WEBVTT 00:00:00.000 --> 00:00:03.020 Welcome back to the Deep Dive. Our mission today 00:00:03.020 --> 00:00:06.179 is, well, it's pretty profound when you think 00:00:06.179 --> 00:00:08.460 about it. We're pulling back the curtain on the 00:00:08.460 --> 00:00:11.380 very architecture of our digital world. And when 00:00:11.380 --> 00:00:13.880 you look closely, you find one person's work 00:00:13.880 --> 00:00:15.839 is just, I mean, it's absolutely everywhere. 00:00:16.160 --> 00:00:18.440 Exactly. We're talking about an American computer 00:00:18.440 --> 00:00:21.699 scientist who is, and I don't think this is an 00:00:21.699 --> 00:00:24.519 exaggeration, quite possibly the most influential 00:00:24.519 --> 00:00:26.980 technologist you've never heard of. He's just 00:00:26.980 --> 00:00:29.980 completely under the radar for most people. It's... 00:00:30.510 --> 00:00:33.390 It's astonishing, really, how central his role 00:00:33.390 --> 00:00:36.149 was. The code he wrote, the concepts he developed, 00:00:36.350 --> 00:00:39.789 they're essential to literally every device you 00:00:39.789 --> 00:00:41.789 use. The phone in your pocket. The phone in your 00:00:41.789 --> 00:00:43.770 pocket, the network that's delivering this audio 00:00:43.770 --> 00:00:46.289 to you right now, the servers hosting the files, 00:00:46.369 --> 00:00:48.270 the operating systems that run Wall Street. It 00:00:48.270 --> 00:00:50.869 all goes back to him. His name is Dennis McAllister 00:00:50.869 --> 00:00:54.509 Ritchie. He lived from 1941 to 2011. So the big 00:00:54.509 --> 00:00:56.090 question we're trying to tackle in this deep 00:00:56.090 --> 00:00:58.429 dive is to really understand the mechanics of 00:00:58.429 --> 00:01:03.030 that influence. one. a relatively quiet researcher 00:01:03.030 --> 00:01:05.609 working almost his entire career at Bell Labs, 00:01:05.790 --> 00:01:08.750 co -create the Unix operating system, and maybe 00:01:08.750 --> 00:01:11.170 even more importantly, single -handedly engineer 00:01:11.170 --> 00:01:14.150 the C programming language. And those two things 00:01:14.150 --> 00:01:17.090 together, Unix and C, they really do form the 00:01:17.090 --> 00:01:20.750 absolute essential DNA of the entire modern digital 00:01:20.750 --> 00:01:23.430 age. It's not a stretch to say that. It really 00:01:23.430 --> 00:01:25.650 isn't. So to unpack all this, we've pulled together 00:01:25.650 --> 00:01:27.769 a lot of material. We've got a comprehensive 00:01:27.769 --> 00:01:30.409 biographical study on his career, some internal... 00:01:30.409 --> 00:01:32.510 histories from Bell Labs, academic papers that 00:01:32.510 --> 00:01:34.950 break down the architecture of sea in Unix, and 00:01:34.950 --> 00:01:37.890 the official citations from the major awards 00:01:37.890 --> 00:01:39.829 he received. Right. And our goal here is to get 00:01:39.829 --> 00:01:42.530 past just the facts, the dates and names, and 00:01:42.530 --> 00:01:44.569 really understand the conceptual breakthroughs 00:01:44.569 --> 00:01:47.409 that made his work, well, immortal. Okay, let's 00:01:47.409 --> 00:01:49.829 do it. Let's start right at the foundation. Dennis 00:01:49.829 --> 00:01:53.310 Ritchie, born September 9, 1941, in Bronxville, 00:01:53.349 --> 00:01:55.790 New York. His professional destiny, it seems, 00:01:55.989 --> 00:01:58.390 was almost predetermined. It really was. You 00:01:58.390 --> 00:02:00.129 can almost call it a generational mandate. And 00:02:00.129 --> 00:02:02.370 that's a key detail that, you know, sometimes 00:02:02.370 --> 00:02:04.469 gets lost when we talk about this myth of the 00:02:04.469 --> 00:02:06.549 solitary genius. But never that simple, is it? 00:02:06.629 --> 00:02:09.889 Never. His father, Alistair E. Ritchie, was a 00:02:09.889 --> 00:02:12.729 longtime Bell Labs scientist himself, a real 00:02:12.729 --> 00:02:15.449 expert in switching circuit theory. He even co 00:02:15.449 --> 00:02:17.310 -authored a pretty significant textbook, The 00:02:17.310 --> 00:02:20.050 Design of Switching Circuits. Wow. So Dennis 00:02:20.050 --> 00:02:22.509 grew up with this deep... tangible connection 00:02:22.509 --> 00:02:25.509 to that world of foundational research, the kind 00:02:25.509 --> 00:02:28.050 of long -term, deep thinking that Bell Labs was 00:02:28.050 --> 00:02:30.349 famous for. Exactly. He wasn't just walking into 00:02:30.349 --> 00:02:33.150 an office for a job. He was stepping into a family 00:02:33.150 --> 00:02:36.349 legacy, a tradition of deep scientific inquiry. 00:02:36.930 --> 00:02:39.250 So after high school in New Jersey, he heads 00:02:39.250 --> 00:02:42.389 to Harvard. And his degrees are in physics and 00:02:42.389 --> 00:02:44.740 applied mathematics. Which, when you think about 00:02:44.740 --> 00:02:46.800 it, is the perfect intellectual tool set for 00:02:46.800 --> 00:02:49.580 inventing a system language. How so? Well, physics 00:02:49.580 --> 00:02:51.780 gives you this rigorous understanding of, you 00:02:51.780 --> 00:02:53.819 know, the underlying reality, the hard constraints 00:02:53.819 --> 00:02:56.840 of the universe. And applied math gives you the 00:02:56.840 --> 00:02:59.360 tools to model and solve real world problems 00:02:59.360 --> 00:03:01.960 within those constraints. It's the what is and 00:03:01.960 --> 00:03:03.659 the what to do with it. That's a great way to 00:03:03.659 --> 00:03:06.930 put it. That mix is absolutely crucial for building 00:03:06.930 --> 00:03:09.050 software that needs to be efficient, elegant, 00:03:09.189 --> 00:03:12.849 and mathematically sound. But there's a fascinating 00:03:12.849 --> 00:03:15.830 academic footnote here that I think speaks volumes 00:03:15.830 --> 00:03:18.810 about the environment at Bell Labs versus, say, 00:03:18.849 --> 00:03:22.310 traditional academia. Ah, yes. The famous missing 00:03:22.310 --> 00:03:25.150 Ph .D. Right. So he completes a draft of his 00:03:25.150 --> 00:03:28.530 Ph .D. thesis at Harvard in 1968. It's titled 00:03:28.530 --> 00:03:31.189 Computational Complexity and Program Structure. 00:03:31.550 --> 00:03:33.930 The work is done. The work is absolutely done. 00:03:34.050 --> 00:03:36.250 His supervisor was Patrick C. Fisher. We have 00:03:36.250 --> 00:03:38.909 the records. But the source material is very 00:03:38.909 --> 00:03:41.490 clear. He never officially received the degree. 00:03:41.750 --> 00:03:44.389 So wait. This is the genius who's about to rewrite 00:03:44.389 --> 00:03:46.629 the rules of computing. And he couldn't be bothered 00:03:46.629 --> 00:03:48.949 to file the final paperwork. What does that tell 00:03:48.949 --> 00:03:51.009 you about the priorities at Bell Labs in the 00:03:51.009 --> 00:03:53.569 late 60s? It tells you everything you need to 00:03:53.569 --> 00:03:55.729 know. The work he was doing was so critical, 00:03:55.810 --> 00:03:58.629 so immediate, and so groundbreaking that chasing 00:03:58.629 --> 00:04:01.430 a formal academic credential just became secondary. 