WEBVTT 00:00.000 --> 00:11.280 Ladies and gentlemen, I'm very pleased to welcome James Bonnet to stage. 00:11.280 --> 00:15.520 He's going to give you a different perspective looking at strategies for trusting your 00:15.520 --> 00:19.960 employer in open source and I'll let James tell you more. 00:19.960 --> 00:27.880 So, thank you very much for the introduction. 00:27.880 --> 00:33.520 As you said, my name is James Bonnet, I've been working in open source now for it depends 00:33.520 --> 00:39.520 how you count at something like 30 years and now have the gray hairs to prove it. 00:39.520 --> 00:46.840 I began my career as a physicist and a developer, but I got on to business models. 00:46.920 --> 00:52.040 I led the effort to create the technical advisory board of the Linux Foundation and I am 00:52.040 --> 00:53.040 a kernel developer. 00:53.040 --> 00:58.960 I still, I'm listed as a maintainer of the Scuzzy subsystem and the peer-risk subsystem. 00:58.960 --> 01:01.840 But the reason I, oh, yeah, I need to give you a disclaimer. 01:01.840 --> 01:06.880 Obviously, I do work for Microsoft as talk is based on personal observation and you can 01:06.880 --> 01:12.360 attribute anything I say to me, blame me for it, but you can't blame my employer. 01:12.360 --> 01:16.800 So let's begin with a little history. 01:16.800 --> 01:19.880 The Linux began life as a hobby project in 1991. 01:19.880 --> 01:24.640 This is basically the logical canon, everybody knows this. 01:24.640 --> 01:29.880 By 1992, and 1991 it was about 100 lines of codeliness posted to the internet. 01:29.880 --> 01:34.400 By 1992, it was actually a self-hosting operating system. 01:34.400 --> 01:38.680 The first distribution started to be widely installed that basically meant that you could 01:38.680 --> 01:44.680 buy a stack of, in those days it was 10 floppy disks and stored it on your PC in 1993. 01:44.680 --> 01:49.760 And it was as effective as the Unix operating system by that. 01:49.760 --> 01:54.000 What wasn't widely reported at the time was that I was in the math department at university 01:54.000 --> 01:57.440 and I was subversively using Linux to replace Unix. 01:57.440 --> 02:02.200 And the reason I was doing this is money because basically the Unix workstations in those 02:02.200 --> 02:04.200 days cost 10,000 quid. 02:04.200 --> 02:07.880 I could go out and buy 100 quid PC and install Linux on it. 02:07.880 --> 02:14.640 I could get 10 of those for every 1000 quid we'd have spent on a Unix machine. 02:14.960 --> 02:17.400 Unix continued its forward progress march. 02:17.400 --> 02:21.600 By 1997, if you ever heard of it, there was this thing called lamp. 02:21.600 --> 02:25.520 It stood for Unix Apache, my SQL and PHP. 02:25.520 --> 02:28.880 And this was actually what began to penetrate the enterprise. 02:28.880 --> 02:33.800 Because the nascent internet had just come along in 1995, two years ago. 02:33.800 --> 02:38.280 The corporations were all starting to try and get on the internet. 02:38.280 --> 02:41.320 And the way you did that was by a web server that could do stuff. 02:41.400 --> 02:44.720 And this lamp stack was basically a web server that could do stuff. 02:44.720 --> 02:49.120 So everybody in the Unix space and of course Microsoft in the Windows space, 02:49.120 --> 02:53.040 we're rushing to produce web servers, but Linux already had one. 02:53.040 --> 02:58.280 And so the easiest way of actually getting your company onto the internet in those days was to install Linux 02:58.280 --> 03:01.080 and just place the machine there. 03:01.080 --> 03:05.760 Lamp was actually subversively installed by Cisadmin's. 03:05.760 --> 03:10.120 This means there was no corporate decision to actually do this installation. 03:10.200 --> 03:14.080 It was all done by the people in the lowest structure of the corporation. 03:14.080 --> 03:17.560 If you think how a corporation works, that's really not what they do today. 03:17.560 --> 03:22.120 You don't have the boss come in and ask the Cisadmin's what they're doing in your data center 03:22.120 --> 03:26.480 because your customers would be very unhappy about the security implications of that today. 03:26.480 --> 03:31.960 But you know, way back 20 years, this was the Wild West everybody did this. 03:31.960 --> 03:36.120 The executives of the company and sort of the CTOC level C Suite, 03:36.120 --> 03:40.440 even the guy in charge of the data center, were often unaware that this was happening. 03:40.440 --> 03:49.480 So Linux was just quietly seeping into all of these industrial data centers with nobody who was actually paying the money knew what was going on. 03:49.480 --> 03:55.240 But would have mass started spreading and sort of by the end of 1998, 1999, 03:55.240 --> 04:00.240 even the C Suite executives were aware of the fact that they actually had Linux in their organization. 04:00.240 --> 04:06.080 They just didn't know how much, but they were also aware of the reasons why it was actually becoming popular. 04:06.560 --> 04:08.800 And it was the cost benefit and the reliability. 04:08.800 --> 04:13.440 It was basically, you bought any old Claptop PC, the 100-quid argument again. 04:13.440 --> 04:17.920 You have still Linux on it and it just runs as a robust internet web server. 04:17.920 --> 04:22.880 You don't need to buy this expensive Unix workstation for a thousand quid to do the same job. 04:22.880 --> 04:28.560 So this 10-to-one cost saving was what really appealed to the C Suite. 04:28.560 --> 04:36.960 And obviously, in 1998, the Unix was just ended, leaving Carnage and Devastation in their wake. 04:36.960 --> 04:40.160 There was pretty much no inheritor to Unix then. 04:40.160 --> 04:44.240 Some thought it would might be able to pick up on the destruction, 04:44.240 --> 04:47.840 but they were basically all in pieces on the floor, 04:47.840 --> 04:49.440 because they'd competed with each other. 04:49.440 --> 04:53.200 They fragmented the operating system and they'd screwed up their business models. 04:53.200 --> 04:57.680 So Microsoft thought that the data center was ripe for the picking. 04:57.680 --> 05:04.640 It already had Windows MT to do this, and it was coming in with IS, which was internet information server. 05:04.640 --> 05:08.080 And as Microsoft tried to penetrate the data center with this, 05:08.080 --> 05:12.720 it found its path blocked pretty much every turn by this free-up start called Linux. 05:12.720 --> 05:16.640 Because even though Microsoft Windows ran on PCs, 05:16.640 --> 05:18.720 you still had to pay for the operating system. 05:18.720 --> 05:21.840 So it was cheaper than the $1,000 Unix machine. 05:21.840 --> 05:25.040 You could still buy the $100 PC, 05:25.040 --> 05:29.920 but you then had to pay Microsoft a couple of $100 to install IS and everything on it. 05:29.920 --> 05:32.960 Actually, Windows Server was pretty expensive in those days. 05:32.960 --> 05:37.600 And so this is when Microsoft started to see Linux as a real threat to its business model. 05:37.600 --> 05:41.280 This is also the year what was called the Halloween documents were published. 05:41.280 --> 05:43.360 I don't really have time to tell you what they are, 05:43.360 --> 05:45.680 but you can go and look them up on the internet there. 05:45.680 --> 05:48.400 Pretty infamous. 05:48.400 --> 05:50.320 But as this was happening, 05:50.320 --> 05:56.240 especially what makes executives pay attention is not really success its failure. 05:56.240 --> 05:59.600 The fact that Microsoft was failing to penetrate the enterprise 05:59.600 --> 06:05.600 with what looked like such a winning proposition made a load of executives in the American software industry 06:05.600 --> 06:07.600 asked what the hell is going on. 06:07.600 --> 06:11.520 You know, and when they saw that it was being blocked by something, a lot of them, 06:11.520 --> 06:13.280 it's still never heard of Linux. 06:13.280 --> 06:16.080 They wanted to know why, and more importantly, 06:16.160 --> 06:18.480 they wanted to know if they could make money from it. 06:18.480 --> 06:22.480 So this was the beginning of what was called the dot com bubble. 06:22.480 --> 06:24.880 Because by the time 2000 came along, 06:24.880 --> 06:29.920 everybody was trying to invest in Linux because obviously it was this little upstart operating system 06:29.920 --> 06:30.960 that was beating Windows. 06:32.080 --> 06:37.120 Problem was that Linux, the kernel, and most of the tools that we used in the operating system 06:37.120 --> 06:41.040 were all three software at that time, fiercely run by volunteers 06:41.040 --> 06:43.520 who really didn't want to give up their independence. 