WEBVTT 00:00:00.000 --> 00:00:01.980 Okay, let's unpack this. We're diving deep today 00:00:01.980 --> 00:00:06.000 into the source material on just a giant in the 00:00:06.000 --> 00:00:09.240 field, James Gosling. Yeah, an absolute architect 00:00:09.240 --> 00:00:11.660 of modern computing. Right. And if you've ever, 00:00:11.759 --> 00:00:13.419 I don't know, run an enterprise application, 00:00:13.740 --> 00:00:16.640 loaded a web page back in the late 90s, or even 00:00:16.640 --> 00:00:19.339 used an Android phone, his work has touched your 00:00:19.339 --> 00:00:22.019 life. Definitely. He gets this title, the father 00:00:22.019 --> 00:00:24.039 of Java, which is true, of course. It's deserved, 00:00:24.300 --> 00:00:26.219 yeah. But our sources really show that it... 00:00:26.699 --> 00:00:28.879 Well, it barely scratches the surface of his 00:00:28.879 --> 00:00:30.660 actual influence. That's what we want to get 00:00:30.660 --> 00:00:33.619 into. The sources stitch together this amazing 00:00:33.619 --> 00:00:36.520 narrative, right? It covers decades of innovation. 00:00:36.700 --> 00:00:39.719 It tracks him from his roots in Canada to Sun 00:00:39.719 --> 00:00:43.340 Microsystems. And then Oracle, Google, and finally 00:00:43.340 --> 00:00:45.799 to AWS right up until his retirement, which was 00:00:45.799 --> 00:00:48.640 just a few months ago in July 2024. So this isn't 00:00:48.640 --> 00:00:51.159 just a biography. It's really, I think, a lens 00:00:51.159 --> 00:00:54.420 into the history of modern computer architecture 00:00:54.420 --> 00:00:56.579 itself. It really is. Our mission today is for 00:00:56.579 --> 00:00:59.240 you, the learner. We want to give you that critical 00:00:59.240 --> 00:01:01.899 shortcut to understanding Gosling's unique path. 00:01:02.219 --> 00:01:04.459 We're going to trace that specific intellectual 00:01:04.459 --> 00:01:07.680 journey, you know, from these surprisingly complex 00:01:07.680 --> 00:01:11.040 high school projects. All the way to this philosophy 00:01:11.040 --> 00:01:13.980 of code independence that led to the famous write 00:01:13.980 --> 00:01:17.120 once, run anywhere paradigm. Exactly. How did 00:01:17.120 --> 00:01:20.079 a student's project this need for academic portability 00:01:20.079 --> 00:01:23.099 eventually somehow morph into the infrastructure 00:01:23.099 --> 00:01:25.900 that powers global commerce? That's what we need 00:01:25.900 --> 00:01:27.900 to figure out. We're looking at a true foundational 00:01:27.900 --> 00:01:30.060 architect. He's a Canadian computer scientist 00:01:30.060 --> 00:01:33.799 born in Calgary, May 19th, 1955. And he just 00:01:33.799 --> 00:01:37.140 he profoundly shaped how programs get distributed 00:01:37.140 --> 00:01:39.420 and executed on all kinds of different hardware. 00:01:39.620 --> 00:01:41.840 Yeah, his early work really defined the philosophy 00:01:41.840 --> 00:01:44.560 that let the Internet and later cloud computing 00:01:44.560 --> 00:01:47.480 scale the way they did. So let's start at the 00:01:47.480 --> 00:01:49.959 very beginning. Calgary, Alberta. Right. The 00:01:49.959 --> 00:01:51.920 sources mention his background, you know, English, 00:01:52.099 --> 00:01:54.900 Welsh, Scottish, Icelandic descent. But what's 00:01:54.900 --> 00:01:56.620 really fascinating is how early he was thrown 00:01:56.620 --> 00:01:59.400 into designing real high -stakes systems. In 00:01:59.400 --> 00:02:01.560 high school. That's the part that's just mind 00:02:01.560 --> 00:02:03.879 -boggling. He's at William Amberhart High School. 00:02:03.959 --> 00:02:06.620 And while other kids are doing, I don't know, 00:02:06.700 --> 00:02:08.780 textbook problems. He's solving problems for 00:02:08.780 --> 00:02:11.460 space physicists, not an exaggeration. No, it's 00:02:11.460 --> 00:02:14.409 not. The sources are clear. While he's still 00:02:14.409 --> 00:02:16.569 in high school, he gets a job writing software 00:02:16.569 --> 00:02:19.810 to analyze data for the University of Calgary's 00:02:19.810 --> 00:02:22.229 physics department. And this wasn't like some 00:02:22.229 --> 00:02:25.610 simulated data set for a class project. This 00:02:25.610 --> 00:02:28.849 was hard, raw telemetry coming straight from 00:02:28.849 --> 00:02:31.669 the ISIS -2 satellite. Okay, wait. An ISIS -2 00:02:31.669 --> 00:02:35.520 satellite? Yeah. How? How common could that have 00:02:35.520 --> 00:02:38.159 been for a high school kid in Calgary in the 00:02:38.159 --> 00:02:41.979 1970s to be trusted with actual space data? Almost 00:02:41.979 --> 00:02:43.919 unheard of. It just shows this incredible level 00:02:43.919 --> 00:02:46.800 of trust in his abilities from a very early age. 00:02:46.919 --> 00:02:48.699 But more than that, it forced him to think about 00:02:48.699 --> 00:02:50.719 system robustness right out of the gate, didn't 00:02:50.719 --> 00:02:53.460 it? It had to. The ISIS -2 satellite, which was 00:02:53.460 --> 00:02:56.740 launched by Canada in 71, was studying the ionosphere. 00:02:57.400 --> 00:03:00.139 Analyzing that kind of data is, well, it's way 00:03:00.139 --> 00:03:02.240 more than just basic coding. What are the challenges 00:03:02.240 --> 00:03:04.780 there? You're dealing with huge volumes of data. 00:03:04.979 --> 00:03:07.919 It's potentially noisy. It's coming in asynchronously 00:03:07.919 --> 00:03:11.900 and it has to be processed accurately. One mistake 00:03:11.900 --> 00:03:14.560 and you could ruin crucial scientific analysis. 00:03:14.919 --> 00:03:18.120 So failure has real consequences. Big ones. That 00:03:18.120 --> 00:03:20.419 early exposure to high stakes data processing 00:03:20.419 --> 00:03:23.800 forced him to prioritize robustness and scalability 00:03:23.800 --> 00:03:26.620 from day one. It just became part of his DNA 00:03:26.620 --> 00:03:29.719 as an engineer. And that experience clearly laid 00:03:29.719 --> 00:03:32.949 the groundwork for a whole career. focused on 00:03:32.949 --> 00:03:36.610 building reliable systems. He stayed close to 00:03:36.610 --> 00:03:38.810 home for his undergraduate degree. Yep, got his 00:03:38.810 --> 00:03:41.009 Bachelor of Science from the University of Calgary. 