00:04:01.770 --> 00:04:04.229 He joined the Bell Labs Computing Science Research 00:04:04.229 --> 00:04:08.750 Center in 1967. By 68, he was already deep into 00:04:08.750 --> 00:04:11.229 projects that were moving way too fast for university 00:04:11.229 --> 00:04:14.150 timelines. The innovation was just outpacing 00:04:14.150 --> 00:04:16.790 the bureaucracy. Completely. The urgency of what 00:04:16.790 --> 00:04:19.250 they were building simply superseded the formality 00:04:19.250 --> 00:04:21.709 of getting the degree. I mean, we only know about 00:04:21.709 --> 00:04:24.050 this because in 2020, researchers found a copy 00:04:24.050 --> 00:04:26.829 of that lost dissertation. The ideas were proven. 00:04:26.910 --> 00:04:29.829 The formal title just wasn't necessary. His attention 00:04:29.829 --> 00:04:32.230 was already on a much bigger, much more practical 00:04:32.230 --> 00:04:34.850 problem. And that problem brings us right into 00:04:34.850 --> 00:04:37.519 our first section. the context at Bell Labs, 00:04:37.740 --> 00:04:40.560 and the genesis of Unix. And it's funny, we often 00:04:40.560 --> 00:04:42.920 talk about innovation, but sometimes failure 00:04:42.920 --> 00:04:45.500 is the best catalyst. And the catalyst for Unix 00:04:45.500 --> 00:04:48.040 was a massive, ambitious failure called Multics. 00:04:48.180 --> 00:04:50.100 That is the perfect place to start the story. 00:04:51.000 --> 00:04:53.540 Multics, which stands for Multiplexed Information 00:04:53.540 --> 00:04:56.600 and Computing Service, was the dominant project 00:04:56.600 --> 00:04:59.079 that Ritchie and his colleague Ken Thompson were 00:04:59.079 --> 00:05:01.800 working on in the mid -60s. This was the moonshot 00:05:01.800 --> 00:05:03.600 of operating systems at the time. It was the 00:05:03.600 --> 00:05:06.800 moonshot. A huge, multi -user, time -sharing 00:05:06.800 --> 00:05:10.379 utility designed for mainframes. It was a massive 00:05:10.379 --> 00:05:13.519 consortium project with Bell Labs, MIT, and General 00:05:13.519 --> 00:05:16.160 Electric. It was supposed to be the ultimate 00:05:16.160 --> 00:05:18.959 operating system. So what went wrong? Too many 00:05:18.959 --> 00:05:21.870 cooks in the kitchen. Too much ambition. Precisely. 00:05:22.129 --> 00:05:24.949 Multics became a sprawling, overly complicated 00:05:24.949 --> 00:05:28.430 beast. It just demanded enormous resources, both 00:05:28.430 --> 00:05:31.490 in hardware and, frankly, in human effort. By 00:05:31.490 --> 00:05:33.990 the late 60s, Bell Labs realized the complexity 00:05:33.990 --> 00:05:36.230 and the cost were just way out of line with the 00:05:36.230 --> 00:05:38.449 returns. It was too unwieldy for the tech they 00:05:38.449 --> 00:05:40.490 had at the time. Exactly. So they made the tough, 00:05:40.629 --> 00:05:42.949 but as it turns out, incredibly transformative 00:05:42.949 --> 00:05:45.329 decision to pull out of the Multics project completely. 00:05:45.930 --> 00:05:48.569 So Bell Labs is out, but Thompson and Ritchie 00:05:48.569 --> 00:05:51.370 are left with this intellectual itch. They still 00:05:51.370 --> 00:05:53.470 want to build a time -sharing system, but one 00:05:53.470 --> 00:05:56.350 that's simple, clean, and, you know, actually 00:05:56.350 --> 00:05:58.910 works. Yes, and this is where that famous anecdote 00:05:58.910 --> 00:06:01.629 comes in, the one about the old hardware. Ken 00:06:01.629 --> 00:06:03.730 Thompson, just determined to keep working on 00:06:03.730 --> 00:06:06.769 this, found an old, basically forgotten mini 00:06:06.769 --> 00:06:09.860 -computer kicking around the lab, a PDP -7. And 00:06:09.860 --> 00:06:11.860 compared to the mainframes they were using for 00:06:11.860 --> 00:06:14.600 Multics, this thing was a toy. An absolute toy. 00:06:14.660 --> 00:06:17.279 It had very limited processing power, tiny memory. 00:06:17.720 --> 00:06:20.839 But that constraint, that limited machine, it 00:06:20.839 --> 00:06:23.839 forced simplicity. It forced elegance. There 00:06:23.839 --> 00:06:26.720 was no room for complexity. None. Thompson decided 00:06:26.720 --> 00:06:29.120 to just start over. He'd develop a new operating 00:06:29.120 --> 00:06:31.519 system and new programs from scratch on this 00:06:31.519 --> 00:06:33.819 machine. And Ritchie and others quickly joined 00:06:33.819 --> 00:06:36.699 in to help. That shift from this huge bureaucratic 00:06:36.699 --> 00:06:39.980 project to a small, nimble team focused on one 00:06:39.980 --> 00:06:42.800 restrictive piece of hardware. That's the real 00:06:42.800 --> 00:06:45.319 origin story. They had to be smart, efficient, 00:06:45.459 --> 00:06:48.279 and just ruthless in their minimalism. And of 00:06:48.279 --> 00:06:50.720 course, the name itself is pure Bell Labs humor. 00:06:50.860 --> 00:06:52.959 A pun on the very system they had just abandoned. 00:06:53.399 --> 00:06:55.720 That's right. Multics, you know, implies many 00:06:55.720 --> 00:06:59.199 complexity. In 1970, another key figure, Brian 00:06:59.199 --> 00:07:01.980 Kernighan, suggested the name Unix, which was 00:07:01.980 --> 00:07:04.279 a bit of a self -deprecating pun. It implied 00:07:04.279 --> 00:07:06.759 the system was the opposite of Multics. Simpler, 00:07:06.899 --> 00:07:09.319 streamlined, focused on one thing at a time. 00:07:09.500 --> 00:07:11.920 The power of that collaboration, Richie and Thompson. 00:07:12.839 --> 00:07:15.500 It just can't be overstated. Doug McElroy, who 00:07:15.500 --> 00:07:17.399 was also there, summed it up perfectly. He wrote 00:07:17.399 --> 00:07:20.079 that the names of Ritchie and Thompson may safely 00:07:20.079 --> 00:07:22.620 be assumed to be attached to almost everything 00:07:22.620 --> 00:07:25.220 not otherwise attributed. They were the core 00:07:25.220 --> 00:07:27.899 architects, absolutely. Thomson built that initial 00:07:27.899 --> 00:07:30.819 OS structure, but Ritchie was already tackling 00:07:30.819 --> 00:07:33.019 the huge fundamental problem that could have 00:07:33.019 --> 00:07:35.040 killed the whole project. The language problem. 00:07:35.100 --> 00:07:37.540 The language problem. Thomson was initially working 00:07:37.540 --> 00:07:40.379 in assembly language, which is incredibly cumbersome 00:07:40.379 --> 00:07:43.420 and tied to one specific machine. To make Unix 00:07:43.420 --> 00:07:45.680 what it needed to be, flexible and portable, 00:07:46.019 --> 00:07:48.500 it couldn't just live in assembly. Which brings 00:07:48.500 --> 00:07:51.040 us naturally to the precursor language. Yeah. 00:07:51.100 --> 00:07:54.019 B. This was Thompson's first attempt to move 00:07:54.019 --> 00:07:56.699 beyond the brutal inefficiency of assembly. Right. 00:07:56.939 --> 00:07:59.079 Assembly language is basically talking directly 00:07:59.079 --> 00:08:01.920 to the hardware. It's fast, but it's a nightmare 00:08:01.920 --> 00:08:04.040 to write for large systems. And if you get a 00:08:04.040 --> 00:08:05.620 new computer, you have to rewrite everything. 