06:43.520 --> 06:46.240 So Linux had no corporate structure at all. 06:46.240 --> 06:49.440 It was basically just his kernel, his a bunch of tools. 06:49.440 --> 06:52.160 Oh look, there's a few guys over there putting a distribution together, 06:52.160 --> 06:54.480 and there's a few guys over there doing the same thing. 06:54.480 --> 06:57.120 There was basically no corporate organization at all. 06:58.320 --> 07:01.360 So OSDL was formed to try and fix this. 07:01.360 --> 07:04.320 It was the first of the foundations, effectively, 07:04.320 --> 07:07.840 who actually knows what OSDL stands for, raise your hand. 07:07.840 --> 07:09.600 One person, and it's okay. 07:09.600 --> 07:12.160 So it's called the open source development laboratory. 07:12.160 --> 07:15.440 It died about 20 years ago, which is why only one person and the audience 07:15.440 --> 07:19.520 has ever heard of it, but it was the precursor of the Linux Foundation. 07:19.520 --> 07:22.240 And because it was formed in the dot com bubble, 07:22.240 --> 07:26.000 companies were falling over themselves to pay money into this foundation. 07:26.000 --> 07:28.960 It's initial seed capital with some around $10 million. 07:29.520 --> 07:31.680 Well, it was at least $0.5 million. 07:31.680 --> 07:34.080 The exact figures are sort of lost in the accounts. 07:35.840 --> 07:40.400 And its purpose, stated purpose of OSDL was to make Linux business ready. 07:40.480 --> 07:43.760 And the reason they formed it is because the company's realized, 07:43.760 --> 07:46.960 they couldn't just show money on a load of scruffy volunteers who 07:46.960 --> 07:49.440 are sitting around at the kernel summit, figuring out 07:49.440 --> 07:52.640 intricacies of memory management subsystems and regulars, 07:52.640 --> 07:55.520 they actually needed some suits somewhere to talk to. 07:55.520 --> 07:56.720 And we didn't have any. 07:56.720 --> 08:00.160 So they very kindly formed the suits for us in OSDL, 08:00.160 --> 08:03.360 and told us this was the organization that was speaking for kernel developers. 08:04.160 --> 08:07.040 They actually hired Linux in 2003, 08:07.040 --> 08:10.160 and actually a bunch of other kernel developers as well. 08:10.240 --> 08:13.920 OSDL actually had quite a large stable of kernel developers at that time. 08:16.320 --> 08:22.080 But the relationship between the Linux kernel and OSDL was not always a happy one. 08:22.080 --> 08:27.200 The reason was basically that to a bunch of volunteers who were working on our 08:27.200 --> 08:30.080 operating system, OSDL looked like a parachute 08:30.080 --> 08:33.360 to do in management infrastructure that was trying to tell us 08:33.360 --> 08:36.080 what we should actually be doing with our operating system. 08:36.080 --> 08:37.040 And we didn't like that. 08:37.920 --> 08:43.360 OSDL saw themselves, basically, as the grownups in the room, 08:43.360 --> 08:46.640 the guys and suits who knew how to talk to industry, 08:46.640 --> 08:48.400 they'd gathered requirements from industry. 08:48.400 --> 08:50.880 They had a load of projects that they wanted to do, 08:50.880 --> 08:54.240 but they'd hired nowhere near enough kernel developers actually to do them. 08:54.240 --> 08:56.960 So they ended up spending a lot of their time sending 08:56.960 --> 09:00.720 their directs of engineering and people to Linux conferences, 09:00.720 --> 09:04.000 primarily the kernel summit, to tell us how we should be running Linux. 09:04.000 --> 09:06.160 And this did not go down very well. 09:06.160 --> 09:09.920 The friction started to build pretty much as soon as OSDL was formed. 09:09.920 --> 09:14.560 By the time it had hired Linux, it basically thought it had found the winning argument. 09:14.560 --> 09:16.160 And this was in 2003. 09:16.160 --> 09:20.080 A year later, we were really getting ticked off by all the friction. 09:20.080 --> 09:25.280 We did spend a year in 2004 trying to explain to OSDL that 09:25.280 --> 09:29.120 we, the kernel developers, needed them to interact with us in a different way. 09:29.120 --> 09:33.280 I mean, we were interested in what the requirements of industry were, 09:33.280 --> 09:36.560 but we wanted them translated into technical problems that we 09:36.560 --> 09:38.800 would decide whether we're interested in trying to solve. 09:38.800 --> 09:42.880 We didn't want the usual architectural roadmap for an operating system, 09:42.880 --> 09:46.080 dumped on us by a suit, he didn't know what they were talking about. 09:46.080 --> 09:48.480 So this friction was really serious. 09:48.480 --> 09:52.640 And by the end of 2004, talking to OSDL just wasn't working, 09:52.640 --> 09:55.440 because the attitude of their management was, yeah, 09:55.440 --> 09:58.880 but you guys, I mean, remember, this is 20 years ago, 09:58.880 --> 10:00.800 even I was a lot younger than, 10:00.800 --> 10:03.280 but I was one of the oldest Linux kernel developers. 10:03.280 --> 10:06.240 We were going from people who were 16, 17. 10:06.240 --> 10:09.920 I mean, Christoph Helwig was about 17 at the time with red hair, 10:09.920 --> 10:13.200 and these suits used to cow when they saw him coming. 10:13.200 --> 10:17.680 So they thought that we just weren't capable of actually running an operating system 10:17.680 --> 10:20.000 successfully, which is of course what we were doing. 10:20.000 --> 10:22.480 And we knew it. 10:22.480 --> 10:26.800 So we had to have a phase where we learned to get along together. 10:26.800 --> 10:30.160 We needed to cooperate. 10:30.160 --> 10:35.200 And the way we did it is, in 2005, the kernel summit recognized, 10:35.200 --> 10:37.120 this was a problem that wasn't going to go away. 10:37.120 --> 10:40.240 I mean, we'd had a year of massive friction with OSDL. 10:40.240 --> 10:42.560 We really need to do something about it. 10:42.560 --> 10:46.240 The real problem was, basically, that we thought, 10:46.240 --> 10:48.800 if we could just talk to the companies directly, 10:48.800 --> 10:50.560 we'd be able to get along a lot better. 10:50.560 --> 10:52.800 They would actually realize that, you know, 10:52.800 --> 10:55.200 we were actually doing service, and we could negotiate 10:55.200 --> 10:56.560 of more equal terms. 10:56.560 --> 11:00.080 Effectively, we just wanted them to recognize us as co-equals 11:00.080 --> 11:01.520 in the power structure. 11:01.520 --> 11:04.080 But OSDL was getting in the way, and it wouldn't let us 11:04.080 --> 11:07.280 talk to the companies directly, because we were scruffy engineers 11:07.280 --> 11:11.600 and corporate suits didn't really want to know who we are, who we were. 11:11.600 --> 11:15.360 So what we did was, we proposed the formation 11:15.360 --> 11:20.880 of the technical advisory board for OSDL as it then was, 11:20.880 --> 11:23.520 and we demanded the board seat on OSDL. 11:23.520 --> 11:27.760 Now, if you know anything about what goes on in corporate structures, 11:27.760 --> 11:31.520 this is basically a power move. 11:31.520 --> 11:34.240 It's an exercise of extreme power. 11:34.240 --> 11:37.920 You usually, even in a corporation, and I hold the majority shareholder, 11:37.920 --> 11:39.600 you wouldn't expect this maneuver to get off. 11:39.600 --> 11:42.400 Because remember, we felt, talked to all of the executives 11:42.400 --> 11:44.000 in the CEO of OSDL. 11:44.000 --> 11:46.880 So we're going to go over that head and make a proposal 11:46.880 --> 11:49.520 to fix this to the board of directors directly. 11:49.520 --> 11:51.680 And if you know any board on any company, 11:51.680 --> 11:54.880 they're really reluctant to actually get their hands dirty 11:54.880 --> 11:57.520 and go and tell, they have the power to tell a CEO 11:57.520 --> 12:00.240 in the company what to do, but they rarely exercise it. 12:00.240 --> 12:02.720 And we were demanding they exercise it. 12:02.720 --> 12:06.320 Right? So on paper, this move shouldn't have worked, 12:06.320 --> 12:09.440 because it's something like what you see in succession. 12:09.440 --> 12:12.640 Board conversations are almost never structured like this. 12:12.640 --> 12:14.800 Board's like to do everything by consensus, 12:14.800 --> 12:17.120 and they usually don't like to cross the CEO. 12:17.120 --> 12:19.280 And we gave them a proposal that said, pretty much, 12:19.280 --> 12:23.680 across the CEO, talked to us, and we'll help you sort it all out. 12:23.680 --> 12:26.640 But they did like a hoodspot. 12:26.640 --> 12:29.520 They figured out that if we could go up to them 12:29.520 --> 12:32.240 and make these phenomenally outrageous demands. 