D boiler solution for whole aircraft, but 1429 01:53:54,880 --> 01:53:57,080 very much paced. We still didn't have enough 1430 01:53:57,080 --> 01:54:02,480 memory in 198586 for the full airplane. 1431 01:54:02,800 --> 01:54:05,960 By the time we got to 1990, that was kind of beginning to be 1432 01:54:05,960 --> 01:54:09,320 fixed. But I mean, I could hardly 1433 01:54:09,320 --> 01:54:14,880 imagine the Rachel Fergus, I mean, this machine that I'm 1434 01:54:15,360 --> 01:54:26,840 talking to on it's about I think 32 gigabytes of core memory, TB 1435 01:54:27,320 --> 01:54:34,520 solid-state disk. And it also actually like 100 1436 01:54:34,520 --> 01:54:36,400 times faster than the Cray 1 was. 1437 01:54:38,720 --> 01:54:42,120 So we're living in a totally different world. 1438 01:54:42,120 --> 01:54:46,160 And now you move on to where do we go next? 1439 01:54:46,160 --> 01:54:51,040 Well, we still then have the solution to the turbulence. 1440 01:54:55,560 --> 01:54:58,880 We have this large area simulation, for example, and 1441 01:55:00,000 --> 01:55:08,360 that still seems to be a late deal open, still open about how 1442 01:55:08,360 --> 01:55:10,360 to set about. I mean, nowadays you have these 1443 01:55:10,360 --> 01:55:13,520 buzzwords like scale, resolving simulations and so on. 1444 01:55:15,480 --> 01:55:18,160 But I don't think anybody can calculate a flow. 1445 01:55:18,680 --> 01:55:21,560 Let's say it was submarine at rental, some intent on the 9th 1446 01:55:21,720 --> 01:55:29,240 in an accurate way. At the present time when I got 1447 01:55:29,240 --> 01:55:34,040 involved partly under due to pressure from funding agencies 1448 01:55:34,040 --> 01:55:42,000 and trying to work on high order methods in the last 15 years, I 1449 01:55:42,000 --> 01:55:49,720 think they can contribute. But I know that you're very 1450 01:55:49,720 --> 01:55:53,720 interested in machine learning approaches. 1451 01:55:58,640 --> 01:56:04,880 I did get to make one stab at looking at machine learning. 1452 01:56:05,880 --> 01:56:08,080 Some work we did just about last year. 1453 01:56:10,160 --> 01:56:13,960 I had to Stanford before I wound up taking this job at Texas A&M 1454 01:56:13,960 --> 01:56:17,040 of the Stanford kind of force me to return. 1455 01:56:17,600 --> 01:56:26,720 But we actually did write a paper with Jeremy Morton and 1456 01:56:27,480 --> 01:56:33,120 Freddie Witherman, who was then my postdoctor, Michael 1457 01:56:33,120 --> 01:56:38,600 Cochendorfer, who was the faculty at Stanford on machine 1458 01:56:38,600 --> 01:56:42,800 learning for trying to control vortex shedding officer under 1459 01:56:42,800 --> 01:56:47,080 the low Reynolds numbers. And we were using this. 1460 01:56:47,080 --> 01:56:49,840 I know of a so-called Koopman operator and I don't know 1461 01:56:49,840 --> 01:56:51,520 whether you're familiar with that or not. 1462 01:56:53,640 --> 01:56:59,640 But what Koopman did was to say well if you have a a non nano 1463 01:56:59,640 --> 01:57:09,680 process, let's say X superscript n + 1 equals some function of X 1464 01:57:09,680 --> 01:57:15,400 or north that's non needed. Then if you look at all possible 1465 01:57:16,560 --> 01:57:23,360 functions of that solution, so-called G of X, it turns out 1466 01:57:25,080 --> 01:57:32,800 that that's if you think of the update to the solution G when 1467 01:57:32,800 --> 01:57:34,680 you do this non needed operation. 1468 01:57:35,960 --> 01:57:38,560 But that's linear in the sense that if you have a one function 1469 01:57:38,560 --> 01:57:44,040 G1 and another one G2, a linear combination still works. 1470 01:57:44,040 --> 01:57:49,200 So what you do is you associate an infinitely dimensional linear 1471 01:57:49,200 --> 01:57:52,400 operator with a finite dimensional non linear operator. 