00:08:05.899 --> 00:08:08.120 So Thompson needed a higher level language. A 00:08:08.120 --> 00:08:10.800 slightly higher level language. Yeah. He created 00:08:10.800 --> 00:08:13.220 B, which was based on an earlier language called 00:08:13.220 --> 00:08:16.339 BCPL. It was meant to give them some structure, 00:08:16.459 --> 00:08:19.160 some speed and development while still being 00:08:19.160 --> 00:08:21.560 low level enough for an operating system. So 00:08:21.560 --> 00:08:25.430 B solved the how fast can we write code? problem, 00:08:25.470 --> 00:08:28.329 but it created new structural problems as Unix 00:08:28.329 --> 00:08:31.600 grew. Why wasn't B good enough? This is a really 00:08:31.600 --> 00:08:33.840 crucial technical point for understanding why 00:08:33.840 --> 00:08:37.500 C was so necessary. B was extremely minimalist, 00:08:37.600 --> 00:08:40.179 which was great for that tiny PDP -7 machine, 00:08:40.419 --> 00:08:43.059 but it had serious limitations when Unix started 00:08:43.059 --> 00:08:44.779 to grow and they got a more powerful machine, 00:08:44.919 --> 00:08:47.059 the PDP -11. What were those limitations? Two 00:08:47.059 --> 00:08:50.320 big ones. First, B was typeless. Everything was 00:08:50.320 --> 00:08:52.340 just treated as a machine word, a chunk of bits. 00:08:52.580 --> 00:08:55.360 This made handling complex data, and especially 00:08:55.360 --> 00:08:57.559 defining the internal data structures of the 00:08:57.559 --> 00:09:00.159 OS, like file information or memory tables, really 00:09:00.159 --> 00:09:02.120 hard. clumsy so it was hard to keep the system's 00:09:02.120 --> 00:09:05.460 data organized exactly but the second and maybe 00:09:05.460 --> 00:09:08.779 bigger problem was that b couldn't handle a key 00:09:08.779 --> 00:09:12.419 feature of the new pdp 11 byte addressing and 00:09:12.419 --> 00:09:15.000 what's that the pdp 11 could address individual 00:09:15.000 --> 00:09:17.879 bytes of memory not just whole words this is 00:09:17.879 --> 00:09:19.700 absolutely essential for working efficiently 00:09:19.700 --> 00:09:22.000 with things like characters and text strings 00:09:22.000 --> 00:09:24.980 b was just incapable of using this fundamental 00:09:24.980 --> 00:09:28.590 feature of the hardware so b was a dead end Richie 00:09:28.590 --> 00:09:31.370 recognized it immediately. If Unix was going 00:09:31.370 --> 00:09:34.210 to evolve and be ported to new machines, B wasn't 00:09:34.210 --> 00:09:36.730 going to cut it. He needed a new language, one 00:09:36.730 --> 00:09:39.409 that kept B's elegance, but added the necessary 00:09:39.409 --> 00:09:41.750 structures and the ability to talk directly and 00:09:41.750 --> 00:09:44.009 efficiently to the hardware. That need right 00:09:44.009 --> 00:09:47.110 there was what gave birth to C. So Thompson starts 00:09:47.110 --> 00:09:49.649 the OS, but it's Richie who builds the architectural 00:09:49.649 --> 00:09:52.690 tool, the language, that guarantees its survival 00:09:52.690 --> 00:09:55.590 and its future. That transition completely changed 00:09:55.590 --> 00:09:57.730 the destiny of Unix. It really did. And that's 00:09:57.730 --> 00:10:01.049 where we get into Richie's masterpiece. C is, 00:10:01.370 --> 00:10:03.970 I think you could argue, his single greatest 00:10:03.970 --> 00:10:08.509 invention. And its genius lies in what we call 00:10:08.509 --> 00:10:11.129 its dual nature. We have to really break this 00:10:11.129 --> 00:10:13.309 down because it's a paradox. It functions as 00:10:13.309 --> 00:10:15.230 both a low -level language and a high -level 00:10:15.230 --> 00:10:17.629 language at the same time. How does that even 00:10:17.629 --> 00:10:20.529 work? This is the core reason Siege has dominated 00:10:20.529 --> 00:10:23.429 computing for decades. Most languages, you know, 00:10:23.450 --> 00:10:24.870 they have to pick a side. You have your high 00:10:24.870 --> 00:10:27.590 -level languages like Python or Java today, which 00:10:27.590 --> 00:10:30.549 are fantastic for abstracting away complex logic. 00:10:30.590 --> 00:10:32.169 They're great for business applications. But 00:10:32.169 --> 00:10:33.710 they're slow, right? They have a lot of overhead. 00:10:33.889 --> 00:10:36.610 They often need a big virtual machine or an interpreter 00:10:36.610 --> 00:10:38.950 to translate their commands into machine code. 00:10:39.169 --> 00:10:41.830 That adds overhead, and it removes your direct 00:10:41.830 --> 00:10:44.210 control over the hardware. And on the other end 00:10:44.210 --> 00:10:45.710 of the spectrum, you have assembly language. 00:10:46.250 --> 00:10:48.950 Super efficient, but basically unmanageable for 00:10:48.950 --> 00:10:51.690 a big project. Exactly. C is the perfect hybrid. 00:10:51.889 --> 00:10:54.509 On the low -level side, C has constructs that 00:10:54.509 --> 00:10:57.049 translate very, very closely to the hardware's 00:10:57.049 --> 00:10:59.629 own instruction set. The big one is pointers. 00:10:59.929 --> 00:11:01.990 Pointers give you direct control over memory. 00:11:02.379 --> 00:11:04.820 Direct surgical control. You can allocate memory. 00:11:05.039 --> 00:11:07.299 You can manipulate it directly. This is essential 00:11:07.299 --> 00:11:09.779 if you're writing a kernel or device driver or 00:11:09.779 --> 00:11:12.559 firmware for a tiny chip. You get the raw speed 00:11:12.559 --> 00:11:15.320 and access you need. So you get that raw machine 00:11:15.320 --> 00:11:17.799 level speed, but then you layer on the structural 00:11:17.799 --> 00:11:20.259 clarity of a high level language. Precisely. 00:11:20.590 --> 00:11:23.289 Richie took the core ideas of B, but he added 00:11:23.289 --> 00:11:26.110 these crucial high -level elements. The most 00:11:26.110 --> 00:11:28.370 important were data types, so you can define 00:11:28.370 --> 00:11:31.470 integers, characters, and so on, and critically 00:11:31.470 --> 00:11:34.330 user -defined data structures. These are called 00:11:34.330 --> 00:11:37.149 structs in C. And structs let you organize complex 00:11:37.149 --> 00:11:39.669 information. They let you map real -world things 00:11:39.669 --> 00:11:42.629 like a user profile, a network packet, a financial 00:11:42.629 --> 00:11:45.490 record into an organized block of memory. That 00:11:45.490 --> 00:11:48.190 ability to organize and map complex data is the 00:11:48.190 --> 00:11:50.610 absolute hallmark of modern software development. 00:11:50.950 --> 00:11:53.029 So C gives you the speed of a low -level language, 00:11:53.210 --> 00:11:55.950 but the organization of a high -level one. And 00:11:55.950 --> 00:11:58.129 for Unix, this duality meant one thing above 00:11:58.129 --> 00:12:01.350 all else, portability. That word, portability, 00:12:01.350 --> 00:12:04.120 is everything. It's the whole ballgame. Before 00:12:04.120 --> 00:12:06.500 C, if you wanted to run your operating system 00:12:06.500 --> 00:12:09.360 on a new machine, say moving from that PDP -7 00:12:09.360 --> 00:12:12.440 to a PDP -11, you had to rewrite most of it. 00:12:12.500 --> 00:12:14.840 A huge undertaking. It basically locked you into 00:12:14.840 --> 00:12:17.419 one hardware vendor. It was a massive commercial 00:12:17.419 --> 00:12:20.700 barrier. C completely shattered that. Because 00:12:20.700 --> 00:12:23.600 C abstracts the code away from the specific hardware, 00:12:23.820 --> 00:12:26.279 you could write Unix in C, and then all you needed 00:12:26.279 --> 00:12:28.700 was a C compiler for the new machine. And you 00:12:28.700 --> 00:12:30.740 could just move the entire OS over. you could 00:12:30.740 --> 00:12:34.279 port 90 % of it with minimal effort. This ability 00:12:34.279 --> 00:12:36.820 to move Unix and all its applications