12:32.240 --> 12:35.120 And then actually, we backed it up with a massive document 12:35.120 --> 12:37.760 explaining why if they did this, it would work. 12:37.760 --> 12:40.480 And they looked at, they all thought that, yeah, actually, 12:40.480 --> 12:43.120 we might as well, I mean, we're failing at this model anyway. 12:43.120 --> 12:45.280 We might as well give these guys a chance. 12:45.280 --> 12:49.520 So it took one round of canvassing and a board vote, 12:49.520 --> 12:51.920 and we had everything we demanded. 12:51.920 --> 12:56.160 And the relationship between the Linux, the kernel, 12:56.160 --> 12:59.440 and the corporations who are interested in using it, 12:59.440 --> 13:01.920 had a sea change of their particular point. 13:01.920 --> 13:06.480 Out of the sea change grew pretty much all of the corporate interactions 13:06.480 --> 13:08.400 you see today in open source. 13:08.400 --> 13:13.760 But this was the point in history where the change began. 13:13.760 --> 13:16.960 And obviously, the document we put together 13:16.960 --> 13:19.200 did not say we have the solutions, you know, 13:19.200 --> 13:20.720 it's sort of a bad document in that sense. 13:20.720 --> 13:23.040 Usually you pretend you know what you're talking about. 13:23.040 --> 13:25.200 What it said is we can set up a structure 13:25.200 --> 13:28.400 where we can both work together mutually to solve the problems. 13:28.400 --> 13:31.040 Because the companies didn't know how to talk to us. 13:31.040 --> 13:32.960 We didn't really know how to talk to them. 13:32.960 --> 13:35.760 But if we establish the direct channel of communication, 13:35.760 --> 13:37.520 our contention was we'd work it out. 13:37.520 --> 13:41.600 And of course, that actually did end up working out pretty well. 13:41.600 --> 13:46.560 And in 2007, OSTL, this is why it doesn't exist anymore, 13:46.560 --> 13:49.840 merge with the free standards group to become the Linux Foundation. 13:49.840 --> 13:53.760 So I imagine everybody now in the room is heard of the Linux Foundation. 13:53.760 --> 13:55.360 This is how it began. 13:55.360 --> 14:00.560 But in recognition of the fact that the power move we did in 2005 14:00.560 --> 14:02.800 by 2007, it was really working. 14:02.800 --> 14:06.160 We were actually establishing some synergy with the company. 14:06.160 --> 14:08.800 And everything was actually going pretty smoothly. 14:08.800 --> 14:12.000 The bylaws of the Linux Foundation, if you read them, 14:12.000 --> 14:14.960 now actually have the technical advisory board 14:14.960 --> 14:17.760 and the board seat dedicated to developers 14:17.760 --> 14:20.640 enshrined sitting in their actual bylaws 14:20.640 --> 14:25.280 as in recognition of the fact that this structure really, really does work. 14:25.280 --> 14:29.120 I have to say that pretty much no other foundation 14:29.120 --> 14:31.680 that's ever been formed has a structure like this. 14:31.680 --> 14:35.120 Because remember, usually, and you saw it in the previous talk, 14:35.120 --> 14:37.760 foundations are formed to manage projects. 14:37.840 --> 14:40.560 The Linux Foundation does not manage Linux. 14:40.560 --> 14:42.800 Linux through the technical advisory board 14:42.800 --> 14:46.320 and its board seat really technically manages the Linux Foundation. 14:46.320 --> 14:48.960 The relationship is entirely inverted. 14:48.960 --> 14:51.280 And this is the way the kernel developers like it. 14:51.280 --> 14:53.520 I realize that now the project runs this way, 14:53.520 --> 14:56.960 but it works very well for us. 14:56.960 --> 15:00.480 Obviously, by 2008, since I'd put the founding, 15:00.480 --> 15:02.160 I basically had the kernel summit. 15:02.160 --> 15:04.080 I was the only person wearing a bow tie. 15:04.080 --> 15:07.520 So I was the one they pushed forwards to try and do all of this. 15:07.520 --> 15:11.520 So I was in the front line as the chair of the technical advisory board 15:11.520 --> 15:14.160 and actually the board member at OSDL 15:14.160 --> 15:18.240 and then the Linux Foundation trying to ensure they did what we told them. 15:18.240 --> 15:20.640 And obviously, then I was also in the front line 15:20.640 --> 15:22.240 of the fight against Microsoft. 15:22.240 --> 15:26.400 Because even ten years later now, the 1998, 15:26.400 --> 15:29.120 where we blocked them from entering the enterprise 15:29.120 --> 15:31.920 and internet information explorer, they still hadn't forgotten. 15:31.920 --> 15:34.000 They still thought the windows was the natural 15:34.000 --> 15:36.320 inheritor of the operating systems. 15:36.320 --> 15:40.400 And so at that time, I can remember getting on stage 15:40.400 --> 15:44.720 with a lot of Microsoft people who came to try and explain their business model. 15:44.720 --> 15:49.040 And in 2008, it was actually monetizing patents against Linux. 15:49.040 --> 15:51.600 And I can remember the open source business conference 15:51.600 --> 15:53.040 that doesn't exist anymore. 15:53.040 --> 15:55.280 Massive conference out in Silicon Valley. 15:55.280 --> 15:59.280 I was on stage with Brad Smith, who's now executive chairman of Microsoft. 15:59.280 --> 16:02.800 But then he was chief corporate counsel going toe to toe 16:02.800 --> 16:06.800 over why the way Microsoft tried to use patents 16:06.800 --> 16:08.560 just wouldn't work for open source. 16:08.560 --> 16:12.400 And it was one of those conversations where the cleaners 16:12.400 --> 16:15.600 mocked the blood off the floor after you finished. 16:15.600 --> 16:18.720 But of course, now I work for Microsoft. 16:18.720 --> 16:22.640 And this is the story of how that came to be in some senses. 16:22.640 --> 16:26.480 The lesson you can get from this story is that there's always scope 16:26.480 --> 16:29.440 for today's enemy to become tomorrow's friend. 16:29.440 --> 16:34.000 And that's actually how you build alliances with industry. 16:34.000 --> 16:39.120 But there's a lot that came out of that initial fracker with OSDL. 16:39.120 --> 16:44.000 One of which was the changing role of engineers in industry entirely. 16:44.000 --> 16:47.200 Not just an open source, but we actually change the statues 16:47.200 --> 16:50.000 at exactly your own engineers, even in proprietary software. 16:50.000 --> 16:51.680 You've heard of this thing called in a source. 16:51.680 --> 16:56.320 It's actually trying to bring open source techniques to proprietary projects. 16:56.320 --> 17:00.560 By 2005, the recent OSDL thought it had to exist, 17:00.560 --> 17:04.640 is that the way corporations talk to each other 17:04.640 --> 17:07.920 is by these things called business development departments. 17:07.920 --> 17:09.680 And obviously, they do a bit of negotiation 17:09.680 --> 17:11.760 at the sales level and the CO level. 17:11.760 --> 17:15.760 But the Linux organization was an unstructured group of volunteers 17:15.760 --> 17:17.600 who just communicated by email. 17:17.600 --> 17:19.520 We had no executive they could talk to. 17:19.520 --> 17:22.240 We had no business development organization. 17:22.240 --> 17:25.280 And without this business development organization, 17:25.280 --> 17:27.280 they just did not know how to talk to us. 17:27.280 --> 17:30.000 We couldn't negotiate corporate to corporate agreements. 17:30.000 --> 17:34.960 We couldn't even send an executive out to sign an agreement on behalf of Linux, 17:34.960 --> 17:37.440 because nobody knew you could speak for Linux. 17:37.440 --> 17:41.520 It was basically what any volunteer had to think at any particular time. 17:41.520 --> 17:46.400 So this was the problem of they didn't really know how they should be communicating with us. 17:46.400 --> 17:50.400 And we didn't really know how we could make promises to them. 17:50.400 --> 17:54.240 But if you think about business development today, 17:54.240 --> 17:56.640 business development has a sort of a dying art. 17:56.640 --> 18:00.240 Engineers in a lot of firms actually fill this role. 18:00.240 --> 18:03.200 You see a lot of people shaking their heads. 18:03.200 --> 18:06.000 But if you think about what you do in open source, 18:06.000 --> 18:09.440 especially open source that shared between a lot of corporations. 18:09.440 --> 18:13.120 Corporations, employee engineers, engineers work on the project. 18:13.120 --> 18:16.960 The engineers pretty much know what their company wants out of this project. 18:16.960 --> 18:22.160 But then negotiate a sort of patch and code level with all of the other engineers. 18:22.160 --> 18:24.480 And you sort of get negotiations to go, 18:24.480 --> 18:26.000 well, my company's interested in this. 18:26.000 --> 18:27.360 We got patches for this. 18:27.360 --> 18:28.000 You want there. 