00:03:41.520 --> 00:03:43.740 But then he follows that classic path for anyone 00:03:43.740 --> 00:03:46.020 wanting to do advanced research. He moves to 00:03:46.020 --> 00:03:48.139 Carnegie Mellon University, CMU. Right, for his 00:03:48.139 --> 00:03:51.020 MA and his PhD in computer science. And CMU at 00:03:51.020 --> 00:03:53.159 that time was, I mean, it still is, but it was 00:03:53.159 --> 00:03:55.780 just this hotbed of computing innovation. His 00:03:55.780 --> 00:03:58.360 time there was focused on theory, sure, but he 00:03:58.360 --> 00:04:00.479 also produced these really practical tools that 00:04:00.479 --> 00:04:03.300 show he was both a builder and a thinker. Let's 00:04:03.300 --> 00:04:04.620 talk about the first one, which is really about 00:04:04.620 --> 00:04:06.819 developer tools. It's something coders would 00:04:06.819 --> 00:04:09.259 know. Yes. While he was working on his doctorate, 00:04:09.319 --> 00:04:12.319 he wrote his own version of the famous text editor 00:04:12.319 --> 00:04:16.579 Emacs. He called it Gosling Emacs or Gosmax for 00:04:16.579 --> 00:04:19.699 short. And Emacs is not a simple program. It's 00:04:19.699 --> 00:04:22.480 one of the most foundational, extensible, complex 00:04:22.480 --> 00:04:25.500 developer tools in history. It is. So for him 00:04:25.500 --> 00:04:28.720 to write his own robust version, it just shows 00:04:28.720 --> 00:04:31.779 this deep, deep engagement with the very process 00:04:31.779 --> 00:04:33.860 of creating software. It's like he wanted to 00:04:33.860 --> 00:04:35.759 improve the foundational environment where all 00:04:35.759 --> 00:04:37.759 the work gets done. But then there's the other 00:04:37.759 --> 00:04:39.680 project from that time, which is maybe even more 00:04:39.680 --> 00:04:41.560 important in terms of what came later with Java. 00:04:41.819 --> 00:04:43.920 It's about system design. This is where it gets 00:04:43.920 --> 00:04:45.420 really interesting, especially when you think 00:04:45.420 --> 00:04:48.279 about architecture, because even before he joined 00:04:48.279 --> 00:04:51.220 Sun, Gosling was dealing with concurrency and 00:04:51.220 --> 00:04:53.629 parallel computing at a level that. Frankly, 00:04:53.689 --> 00:04:55.709 very few people were. And the sources have a 00:04:55.709 --> 00:04:58.129 specific example of this. They do. They detail 00:04:58.129 --> 00:05:00.810 how he built a multiprocessor version of Unix, 00:05:00.889 --> 00:05:04.350 specifically for a 16 -way computer system that 00:05:04.350 --> 00:05:07.649 they had at CMU. 16 processors. Let's just pause 00:05:07.649 --> 00:05:11.689 on that. In the early 1980s, 16 processors all 00:05:11.689 --> 00:05:14.769 trying to cooperate on one system. That's bleeding 00:05:14.769 --> 00:05:17.089 edge stuff. Absolutely. It's a massive engineering 00:05:17.089 --> 00:05:18.970 challenge. You're designing an operating system 00:05:18.970 --> 00:05:21.670 kernel that has to manage all of that communication, 00:05:21.949 --> 00:05:24.870 all 16 processors talking at once. He forces 00:05:24.870 --> 00:05:28.189 you to internalize things like concurrency, thread 00:05:28.189 --> 00:05:30.670 safety, resource synchronization. All of it. 00:05:30.689 --> 00:05:33.110 If you're building a system to manage that kind 00:05:33.110 --> 00:05:35.889 of parallelism, you are inherently forced to 00:05:35.889 --> 00:05:39.209 think about how code behaves in these messy, 00:05:39.350 --> 00:05:42.029 nonlinear distributed environments. Which are 00:05:42.029 --> 00:05:45.060 the exact problems. that Java years later was 00:05:45.060 --> 00:05:47.459 designed to solve for everyone else. Precisely. 00:05:47.459 --> 00:05:49.639 It's that old saying, you know, necessity is 00:05:49.639 --> 00:05:51.959 the mother of invention. A high -level theoretical 00:05:51.959 --> 00:05:54.439 side of all this work then culminated in his 00:05:54.439 --> 00:05:57.620 doctoral thesis in 1983. And the title was Algebraic 00:05:57.620 --> 00:05:59.639 Constraints. It sounds super academic, I know, 00:05:59.720 --> 00:06:02.199 but the core idea is central to modern software 00:06:02.199 --> 00:06:04.600 engineering. It's all about defining and managing 00:06:04.600 --> 00:06:07.199 complex relationships within a structured system. 00:06:08.000 --> 00:06:10.839 Think of constraints as just, strict rules. If 00:06:10.839 --> 00:06:13.620 A happens, B must follow, or C must never happen. 00:06:13.879 --> 00:06:16.800 So how does an abstract theory about relationships 00:06:16.800 --> 00:06:19.720 and rules actually show up in a practical language 00:06:19.720 --> 00:06:22.339 like Java? It's everywhere in Java's design. 00:06:22.480 --> 00:06:24.399 If you look at Java, what are its defining features? 00:06:24.660 --> 00:06:27.720 It has strict type safety. It has compulsory 00:06:27.720 --> 00:06:30.800 exception handling. It has a managed memory model. 00:06:30.899 --> 00:06:33.180 These are all ways of algebraically constraining 00:06:33.180 --> 00:06:35.300 the programmer. Right, so unlike older languages 00:06:35.300 --> 00:06:37.600 like C. which give you enough rope to hang yourself. 00:06:37.879 --> 00:06:40.699 Java enforces rules that limit how much chaos 00:06:40.699 --> 00:06:43.879 you can introduce. His thesis on algebraic constraints 00:06:43.879 --> 00:06:46.560 is the intellectual foundation for Java's whole 00:06:46.560 --> 00:06:49.639 design philosophy. Predictable, rule -based, 00:06:49.839 --> 00:06:53.240 safe execution. Okay, so his student career creates 00:06:53.240 --> 00:06:56.389 this perfect trifecta. You've got... deep engagement 00:06:56.389 --> 00:06:59.569 with developer tools with gosmax mastery of high 00:06:59.569 --> 00:07:02.149 -level parallel systems with that 16 -way unix 00:07:02.149 --> 00:07:04.589 build and a theoretical foundation built on strict 00:07:04.589 --> 00:07:06.990 rule -based design from his thesis yeah that 00:07:06.990 --> 00:07:09.589 exact skill set just seems tailor -made for a 00:07:09.589 --> 00:07:11.490 place like sun microsystems and that's exactly 00:07:11.490 --> 00:07:14.189 where he went that trajectory led him straight 00:07:14.189 --> 00:07:17.310 to sun microsystems in 1984 and he stayed there 00:07:17.310 --> 00:07:20.490 for an incredible 26 years 26 years i mean that's 00:07:20.490 --> 00:07:24.379 an entire career and in that time He didn't just 00:07:24.379 --> 00:07:26.360 define his own career, he really defined the 00:07:26.360 --> 00:07:29.160 identity of Sun, which for so long was all about 00:07:29.160 --> 00:07:31.579 innovation and networking and open systems. It's 00:07:31.579 --> 00:07:33.420 really important to remember, though, that before 00:07:33.420 --> 00:07:36.920 he ever got to Java, there were other huge projects. 00:07:37.259 --> 00:07:39.899 And some of them failed strategically, but were 00:07:39.899 --> 00:07:42.720 still brilliant technically. The prime example, 00:07:42.959 --> 00:07:46.529 absolutely, is New WS. the network extensible 00:07:46.529 --> 00:07:49.850 window system. What was NewWS trying to do? NewWS 00:07:49.850 --> 00:07:53.350 was Gosling's big idea for an early Unix windowing 00:07:53.350 --> 00:07:55.829 system. And on a technical level, it was brilliant. 