easily 00:12:36.820 --> 00:12:40.000 across different processor families, from mainframes 00:12:40.000 --> 00:12:42.779 down to the first microprocessors, that is what 00:12:42.779 --> 00:12:45.500 made Unix a global force and C, the master key 00:12:45.500 --> 00:12:48.019 to hardware independence. That's a massive intellectual 00:12:48.019 --> 00:12:50.659 leap. It really defined how we think about software 00:12:50.659 --> 00:12:53.360 development from that point on. But even a brilliant 00:12:53.360 --> 00:12:55.600 invention needs a standard to really take hold 00:12:55.600 --> 00:12:58.340 globally. And that brings us to the famous book 00:12:58.340 --> 00:13:00.480 he wrote with Brian Kernighan, known simply by 00:13:00.480 --> 00:13:03.139 their initials. The C programming language, published 00:13:03.139 --> 00:13:06.799 in 1978, famously known to every programmer as 00:13:06.799 --> 00:13:09.519 KNR. And this wasn't just a book. For decades, 00:13:09.580 --> 00:13:12.139 it was the constitution of the C language. Why 00:13:12.139 --> 00:13:14.419 was KNR so important? What made it so influential? 00:13:14.740 --> 00:13:16.480 Well, a couple of things. First, the language 00:13:16.480 --> 00:13:19.559 was still evolving and KNR provided a definitive, 00:13:19.679 --> 00:13:22.740 elegant standard. It defined what became known 00:13:22.740 --> 00:13:25.840 as KNRC, which was the de facto benchmark for 00:13:25.840 --> 00:13:27.960 years. So it got everyone on the same page. Exactly. 00:13:27.960 --> 00:13:31.259 And second, the style of the book perfectly mirrored 00:13:31.259 --> 00:13:34.200 the style of the language itself. It was compact, 00:13:34.519 --> 00:13:37.480 efficient, and crystal clear. It wasn't a huge, 00:13:37.539 --> 00:13:40.159 verbose textbook. It was a reference manual and 00:13:40.159 --> 00:13:43.000 a tutorial all in one. It let programmers all 00:13:43.000 --> 00:13:44.879 over the world learn the standard quickly and 00:13:44.879 --> 00:13:47.179 know that their code would work everywhere. It 00:13:47.179 --> 00:13:49.840 basically professionalized the language. So KNR 00:13:49.840 --> 00:13:52.519 gives it consistency, which combined with C's 00:13:52.519 --> 00:13:55.039 portability just leads to its absolute dominance. 00:13:55.080 --> 00:13:57.679 It cements its reputation as the architectural 00:13:57.679 --> 00:14:00.139 scaffolding of modern software. Let's really 00:14:00.139 --> 00:14:02.700 try to quantify that DNA claim. Where do we see 00:14:02.700 --> 00:14:06.120 C's the legacy today? That claim that C is the 00:14:06.120 --> 00:14:08.320 genetic material of the modern software world 00:14:08.320 --> 00:14:11.279 is, well, it's provably true. I mean, C itself 00:14:11.279 --> 00:14:14.080 is still highly visible. It's used everywhere 00:14:14.080 --> 00:14:16.620 in operating system kernels, embedded systems, 00:14:16.840 --> 00:14:20.080 firmware for everything from your car to your 00:14:20.080 --> 00:14:22.720 smart fridge, supercomputing. Anything that needs 00:14:22.720 --> 00:14:25.080 to be fast and close to the metal. Anything that 00:14:25.080 --> 00:14:27.039 needs raw speed and direct hardware control is 00:14:27.039 --> 00:14:29.620 still often written in C. But the influence is 00:14:29.620 --> 00:14:32.039 so much deeper than just the code that is still 00:14:32.039 --> 00:14:34.980 C. It's the language that basically spawned every 00:14:34.980 --> 00:14:37.700 other major modern language. It's the linguistic 00:14:37.700 --> 00:14:41.360 root. It is. C laid the structural and the philosophical 00:14:41.360 --> 00:14:43.360 groundwork for almost everything that followed. 00:14:43.700 --> 00:14:45.980 When you look at the syntax, the curly braces, 00:14:46.159 --> 00:14:47.919 the control flow, the way you define functions, 00:14:48.279 --> 00:14:50.620 that's carried forward into the derivatives that 00:14:50.620 --> 00:14:52.500 dominate the commercial world today. You're talking 00:14:52.500 --> 00:14:55.269 about things like C++. Direct descendants, yes. 00:14:56.250 --> 00:14:58.889 C++C, which added object -oriented programming 00:14:58.889 --> 00:15:02.210 on top of C's core. Objective -C, which was famously 00:15:02.210 --> 00:15:05.070 used by Apple for years. And C -Shark, which 00:15:05.070 --> 00:15:06.970 is the foundational language for Microsoft's 00:15:06.970 --> 00:15:09.450 entire .NET framework. So the biggest players 00:15:09.450 --> 00:15:12.850 in tech built their empires on languages derived 00:15:12.850 --> 00:15:15.529 directly from C. They did. And even languages 00:15:15.529 --> 00:15:17.409 that look pretty different on the surface, like 00:15:17.409 --> 00:15:20.659 Java and Python, owe a massive debt. Java syntax 00:15:20.659 --> 00:15:24.419 is directly derived from C and C++. It's used 00:15:24.419 --> 00:15:26.559 everywhere in big corporate environments, and 00:15:26.559 --> 00:15:28.240 it's the foundation of the Android operating 00:15:28.240 --> 00:15:31.019 system. And Python. The core interpreter for 00:15:31.019 --> 00:15:33.259 Python, the engine that actually runs the code, 00:15:33.519 --> 00:15:36.840 is itself written in C for performance. So even 00:15:36.840 --> 00:15:38.960 when you're using a modern, high -level abstract 00:15:38.960 --> 00:15:41.519 language, the high -performance engine running 00:15:41.519 --> 00:15:44.679 under the hood is almost certainly C code, built 00:15:44.679 --> 00:15:46.580 on the design principles that Richie laid out. 00:15:46.720 --> 00:15:49.789 Wow. So without C's ability to bridge that gap 00:15:49.789 --> 00:15:51.830 between efficient hardware access and high level 00:15:51.830 --> 00:15:54.669 structure, the modern software ecosystem just 00:15:54.669 --> 00:15:57.269 it wouldn't exist as we know it. It was the essential 00:15:57.269 --> 00:16:00.470 pivot point. It really was. So C may be Ritchie's 00:16:00.470 --> 00:16:02.529 defining technical achievement, but it was a 00:16:02.529 --> 00:16:04.889 tool that he forged for a very specific purpose 00:16:04.889 --> 00:16:08.710 to enable Unix. Let's talk about Unix as a foundational 00:16:08.710 --> 00:16:10.669 system, because this is where that portability 00:16:10.669 --> 00:16:13.429 innovation really paid off. This was the strategic 00:16:13.429 --> 00:16:18.299 masterstroke. The decision around 1973 to rewrite 00:16:18.299 --> 00:16:22.279 almost the entire Unix kernel in C. Before that, 00:16:22.440 --> 00:16:24.460 operating systems were written in assembly. They 00:16:24.460 --> 00:16:26.360 were proprietary. They were bound to the hardware 00:16:26.360 --> 00:16:29.179 they were sold with. But by writing Unix in C, 00:16:29.500 --> 00:16:32.220 Ritchie and Thompson guaranteed its flexibility. 00:16:32.519 --> 00:16:34.519 And its ability to be ported across different 00:16:34.519 --> 00:16:37.840 processors. Easily, efficiently, and quickly. 