1472 01:57:54,480 --> 01:57:58,800 And then infinitely dimensional operators might have continuous 1473 01:57:58,800 --> 01:58:02,520 spectrum. But it could be that there might 1474 01:58:02,520 --> 01:58:09,800 be an invariant subset where anything in that subset in the 1475 01:58:09,800 --> 01:58:12,080 initial closure subset stays in the subset. 1476 01:58:12,840 --> 01:58:16,320 Sure is non linear updates. And if you can detect a 1477 01:58:20,680 --> 01:58:24,640 invariant subspace, it might have discrete eigenvalues and 1478 01:58:24,640 --> 01:58:27,400 eigenfunctions and that can be used as a model. 1479 01:58:27,400 --> 01:58:30,000 And we think that vortex shedding is an example of 1480 01:58:30,360 --> 01:58:32,280 something where that might happen. 1481 01:58:33,360 --> 01:58:38,680 We did manage to come up with a control system based on that 1482 01:58:38,680 --> 01:58:44,280 that kind of worked and got into this nips neuro. 1483 01:58:44,320 --> 01:58:52,520 Oh yeah. It has quite a few citations 1484 01:58:52,520 --> 01:58:57,600 already, but when I got to Texas A and MI couldn't find anybody 1485 01:58:58,120 --> 01:58:59,920 on the machine learning side to work with. 1486 01:58:59,920 --> 01:59:01,880 And it's sort of by the wayside it. 1487 01:59:02,680 --> 01:59:09,320 Well, I, I do think that somebody, maybe you will succeed 1488 01:59:09,320 --> 01:59:12,600 in developing some kind of giant neural network that will 1489 01:59:13,080 --> 01:59:18,240 probably predict turbulent flows reasonably well within a few 1490 01:59:18,240 --> 01:59:20,720 years. Maybe not. 1491 01:59:20,720 --> 01:59:27,160 While I'm around the trouble, assuming we'd have some giant 1492 01:59:27,280 --> 01:59:32,400 neural network like these large language models of it could 1493 01:59:32,600 --> 01:59:38,000 predict a turbulent separated flow, then the trouble is you 1494 01:59:38,240 --> 01:59:40,720 might not get any understanding out of it. 1495 01:59:40,720 --> 01:59:43,360 I mean, you can run the neural network and it gives you some 1496 01:59:43,360 --> 01:59:45,360 numbers. It doesn't really give you an 1497 01:59:45,360 --> 01:59:47,600 understanding of what's going on necessarily. 1498 01:59:48,600 --> 01:59:53,000 So that's my. Yes, that is the. 1499 01:59:54,600 --> 01:59:57,040 I don't know if you've come across this, but at least what I 1500 01:59:57,040 --> 02:00:02,240 was told and for, yeah, as you know, one of the dreams for a 1501 02:00:02,240 --> 02:00:06,160 lot of these aerospace companies is digital certification, being 1502 02:00:06,160 --> 02:00:09,440 able to actually get a, you know, the approval to certify. 1503 02:00:09,440 --> 02:00:15,200 But you have to understand how the code and the method has got 1504 02:00:15,200 --> 02:00:18,520 to the answer. And I guess the risk with the 1505 02:00:18,520 --> 02:00:21,640 machine learning is it may be able to get the answer, but if 1506 02:00:21,640 --> 02:00:23,840 you don't understand exactly how it did it. 1507 02:00:24,360 --> 02:00:29,240 Then it's probably not going to be suitable for a certification. 1508 02:00:30,280 --> 02:00:32,840 So that still is maybe a challenge. 1509 02:00:33,600 --> 02:00:37,120 I think that's true. But of course there are. 1510 02:00:37,120 --> 02:00:46,960 I mean, there are people like this, Stephen Branton at Nathan 1511 02:00:46,960 --> 02:00:49,680 Goods at University of Washington, who've made a lot of 1512 02:00:49,680 --> 02:00:54,280 mind about detecting PDS based on just looking at data such 1513 02:00:54,280 --> 02:00:59,440 like, but I don't really, well, I don't know, I'm not too 1514 02:00:59,440 --> 02:01:06,400 worried because I think I don't really feel that I'm going to be 1515 02:01:06,400 --> 02:01:09,040 replaced by some kind of neural network while I'm alive. 1516 02:01:14,240 --> 02:01:17,640 Problem with the next generation to grapple with it. 1517 02:01:20,680 --> 02:01:24,880 So what would be your, what would be maybe your advice? 