18:28.000 --> 18:28.960 You got patches for that. 18:28.960 --> 18:30.720 How would we combine the patch set? 18:30.720 --> 18:34.960 That's pretty much a business business agreement between two corporations, 18:34.960 --> 18:37.920 except it's just been negotiated by two engineers. 18:37.920 --> 18:44.000 So you, the engineers and the room, act effectively as business development and agreement for your corporations. 18:44.400 --> 18:47.360 You also act as copyright agents. 18:47.360 --> 18:50.240 This has been a seismic shock to the industry. 18:50.240 --> 18:58.640 And it's still today, you'll find a lot of companies who think that intellectual property is the life blood of any software company. 18:58.640 --> 19:07.840 And it's jellously guarded by the legal department who usually appoints one single vice president as the guy who is authorised to buying that intellectual property. 19:07.840 --> 19:11.840 So there's usually one vice president in the company who can sign a contract that says, 19:12.000 --> 19:16.800 you know, we give you a license to our patents, so we give you a license to our copyrights or whatever. 19:16.800 --> 19:22.000 You as open source developers whenever you sign off on the DCO in any open source project, 19:22.000 --> 19:27.840 you are now binding your company to both its copyrights and its patents on behalf of that company. 19:27.840 --> 19:33.760 Effectively, you are all now agents for copyrights and patents in open source. 19:33.760 --> 19:38.880 So the job that used to belong to one person because it was so jellously guarded, 19:38.960 --> 19:42.800 is now spread around all of you lots sitting in this room. 19:42.800 --> 19:48.080 That also means that you've taken over functions of the legal department used to jellously guard. 19:48.080 --> 19:49.760 That's pretty good. 19:49.760 --> 19:56.000 And effectively, in an open source project, you actually act as your company's representative to the project. 19:56.000 --> 20:04.880 People judge your company by the way you behave in that open source project because for some of the time you actually are reflecting the interest of your corporation. 20:04.880 --> 20:09.600 And this is right because an open source, we're looking for the problems we want to solve. 20:09.600 --> 20:14.000 We don't want the corporate corporation telling us through its engineer what to do, 20:14.000 --> 20:20.160 but we do want the corporation to describe its problems to us so we can come with a mutually satisfactory solution. 20:20.160 --> 20:25.280 And use the engineer representing the corporation to the project are describing the problem. 20:26.480 --> 20:33.280 And obviously, you act as the representative of whatever project you're working on inside your company. 20:33.280 --> 20:37.600 Because if your company is so sometimes you describe the problem, 20:37.600 --> 20:42.560 you come back from the negotiation and the community goes, well, it's an interesting problem, 20:42.560 --> 20:46.240 but we did it here and we did it this way. Why do you want to solve it that way? 20:46.240 --> 20:50.080 And you might look at it as an engineer and go, oh gosh, we hadn't thought of that. 20:50.080 --> 20:56.960 And then your job is to go to the back of the company and say, well, the community just told us that we couldn't do it the way we were expecting to do it. 20:56.960 --> 21:02.400 We have to do it a different way instead. And now you have to alter the business plan to accommodate this different way. 21:02.480 --> 21:08.320 So this is how open source projects reflect influence back onto the corporations. 21:08.320 --> 21:13.760 And again, they do it through their engineers. And the statue and influence of their engineers through that project. 21:14.720 --> 21:23.040 So a lot rides on your shoulders. Effectively, you are an ambassador, both from the project to your company and from the company to your project. 21:25.680 --> 21:31.760 Engineers are also no longer fungible. So fungible means that, let's say, the five-year-old note, 21:31.760 --> 21:38.640 I haven't my pocket. I could swap it with any five-year-old note in the audience and it would still retain the same capability of buying stuff. 21:38.640 --> 21:43.520 That's what fungibility is. You can just interchange all these things arbitrarily. 21:44.080 --> 21:53.360 In the early days in proprietary projects, engineers were deemed to be fungible because all I really needed was a set of engineers with a particular skill set. 21:53.360 --> 21:57.200 And it didn't matter who they were. Any old engineers with that skill set would do. 21:58.080 --> 22:05.840 But nowadays, in open source, a lot of you who work on open source, if you change company, 22:05.840 --> 22:09.920 your influence in the open source would go with you and the corporation would lose it. 22:09.920 --> 22:16.560 So the fungibility of engineers is gone because open source makes you different. It means that 22:16.560 --> 22:21.680 the project goes with you when you leave and also the contribution stream. 22:22.320 --> 22:27.200 And the community development of that project continues regardless of what the company did. 22:27.200 --> 22:30.320 I mean, if you were working on a proprietary project and they fired you, 22:30.320 --> 22:34.000 you wouldn't get to work on that project anymore. But if you're working on an open source project, 22:34.000 --> 22:38.560 it can go for a bid they fired you. You can still go to another company and continue working on that project. 22:40.560 --> 22:47.280 And this non-fungibility of engineers results in massive empowerment. So it gives you way more power 22:47.280 --> 22:52.720 than the engineers had 20 years ago when they were just regarded as fungible cogs in the corporate 22:52.720 --> 22:58.480 machinations of building proprietary products. And obviously, to quote Marvel Comics, 22:58.480 --> 23:03.440 with great power comes great responsibility. So part of this talk is how you should use that 23:03.440 --> 23:08.320 power and the responsibility. One of the biggest problems that people have in open source to go 23:08.320 --> 23:14.400 back to the theme that was actually on the previous talk is that a lot of times people are afraid 23:14.400 --> 23:18.560 of using that power. You're afraid that if you stand up and speak in your company they'll 23:18.560 --> 23:23.520 fire you because that's not what they want to do. Corporations actually do try and project 23:23.520 --> 23:28.880 this sort of aura of corporate power to keep you in line. It's not really a one-way street. 23:28.880 --> 23:35.920 It's partly their fault as well. But your job as an open source engineer employed by a company is 23:35.920 --> 23:41.120 to guide that company through its adoption of open source. And there are always problems in the 23:41.120 --> 23:45.840 where corporation adopts open source. And if your company is screwing up, you can't just sit 23:45.840 --> 23:53.280 back and go hard. It's partly your fault. You need to fix it. And obviously, having this greater 23:53.280 --> 24:01.680 power means that you can more easily affect change inside your company because unlike 20 years ago, 24:01.680 --> 24:05.440 they are a bit more afraid of firing you if you say something that they don't like. 24:06.080 --> 24:12.240 So to get on to the meat of the talk with 25 minutes to go, how do you actually trust corporations? 24:14.800 --> 24:19.360 Firstly, you do have to understand how they work. You can't just treat a corporation as a black 24:19.360 --> 24:23.840 box and expect to do what you want. You actually need to understand the mechanics of what it does. 24:26.080 --> 24:31.200 And obviously, the reason for this is if your job as an engineer is to affect change within 24:31.200 --> 24:36.160 your organization when they're screwing up, you have to know how to affect their change. 24:36.160 --> 24:39.840 There are no ifs and buts about this, you have to understand how to do it. 24:42.720 --> 24:47.680 And part of the problem and the reason everybody criticizes it, especially in today's world of 24:47.680 --> 24:53.120 Donald Trump, transactionalism is a dirty word. But corporations, all their relationships are 24:53.120 --> 24:58.480 completely transactional, whether it's with other companies, with our employees, with you, 24:58.480 --> 25:03.120 with open source projects. The question they always ask is, what is the mutual benefit? 25:03.120 --> 25:07.280 You know, if I'm not getting something out of this relationship, this relationship is no good to me, 25:07.280 --> 25:12.640 as far as the corporation sees it. And this is right because corporations' job is to make money. 