00:16:38.120 --> 00:16:40.000 Thompson got the ball rolling, but it sounds 00:16:40.000 --> 00:16:41.659 like Ritchie was the one who made sure it wasn't 00:16:41.659 --> 00:16:44.190 chained to one machine. Richie was crucial in 00:16:44.190 --> 00:16:46.190 those porting efforts. He took on the really 00:16:46.190 --> 00:16:48.870 detailed, difficult work of moving Unix to different 00:16:48.870 --> 00:16:51.529 platforms. And this work didn't just validate 00:16:51.529 --> 00:16:54.509 C as a language. It demonstrated a whole new 00:16:54.509 --> 00:16:57.629 economic model for software. Companies no longer 00:16:57.629 --> 00:16:59.830 had to rebuild their entire system if they bought 00:16:59.830 --> 00:17:01.429 new hardware. They just needed a C compiler. 00:17:01.840 --> 00:17:03.820 They just needed a C compiler for the new machine 00:17:03.820 --> 00:17:06.720 and they could bring Unix with them. For IT departments 00:17:06.720 --> 00:17:09.539 in the 70s and 80s, this was absolutely revolutionary. 00:17:09.980 --> 00:17:12.960 But Unix was influential beyond just the code, 00:17:12.980 --> 00:17:15.799 right? It established entirely new ways of thinking 00:17:15.799 --> 00:17:17.619 about computing, principles that are now just 00:17:17.619 --> 00:17:20.099 baked into everything. Oh, absolutely. We can 00:17:20.099 --> 00:17:22.539 point to several huge conceptual ideas that came 00:17:22.539 --> 00:17:25.539 out of Unix. First, the hierarchical file system. 00:17:26.109 --> 00:17:29.430 the idea of directories and subdirectories. Before 00:17:29.430 --> 00:17:32.410 Unix, data organization was often flat and just 00:17:32.410 --> 00:17:35.349 chaotic. This gave it a logical, manageable structure. 00:17:35.650 --> 00:17:37.450 And then there's the philosophy of small tools. 00:17:37.829 --> 00:17:39.990 Right. The idea that instead of building one 00:17:39.990 --> 00:17:41.890 massive program that tries to do everything, 00:17:42.069 --> 00:17:44.670 you build lots of small, specialized, reliable 00:17:44.670 --> 00:17:47.890 programs, utilities that each do one thing and 00:17:47.890 --> 00:17:50.140 do it well. And you chain them together. You 00:17:50.140 --> 00:17:51.799 chain them together using the invention of pipes. 00:17:51.980 --> 00:17:54.720 The pipe mechanism lets you redirect the output 00:17:54.720 --> 00:17:57.140 of one program to be the input of another right 00:17:57.140 --> 00:17:59.519 from the command line. This ability to just combine 00:17:59.519 --> 00:18:02.660 small, efficient programs on the fly was immensely 00:18:02.660 --> 00:18:05.599 powerful. It's the foundation of all modern scripting 00:18:05.599 --> 00:18:08.099 and automation. And finally, you have the model 00:18:08.099 --> 00:18:11.240 for processes and users. Yes. Unix defined a 00:18:11.240 --> 00:18:13.460 really robust model for managing multiple users, 00:18:13.720 --> 00:18:17.210 concurrent processes, memory protection. the 00:18:17.210 --> 00:18:19.450 very structure of a modern timesharing operating 00:18:19.450 --> 00:18:22.279 system. These weren't just implementation details. 00:18:22.480 --> 00:18:25.059 They were philosophical principles of modularity, 00:18:25.099 --> 00:18:27.680 simplicity, and composability. And they became 00:18:27.680 --> 00:18:29.960 the gold standard for every OS that followed, 00:18:30.119 --> 00:18:34.099 whether it was Windows NT, macOS, or Linux. Richie 00:18:34.099 --> 00:18:36.480 was also key in documenting all of this, right? 00:18:36.559 --> 00:18:39.059 Spreading the philosophy. Documentation was vital, 00:18:39.200 --> 00:18:41.319 especially for getting Unix adopted in academia. 00:18:41.700 --> 00:18:43.920 He co -authored the Unix Programmer's Manual 00:18:43.920 --> 00:18:47.359 way back in 1971. Later, the foundational academic 00:18:47.359 --> 00:18:50.250 paper, the UNIX Timesharing System, which he 00:18:50.250 --> 00:18:52.470 wrote with Thompson, helped formalize the whole 00:18:52.470 --> 00:18:55.009 model for computer science and everywhere. These 00:18:55.009 --> 00:18:56.990 documents didn't just explain the code, they 00:18:56.990 --> 00:18:59.470 institutionalized the philosophy. So let's bring 00:18:59.470 --> 00:19:02.009 that legacy into the modern era. The word Unix 00:19:02.009 --> 00:19:04.630 today refers to this huge family of derivatives. 00:19:04.930 --> 00:19:07.069 And Ritchie himself actually commented on this 00:19:07.069 --> 00:19:09.859 in a 1999 interview. Right. And his perspective 00:19:09.859 --> 00:19:11.839 is so important because it validates the whole 00:19:11.839 --> 00:19:14.440 open source movement that came later. He consistently 00:19:14.440 --> 00:19:16.859 said that systems like Linux and the various 00:19:16.859 --> 00:19:20.140 BSD variants weren't just clones. He saw them 00:19:20.140 --> 00:19:23.059 as a vital, robust continuation of the fundamental 00:19:23.059 --> 00:19:26.460 basis of Unix. The idea was bigger than the original 00:19:26.460 --> 00:19:29.980 code. The idea was bigger than the code. And 00:19:29.980 --> 00:19:32.140 his comments on Linux are particularly telling 00:19:32.140 --> 00:19:34.539 because that was the phenomenon that really changed 00:19:34.539 --> 00:19:37.490 enterprise computing. What did he say about it? 00:19:37.569 --> 00:19:40.529 He was incredibly supportive. He said, and I'm 00:19:40.529 --> 00:19:43.390 quoting here, the Linux phenomenon is quite delightful 00:19:43.390 --> 00:19:46.250 because it draws so strongly on the basis that 00:19:46.250 --> 00:19:49.869 Unix provided. He called Linux among the healthiest 00:19:49.869 --> 00:19:52.970 of the direct Unix derivatives. He really understood 00:19:52.970 --> 00:19:55.269 the power of the open source model to advance 00:19:55.269 --> 00:19:57.710 the core ideas he and Thompson had pioneered. 00:19:57.789 --> 00:19:59.630 Which reflects this core belief that it was about 00:19:59.630 --> 00:20:01.450 the evolution of the ideas, not who owned the 00:20:01.450 --> 00:20:05.579 code. Exactly. He saw Unix and Linux as, in his 00:20:05.579 --> 00:20:08.259 words, the continuation of ideas that were started 00:20:08.259 --> 00:20:10.619 by Ken and me and many others many years ago. 00:20:10.779 --> 00:20:13.440 It was never about proprietary control for him. 00:20:13.619 --> 00:20:15.059 And if you look at mission critical computing 00:20:15.059 --> 00:20:17.380 today, I mean, that legacy is just undeniable. 00:20:17.519 --> 00:20:19.359 Completely. The principles of Unix stability, 00:20:19.720 --> 00:20:22.559 security, scalability, they're still the standard. 00:20:22.740 --> 00:20:25.119 While Windows might dominate the consumer desktop, 00:20:25.420 --> 00:20:27.579 the source material is clear that the commercial 00:20:27.579 --> 00:20:31.720 derivatives, things like IBM's AIX, HPUX. They 00:20:31.720 --> 00:20:35.680 are the de facto OS for anything that is mission 00:20:35.680 --> 00:20:37.539 critical. They're running the world's infrastructure. 00:20:37.799 --> 00:20:40.880 They're running global stock exchanges, massive 00:20:40.880 --> 00:20:43.640 corporate databases, telecommunications networks, 00:20:44.099 --> 00:20:47.480 supercomputing clusters. You simply cannot build 00:20:47.480 --> 00:20:50.200 a modern data center without relying on the architectural 00:20:50.200 --> 00:20:52.480 principles that Ritchie and Thompson laid down 00:20:52.480 --> 00:20:55.940 50 years ago. That reliance speaks volumes. Now, 00:20:55.960 --> 00:20:58.259 it's easy to just focus on C and Unix, but Ritchie 00:20:58.259 --> 00:21:00.299 was innovating at Bell Labs for his whole career. 