00:07:56.009 --> 00:07:58.110 It was based on the PostScript language. The 00:07:58.110 --> 00:08:00.149 same language used for printing documents. The 00:08:00.149 --> 00:08:02.810 very same. Which made it incredibly powerful 00:08:02.810 --> 00:08:05.829 for rendering complex graphics and text consistently, 00:08:06.009 --> 00:08:08.730 especially across a network. It let you display 00:08:08.730 --> 00:08:11.410 applications remotely, which was a crucial feature 00:08:11.410 --> 00:08:13.290 for the distributed computing world that Sun 00:08:13.290 --> 00:08:16.379 was championing. But technical superiority, as 00:08:16.379 --> 00:08:19.600 we see time and time again, often loses out to 00:08:19.600 --> 00:08:21.519 business strategy. And this is where it gets 00:08:21.519 --> 00:08:23.259 really interesting because new WS completely 00:08:23.259 --> 00:08:26.259 failed to become the standard. It lost the war. 00:08:26.360 --> 00:08:28.800 It lost the war for the desktop to the X Windows 00:08:28.800 --> 00:08:32.200 system, which is still in some ways with us today. 00:08:32.259 --> 00:08:35.980 So why? Why did a technically superior solution 00:08:35.980 --> 00:08:39.710 from a genius like Gosling just... Fizzle out. 00:08:39.830 --> 00:08:43.529 It boils down to a classic and very costly strategic 00:08:43.529 --> 00:08:46.409 mistake. The sources are unanimous on this. It 00:08:46.409 --> 00:08:48.730 was all about licensing. Sun kept it proprietary. 00:08:49.190 --> 00:08:52.309 They did. Sun did not grant New WS an open source 00:08:52.309 --> 00:08:54.889 license. It was a closed proprietary system. 00:08:55.009 --> 00:08:57.570 And that was a fatal miscalculation in the world 00:08:57.570 --> 00:08:59.870 of network Unix computing. Right. Because the 00:08:59.870 --> 00:09:02.230 X Windows system was managed by the X Consortium. 00:09:02.230 --> 00:09:04.830 Its license was super permissive. Basically free 00:09:04.830 --> 00:09:08.649 for anyone to use, adapt, extend. Exactly. yourself 00:09:08.649 --> 00:09:11.190 in the shoes of a hardware vendor in the 80s, 00:09:11.210 --> 00:09:13.929 you're HP, you're DEC, you're an independent 00:09:13.929 --> 00:09:16.029 software developer. Are you going to bet your 00:09:16.029 --> 00:09:18.529 entire future on a closed system that's completely 00:09:18.529 --> 00:09:21.070 controlled by your competitor, son? No way. You're 00:09:21.070 --> 00:09:23.809 going with the open consortium governed standard 00:09:23.809 --> 00:09:25.549 that you can actually use without asking for 00:09:25.549 --> 00:09:27.549 permission or paying a fee. And that's what everyone 00:09:27.549 --> 00:09:29.990 did. The proprietary barrier just stopped the 00:09:29.990 --> 00:09:32.409 ecosystem from ever growing. This was a multi 00:09:32.409 --> 00:09:34.809 -million dollar battle for the future of how 00:09:34.809 --> 00:09:37.879 we interact with networked computers. And the 00:09:37.879 --> 00:09:40.039 better engineering lost to the better licensing 00:09:40.039 --> 00:09:42.820 strategy. It was a powerful and I'm sure very 00:09:42.820 --> 00:09:45.799 painful lesson for Gosling and for Sun's leadership. 00:09:46.200 --> 00:09:48.820 The sources strongly suggest this failure directly 00:09:48.820 --> 00:09:51.200 informed their strategy for Java a few years 00:09:51.200 --> 00:09:54.019 later. How so? The initial widespread availability 00:09:54.019 --> 00:09:56.840 and the permissive licensing of Java were a direct 00:09:56.840 --> 00:09:59.879 reaction to proprietary trap that swallowed new 00:09:59.879 --> 00:10:03.559 WS. The lesson was clear. If you want to create 00:10:03.559 --> 00:10:06.320 a global standard, pervasive adoption is everything, 00:10:06.580 --> 00:10:08.879 even if it means giving up some of that initial 00:10:08.879 --> 00:10:11.559 proprietary control. He even co -authors the 00:10:11.559 --> 00:10:14.600 New W .S. book in 1989. But by then, the writing 00:10:14.600 --> 00:10:16.460 was on the wall. The market had already made 00:10:16.460 --> 00:10:19.200 its decision. OK, that context is so vital. Now 00:10:19.200 --> 00:10:21.740 let's pivot from that very visible failure of 00:10:21.740 --> 00:10:25.460 New W .S. to a more obscure, but maybe equally 00:10:25.460 --> 00:10:28.220 crucial. Little technical success. It really 00:10:28.220 --> 00:10:30.279 highlights his mastery of the low -level plumbing 00:10:30.279 --> 00:10:32.340 of systems. You're talking about SHAR? The bundle 00:10:32.340 --> 00:10:34.840 program known as SHAR. SHAR stands for Shell 00:10:34.840 --> 00:10:37.740 Archive. It's a low -level Unix utility that 00:10:37.740 --> 00:10:40.379 Gosling co -wrote. And to understand why it was 00:10:40.379 --> 00:10:42.399 so important, you have to remember what that 00:10:42.399 --> 00:10:45.419 era was like. This is before reliable tools like 00:10:45.419 --> 00:10:48.320 Xyp, before stable binary transfer protocols 00:10:48.320 --> 00:10:50.840 were everywhere. So just sending a group of files 00:10:50.840 --> 00:10:53.120 from one Unix machine to another could be a huge 00:10:53.120 --> 00:10:55.259 pain. A huge pain. You had all these different... 00:10:55.529 --> 00:10:58.169 Potentially incompatible systems. How do you 00:10:58.169 --> 00:11:00.389 get a whole directory of source code from one 00:11:00.389 --> 00:11:02.850 to another reliably? It seems like such a mundane 00:11:02.850 --> 00:11:05.570 problem, but it's foundational. So how did Char 00:11:05.570 --> 00:11:07.919 solve it? It was so simple. It was brilliant. 00:11:08.159 --> 00:11:10.799 It took a bunch of files and bundled them into 00:11:10.799 --> 00:11:13.460 a single massive shell script. So it's not a 00:11:13.460 --> 00:11:15.679 compressed file. It's a script. Exactly. When 00:11:15.679 --> 00:11:18.000 the person on the other end received this single 00:11:18.000 --> 00:11:20.580 file and executed it, the shell archive would 00:11:20.580 --> 00:11:23.360 just unpack itself, recreating all the original 00:11:23.360 --> 00:11:26.360 files and directories. It leveraged the one thing 00:11:26.360 --> 00:11:28.440 you could count on being compatible across every 00:11:28.440 --> 00:11:30.919 single Unix system. The born shell interpreter. 00:11:31.159 --> 00:11:33.720 Yeah, the shell. It turns your entire package 00:11:33.720 --> 00:11:36.580 of files into a self -executing text. file, which 00:11:36.580 --> 00:11:39.220 was way easier and more reliable to send through 00:11:39.220 --> 00:11:42.700 primitive email systems or early FTP, it completely 00:11:42.700 --> 00:11:45.120 avoided the nightmare of inconsistent binary 00:11:45.120 --> 00:11:47.639 formats. It's another example of that architecture 00:11:47.639 --> 00:11:50.500 -neutral thinking, isn't it? Find the one common 00:11:50.500 --> 00:11:53.159 layer that exists everywhere and write for that. 00:11:53.419 --> 00:11:55.820 That's a great way to put it. And the sources 00:11:55.820 --> 00:11:58.919 confirm its importance. This little utility was 00:11:58.919 --> 00:12:01.059 detailed in the absolute bible of the field, 00:12:01.220 --> 00:12:04.220 the Unix programming environment, by Brian Kernaghan 00:12:04.220 --> 00:12:06.879 and Rob Pike. Pages 97 to 100, specifically. 