25:12.640 --> 25:17.520 If a relationship is going bad and I'm not making money from it, it's easier to dump it and not have the 25:17.520 --> 25:23.760 to cost-rain. But the thing that a lot of open source people don't appreciate when you talk about the 25:23.760 --> 25:29.120 ideals and open sources, you can laugh at the companies for being transactional. But you, 25:29.120 --> 25:34.080 as open source developers, are effectively just as transactional as they are. If you think about it, 25:34.800 --> 25:39.040 a corporation only invests in things that's relevant to its mission. And its mission can change. 25:39.040 --> 25:43.600 I mean, if it's screwing up and the board tells it that you need to go into a different segment, 25:43.600 --> 25:48.400 it often will, with particular ruthlessness, it'll fire all its engineers, move into the 25:48.400 --> 25:54.480 new segment, acquire load of new ones, do lots of other stuff. But if that happened and the 25:54.480 --> 25:58.880 corporation stopped investing in your project, you probably leave the corporation going work for 25:58.880 --> 26:04.320 another one that actually would invest in your project. You, therefore, are not faithful to the 26:04.320 --> 26:10.160 company's mission statement, you're faithful to your project. That's actually a relationship transactionalism. 26:10.160 --> 26:15.360 The fact that you'd upstix and leave if they didn't do what you wanted them to do. That's transactionalism 26:15.440 --> 26:22.880 on your behalf. Being transparent about this transactionalism can help. Being willing to leave 26:22.880 --> 26:28.560 a company, being transactional about it, also increases your power as well, because it's your negotiating 26:28.560 --> 26:33.360 strength with the company. Your threat to leave if you want to make it is what they have to take 26:33.360 --> 26:41.120 seriously. But the fact that both of you have the capability of hurting each other and you're 26:41.120 --> 26:46.320 in a position to exercise that capability is the basis for a mutual understanding. It's pretty 26:46.320 --> 26:53.360 much how all peace treaties work and how all contracts between competing companies work in the industry. 26:53.360 --> 26:58.160 This is the basis, this transactionalism is the basis of actually how relationships work in 26:58.160 --> 27:03.040 industry. So the good news is, pretty much, you already understand all of this anyway, because 27:03.040 --> 27:08.400 you've been doing it. It's exactly the same transactionalism as the famous phrase an open source 27:08.400 --> 27:13.200 scratching your own edge. It's basically in order to come to the project and try and do something 27:13.200 --> 27:18.640 with it, you already have to be using the project. And when you try and fix something in a project, 27:18.640 --> 27:22.960 if they really, really told you to go away and go to some other project, effectively you would 27:22.960 --> 27:28.240 give up and go away, stop using that product and start using a different one. This also scratching 27:28.240 --> 27:33.760 your own edge is transactionalism. We practice transactionalism at open source every day. 27:34.480 --> 27:41.840 Obviously, you have to understand that the main corporate driver is revenue. How much revenue 27:41.840 --> 27:46.800 you produce is what the corporation wants to get out of you. There are no ifs and buts about this. 27:46.800 --> 27:51.840 Corporations make lofty statements about their missions, but a base their mission is to make money. 27:52.800 --> 27:57.520 And obviously, if they don't make money, there's very little chance of them paying you, 27:57.520 --> 28:02.240 because you need revenue coming in and profit off that revenue to actually pay your engineers. 28:02.880 --> 28:08.320 Obviously, they do it via some sort of business plan. But the point about business plans, 28:08.320 --> 28:14.080 and this is the dirty secret of all corporations, is pretty much a lot of business plans just 28:14.080 --> 28:18.160 don't work. You've seen a lot of failures in open source. The number of people who tried to 28:18.160 --> 28:23.680 adapt licenses, it's all in pursuit of money. And a lot of the business plans that all the people 28:23.680 --> 28:28.480 you think are executives who know what they're doing come up with simply do not work in the marketplace. 28:29.440 --> 28:34.640 When a corporation finds a plan that actually works, and by working, it's like Mr. Macauber, 28:34.640 --> 28:41.520 where in Charles Dickens he said, if your income is 20 quid and your outgoing is 19 quid and 19 28:41.520 --> 28:46.960 shillings, you're happy you're making a profit. If your outgoing is 20 quid and five and one shilling, 28:46.960 --> 28:52.160 you're unhappy because you're losing money. Just making that extra shilling above your expenses 28:52.160 --> 28:57.760 is enough to keep a company in business. So as far as business plans go, a corporation isn't 28:57.760 --> 29:02.320 a magic quest for the best business plan you could ever have. It's a quest for something that 29:02.320 --> 29:07.920 actually works because most business plans don't. And most executives know this. Once a company 29:07.920 --> 29:12.160 finds a business plan that actually works, they often stick to it for grim death. And it doesn't 29:12.160 --> 29:18.400 matter how inefficient it is, they will fight for that plan. This is the point, it doesn't have to 29:18.400 --> 29:23.440 be optimal. Almost nothing in the where corporation works is optimal and you have to live with that. 29:23.440 --> 29:32.480 And you have to understand how to make that work for you. Obviously also, corporations understand 29:32.480 --> 29:39.200 that decisions have to be made in quick time. You can't do the search for optimisation, 29:39.200 --> 29:43.520 while you're making a business decision. It's rather like, let's say you're in Olympic athlete 29:43.520 --> 29:48.480 on the blocks for a hundred meter race. When the starting gun goes off, you don't sit in the 29:48.560 --> 29:52.720 blocks and think how you're going to win the race. You basically race ahead and try and beat 29:52.720 --> 29:57.200 the competition. Whatever plan you had at that point the gun went off is the one you have to run 29:57.200 --> 30:04.560 with. There's no time to rethink. All corporations are in a race like that. So in order to 30:04.560 --> 30:11.600 influence the corporation, you effectively have to make yourself useful. And to go back, I suppose 30:11.600 --> 30:16.400 the previous talk and present a contrary in position. If a company is just paying you to work 30:16.480 --> 30:21.840 on open source and you're doing nothing but working on that open source project, you are not 30:21.840 --> 30:27.440 being useful to that company. All you are is a figurehead that they like, they employ, 30:27.440 --> 30:32.800 but you're not actually doing anything to sort of move them forward or help them achieve their goals. 30:34.320 --> 30:39.840 And the point for you is you could equally well be paid to do the job you're doing by another company. 30:39.840 --> 30:44.320 You've effectively made yourself completely fungible because they could fire you and you can go 30:44.320 --> 30:48.320 work for another company, take your project with you. But assuming they're benefiting from the 30:48.320 --> 30:52.960 work you were doing on that project, they're just as equally benefit from the work you do with 30:52.960 --> 30:58.320 your paid by another company. Unless you're doing something that makes the work you're doing, 30:58.320 --> 31:03.200 peculiarly relevant to their business plan. That's what you need to find inside the company. 31:03.760 --> 31:11.040 You need to figure out your value and this isn't always easy. But once you have your value add, 31:11.120 --> 31:16.320 that value add is the value you provide to the company over and above actually working on open source. 31:17.040 --> 31:22.160 This means that you're always going to be spending some of your time involved in corporate politics 31:22.160 --> 31:28.480 if you do your job well. No one is going to be paid to work 100% on open source. That doesn't 31:28.480 --> 31:34.080 mean that you can't work on it and you can't get your percentage of time up to 50-70%. But they're 31:34.080 --> 31:39.040 still some percentage of time that you should be spending trying to fix your company because if you're 31:39.040 --> 31:47.520 not, you're not providing value to them. So how do you do this? The key to doing this is to fix 31:47.520 --> 31:54.080 broken stuff first. Like I said, most corporations have a wrong business plan. If you find the wrong 31:54.080 --> 31:58.720 ness in the business plan is often very easy to point it out and suggest how to make it right. 