00:21:00.809 --> 00:21:02.809 After those huge successes, he kept pushing, 00:21:02.930 --> 00:21:05.349 trying to move beyond Unix's own limitations. 00:21:05.930 --> 00:21:08.829 What came next? Right. Well into the 80s and 00:21:08.829 --> 00:21:11.390 90s, he was a central figure in two major successor 00:21:11.390 --> 00:21:15.710 projects at Bell Labs, Plan 9 and later the Inferno 00:21:15.710 --> 00:21:18.349 operating system. And these weren't just, you 00:21:18.349 --> 00:21:20.990 know, Unix 2 .0. They were radical attempts to 00:21:20.990 --> 00:21:23.569 reinvent the whole OS model for the age of networking. 00:21:23.970 --> 00:21:26.589 So what was the big conceptual shift from Unix? 00:21:26.730 --> 00:21:29.210 Well, Unix was fundamentally built for a single 00:21:29.210 --> 00:21:32.190 machine that hosted multiple users. Plan 9 and 00:21:32.190 --> 00:21:34.369 Inferno tried to make the entire network function 00:21:34.369 --> 00:21:37.539 as a single, cohesive computer. They extended 00:21:37.539 --> 00:21:40.200 the Unix philosophy of everything is a file to 00:21:40.200 --> 00:21:42.359 include network resources. What does that mean 00:21:42.359 --> 00:21:45.059 in practice? It means that in Plan 9, for example, 00:21:45.200 --> 00:21:48.440 accessing a remote server CPU or a printer down 00:21:48.440 --> 00:21:50.980 the hall or a file on another continent, it was 00:21:50.980 --> 00:21:52.920 all treated as if you were just reading or writing 00:21:52.920 --> 00:21:55.180 to a local file in your own file system. It was 00:21:55.180 --> 00:21:58.220 this beautiful, unified model. So if Unix defined 00:21:58.220 --> 00:22:01.299 local timesharing, Plan 9 and Inferno were trying 00:22:01.299 --> 00:22:03.700 to define true, transparent networked computing. 00:22:04.240 --> 00:22:06.839 Precisely. Now, they never achieved the same 00:22:06.839 --> 00:22:09.779 widespread adoption as Unix. Commercially, they 00:22:09.779 --> 00:22:12.440 were probably too far ahead of their time, but 00:22:12.440 --> 00:22:14.920 they were intellectually groundbreaking and hugely 00:22:14.920 --> 00:22:17.539 influential in academic circles. A lot of their 00:22:17.539 --> 00:22:20.319 concepts later found their way into modern distributed 00:22:20.319 --> 00:22:22.599 systems. And he was involved in the programming 00:22:22.599 --> 00:22:25.269 languages for those, too. Yes, he contributed 00:22:25.269 --> 00:22:28.349 to the language limbo, which was part of Inferno. 00:22:28.450 --> 00:22:30.950 It just shows he was always searching for that 00:22:30.950 --> 00:22:33.210 next foundational architecture, even when the 00:22:33.210 --> 00:22:34.890 one he already built was completely dominant. 00:22:35.130 --> 00:22:38.269 Okay, let's turn to a really fascinating, almost 00:22:38.269 --> 00:22:41.990 hidden chapter in his career. A piece of classified 00:22:41.990 --> 00:22:45.329 history that shows his depth in fields way beyond 00:22:45.329 --> 00:22:48.339 operating systems. specifically cryptography. 00:22:48.420 --> 00:22:50.940 This part just adds so much texture to his profile. 00:22:51.259 --> 00:22:53.960 During the 1970s, Ritchie collaborated with two 00:22:53.960 --> 00:22:56.140 other researchers, James Reeds and Robert Morris, 00:22:56.259 --> 00:22:58.619 on a highly technical cryptographic attack. And 00:22:58.619 --> 00:23:02.160 the target was the M209 US cipher machine. For 00:23:02.160 --> 00:23:04.000 anyone who isn't a history buff, what was that? 00:23:04.279 --> 00:23:06.920 It was a mechanical, portable cipher machine 00:23:06.920 --> 00:23:10.759 widely used during and after World War Two. It 00:23:10.759 --> 00:23:13.180 used a series of adjustable rotors and a paper 00:23:13.180 --> 00:23:16.339 tape key to encrypt messages. It was very secure 00:23:16.339 --> 00:23:18.420 for its time, but, you know, purely mechanical. 00:23:18.680 --> 00:23:21.000 And the attack they developed was a ciphertext 00:23:21.000 --> 00:23:23.720 only attack. That's the hardest kind, isn't it? 00:23:23.779 --> 00:23:27.059 It is the holy grail of cryptanalysis. It means 00:23:27.059 --> 00:23:29.880 they only had access to the encrypted message, 00:23:30.000 --> 00:23:32.920 the scrambled text. They had no idea what the 00:23:32.920 --> 00:23:35.210 original message said and from that alone they 00:23:35.210 --> 00:23:37.230 could figure out the original message and crucially 00:23:37.230 --> 00:23:39.769 the machine settings used to encrypt it. That 00:23:39.769 --> 00:23:42.430 requires some serious mathematical skill. Immense 00:23:42.430 --> 00:23:44.910 computational skill, a deep understanding of 00:23:44.910 --> 00:23:48.369 statistics, probability theory, basically finding 00:23:48.369 --> 00:23:50.690 the subtle mathematical weaknesses in the machine's 00:23:50.690 --> 00:23:53.329 design and their attack was devastatingly effective. 00:23:53.589 --> 00:23:55.470 How effective? The source material says their 00:23:55.470 --> 00:23:58.150 method could solve messages of at least 2 ,000 00:23:58.150 --> 00:24:01.250 to 2 ,500 letters just from analyzing the ciphertext. 00:24:01.470 --> 00:24:03.450 This wasn't just an academic exercise, it was 00:24:03.450 --> 00:24:05.990 a practical demonstration of a major vulnerability 00:24:05.990 --> 00:24:09.369 in an established cipher system. And yet, this 00:24:09.369 --> 00:24:12.109 research was never published. It stayed secret 00:24:12.109 --> 00:24:15.390 for decades. Why? This is where Bell Labs' research 00:24:15.390 --> 00:24:18.079 intersects with global security. After they did 00:24:18.079 --> 00:24:19.980 the work, they had discussions with the National 00:24:19.980 --> 00:24:22.799 Security Agency, the NSA, and the NSA pointed 00:24:22.799 --> 00:24:25.019 out that the principles they had exploited still 00:24:25.019 --> 00:24:27.519 applied to similar mechanical cipher machines 00:24:27.519 --> 00:24:30.079 that were still in active use by foreign governments 00:24:30.079 --> 00:24:32.599 around the world. So they were sitting on a live. 00:24:33.200 --> 00:24:35.160 intelligence vulnerability. If they published 00:24:35.160 --> 00:24:37.660 it, they could compromise ongoing global communications. 00:24:38.059 --> 00:24:40.440 Exactly. So out of a sense of responsibility 00:24:40.440 --> 00:24:44.240 for national and really global stability, the 00:24:44.240 --> 00:24:46.839 authors just collectively decided not to publish 00:24:46.839 --> 00:24:49.039 it. They suppressed their own research. It's 00:24:49.039 --> 00:24:51.380 a remarkable insight into the high stakes ethical 00:24:51.380 --> 00:24:54.579 environment at Bell Labs. It shows Ritchie wasn't 00:24:54.579 --> 00:24:57.240 just a genius coder. He was a responsible scientist 00:24:57.240 --> 00:24:59.680 who understood the real world implications of 00:24:59.680 --> 00:25:01.680 his work. Moving from secrets to recognition. 