1518 02:01:25,120 --> 02:01:29,240 So if, if somebody now is listening or watching to this in 1519 02:01:29,240 --> 02:01:33,920 their 20s or their 30s or, or really at any age and we're 1520 02:01:33,920 --> 02:01:36,920 going to, you know, wanted to make an impact in the field of 1521 02:01:36,920 --> 02:01:40,200 aeronautics or CFD. Is there any things that you've 1522 02:01:40,200 --> 02:01:43,680 learnt along the way in terms of your mentality, the way you 1523 02:01:43,680 --> 02:01:45,640 approach things? You, you talked about the 1524 02:01:45,640 --> 02:01:50,480 British Army example, but is there any sort of yeah, advice 1525 02:01:50,480 --> 02:01:57,280 you would give to people? Well, I think really one of the 1526 02:01:57,280 --> 02:02:02,280 problems is that there are an awful lot of people doing 1527 02:02:03,400 --> 02:02:09,760 doctoral studies in CFD at the moment and very hard to stand 1528 02:02:09,760 --> 02:02:15,320 out from the crowd. And I think what I would say is, 1529 02:02:15,320 --> 02:02:20,600 in principle, if the crowd is going one way, you need to go in 1530 02:02:20,600 --> 02:02:23,160 a different direction. But of course, you've got to 1531 02:02:23,160 --> 02:02:26,280 find a different direction that actually works, right. 1532 02:02:26,280 --> 02:02:28,400 Because if you go the same direction as everybody else, 1533 02:02:28,400 --> 02:02:30,480 you're going to do the same thing as everybody else, and 1534 02:02:30,880 --> 02:02:32,680 nobody's going to care all that much. 1535 02:02:34,280 --> 02:02:38,000 So the trick is to find a different direction, which, if 1536 02:02:38,000 --> 02:02:41,760 you get it to work, will change the direction of the field. 1537 02:02:42,520 --> 02:02:45,560 But that's not easy. Maybe. 1538 02:02:46,720 --> 02:02:50,640 Yeah. Well, as matter of luck also 1539 02:02:50,640 --> 02:02:56,680 getting into that just I think in my case I arrived at 1540 02:02:56,880 --> 02:03:01,080 shantonic fluid at the moment when there was a real 1541 02:03:01,080 --> 02:03:06,160 opportunity to do something and have an impact and kind of hard. 1542 02:03:06,160 --> 02:03:09,200 Now. I mean I can tell you that as a 1543 02:03:09,200 --> 02:03:18,920 practical problem, we have all these electric vitae things like 1544 02:03:18,920 --> 02:03:25,680 the joby and they have to do a transition from hovering to 1545 02:03:25,680 --> 02:03:29,880 forward flight typically by tilting their oceans over. 1546 02:03:31,400 --> 02:03:34,360 And we're way short of being able to calculate that by any 1547 02:03:34,360 --> 02:03:36,320 numerical method at the present time. 1548 02:03:36,320 --> 02:03:42,800 And seems to me so if that's somebody could really solve that 1549 02:03:42,800 --> 02:03:46,080 problem, we would have a pretty big impact. 1550 02:03:50,640 --> 02:03:54,640 So tackle the hard problems, I guess you're saying look for the 1551 02:03:54,640 --> 02:03:59,360 disruptive ideas, the difficult problems, and if you can solve 1552 02:03:59,360 --> 02:04:02,200 it, you will have an impact. Rather than taking the easy 1553 02:04:02,200 --> 02:04:06,680 incremental work, go for the, go for the more challenging things 1554 02:04:06,680 --> 02:04:10,520 that yeah. It would also worthwhile, 1555 02:04:11,240 --> 02:04:13,040 actually, we're willing to say. All right. 1556 02:04:15,600 --> 02:04:19,160 Well, firstly, of course, you know, the whole issue of funding 1557 02:04:19,160 --> 02:04:23,760 is, I mean a lot of universities specifically require people to 1558 02:04:24,960 --> 02:04:27,640 to bring in money. Texas, their name, they don't 1559 02:04:27,920 --> 02:04:32,840 even have have prizes for people who bring in more than $1 1560 02:04:32,840 --> 02:04:34,280 million of the stuff or the other. 1561 02:04:34,280 --> 02:04:40,800 But The thing is, if you solve a research problem, then that 1562 02:04:41,160 --> 02:04:42,760 basically determines the funding. 