31:58.720 --> 32:03.120 And a wrong business plan is easy to spot for corporations because it's losing the money in some 32:03.120 --> 32:07.440 sense they want to fix it anyway. And so this is the key to actually doing it. 32:09.360 --> 32:14.400 Suggesting improvements is not the way you do it. Remember, corporations always have a 32:14.400 --> 32:20.000 suboptimal business plan but it's having a suboptimal business plan is better than having a business 32:20.000 --> 32:23.760 plan that doesn't work and so they're happy with it and they'll fight tooth and nail against 32:23.760 --> 32:28.880 changing it. So the way you affect changes you fix the broken stuff first and you work on the 32:28.880 --> 32:34.640 feature improvements later. And if this sounds familiar, it should because it's basically the way 32:34.640 --> 32:40.560 that new contributors actually gain trust in an open source project. Amazingly enough, 32:40.560 --> 32:45.920 it's an exact parallel. The way a new contributor gains trust in an open source project is exactly 32:45.920 --> 32:50.560 the same way as you could gain trust within your company. You're just sort of working on different 32:50.560 --> 32:58.640 things than code. Clearly, you have to not be afraid of speaking truth to power. If you see a place 32:58.640 --> 33:03.680 where the company is screwing up, speak up and say so. Tell your manager, they don't respond, 33:03.680 --> 33:08.720 tell his manager and so on. See if you can get all the way up the corporate chain. Ultimately, 33:08.720 --> 33:13.360 you will find somebody as long as you've identified the correct problem. As in, this business 33:13.360 --> 33:18.240 plan is not working. Somebody somewhere will actually listen to that because it's true. 33:18.240 --> 33:22.880 And somebody in the corporate hierarchy has to recognize that truth, pretty much otherwise the 33:22.880 --> 33:26.800 corporation is doomed. If they're never going to recognize the fact they have a failing business 33:26.880 --> 33:30.400 plan, they're going to continue doing it. And that means the company at some point is going 33:30.400 --> 33:35.600 to go under. So they are always very attuned to people coming along and saying this bit's failing, 33:35.600 --> 33:42.080 we should improve it. And that's actually the primary value add of an individual contributor 33:42.080 --> 33:47.040 at a company is finding stuff that doesn't work and making it work. That's pretty much your job. 33:47.040 --> 33:52.880 It's the same description of a job you have in open source often as contributors. And actually, 33:52.960 --> 33:58.000 also, once you're on this ladder, this is the slippery slope. It's also the same ladder that 33:58.000 --> 34:03.120 gets you to promotion. We do need open source engineers to go for promotion. Because the higher 34:03.120 --> 34:08.160 you stand inside your company, the more influence you actually naturally have. And we need you to 34:08.160 --> 34:13.680 expand this influence. The percentage of open source developers who are actually executives with 34:13.680 --> 34:18.800 them, their company is tiny. And it's much smaller than for proprietary engineers. Most of the 34:18.800 --> 34:23.840 corporations you work with are staffed by engineers who are actually working at that corporation 34:23.840 --> 34:28.720 more than 20 years ago before the dawn of open source. Most of them are the people who are now 34:28.720 --> 34:33.520 in the fellow architect, distinguished engineer position. Most of them are the executives. 34:33.520 --> 34:39.040 We need many more open source developers to become executives because the scrubs in the business 34:39.040 --> 34:47.120 plan can only be fixed at the executive level. And obviously open source who really really needs 34:47.120 --> 34:53.360 influence in the C Suite. C Suite means C, O, C, F, O, C, T, O. Anybody with C in front of that title 34:53.360 --> 34:59.920 is often in the C Suite. So you need to do things that individual contributors often do inside 34:59.920 --> 35:05.600 the company in proprietary companies. Architecture reviews, code reviews, that sort of thing. Just 35:05.600 --> 35:10.800 make it look like you're helping your corporation. But as part of this, you can also form alliances 35:10.800 --> 35:16.160 within your company. Because there's always somebody else sitting in the executive chain who would 35:16.160 --> 35:20.240 look at you saying, you know, this business plans broken, who would think, you know, if I could 35:20.240 --> 35:24.800 fix that business plan, the company might promote me as well. And obviously, if they're, if 35:24.800 --> 35:29.120 they're slimy management type promotion is exactly what they're after, you can form an alliance 35:29.120 --> 35:32.560 with them. The alliance will be transactional. At some point they'll stab you in the back, 35:32.560 --> 35:38.960 but it'll work for a while, like all open source things do. Corporate politics is another thing 35:38.960 --> 35:44.480 that you really really need to pay attention to. As I said, corporate corporations often don't 35:44.560 --> 35:49.200 exactly know what the right thing to do is. This leads to a lot of debate internally at the 35:49.200 --> 35:54.480 executive level. And that makes them speak with many in factionalised voices. Corporations 35:54.480 --> 35:58.640 through their marketing departments always try to appear as though they have the same message. 35:58.640 --> 36:02.880 But under the covers behind the scenes, they're usually in a firm end with lots and lots of 36:02.880 --> 36:07.680 different opinions by what they should be doing. The only time that marketing can actually get 36:07.680 --> 36:12.240 corporations to agree to anything is usually when it's announced in the press. Marketing often 36:12.320 --> 36:16.560 tries to drive corporate decisions simply by doing that. You can't get agreement with all 36:16.560 --> 36:20.000 of the executives. You might as well just go out and announce something. And then they'll all 36:20.000 --> 36:26.800 be forced to stick to whatever you announced. And as long as something works, as in it's making 36:26.800 --> 36:32.240 money, the Mr McCorber, five shillings above instead of five shillings below, a corporation is happy. 36:32.800 --> 36:36.640 If you come along with a plan for making more money and they look at it and go, 36:36.640 --> 36:42.000 but it's risky, whatever it failed, you know, they might be unhappy. So once they find something 36:42.000 --> 36:49.600 works, they will often stick to it. But they also care about something that either damages 36:49.600 --> 36:54.880 revenue or reputation. Reputation to a corporation is lifeblood because it's what gets people 36:54.880 --> 37:00.160 to pay for money over for trusting you. And obviously revenue is key to actually making the 37:00.160 --> 37:06.080 corporation function in future. Anything that damages these two functions that you can point to 37:06.080 --> 37:13.760 is something that's broken that you can actually fix. And one of the big problems, especially 37:13.760 --> 37:18.320 in large companies, is that often a lot of the factions within the company just don't get open 37:18.320 --> 37:25.120 source. Usually there is a small faction within a company that's made its reputation by driving 37:25.120 --> 37:30.080 open source forwards. And that's become the dominant faction, but it's not the only factions. 37:30.160 --> 37:36.640 There are many different voices within the company. And so you need to form alliances within 37:36.640 --> 37:40.880 the factions within the corporation. Obviously your natural allies are going to be the people 37:40.880 --> 37:45.600 who are actually doing open source, but don't forget the underdogs. The underdogs are always the 37:45.600 --> 37:51.120 people who are jealous and have been overlooked through promotion. And often if you can give them a plan 37:51.120 --> 37:56.320 which looks like, oh, by the way, if I did this, I might get more influence in the company, 37:56.400 --> 38:00.880 they might actually become your natural allies as well. So don't overlook the fact that non-open 38:00.880 --> 38:06.000 source people can become your allies. Yesterday's enemies can become tomorrow's friends. 