00:12:07.240 --> 00:12:09.240 If your tool makes it into that book, you've 00:12:09.240 --> 00:12:11.659 achieved a kind of permanent foundational status. 00:12:12.019 --> 00:12:14.840 This dual capability is what's so striking. He 00:12:14.840 --> 00:12:17.100 could design a huge, complex, network -aware 00:12:17.100 --> 00:12:19.620 graphics system like New WS. And at the same 00:12:19.620 --> 00:12:22.639 time, engineer a simple, totally robust, low 00:12:22.639 --> 00:12:25.389 -level utility like Char. He understood the whole 00:12:25.389 --> 00:12:27.529 stack from the shell script all the way up to 00:12:27.529 --> 00:12:29.789 the network graphics layer. And that duality 00:12:29.789 --> 00:12:31.710 is the perfect bedrock for building something 00:12:31.710 --> 00:12:34.110 like Java, a language that was designed to span 00:12:34.110 --> 00:12:37.110 all of those layers. And so now we get to the 00:12:37.110 --> 00:12:42.269 main event, the Magnum Opus. In 1994, James Gosling 00:12:42.269 --> 00:12:44.769 is formally credited with inventing the Java 00:12:44.769 --> 00:12:47.409 programming language. And he was the lead designer. 00:12:47.889 --> 00:12:50.370 This all came out of an internal project at Sun 00:12:50.370 --> 00:12:53.049 called the Green Project, right? Which was originally 00:12:53.049 --> 00:12:55.799 aimed at... like smart toasters and consumer 00:12:55.799 --> 00:12:58.740 electronics. It was, yeah. But the true potential 00:12:58.740 --> 00:13:00.980 of the language became clear very, very quickly. 00:13:01.240 --> 00:13:03.600 And crucially, he wasn't just the guy with the 00:13:03.600 --> 00:13:06.639 idea. He wasn't just the architect drawing blueprints. 00:13:06.879 --> 00:13:09.980 The sources confirm he did the hard work. He 00:13:09.980 --> 00:13:12.059 implemented the language's original compiler 00:13:12.059 --> 00:13:15.500 and virtual machine, VM, himself. That's so critical. 00:13:15.620 --> 00:13:17.799 He owned the blueprint and the first working 00:13:17.799 --> 00:13:21.340 model. He built it. So Java was created to solve 00:13:21.340 --> 00:13:24.179 the ultimate problem of a networked world. Platform 00:13:24.179 --> 00:13:27.179 incompatibility. Software for a SunSpark machine 00:13:27.179 --> 00:13:30.179 wouldn't run on an Intel PC, which wouldn't run 00:13:30.179 --> 00:13:32.440 on an Apple computer. And Java's big idea was 00:13:32.440 --> 00:13:34.860 to introduce this concept of architecture -neutral 00:13:34.860 --> 00:13:37.600 execution to just wipe that problem away. Okay, 00:13:37.639 --> 00:13:39.399 let's break that down. What does architecture 00:13:39.399 --> 00:13:41.559 -neutral execution actually mean for a developer 00:13:41.559 --> 00:13:44.419 writing code? It means you, the developer, you 00:13:44.419 --> 00:13:47.139 stop writing code for a specific chip. You're 00:13:47.139 --> 00:13:49.360 not targeting the Intel instruction set or the 00:13:49.360 --> 00:13:51.419 Spark instruction set. You just write your Java 00:13:51.419 --> 00:13:54.080 source code. And then the Java compiler... takes 00:13:54.080 --> 00:13:56.840 that source code and translates it into an intermediate 00:13:56.840 --> 00:13:59.610 format. Right. Called bytecode. And this bytecode 00:13:59.610 --> 00:14:01.850 is the magic ingredient. It's not machine specific 00:14:01.850 --> 00:14:05.190 at all. It's specific to this standardized imaginary 00:14:05.190 --> 00:14:07.710 computer that we call the Java Virtual Machine 00:14:07.710 --> 00:14:11.429 or the JVM. So the JVM is like a universal translator 00:14:11.429 --> 00:14:13.690 that you install on every single physical computer. 00:14:14.009 --> 00:14:17.370 The bytecode, your program, then runs exactly 00:14:17.370 --> 00:14:19.970 the same on a giant server or a little desktop. 00:14:20.090 --> 00:14:22.330 Because it's always executing on the exact same 00:14:22.330 --> 00:14:25.899 abstract machine. And that is the core of the 00:14:25.899 --> 00:14:28.340 write once, run anywhere philosophy. That's the 00:14:28.340 --> 00:14:30.220 whole game. And what makes this deep dive so 00:14:30.220 --> 00:14:33.000 fascinating is that this revolutionary idea, 00:14:33.240 --> 00:14:36.360 it wasn't a sudden flash of genius. It didn't 00:14:36.360 --> 00:14:38.960 come out of nowhere. The sources let us trace 00:14:38.960 --> 00:14:42.340 the idea back to two very specific earlier technical 00:14:42.340 --> 00:14:44.659 challenges he had faced. These were like proofs 00:14:44.659 --> 00:14:47.620 of concept for virtualization. Exactly. Let's 00:14:47.620 --> 00:14:49.340 start with the immediate trigger. The thing that 00:14:49.340 --> 00:14:51.659 happened at Sun that directly inspired the Java 00:14:51.659 --> 00:14:55.110 VM design. Okay. The idea came to him while he 00:14:55.110 --> 00:14:58.009 was working on a really tedious task, porting 00:14:58.009 --> 00:15:02.090 software from a niche piece of hardware, a PERQ 00:15:02.090 --> 00:15:04.409 computer. A PERQ computer. Never heard of it. 00:15:04.620 --> 00:15:06.379 Not many people have. It was a high -performance 00:15:06.379 --> 00:15:09.399 graphics workstation, but it had its own proprietary 00:15:09.399 --> 00:15:13.259 system. And Sun needed to move its software off 00:15:13.259 --> 00:15:15.539 of the PERQ and onto more standard computers. 00:15:15.879 --> 00:15:17.419 So rewriting everything from scratch would be 00:15:17.419 --> 00:15:19.720 a nightmare. A total nightmare. The challenge 00:15:19.720 --> 00:15:22.259 was that the PERQ had its own internal code. 00:15:22.480 --> 00:15:25.580 So Gosling's solution was essentially to build 00:15:25.580 --> 00:15:29.879 an emulator. The PERQ used this internal language 00:15:29.879 --> 00:15:32.440 called Q -code. So Q -code was its machine logic. 