1563 02:04:45,040 --> 02:04:48,680 If you were to solve it, why should they continue to fund it? 1564 02:04:48,680 --> 02:04:50,560 All right. Yeah. 1565 02:04:50,720 --> 02:04:54,880 So a lot of people have a sort of strategy that they never 1566 02:04:54,880 --> 02:04:57,160 solve a problem. They're always light at the end 1567 02:04:57,160 --> 02:05:00,800 of the tunnel where you might begin to solve it and another 1568 02:05:00,800 --> 02:05:02,480 couple of beers, you keep on funding it. 1569 02:05:04,440 --> 02:05:08,000 What's amazing is you should go ahead and solve it and move on 1570 02:05:08,000 --> 02:05:13,160 to something else. But then yeah, the trick there 1571 02:05:13,160 --> 02:05:23,960 is all the work I've done that had maximum impact was never 1572 02:05:23,960 --> 02:05:28,240 actually directly funded. At the time I did it, we didn't 1573 02:05:28,240 --> 02:05:32,760 have a grant relevant tetrahedral mesh solvent for the 1574 02:05:32,760 --> 02:05:36,040 Euler equations that we wrote that airplane code. 1575 02:05:36,720 --> 02:05:39,160 We had grants and Nassilagni and other places. 1576 02:05:40,920 --> 02:05:42,880 All right. Well, it wasn't to do that. 1577 02:05:44,000 --> 02:05:48,160 And I didn't have any grant to do our dynamic design by control 1578 02:05:48,160 --> 02:05:51,360 theory. I got support later on from the 1579 02:05:51,360 --> 02:05:56,200 Air Force, but at the time I did it, it was not funded. 1580 02:05:58,320 --> 02:06:02,960 So there's something I think I learned at Grumman a little bit. 1581 02:06:03,560 --> 02:06:08,480 So my general approach at Grumman was that I always did 1582 02:06:08,480 --> 02:06:12,400 whatever they asked me to do, even if I thought it wasn't 1583 02:06:12,440 --> 02:06:18,840 necessary or that useful. But I could do it fast enough 1584 02:06:19,960 --> 02:06:24,200 that I could still spend half the time, or maybe even more 1585 02:06:24,200 --> 02:06:26,600 than half my time doing something that I thought was 1586 02:06:27,120 --> 02:06:31,880 actually going to work. But I could both so. 1587 02:06:33,240 --> 02:06:35,040 That's a good. I think it's a good tip. 1588 02:06:35,760 --> 02:06:38,440 So be be fast so you can do the other stuff as well. 1589 02:06:39,000 --> 02:06:43,040 Yeah, Yeah, exactly. So in fact, if you look at sort 1590 02:06:43,040 --> 02:06:45,600 of young faculty, I mean, what you should do is, all right, 1591 02:06:45,600 --> 02:06:52,240 you, you have a project that's safe when you've got funding. 1592 02:06:53,560 --> 02:06:58,240 But now do that with say half the time and spend the other 1593 02:06:58,240 --> 02:07:01,560 half trying something that's high risk and way out there. 1594 02:07:01,560 --> 02:07:03,600 And if it doesn't work, you haven't lost anything. 1595 02:07:03,600 --> 02:07:05,480 If it does work, more than a big impact. 1596 02:07:07,080 --> 02:07:07,840 Yeah. No, that's good. 1597 02:07:08,160 --> 02:07:13,280 That is very good. That is, that is good advice. 1598 02:07:14,960 --> 02:07:18,480 Well, I am, I want to be appreciative of your time. 1599 02:07:18,480 --> 02:07:22,640 I I really do Thank you so much. And I think people will be able 1600 02:07:22,640 --> 02:07:26,680 to tell that your journey as you said, I think that was a very 1601 02:07:26,680 --> 02:07:30,840 interesting point. I would argue that if you had 1602 02:07:30,840 --> 02:07:37,280 done what you did shifted by five years, your code may well 1603 02:07:37,280 --> 02:07:41,000 have been the fluent, the star CD of the world. 1604 02:07:41,080 --> 02:07:44,600 You know, I think the fact that you did a lot of your work, I 1605 02:07:44,600 --> 02:07:49,240 guess just before the boom in CFD, but that does not mean that 1606 02:07:49,240 --> 02:07:55,120 your codes were not in some ways massively the, the influence 1607 02:07:55,120 --> 02:07:57,800 for, for the practically the codes that people use today. 