00:25:02.440 --> 00:25:04.940 Dennis Ritchie's work, while, you know, often 00:25:04.940 --> 00:25:07.279 invisible to the average person, was recognized 00:25:07.279 --> 00:25:09.920 for decades by the highest institutions in science 00:25:09.920 --> 00:25:12.099 and tech. And it really started with the computing 00:25:12.099 --> 00:25:15.420 world's ultimate honor. The pinnacle, yeah. In 00:25:15.420 --> 00:25:18.859 1983, he and Ken Thompson jointly received the 00:25:18.859 --> 00:25:21.900 ACM Turing Award. It's often called the Nobel 00:25:21.900 --> 00:25:24.700 Prize of Computing. And the citation was very 00:25:24.700 --> 00:25:27.640 specific. It was for their development of generic 00:25:27.640 --> 00:25:30.160 operating systems theory and specifically for 00:25:30.160 --> 00:25:32.660 the implementation of the UNIX operating system. 00:25:32.900 --> 00:25:35.819 That phrasing is so important. They weren't just 00:25:35.819 --> 00:25:37.839 awarded for building a thing. They were honored 00:25:37.839 --> 00:25:40.039 for advancing the theory of how operating systems 00:25:40.039 --> 00:25:42.980 should even work. That's the key. Ritchie's acceptance 00:25:42.980 --> 00:25:45.859 lecture was titled Reflections on Software Research. 00:25:46.160 --> 00:25:48.900 And it's this profound look at the thinking behind 00:25:48.900 --> 00:25:52.279 C and Unix, the need for simplicity, the power 00:25:52.279 --> 00:25:54.640 of modularity, the importance of bridging hardware 00:25:54.640 --> 00:25:57.380 and human logic. It set the standard for what 00:25:57.380 --> 00:25:59.619 system theory should be. And that pattern of 00:25:59.619 --> 00:26:01.640 joint recognition for the two creations, the 00:26:01.640 --> 00:26:04.240 OS and the language, just continued through the 00:26:04.240 --> 00:26:07.839 90s. It did. Seven years later, in 1990, the 00:26:07.839 --> 00:26:10.400 IEEE gave them the Richard W. Hamming Medal. 00:26:10.890 --> 00:26:13.089 And again, the citation focused on both halves 00:26:13.089 --> 00:26:15.849 of his contribution for the origination of the 00:26:15.849 --> 00:26:18.450 Unitex operating system and the C programming 00:26:18.450 --> 00:26:21.609 language. It's just so rare for one person to 00:26:21.609 --> 00:26:23.569 have that level of influence on both the architecture 00:26:23.569 --> 00:26:26.269 and the tool used to build it. And their status 00:26:26.269 --> 00:26:28.670 as foundational figures was cemented by the Computer 00:26:28.670 --> 00:26:31.809 History Museum. In 1997, yeah. They were inducted 00:26:31.809 --> 00:26:34.359 as Computer History Museum fellows. Again, for 00:26:34.359 --> 00:26:36.819 co -creation of the UNICEF operating system and 00:26:36.819 --> 00:26:38.759 for development of the C programming language, 00:26:39.039 --> 00:26:42.400 this trio of major technical awards really placed 00:26:42.400 --> 00:26:45.259 them in the pantheon. But the biggest, most high 00:26:45.259 --> 00:26:47.240 -profile honor came from the U .S. government, 00:26:47.500 --> 00:26:49.920 recognizing not just the academic achievement, 00:26:50.039 --> 00:26:52.680 but the massive global impact. That would be 00:26:52.680 --> 00:26:54.819 the National Medal of Technology. Thompson and 00:26:54.819 --> 00:26:57.400 Ritchie received the 1998 award, and it was presented 00:26:57.400 --> 00:27:00.619 by President Bill Clinton in 1999. And the citation 00:27:00.619 --> 00:27:03.150 for that medal just spelled out the... scope 00:27:03.150 --> 00:27:05.890 of their influence far beyond the lab. What did 00:27:05.890 --> 00:27:07.930 the government citations say? How did they frame 00:27:07.930 --> 00:27:11.210 the impact? It was incredibly clear. It said 00:27:11.210 --> 00:27:14.089 their work led to enormous advances in computer 00:27:14.089 --> 00:27:16.569 hardware, software, and networking systems and 00:27:16.569 --> 00:27:19.630 stimulated growth of an entire industry, thereby 00:27:19.630 --> 00:27:22.410 enhancing American leadership in the information 00:27:22.410 --> 00:27:25.869 age. Wow. So that's a direct line from their 00:27:25.869 --> 00:27:29.210 lab work to global economic stimulus and American 00:27:29.210 --> 00:27:31.309 technological leadership? There's no ambiguity 00:27:31.309 --> 00:27:34.259 there. Few pieces of software have ever received 00:27:34.259 --> 00:27:36.859 that kind of sweeping recognition for their societal 00:27:36.859 --> 00:27:39.200 and commercial impact. And he kept getting honors 00:27:39.200 --> 00:27:42.559 well into the 2000s. He did. In 2005, the Industrial 00:27:42.559 --> 00:27:45.359 Research Institute Achievement Award. And finally, 00:27:45.400 --> 00:27:47.599 in 2011, right before he passed away, he and 00:27:47.599 --> 00:27:49.960 Thompson were awarded the Japan Prize for Information 00:27:49.960 --> 00:27:53.000 and Communications. This three -decade string 00:27:53.000 --> 00:27:55.440 of honors just confirms that his work wasn't 00:27:55.440 --> 00:27:58.099 some transient innovation. It's a permanent, 00:27:58.220 --> 00:28:01.839 growing foundation for the modern world. Incredible 00:28:01.839 --> 00:28:04.359 career. How did his long tenure at Bell Labs 00:28:04.359 --> 00:28:06.480 and its successor companies finally wrap up? 00:28:06.539 --> 00:28:09.019 His career was almost entirely within that one 00:28:09.019 --> 00:28:12.559 research ecosystem. During an AT &T restructuring 00:28:12.559 --> 00:28:15.119 in the mid 90s, his research center was transferred 00:28:15.119 --> 00:28:18.119 over to Lucent Technologies. He continued his 00:28:18.119 --> 00:28:20.319 work there, eventually moving into a leadership 00:28:20.319 --> 00:28:24.049 role. He retired in 2007 as the head of the Systems 00:28:24.049 --> 00:28:26.849 Software Research Department at Lucent. Just 00:28:26.849 --> 00:28:29.869 a long, distinguished career spent quietly building 00:28:29.869 --> 00:28:32.109 the world's infrastructure, largely shielded 00:28:32.109 --> 00:28:33.829 from all the commercial noise that his inventions 00:28:33.829 --> 00:28:36.170 had created. And the final chapter of his life, 00:28:36.269 --> 00:28:38.390 unfortunately, really illustrates his role as 00:28:38.390 --> 00:28:41.829 the unseen architect. He was found dead on or 00:28:41.829 --> 00:28:45.869 about October 12, 2011. He was 70 years old, 00:28:45.950 --> 00:28:48.160 living at his home in New Jersey. He'd been in 00:28:48.160 --> 00:28:49.819 frail health for a few years after treatment 00:28:49.819 --> 00:28:52.140 for prostate cancer and heart disease. And the 00:28:52.140 --> 00:28:54.500 timing of his death is just so critical to understanding 00:28:54.500 --> 00:28:56.720 his public profile. The news of Richie's death 00:28:56.720 --> 00:28:59.099 was almost completely overshadowed by the intense 00:28:59.099 --> 00:29:01.259 global media coverage of the passing of Apple 00:29:01.259 --> 00:29:03.559 co -founder Steve Jobs, which had happened just 00:29:03.559 --> 00:29:05.920 the week before. Which created this stark, almost 00:29:05.920 --> 00:29:08.559 poetic contrast. Yeah. You had the high profile 00:29:08.559 --> 00:29:11.759 consumer facing icon whose products defined user 00:29:11.759 --> 00:29:15.240 experience. Versus the foundational internal 00:29:15.240 --> 00:29:18.059 architect whose work made. all of those products 00:29:18.059 --> 00:29:21.559 even possible. The media really struggled to 00:29:21.559 --> 00:29:23.759 explain Ritchie's influence to the general public. 