00:15:32.559 --> 00:15:34.649 It was its instruction set, yeah. So Gosling 00:15:34.649 --> 00:15:36.769 wrote a program that would translate that Q code 00:15:36.769 --> 00:15:39.070 into instructions for a more standard machine 00:15:39.070 --> 00:15:41.570 at the time, the VX assembler. And then on top 00:15:41.570 --> 00:15:44.029 of that, he built a layer that emulated the PERQ 00:15:44.029 --> 00:15:47.879 hardware on the new VX machine. Wow. So he was 00:15:47.879 --> 00:15:50.620 translating and mimicking one machine inside 00:15:50.620 --> 00:15:53.200 of another. That whole process, translating one 00:15:53.200 --> 00:15:56.299 instruction set, cue code into another, VAC -X 00:15:56.299 --> 00:15:58.379 assembler with an emulation layer in the middle, 00:15:58.480 --> 00:16:00.659 that sounds like a prototype for the JVM. It 00:16:00.659 --> 00:16:03.080 was the industrial prototype, exactly. He saw 00:16:03.080 --> 00:16:05.840 firsthand the power of writing code for an abstract 00:16:05.840 --> 00:16:08.620 machine instead of a real one. But the sources 00:16:08.620 --> 00:16:10.820 confirmed that this idea, this philosophy, it 00:16:10.820 --> 00:16:13.159 was a much deeper pattern for him. It went all 00:16:13.159 --> 00:16:16.139 the way back to his academic days at CMU. This 00:16:16.139 --> 00:16:18.539 is the critical connection. So the impulse for 00:16:18.539 --> 00:16:20.679 virtualization was already there years before. 00:16:20.820 --> 00:16:24.059 Years before. Back in his student days at CMU, 00:16:24.120 --> 00:16:26.740 he created a P -code virtual machine for the 00:16:26.740 --> 00:16:30.299 lab's DEC VAX computer. So this was for a class, 00:16:30.360 --> 00:16:33.700 for a professor. The goal was academic. His professor 00:16:33.700 --> 00:16:35.700 wanted to run programs that were written in a 00:16:35.700 --> 00:16:38.259 language called UCSD Pascal. And why couldn't 00:16:38.259 --> 00:16:40.179 they just run the Pascal programs directly on 00:16:40.179 --> 00:16:42.740 the VAX? Because UCSD Pascal was specifically 00:16:42.740 --> 00:16:45.370 designed to be portable. When you compiled a 00:16:45.370 --> 00:16:48.669 program in UCSD Pascal, it didn't generate VX 00:16:48.669 --> 00:16:50.690 machine code. It generated something called P 00:16:50.690 --> 00:16:52.830 -code portable code. An intermediate language, 00:16:53.049 --> 00:16:55.769 just like Java's bytecode. Exactly like it. And 00:16:55.769 --> 00:16:58.009 to run that P -code, you needed an interpreter, 00:16:58.250 --> 00:17:00.590 a virtual machine running on the target hardware. 00:17:00.950 --> 00:17:04.230 So Gosling built that P -code VM for the VX so 00:17:04.230 --> 00:17:06.789 that his professor and the lab could easily run 00:17:06.789 --> 00:17:09.190 programs written in this cross -platform academic 00:17:09.190 --> 00:17:12.089 language. So the same core principle was tested 00:17:12.089 --> 00:17:15.059 in two totally different contexts. First in academia, 00:17:15.380 --> 00:17:17.980 to make teaching and research portable with P 00:17:17.980 --> 00:17:20.700 -code. And then second, in industry, to solve 00:17:20.700 --> 00:17:23.200 a messy commercial porting problem with Q -code. 00:17:23.359 --> 00:17:25.519 And when you put those two historical dots together. 00:17:26.109 --> 00:17:27.990 You get the Java virtual machine. That's the 00:17:27.990 --> 00:17:30.470 synthesis. It allowed him to articulate the ultimate 00:17:30.470 --> 00:17:33.170 goal for Java, achieving architecture -neutral 00:17:33.170 --> 00:17:36.369 execution for widely distributed programs. By 00:17:36.369 --> 00:17:39.390 hiding the hardware behind the JVM, and this 00:17:39.390 --> 00:17:41.990 is key, giving that JVM away with a permissive 00:17:41.990 --> 00:17:44.750 license, remembering the lesson from UWS. He 00:17:44.750 --> 00:17:46.990 achieved the pervasive adoption he needed to 00:17:46.990 --> 00:17:50.529 define an entire era of computing. That layer 00:17:50.529 --> 00:17:53.130 of abstraction is why Java became the universal 00:17:53.130 --> 00:17:56.789 backbone for large, complex, distributed systems. 00:17:57.269 --> 00:18:00.970 The VM is really the whole story. Okay, so now 00:18:00.970 --> 00:18:03.210 the story takes a sharp turn. We're moving from 00:18:03.210 --> 00:18:05.970 this period of intense architectural creation 00:18:05.970 --> 00:18:08.990 to one of corporate transitions and, frankly, 00:18:09.130 --> 00:18:12.410 a lot of conflict. A huge shift. After 26 years 00:18:12.410 --> 00:18:15.710 at Sun Microsystems, a lifetime in tech, James 00:18:15.710 --> 00:18:19.559 Gosling leaves. April 2nd, 2010. And this happens 00:18:19.559 --> 00:18:21.920 right after Sun's massive acquisition by Oracle 00:18:21.920 --> 00:18:24.140 Corporation. And this departure isn't just a 00:18:24.140 --> 00:18:25.640 footnote in a corporate history book. This is 00:18:25.640 --> 00:18:27.039 the moment where he goes from being the chief 00:18:27.039 --> 00:18:29.440 architect to being a very candid, very public 00:18:29.440 --> 00:18:31.400 industry critic. He didn't just fade away. He 00:18:31.400 --> 00:18:33.119 was very clear about why he left. Oh, extremely. 00:18:33.380 --> 00:18:35.980 He didn't issue one of those bland, sanitized 00:18:35.980 --> 00:18:38.519 press releases. Our sources detail the four specific 00:18:38.519 --> 00:18:40.640 reasons he gave for walking away from the technology 00:18:40.640 --> 00:18:42.799 he created and the company he built. And when 00:18:42.799 --> 00:18:44.539 you look at them together, they just paint this 00:18:44.539 --> 00:18:46.400 picture of a total breakdown of trust and culture. 00:18:46.579 --> 00:18:48.869 So what were they? What were the four reasons 00:18:48.869 --> 00:18:50.849 that pushed him out the door after more than 00:18:50.849 --> 00:18:55.569 two decades? He listed reductions in pay, reductions 00:18:55.569 --> 00:18:58.329 in status, reductions in his decision making 00:18:58.329 --> 00:19:01.660 ability, and the real kicker. Ethical challenges. 00:19:01.839 --> 00:19:04.559 Ethical challenges. That's a strong phrase. It's 00:19:04.559 --> 00:19:06.640 a very strong phrase. It really paints a picture 00:19:06.640 --> 00:19:08.819 of a culture changing overnight. It went from 00:19:08.819 --> 00:19:11.140 a place that valued long term research and the 00:19:11.140 --> 00:19:14.059 intellectual freedom of its top engineers to 00:19:14.059 --> 00:19:16.380 a place focused on immediate cost cutting and 00:19:16.380 --> 00:19:18.299 corporate control. It sounds like the culture 00:19:18.299 --> 00:19:20.880 shock was just too much. Moving from a company 00:19:20.880 --> 00:19:24.039 like Sun, which was very R &D driven, to a much 00:19:24.039 --> 00:19:26.599 more sales and lawyer driven enterprise like 00:19:26.599 --> 00:19:29.849 Oracle. He felt disempowered and it seems morally 00:19:29.849 --> 00:19:32.289 conflicted. And that tension, that underlying 00:19:32.289 --> 00:19:34.670 conflict between the old Sun engineers and the 00:19:34.670 --> 00:19:37.170 new Oracle management, it all came to a head 00:19:37.170 --> 00:19:40.529 in the famous Oracle v. Google lawsuit. The lawsuit 00:19:40.529 --> 00:19:44.450 over Android's use of Java APIs. And Gosling's 00:19:44.450 --> 00:19:47.349 position in all of this was... Well, it was really 00:19:47.349 --> 00:19:50.109 complex, almost paradoxical. It was. He was caught 00:19:50.109 --> 00:19:52.490 between his loyalty to his creation and his clear 00:19:52.490 --> 00:19:54.930 aversion to his new employer. Let's start with 00:19:54.930 --> 00:19:58.549 his view on the acquisition itself. He was pretty 00:19:58.549 --> 00:20:01.049 dismayed by the aggressive legal posture Oracle 00:20:01.049 --> 00:20:03.289 took right away, wasn't he? He was highly critical. 