38:08.080 --> 38:14.240 And effectively, this job is really like leading a community. This is like hurting cats. 38:14.240 --> 38:18.400 Getting engineers to do stuff that you think they should be doing, except you're trying to get 38:18.400 --> 38:23.840 corporate executives to do stuff you should be doing. You know, it's the same people you're 38:23.920 --> 38:29.200 arguing with, same arguments, they're just wearing suits instead of jeans and trainers. That's it. 38:31.920 --> 38:36.800 So, and I think I'm going to be able to leave about ten minutes for questions. The conclusions of all 38:36.800 --> 38:44.240 of this. Well, understanding the transactional nature of corporations isn't as hard as you think. 38:44.240 --> 38:49.680 Mainly because almost all of your interactions with open source are somewhat transactional in nature. 38:49.680 --> 38:55.280 You already understand how to do this. So embrace the new transactionalism of the modern age because 38:55.280 --> 39:01.920 you'll actually find the most of you are good at it. As a community leader or somebody who 39:01.920 --> 39:06.000 aspires to be a community leader, I know there won't be that many community leaders in the room, 39:06.000 --> 39:10.560 but I imagine a hundred percent of you who work in open source all aspires to be community leaders. 39:11.600 --> 39:16.560 You already possess or are developing most of the skills you need to do this in a corporation. 39:16.560 --> 39:22.320 So, you already know how to make yourself useful in some part to the company you work for. 39:22.320 --> 39:27.360 You just don't often know how to exercise that utility. And you need to get in there and try and 39:27.360 --> 39:33.520 exercise the utility. And now, of course, I've got a vision of my head. When I was at school and 39:33.520 --> 39:38.560 really hated rugby and they used to sort of pick me up by the scruffists. It get into the scrum. 39:38.560 --> 39:41.520 But that's what you need to do in corporations as well. 39:41.600 --> 39:49.520 Every company is always especially a big company that always doing something wrong in open source. 39:49.520 --> 39:55.280 You don't have to look very far to find it. So it's easy to find, go out there and find it and 39:55.280 --> 40:00.720 do something about it. That will make you initially useful to your company. This is what you need to do. 40:03.120 --> 40:07.920 And of course, if once you've found it, you can start proving your worth to them by actually 40:07.920 --> 40:12.880 trying to fix it. And often it's just, you know, adoption of some open source projects, 40:12.880 --> 40:17.600 slight modification in the way the business plan works. There are usually easy fixes to this. 40:17.600 --> 40:22.640 So it's not usually that difficult for you to actually find and fix this stuff on your own. 40:25.360 --> 40:30.640 And of course, if we come back to the original question for me, remember, I used to be 40:30.640 --> 40:36.080 Microsoft's number one enemy. Now, I'm actually working for them. How did this come about? 40:36.080 --> 40:41.520 Well, when I was Microsoft's enemy, Microsoft was actually going about trying to 40:43.120 --> 40:49.600 move Windows into the enterprise. You know, they saw it as a multi-billion dollar market 40:49.600 --> 40:54.080 place that was their natural inheritance. They were effectively an operating system company. 40:54.960 --> 41:00.720 In 2012, Microsoft visited pivoted because the operating system business wasn't doing very well 41:00.800 --> 41:06.880 and they became a cloud company. And part of what happened in that transition from operating 41:06.880 --> 41:12.560 system company to cloud company is that now, most of Microsoft's revenue does not come from selling 41:12.560 --> 41:17.760 operating systems. If you look at our recent quarterly figures, most of the revenue comes from 41:18.400 --> 41:23.920 orchestrating workloads in the Azure cloud. And now, obviously, since Azure is sort of the Windows 41:23.920 --> 41:28.560 cloud, we have a massive amount of Windows workloads still running in the cloud. It's still out there, 41:28.880 --> 41:34.240 some corporations still use it. But we also have a massive amount of Linux workloads. 41:35.040 --> 41:39.680 I'm hoping we've published this figure and it's not proprietary, but I believe the current split in 41:39.680 --> 41:48.960 Azure is 70% Linux 30% Windows. Now, the secret that fortunately is sitting in Microsoft's figures, 41:48.960 --> 41:56.640 they don't often emphasize, is that a cloud company does not depend on its customers remaining static. 41:56.720 --> 42:02.560 It depends on its customers growing. The marketplace only likes Microsoft because we're posting 42:02.560 --> 42:09.200 revenue growth figures at about 30%. Pretty much the entirety of that revenue and 30% of our current 42:09.200 --> 42:16.720 operating profit, which are operating revenue is about $250 billion. 30% of that is we need to 42:16.720 --> 42:23.280 grow at about $75 billion a year. That growth is all coming from orchestrating workloads in the cloud. 42:23.360 --> 42:27.920 There's some slight question about how much of that might be open AI and so on, but we won't get 42:27.920 --> 42:35.040 into that. The point is, if you look at the Windows cloud and the Linux cloud, the entirety of our 42:35.040 --> 42:43.280 revenue growth that Microsoft comes from additional Linux workloads, that's not 80% or 50% that's 42:43.280 --> 42:50.640 about 99.9% of our entire revenue growth comes from Linux clouds. All of Microsoft's future revenue 42:50.720 --> 42:57.280 growth is title Linux. That's a fantastically useful figure for me to use in the boardrooms whenever 42:57.280 --> 43:03.600 I sit there. I'm going to do a talk in the curl track about UFI. One of the things I do to make 43:03.600 --> 43:09.520 myself useful in Microsoft is I end up negotiating with them about what features Windows wants to see 43:09.520 --> 43:15.040 in UFI. I just sat in a meeting two days ago where I got a load of Windows engineers who looked 43:15.040 --> 43:20.480 at this feature proposed for the UFI security sub team to do with secure boot. They looked at it and 43:20.480 --> 43:26.320 said, Windows will not implement this feature if it's done, so we think this feature should not 43:26.320 --> 43:31.760 be implemented. I just stood up in the meeting and I said, well, this feature was proposed by Linux. 43:31.760 --> 43:38.480 Linux is now the majority operating system. It's going to get done. Of course, this is quite a 43:38.480 --> 43:45.840 shock to a lot of Windows engineers, but the fact that I can stand there and they will listen to 43:45.840 --> 43:52.880 me is me proving my value to the company. I'm forcing them to change their attitudes. Effectively, 43:52.880 --> 43:59.120 I'm becoming an agent for change even within the old fashion Windows world of Microsoft. The point 43:59.120 --> 44:06.000 is that you all can become agents for change inside your company as well. That's the way you get success. 44:06.080 --> 44:11.680 That's the way you bind corporations to open source. In fact, to go back to the other talk, 44:11.680 --> 44:17.200 that's the way you make a lot of money. If you actually get yourself promoted to be an executive 44:17.200 --> 44:22.240 at a software company, you'll find money falls out of trees on top of you. This is your mission 44:22.240 --> 44:28.000 in life, you should use to accept it. I should have five minutes left for questions, so I'll 44:28.000 --> 44:32.800 just say that if you enjoy this presentation, it's entirely web-based, so I can't upload a PDF, 44:32.800 --> 44:37.520 but I will post a URL to it. Obviously, that doesn't make me a web developer, which since 44:37.520 --> 44:44.320 I'm a kernel engineer is a really unfortunate thing to have said of you. This is the URL for the 44:44.320 --> 44:49.360 presentation. I've got a funny feeling since I was editing it up front. I need to re-upload it, 44:49.360 --> 44:54.720 but it hasn't changed that much. I will now say thank you and call for questions if there are any. 45:03.280 --> 45:07.440 I thank you James, so Henny's going to run around with the microphone. Can I remind you 45:07.440 --> 45:13.280 it's for questions, not statements? 45:19.520 --> 45:27.520 So can you give any advice if somebody finds himself in a company that already has business 45:27.520 --> 45:31.280 plan that's working, how to gain trust from the company? 