00:29:23.980 --> 00:29:26.160 And that contrast became the key to summarizing 00:29:26.160 --> 00:29:29.240 his legacy. The computer historian Paul Ceruzzi 00:29:29.240 --> 00:29:32.059 captured it perfectly. He said Ritchie was completely 00:29:32.059 --> 00:29:35.160 under the radar to the public. But then he offered 00:29:35.160 --> 00:29:37.960 this powerful analogy. If you had a microscope 00:29:37.960 --> 00:29:39.940 and could look in a computer, you'd see his work 00:29:39.940 --> 00:29:43.039 everywhere inside. He was the quiet, fundamental 00:29:43.039 --> 00:29:45.759 structural pillar. That's the perfect image. 00:29:46.059 --> 00:29:48.299 Even those closest to him, like Brian Kearney 00:29:48.299 --> 00:29:50.079 and his co -author, said that Ritchie himself 00:29:50.079 --> 00:29:53.380 never expected C to become so significant. He 00:29:53.380 --> 00:29:55.900 just built it as an elegant tool for system programming, 00:29:56.140 --> 00:29:58.900 not as a universal language. And Kernighan was 00:29:58.900 --> 00:30:01.259 the one who provided that clear bridge for everyone, 00:30:01.400 --> 00:30:03.680 connecting Richie's unseen work to the consumer 00:30:03.680 --> 00:30:06.799 world. He said, the tools that Dennis built and 00:30:06.799 --> 00:30:09.079 their direct descendants run pretty much everything 00:30:09.079 --> 00:30:11.779 today. And he specifically pointed out the essential 00:30:11.779 --> 00:30:14.160 role of C and Unix principles in the code that 00:30:14.160 --> 00:30:17.099 underpins the iPhone. The link is just undeniable. 00:30:17.279 --> 00:30:19.819 When commentators look back at the relative importance 00:30:19.819 --> 00:30:23.019 of Richie versus Jobs, the consensus was clear. 00:30:23.799 --> 00:30:26.660 Richie's work. played a key role in spawning 00:30:26.660 --> 00:30:29.140 the technological revolution of the last 40 years, 00:30:29.279 --> 00:30:32.240 including technology on which Apple went on to 00:30:32.240 --> 00:30:34.460 build its fortune. Which brings us back to that 00:30:34.460 --> 00:30:37.500 idea of the foundational scaffolding. One commentator 00:30:37.500 --> 00:30:41.200 put it perfectly. Ritchie invented and co -invented 00:30:41.200 --> 00:30:44.039 two key software technologies which make up the 00:30:44.039 --> 00:30:46.480 architectural underpinning of effectively every 00:30:46.480 --> 00:30:49.759 single computer software product we use. And 00:30:49.759 --> 00:30:52.569 that's not hype. It's just a statement of technical 00:30:52.569 --> 00:30:54.750 fact. The computer science community certainly 00:30:54.750 --> 00:30:57.390 understood the loss. We saw some really moving 00:30:57.390 --> 00:30:59.349 tributes from the open source world, the same 00:30:59.349 --> 00:31:01.269 movement that Richie had praised so strongly. 00:31:01.490 --> 00:31:03.289 Oh, they knew. They knew where their freedom 00:31:03.289 --> 00:31:06.029 and flexibility came from. Two major operating 00:31:06.029 --> 00:31:08.549 systems dedicated releases to his memory, the 00:31:08.549 --> 00:31:12.410 Fedora 16 Linux distribution and FreeBSD 9 .0. 00:31:12.450 --> 00:31:14.529 It was this acknowledgement of the deep philosophical 00:31:14.529 --> 00:31:17.609 debt they owed him. And there's one lasting tribute 00:31:17.609 --> 00:31:20.049 that really places his legacy among the greats. 00:31:20.109 --> 00:31:23.309 thinkers. His name is now literally in the stars. 00:31:23.569 --> 00:31:27.869 It is. Asteroid 294727, Dennis Ritchie, discovered 00:31:27.869 --> 00:31:30.569 in 2008, was officially named in his honor in 00:31:30.569 --> 00:31:33.769 2012. The quiet architect whose work underpins 00:31:33.769 --> 00:31:36.390 the systems we use to explore the universe now 00:31:36.390 --> 00:31:38.890 has his own place within it. It's a beautiful, 00:31:38.970 --> 00:31:41.089 fitting end to a life defined by foundational 00:31:41.089 --> 00:31:44.529 work. So as we wrap up, what stands out to you 00:31:44.529 --> 00:31:46.670 as the ultimate takeaway from this deep dive 00:31:46.670 --> 00:31:49.519 into Dennis Ritchie's unseen empire? For me, 00:31:49.539 --> 00:31:51.640 it's the recognition that the truly revolutionary 00:31:51.640 --> 00:31:54.460 changes in technology are often structural and 00:31:54.460 --> 00:31:57.119 they're often invisible. Richie was the quiet 00:31:57.119 --> 00:31:59.640 genius. He was the indispensable architect who, 00:31:59.759 --> 00:32:02.819 with Thompson, birthed Unix and then single -handedly 00:32:02.819 --> 00:32:05.140 engineered C, that critical bridge that lets 00:32:05.140 --> 00:32:07.339 software be portable and efficient and just fueled 00:32:07.339 --> 00:32:09.339 the global explosion of software development. 00:32:09.700 --> 00:32:12.299 His inventions are the absolute bedrock of modern 00:32:12.299 --> 00:32:14.920 operating systems, running everything from proprietary 00:32:14.920 --> 00:32:17.160 workstations to the Linux kernel and the fundamental 00:32:17.160 --> 00:32:20.099 code inside virtually every smartphone. He retired 00:32:20.099 --> 00:32:22.859 from Lucent in 2007, leaving this legacy that 00:32:22.859 --> 00:32:24.779 was acknowledged by the highest honors. From 00:32:24.779 --> 00:32:26.799 the Turing Award to the National Medal of Technology, 00:32:27.200 --> 00:32:29.460 he abled the information age by building its 00:32:29.460 --> 00:32:32.359 tools. Right. Now, we talked a lot about how 00:32:32.359 --> 00:32:34.960 his work on Unix established these core computing 00:32:34.960 --> 00:32:37.559 concepts, the file system, process management, 00:32:37.599 --> 00:32:40.240 and so on. We also know he spent his later career 00:32:40.240 --> 00:32:43.579 trying to move past Unix's limitations with systems 00:32:43.579 --> 00:32:46.160 like Plan 9 and Inferno, which proposed radical 00:32:46.160 --> 00:32:49.619 new models for networking. which raises an important 00:32:49.619 --> 00:32:52.079 final question for you, the listener, to think 00:32:52.079 --> 00:32:55.559 about as you use technology today. If C and Unix 00:32:55.559 --> 00:32:58.299 defined the last 50 years of computing by establishing 00:32:58.299 --> 00:33:00.539 principles like the hierarchical file system, 00:33:00.759 --> 00:33:03.380 what foundational concept, hidden deep inside 00:33:03.380 --> 00:33:05.400 our modern systems and derived directly from 00:33:05.400 --> 00:33:08.359 1970s Unix, are we still relying on today that 00:33:08.359 --> 00:33:11.339 is now becoming a constraint? What specific structural 00:33:11.339 --> 00:33:13.740 part of the Unix philosophy needs to be replaced 00:33:13.740 --> 00:33:16.019 or re -engineered? Maybe it's about security 00:33:16.019 --> 00:33:18.700 or distributed identity or data flow to enable 00:33:18.700 --> 00:33:20.960 the next truly revolutionary shift in computing. 00:33:21.279 --> 00:33:23.759 We rely on these structures every day. but are 00:33:23.759 --> 00:33:25.559 they secretly the ceiling for the next great 00:33:25.559 --> 00:33:28.339 innovation? A profound thought on the architecture 00:33:28.339 --> 00:33:30.619 we all take for granted. We hope this deep dive 00:33:30.619 --> 00:33:32.400 encourages you to look a little closer at the 00:33:32.400 --> 00:33:35.259 code that runs your world. Thank you for joining 00:33:35.259 --> 00:33:36.299 us on The Deep Dive.