00:20:03.450 --> 00:20:06.230 He gave this very memorable quote about the early 00:20:06.230 --> 00:20:08.089 integration meetings between the two companies. 00:20:08.170 --> 00:20:10.910 He said, and I'm quoting here, that during the 00:20:10.910 --> 00:20:13.109 integration meetings between Sun and Oracle, 00:20:13.410 --> 00:20:16.509 we could see the Oracle lawyers' eyes sparkle 00:20:16.509 --> 00:20:18.650 when they talked about Sun's patent situation. 00:20:19.109 --> 00:20:21.980 Their eyes sparkle. That image just perfectly 00:20:21.980 --> 00:20:24.339 captures that shift from an engineering culture 00:20:24.339 --> 00:20:28.160 to a revenue through litigation culture. It does. 00:20:28.279 --> 00:20:30.579 He clearly felt the acquisition was less about 00:20:30.579 --> 00:20:32.859 leveraging Sun's amazing technology and more 00:20:32.859 --> 00:20:35.299 about aggressively monetizing his portfolio of 00:20:35.299 --> 00:20:37.559 patents and copyrights. And that had to be the 00:20:37.559 --> 00:20:40.559 core of his ethical challenges. And yet, when 00:20:40.559 --> 00:20:43.099 the legal fight with Google got going, specifically 00:20:43.099 --> 00:20:45.779 over whether Google's implementation of the Java 00:20:45.779 --> 00:20:49.529 APIs in Android was copyright infringement. He 00:20:49.529 --> 00:20:52.130 publicly sided with Oracle. Which surprised everyone, 00:20:52.250 --> 00:20:54.490 given how he laughed and what he'd said. He had 00:20:54.490 --> 00:20:57.069 to clarify it. He said explicitly, while I have 00:20:57.069 --> 00:20:59.490 differences with Oracle, in this case they are 00:20:59.490 --> 00:21:04.410 in the right, Google totally slimed Sun. So what 00:21:04.410 --> 00:21:07.230 was behind that? Why did he feel Google had slimed 00:21:07.230 --> 00:21:09.940 Sun? This is so important to understand. His 00:21:09.940 --> 00:21:12.140 loyalty wasn't to Oracle's management. It was 00:21:12.140 --> 00:21:15.019 to the 26 years of work that he and his colleagues 00:21:15.019 --> 00:21:17.480 had poured into Java at Sun. He was defending 00:21:17.480 --> 00:21:20.900 the work itself. Yes. He said that many at Sun 00:21:20.900 --> 00:21:23.980 were disturbed by Google's actions. The feeling 00:21:23.980 --> 00:21:27.000 was... essentially one of betrayal that Google, 00:21:27.140 --> 00:21:29.180 instead of building their own platform from scratch 00:21:29.180 --> 00:21:31.559 or paying for a full license, had shortcut the 00:21:31.559 --> 00:21:34.059 process by just copying the architecture of the 00:21:34.059 --> 00:21:37.339 Java APIs. They were piggybacking on years of 00:21:37.339 --> 00:21:40.880 work and the entire ecosystem that Sun had painstakingly 00:21:40.880 --> 00:21:43.210 built. So he was defending the intellectual property 00:21:43.210 --> 00:21:45.589 created by the original Sun team, no matter who 00:21:45.589 --> 00:21:47.710 legally owned it at that moment. Precisely. But 00:21:47.710 --> 00:21:50.309 here's where his position gets so nuanced and 00:21:50.309 --> 00:21:52.910 I think so principled. He added a critical caveat 00:21:52.910 --> 00:21:55.289 later on regarding the court's final rulings 00:21:55.289 --> 00:21:57.650 in the case. Right. He made a powerful philosophical 00:21:57.650 --> 00:22:00.289 point. He said he approved of the court's later 00:22:00.289 --> 00:22:03.490 ruling that APIs themselves should not be copyrightable. 00:22:04.109 --> 00:22:06.250 And that's the true architect seeking, right? 00:22:06.529 --> 00:22:09.250 Despite his personal anger at how Google treated 00:22:09.250 --> 00:22:12.150 Sun, his long -term principles were aligned with 00:22:12.150 --> 00:22:15.190 the health of the entire software world. He understood 00:22:15.190 --> 00:22:17.690 that letting companies copyright basic interface 00:22:17.690 --> 00:22:20.490 specifications would just stifle innovation for... 00:22:21.120 --> 00:22:23.859 So he is prioritizing the future of computing 00:22:23.859 --> 00:22:26.920 over a single corporate victory. It's an incredible 00:22:26.920 --> 00:22:29.900 display of intellectual integrity. He could support 00:22:29.900 --> 00:22:33.579 a specific claim that Google had slimed Sun while 00:22:33.579 --> 00:22:35.960 at the same time supporting the legal precedent 00:22:35.960 --> 00:22:38.599 that ensures open collaborative innovation can 00:22:38.599 --> 00:22:41.240 continue. After this whole tumultuous period, 00:22:41.319 --> 00:22:44.259 he embarks on this really diverse post -Sun career. 00:22:44.500 --> 00:22:47.079 He doesn't just retire. He looks for new hard 00:22:47.079 --> 00:22:49.579 problems. His first move was interesting. He 00:22:49.579 --> 00:22:52.660 joins Google in March 2011, maybe trying to find 00:22:52.660 --> 00:22:54.740 a better cultural fit, but he only stays for 00:22:54.740 --> 00:22:56.920 six months. And then he makes a really fascinating 00:22:56.920 --> 00:23:00.019 choice. He follows a former colleague, Bill Voss, 00:23:00.119 --> 00:23:02.740 to a startup called Liquid Robotics. Which is 00:23:02.740 --> 00:23:05.900 such a cool move. Liquid Robotics designed and 00:23:05.900 --> 00:23:10.380 built autonomous, ocean -going robots. wave gliders 00:23:10.380 --> 00:23:12.799 that's a huge shift from enterprise software 00:23:12.799 --> 00:23:15.619 and server platforms but it's a perfect fit for 00:23:15.619 --> 00:23:18.680 a master systems architect autonomous robots 00:23:18.680 --> 00:23:21.779 require incredibly robust complex real -time 00:23:21.779 --> 00:23:24.859 distributed systems they're operating in harsh 00:23:24.859 --> 00:23:28.200 remote unreliable environments it was a chance 00:23:28.200 --> 00:23:30.339 for him to apply all his knowledge about concurrency 00:23:30.339 --> 00:23:32.740 and resilience to a physical system and liquid 00:23:32.740 --> 00:23:34.980 robotics was eventually acquired by boeing in 00:23:34.980 --> 00:23:38.019 late 2016. it was and after that he made his 00:23:38.019 --> 00:23:41.769 final big corporate move. In May 2017, he joins 00:23:41.769 --> 00:23:44.829 Amazon Web Services, AWS, as a distinguished 00:23:44.829 --> 00:23:47.170 engineer. Which feels like a fitting capstone 00:23:47.170 --> 00:23:50.250 to his career. It really does. AWS is, you could 00:23:50.250 --> 00:23:52.670 argue, the ultimate distributed computing environment 00:23:52.670 --> 00:23:55.309 on the planet. It's where global systems run 00:23:55.309 --> 00:23:57.390 at a scale that's almost impossible to comprehend. 