45:31.280 --> 45:35.440 Sorry, find some self in a company which has already a working business plan. 45:36.080 --> 45:44.160 Well, like I said, business plan is like a curates egg. A working business plan usually only works 45:44.160 --> 45:49.520 in parts, right? Because there's not one signal business plan that covers everything your company is 45:49.520 --> 45:55.040 doing. Find the parts that aren't working. There is no company on earth today who's business 45:55.120 --> 46:01.360 plan is 100% working. There will always be pieces that are not actually working, that are actually 46:01.360 --> 46:07.200 possible. You just have to find the pieces that are not working. And if on the face of it, it looks 46:07.200 --> 46:11.600 like everything's working, it's probably your job to dig a little deeper and try and find the 46:11.600 --> 46:17.760 bits that work. But I promise you, there are there. No company has 100% optimal business plan. 46:17.760 --> 46:23.120 They all have a business plan which is, you know, it's about 80% optimal, 20% decimal. Just find 46:23.120 --> 46:34.160 the 20% decimal but it's some fixed. Hi. Hi. Thank you for the presentation. Somebody told me 46:34.160 --> 46:40.400 once that technology is always about community. Doesn't matter if it's open source or proprietary. 46:40.400 --> 46:45.600 It's always a community behind the technology itself either through the vendors or the people 46:45.600 --> 46:53.040 buy the users, the contributors. So how do you think that fits with your, that statement fits 46:53.040 --> 46:57.920 with what you just told us? Is it always a community and you can always find yourself a space within 46:57.920 --> 47:03.200 that community, even if you feel your open source and the community you're dealing with is more 47:03.200 --> 47:09.760 close source, you can still somehow contribute to that community. Doesn't violate your code of 47:09.760 --> 47:15.600 conduct of being open source to integrate the close source community? Okay. So the statement was 47:15.600 --> 47:21.760 effectively that all engineers tend to form communities which is true. So you always find communities 47:21.760 --> 47:28.080 even within proprietary companies and communities sort of an open source as well. But the key point 47:28.080 --> 47:34.160 to the message I had here is that communities that form in proprietary projects tend to be 47:34.240 --> 47:38.960 weak communities because they usually only have people within that company. Communities 47:38.960 --> 47:44.080 the form in open source are powerful communities because they're very broad. And the point there 47:44.080 --> 47:50.240 is that the power that comes from being part of an open source community can serve you far better 47:50.240 --> 47:55.920 inside your company than the power that comes from being part of the proprietary project. 47:55.920 --> 48:01.920 And using this power is what you need to do to help move open source forwards inside your company. 48:02.000 --> 48:07.120 Now it is true that the less powerful community has been there a lot longer usually inside a 48:07.120 --> 48:12.160 company if it has loaded proprietary engineers. They usually have a lot of power just because 48:12.160 --> 48:16.560 they accumulate it over time. Like I said, they usually populate the distinguished engineers, 48:16.560 --> 48:21.760 the fellows, they often all come from this. But that doesn't mean that you can't take the power 48:21.760 --> 48:27.760 of an open source project which is 10 to 100 times bigger than the power they have inside their 48:27.840 --> 48:32.720 small community and leverage it to your advantage. Is that sort of roughly answer your question? 48:44.080 --> 48:52.160 In the process you described gaining trust and so on, is there a difference between a small 48:52.160 --> 49:00.480 company like maybe 10 people and Microsoft a big company? Well the difference between a small 49:00.480 --> 49:06.000 and a big company is the number of things that are going on. A small company usually has a fairly 49:06.000 --> 49:11.680 compact and focused mission so it usually has a fairly simple business plan so it doesn't have as many 49:11.680 --> 49:18.640 moving parts as a large company. And this is good and bad because it means that usually a small 49:18.640 --> 49:25.680 company's business plan is so sort of small that it will either succeed or fail and pretty much 49:25.680 --> 49:30.880 at the nascent startup level you're looking at about 70% of them actually do fail. This is about 49:30.880 --> 49:37.120 the percentage of companies that have bad business plans. So usually your job if you're working 49:37.120 --> 49:43.040 for a small company is to tip the entirety of their business plan over from failure to success. 49:43.040 --> 49:47.760 It's actually a very big job and a very difficult one. If you're working in a bigger company 49:47.760 --> 49:52.000 there are many more moving parts. The business plan you can't just write it on the back of a 49:52.000 --> 49:57.200 cigarette packet is probably sitting in a 10-page document somewhere. But that business plan and a 49:57.200 --> 50:02.880 large company has so many moving parts. There are always pieces of that plan that are passable. 50:02.880 --> 50:07.440 And so if you work in a large company you can just find the pieces of that plan and influence 50:07.440 --> 50:13.360 their change. So actually your pathway in a larger company is easier because you're only 50:13.360 --> 50:18.160 changing a small part of the business plan than it is in a small company because usually your 50:18.160 --> 50:23.200 job there will be to tip the business plan from failure to success. I mean I didn't say this but 50:23.840 --> 50:30.880 the reason I was actually the guy put forward to go to OSDL in 2005 was because I was already 50:30.880 --> 50:36.480 chief technology officer and vice president of a startup company. The reason I was that is because 50:36.480 --> 50:42.480 I found we were formed in 2000 to bring high availability to Linux and I was the one who said 50:42.480 --> 50:46.800 you know this business plan from a technical point of view just doesn't work. We have to fix this. 50:46.800 --> 50:52.320 That's why I became chief technology officer. But I had to tackle the entirety of the business plan 50:52.320 --> 50:57.200 there not just the pieces of it. At a large company you can tackle the pieces of it and it's usually 50:57.200 --> 51:08.880 much easier. One very long question. Sorry I presume this is not the first time you talk about 51:08.960 --> 51:17.120 a topic. Do you feel that if you talk to a technical audience you also get a well reception because 51:17.120 --> 51:23.200 this sounds very much like something that would be well received but people who are already in a 51:23.200 --> 51:30.480 more leadership mentality and I think that if it's a really technical audience it might not go 51:30.480 --> 51:37.280 down as well. Do you have an experience on that or not at all? Well firstly are you saying that the 51:37.360 --> 51:42.160 audience in Foz Dam isn't technical because that's not going to go down very well. The point is 51:42.160 --> 51:48.880 I am here speaking to technical people about business plans, right? So and I do this at a lot of 51:48.880 --> 51:54.080 other conferences as well. Foz Dam is actually my favorite conference largely because there are 51:54.080 --> 52:00.720 pretty much no suits in the room. So I'm not talking to I'm not telling I also I do go to business 52:00.720 --> 52:06.480 conferences or sometimes I get invited less so than I used to when Linux was to start up. 52:06.480 --> 52:11.840 So I have talked to suits giving them roughly the same talk except basically in my message to 52:11.840 --> 52:16.480 them is you have to listen to your engineers because they know what they're doing and your 52:16.480 --> 52:21.120 engineers are doing way more for your company than you think going back to being you know the 52:21.120 --> 52:25.840 business development people for the company being the lawyers for the company and so on. But the 52:25.840 --> 52:32.000 point about this talk is for technical people and it's basically that the power lies in your hands 52:32.000 --> 52:37.760 to affect change within the company and there are many things you can do. Power is only effective 52:37.760 --> 52:42.800 if you exercise it. If you just sit there with the power and do nothing you will achieve nothing 52:42.800 --> 52:47.920 within your company. So the message to the technical people as I hope everybody in this audience 52:47.920 --> 52:54.240 is that the changing role of the engineer over the years has given you the power to affect the 52:54.240 --> 52:59.920 change. You just need to find out what is going wrong for this company and use the power that you 53:00.000 --> 53:04.880 already have through your involvement in open source to make that change effective. That's the 53:04.880 --> 53:12.480 technical argument. It's not an argument for suits. Thank you very much James. Great talk.