1608 02:07:58,800 --> 02:08:02,240 So I think probably everybody who does CFD today, particularly 1609 02:08:02,240 --> 02:08:05,720 in the aerospace sector owes a lot of what they do to what you 1610 02:08:05,720 --> 02:08:07,560 did. And that, you know, I think 1611 02:08:07,560 --> 02:08:12,520 everybody should be and, and I'm sure are very thankful for all 1612 02:08:12,520 --> 02:08:14,920 your work. And of course, all the, I didn't 1613 02:08:14,920 --> 02:08:20,160 mention it, but you have supervised so many students who 1614 02:08:20,160 --> 02:08:24,080 have now gone on to be leading professors and engineers in 1615 02:08:24,080 --> 02:08:27,280 their field who I know, who are appreciative of you. 1616 02:08:27,280 --> 02:08:32,480 And probably by the time this comes out, you will have reached 1617 02:08:32,480 --> 02:08:35,280 your 90th birthday. So I also wanted to say happy 1618 02:08:35,280 --> 02:08:37,880 birthday. Maybe you're not now, but by the 1619 02:08:37,880 --> 02:08:40,560 time it comes out, it will be. So I hope that when people are 1620 02:08:40,560 --> 02:08:44,600 listening to this, they will also say the same thing. 1621 02:08:44,600 --> 02:08:45,400 But. Thank. 1622 02:08:45,720 --> 02:08:48,040 Thank you. One other thing I'm going to 1623 02:08:48,040 --> 02:08:52,360 mention, you know, I, I didn't finally write a book and I'm not 1624 02:08:53,040 --> 02:08:55,480 completely satisfied with the book because it's got an awful 1625 02:08:55,480 --> 02:09:00,560 lot of errors of typos and things like that in it. 1626 02:09:01,000 --> 02:09:03,680 Somehow trying to straighten out later. 1627 02:09:03,680 --> 02:09:07,240 I can get the same fonts all over the place, wound up 1628 02:09:07,680 --> 02:09:10,240 introducing mistakes as fast as trying to correct them. 1629 02:09:10,240 --> 02:09:13,440 But still, the book's fairly comprehensive. 1630 02:09:13,520 --> 02:09:22,040 And I, when I look back, Horace Lamb's Agent Dynamics book, it's 1631 02:09:22,040 --> 02:09:24,080 still useful. You know, it was actually 1632 02:09:24,200 --> 02:09:29,400 written in the 19th century. But if you still some aspects of 1633 02:09:30,880 --> 02:09:36,240 aerodynamic theory for at least for potential flair were all LED 1634 02:09:36,240 --> 02:09:42,040 out in Lamb's book. So whether I've done will 1635 02:09:42,040 --> 02:09:45,320 actually have any be of any interest 100 years from now, 1636 02:09:45,320 --> 02:09:49,720 it's not at all clear to me. But it would be nice if I was 1637 02:09:49,800 --> 02:09:55,680 sure, yeah. I'm sure it will be and I'll put 1638 02:09:55,680 --> 02:09:58,560 a link to your It's computational aerodynamics, 1639 02:09:58,560 --> 02:10:00,880 isn't it? I think that's the book from 1640 02:10:01,160 --> 02:10:03,080 Cambridge University Prep. Yeah, I just got it. 1641 02:10:03,400 --> 02:10:09,160 I'll put a link into the. Regrettable number of kind of 1642 02:10:09,160 --> 02:10:13,400 typos and things like that still, but it's, it does try to 1643 02:10:13,400 --> 02:10:19,320 lay out, you know, a kind of theoretical analysis of CFD 1644 02:10:19,320 --> 02:10:25,000 methods, not just practical how you arrive at schemes. 1645 02:10:25,720 --> 02:10:28,440 No, yeah. Building quite heavily on. 1646 02:10:28,440 --> 02:10:34,000 And people should those books, yeah. 1647 02:10:34,000 --> 02:10:37,000 And those books are great for people to get a whole. 1648 02:10:37,640 --> 02:10:38,680 Yeah. As you say, there's some 1649 02:10:38,680 --> 02:10:41,720 textbooks that people read that influence what they did. 1650 02:10:41,720 --> 02:10:44,160 And I'm sure the same is true for yours. 1651 02:10:44,160 --> 02:10:46,800 So yes, I'm. I'm sure people will be looking 1652 02:10:46,800 --> 02:10:49,880 at that in 100 years and all the other papers that you've. 1653 02:10:50,600 --> 02:10:57,960 You will. Well, thank you again. 1654 02:10:57,960 --> 02:10:59,640 Like I said, I really appreciate it. 1655 02:10:59,680 --> 02:11:02,920 And yeah, it was a pleasure, a pleasure to speak to you all. 1656 02:11:04,400 --> 02:11:05,880 Right. Well, thank you. 1657 02:11:06,000 --> 02:11:07,280 I've enjoyed it actually.