00:23:57.690 --> 00:23:59.230 So you have the architect of the Java virtual 00:23:59.230 --> 00:24:01.950 machine, the technology that enabled distributed 00:24:01.950 --> 00:24:04.869 computing for a generation, now working at the 00:24:04.869 --> 00:24:07.230 very heart of the world's largest cloud platform. 00:24:07.529 --> 00:24:10.190 It's his career coming full circle. Exactly. 00:24:10.309 --> 00:24:12.589 He was designing for the ultimate realization 00:24:12.589 --> 00:24:15.529 of the networked architecture neutral world that 00:24:15.529 --> 00:24:18.089 he helped create in the first place. And even 00:24:18.089 --> 00:24:20.609 outside those big jobs, he stayed influential. 00:24:21.009 --> 00:24:24.130 He took on advisory and board roles. He advised 00:24:24.130 --> 00:24:26.829 Lightbend, the Scala company, which shows his 00:24:26.829 --> 00:24:29.569 interest in new languages for the JVM. He was 00:24:29.569 --> 00:24:32.670 a director at Jelastic, a hosting platform, a 00:24:32.670 --> 00:24:35.269 strategic advisor for Eucalyptus, an early private 00:24:35.269 --> 00:24:37.880 cloud company. He never stopped engaging with 00:24:37.880 --> 00:24:40.660 these deep foundational infrastructure problems 00:24:40.660 --> 00:24:44.220 and not just in software. His board seat at Dirty 00:24:44.220 --> 00:24:46.119 Environmental Solutions shows he was thinking 00:24:46.119 --> 00:24:49.039 about how to apply systems thinking to all sorts 00:24:49.039 --> 00:24:52.380 of complex real world problems. So given the 00:24:52.380 --> 00:24:54.720 scale of his impact, it's no surprise that James 00:24:54.720 --> 00:24:56.640 Gosling has received the highest honors you can 00:24:56.640 --> 00:24:59.299 get, both from a civic standpoint and a technical 00:24:59.299 --> 00:25:01.759 one. These awards really emphasize that his contribution 00:25:01.759 --> 00:25:04.160 wasn't just about programming. It was about building 00:25:04.160 --> 00:25:06.220 national and global infrastructure. structure. 00:25:06.259 --> 00:25:07.900 Let's start with a recognition from his home 00:25:07.900 --> 00:25:10.920 country, Canada. In 2007, he was named an Officer 00:25:10.920 --> 00:25:13.180 of the Order of Canada, which is Canada's second 00:25:13.180 --> 00:25:15.759 highest civilian honor. It is. And to be named 00:25:15.759 --> 00:25:17.980 an officer, which is the second highest grade 00:25:17.980 --> 00:25:20.380 in the order, is reserved for people who've shown 00:25:20.380 --> 00:25:22.980 a lifetime of outstanding achievement and service 00:25:22.980 --> 00:25:25.579 to the nation. This isn't just a tech award. 00:25:25.819 --> 00:25:28.140 It's the country acknowledging that Java's global 00:25:28.140 --> 00:25:31.380 impact is a major Canadian contribution to computer 00:25:31.380 --> 00:25:33.859 science. And his peers in the engineering and 00:25:33.859 --> 00:25:36.220 scientific communities have been just as clear. 00:25:36.359 --> 00:25:39.940 In 2015, he got the I .E. John von Neumann Medal. 00:25:40.359 --> 00:25:42.660 The Wein -Neumann Medal is for outstanding achievements 00:25:42.660 --> 00:25:45.599 in computer -related science and tech. It basically 00:25:45.599 --> 00:25:48.119 puts him in the same category as the giants who 00:25:48.119 --> 00:25:51.099 shaped the entire field. It recognizes seminal 00:25:51.099 --> 00:25:53.460 work that transforms the landscape. And even 00:25:53.460 --> 00:25:55.940 before that, in 2004, he was elected a foreign 00:25:55.940 --> 00:25:57.700 member of the National Academy of Engineering, 00:25:57.960 --> 00:26:01.279 the NAE. The election citation itself is basically 00:26:01.279 --> 00:26:03.160 a perfect summary of his career's importance. 00:26:03.559 --> 00:26:05.740 It's a great citation. It recognizes him for 00:26:05.740 --> 00:26:08.180 two things. The conception and development of 00:26:08.180 --> 00:26:10.009 the architecture for the Java program. programming 00:26:10.009 --> 00:26:12.230 language. And for his contributions to windowing 00:26:12.230 --> 00:26:14.710 systems. Right. It's a formal acknowledgement 00:26:14.710 --> 00:26:17.450 that the technical brilliance of new W .S., even 00:26:17.450 --> 00:26:19.990 though it failed in the market, was still foundational 00:26:19.990 --> 00:26:22.529 work recognized at the absolute highest level. 00:26:22.670 --> 00:26:24.950 And more recently, he was named a Computer History 00:26:24.950 --> 00:26:28.690 Museum Fellow in 2019. That fellowship was awarded 00:26:28.690 --> 00:26:31.650 specifically for the conception, design and implementation 00:26:31.650 --> 00:26:34.750 of the Java programming language. It leaves no 00:26:34.750 --> 00:26:37.210 doubt about his historical role. He was also 00:26:37.210 --> 00:26:40.329 made an ACM Fellow in 2013, got the Economist 00:26:40.329 --> 00:26:43.250 Innovation Award and the Flame Award UCNx Lifetime 00:26:43.250 --> 00:26:46.390 Achievement Award both way back in 2002. The 00:26:46.390 --> 00:26:48.309 recognition across all these different institutions 00:26:48.309 --> 00:26:50.710 just proves Java wasn't just a market success. 00:26:51.130 --> 00:26:54.109 It was a true architectural triumph. But his 00:26:54.109 --> 00:26:56.690 most tangible legacy, the thing he left for millions 00:26:56.690 --> 00:26:59.329 of developers, is his literary one, the foundational 00:26:59.329 --> 00:27:01.650 books that actually defined the language. This 00:27:01.650 --> 00:27:04.390 is so important. For a technology to scale globally, 00:27:04.569 --> 00:27:07.250 it has to be documented perfectly. And Gosling 00:27:07.250 --> 00:27:09.690 was the principal author on the books that were 00:27:09.690 --> 00:27:11.829 basically the constitution for the entire Java 00:27:11.829 --> 00:27:14.410 ecosystem. Let's touch on the two most important 00:27:14.410 --> 00:27:17.049 ones because they serve two very different purposes. 00:27:17.410 --> 00:27:19.730 First, there's the Java programming language, 00:27:19.910 --> 00:27:21.990 which he co -authored with Ken Arnold and David 00:27:21.990 --> 00:27:24.329 Holmes. This is the user manual. It's the instruction 00:27:24.329 --> 00:27:27.029 manual. It's the book that taught a generation 00:27:27.029 --> 00:27:30.210 of programmers how to use the language. The syntax, 00:27:30.390 --> 00:27:33.410 the libraries, the best practices, it ensured 00:27:33.410 --> 00:27:35.609 everyone learned Jada in a standardized way. 00:27:36.009 --> 00:27:38.130 But even more critical for that write once, run 00:27:38.130 --> 00:27:41.029 anywhere promise is the other one, the Java language 00:27:41.029 --> 00:27:44.130 specification. He co -authored this with other 00:27:44.130 --> 00:27:46.710 giants like Bill Joy and Guy Steele. And the 00:27:46.710 --> 00:27:49.230 specification is not an instruction manual. It's 00:27:49.230 --> 00:27:52.069 the rulebook. It's the legal technical constitution 00:27:52.069 --> 00:27:56.140 of Java. It defines with absolute precision how 00:27:56.140 --> 00:27:58.480 the compiler has to behave, how the bytecode 00:27:58.480 --> 00:28:01.220 must execute, what every single command means. 00:28:01.579 --> 00:28:03.839 It's the document that technically and legally 00:28:03.839 --> 00:28:06.359 guarantees that a program compiled on one machine 00:28:06.359 --> 00:28:08.900 will behave identically on any other machine 00:28:08.900 --> 00:28:10.960 that follows the standard. Without that specification, 00:28:11.400 --> 00:28:13.980 the write -once -run -anywhere promise just dissolves 00:28:13.980 --> 00:28:16.480 into total chaos. It's the anchor for the entire 00:28:16.480 --> 00:28:19.019 ecosystem. It's the difference between a textbook 00:28:19.019 --> 00:28:22.019 and a legal code. And it shows how complete his 00:28:22.019 --> 00:28:25.430 legacy is. He didn't just invent the tech. He 00:28:25.430 --> 00:28:27.829 provided the documentary infrastructure for it 00:28:27.829 --> 00:28:29.490 to become a global standard. And you can even 00:28:29.490 --> 00:28:31.910 see his interest in specialized systems in his 00:28:31.910 --> 00:28:34.089 co -authorship of the real -time specification 00:28:34.089 --> 00:28:37.250 for Java. That's a fascinating corner of the 00:28:37.250 --> 00:28:40.289 Java world. Java, because of its garbage collection, 00:28:40.529 --> 00:28:42.869 was traditionally not great for real -time systems 00:28:42.869 --> 00:28:45.789 like factory robots or flight controls where 00:28:45.789 --> 00:28:48.710 timing has to be perfect. And the RTSJ was an 00:28:48.710 --> 00:28:50.990 attempt to push Java into those high -stakes 00:28:50.990 --> 00:28:53.990 areas by defining ways to manage memory and concurrency 00:28:53.990 --> 00:28:57.190 with guaranteed predictable timing. It just shows 00:28:57.190 --> 00:28:59.230 that even after defining the standard for general 00:28:59.230 --> 00:29:01.430 computing, he was still trying to solve the next 00:29:01.430 --> 00:29:03.490 hardest problem. It's just a completely comprehensive 00:29:03.490 --> 00:29:07.269 body of work. Hashtag tag outro. So when we step 00:29:07.269 --> 00:29:09.430 back, what does this all mean? We've traced this 00:29:09.430 --> 00:29:11.450 incredible journey from a kid in high school 00:29:11.450 --> 00:29:14.650 analyzing ISIS -2 satellite data, which demanded 00:29:14.650 --> 00:29:17.670 robustness from day one. All the way to his architectural 00:29:17.670 --> 00:29:20.609 genius in creating the Java virtual machine and 00:29:20.609 --> 00:29:22.990 standardizing this idea of architecture -neutral 00:29:22.990 --> 00:29:26.049 execution. He took these niche ideas from academic 00:29:26.049 --> 00:29:30.130 research, like p -code for UCSD Pascal, and he 00:29:30.130 --> 00:29:32.589 successfully scaled them to define how programs 00:29:32.589 --> 00:29:35.009 run across the entire planet. Yeah, the software 00:29:35.009 --> 00:29:37.150 we all rely on today, whether it's running a 00:29:37.150 --> 00:29:40.349 massive data center at AWS or a complex financial 00:29:40.349 --> 00:29:43.029 trading system or even the early foundations 00:29:43.029 --> 00:29:45.990 of the mobile web. It all rests directly on the 00:29:45.990 --> 00:29:48.410 principles of virtualization and rule based design 00:29:48.410 --> 00:29:50.589 that he established back in the 80s and 90s. 00:29:50.630 --> 00:29:52.309 He created one of the most successful commercial 00:29:52.309 --> 00:29:55.269 technologies ever. He stayed with Sun for 26 00:29:55.269 --> 00:29:57.990 years, literally defining the company and the 00:29:57.990 --> 00:30:00.329 language. And yet maybe the most telling part 00:30:00.329 --> 00:30:02.769 of his whole story isn't what he created, but 00:30:02.769 --> 00:30:05.640 why he left. The source material shows this consistent 00:30:05.640 --> 00:30:08.619 pattern, his walkout from Oracle, citing ethics 00:30:08.619 --> 00:30:11.079 and a loss of control. And then his choices after 00:30:11.079 --> 00:30:13.940 that. To work on really hard, high -impact problems 00:30:13.940 --> 00:30:17.059 at places like Liquid Robotics or AWS. I mean, 00:30:17.079 --> 00:30:19.279 imagine spending more than two decades building 00:30:19.279 --> 00:30:22.000 a company's flagship product only to walk away 00:30:22.000 --> 00:30:24.440 because of a moral conflict. It just speaks volumes 00:30:24.440 --> 00:30:26.680 about his priorities. It was always principle 00:30:26.680 --> 00:30:28.740 and intellectual freedom over just sticking with 00:30:28.740 --> 00:30:31.160 a corporate legacy. It seems like for an architect 00:30:31.160 --> 00:30:34.960 whose real passion was solving these deep, complex 00:30:34.960 --> 00:30:39.380 problems of how code runs everywhere. The aggressive 00:30:39.380 --> 00:30:41.539 corporate machine that grew up around his invention 00:30:41.539 --> 00:30:44.380 eventually became an ethical barrier he just 00:30:44.380 --> 00:30:47.619 couldn't accept. The sources show he always prioritized 00:30:47.619 --> 00:30:50.299 engineering integrity. When he felt his creation 00:30:50.299 --> 00:30:52.759 was being misused, like with Google and Android, 00:30:53.019 --> 00:30:55.440 he defended the original work. But when his new 00:30:55.440 --> 00:30:58.259 employer, Oracle, tried to use copyright law 00:30:58.259 --> 00:31:01.000 to strangle future innovation. He spoke out against 00:31:01.000 --> 00:31:03.099 them. He prioritized the health of the whole 00:31:03.099 --> 00:31:05.420 ecosystem. He did. And that leaves us with a 00:31:05.420 --> 00:31:07.400 final provocative thought for you to think about. 00:31:07.799 --> 00:31:10.519 Despite creating a technology that became a pillar 00:31:10.519 --> 00:31:13.359 of global commerce, James Gosling ultimately 00:31:13.359 --> 00:31:15.420 chose to walk away from his creation's corporate 00:31:15.420 --> 00:31:18.240 home to stay true to his principles of engineering 00:31:18.240 --> 00:31:21.640 freedom. So does the greatest innovation sometimes 00:31:21.640 --> 00:31:24.200 require the innovator to prioritize those philosophical 00:31:24.200 --> 00:31:26.980 principles over decades of corporate dedication, 00:31:27.279 --> 00:31:30.059 even if it means cutting ties with the very thing 00:31:30.059 --> 00:31:32.980 they built? It's a profound question about where 00:31:32.980 --> 00:31:35.680 the true value really lies. in the structure 00:31:35.680 --> 00:31:37.779 of the code or in the ethical environment where 00:31:37.779 --> 00:31:39.220 that code is allowed to grow and prime.