WEBVTT 00:00.000 --> 00:09.640 Okay, we can go for the next. 00:09.640 --> 00:14.640 So my talk's a shell script, I'm sorry, should we play game? 00:14.640 --> 00:21.000 It requires audience participation, so I hate that too, but I thought I'd be really 00:21.000 --> 00:23.800 so much of a look. 00:23.800 --> 00:26.920 Is this a networking commit to freebiscity? 00:26.920 --> 00:27.920 Yes or no? 00:27.920 --> 00:28.920 Easy question. 00:28.920 --> 00:29.920 Is that a networking commit? 00:30.920 --> 00:32.920 Okay everyone says yes. 00:32.920 --> 00:34.920 Is this a networking commit? 00:34.920 --> 00:35.920 No? 00:35.920 --> 00:36.920 Okay. 00:36.920 --> 00:38.920 Is this a networking commit? 00:38.920 --> 00:42.920 Like, I know the top strength hit, but it's a great commit, you should go and read it, but 00:42.920 --> 00:43.920 no. 00:43.920 --> 00:45.920 Is this a networking commit? 00:45.920 --> 00:46.920 No. 00:46.920 --> 00:47.920 This? 00:47.920 --> 00:48.920 Probably not. 00:48.920 --> 00:49.920 No. 00:49.920 --> 00:50.920 No. 00:50.920 --> 00:51.920 No. 00:51.920 --> 00:53.920 Is this a networking commit? 00:53.920 --> 00:55.920 I don't think so. 00:55.920 --> 00:58.920 Is this a networking commit? 00:58.920 --> 00:59.920 It's got fibs in it. 00:59.920 --> 01:02.920 That's a networking thing. 01:02.920 --> 01:03.920 Does it matter? 01:03.920 --> 01:06.920 If we're talking about the networking stack should we highlight this? 01:06.920 --> 01:08.920 It's hard to say. 01:08.920 --> 01:09.920 I can say no. 01:09.920 --> 01:10.920 I don't think this is either. 01:10.920 --> 01:12.920 It doesn't have any magic words in it. 01:12.920 --> 01:14.920 Is this a networking commit? 01:14.920 --> 01:15.920 Yes. 01:15.920 --> 01:17.920 Super easy. 01:17.920 --> 01:19.920 Is this a networking commit? 01:19.920 --> 01:20.920 No? 01:20.920 --> 01:22.920 Well, let's see. 01:22.920 --> 01:27.920 So, you know, we've looked at 12 commits, now very quickly, but 12 commits. 01:27.920 --> 01:33.920 And 3BSD commits have a nice format, mostly, where we get an area highlighted. 01:33.920 --> 01:35.920 So, this is NVMe control, like, straightforward. 01:35.920 --> 01:42.920 We get a subject line, and then we get a body that says some things. 01:42.920 --> 01:45.920 This body uses the word TCP, which makes us all think it's networking, but it says NVMe 01:45.920 --> 01:47.920 is where, like, oh, no. 01:47.920 --> 01:50.920 Is NFS a networking system? 01:50.920 --> 01:51.920 Yes. 01:51.920 --> 01:53.920 So, this is probably, so this is NVMe over TCP. 01:53.920 --> 01:55.920 So, this is probably a networking commit. 01:55.920 --> 01:57.920 So, we can talk about that. 01:57.920 --> 01:59.920 See, we can take it. 01:59.920 --> 02:04.920 I started with a demo, but the demo is a four loop, so it's not as impressive as emails. 02:04.920 --> 02:06.920 It's just a shell script. 02:06.920 --> 02:09.920 It gives us how many commits there've been in the last window. 02:09.920 --> 02:11.920 How many we've taken, how many we've rejected. 02:11.920 --> 02:14.920 If I just press enter, it will take it because there's quite a lot of networking commits. 02:14.920 --> 02:16.920 Do you think this is a networking commit? 02:16.920 --> 02:20.920 If you saw it without seeing the previous commit, would you think it would have networking impact? 02:20.920 --> 02:24.920 It's still Chelsea, although, but they might do other stuff. 02:24.920 --> 02:26.920 It's a guess. 02:26.920 --> 02:28.920 Oh, take it. 02:28.920 --> 02:30.920 I don't think this is a networking commit. 02:30.920 --> 02:32.920 I don't think this is an NV list. 02:32.920 --> 02:34.920 Is this a networking commit? 02:34.920 --> 02:35.920 Yeah. 02:35.920 --> 02:36.920 Yeah, Chris loved it. 02:36.920 --> 02:37.920 He doesn't do anything else. 02:37.920 --> 02:39.920 Definitely take that. 02:39.920 --> 02:44.920 So, there's quite a lot going on. 02:44.920 --> 02:51.920 In any given week, you free BSD, there's one to 200 commits coming through. 02:51.920 --> 02:56.920 And this chart is a summary of the reports I've done for the end of last year, 02:56.920 --> 02:59.920 but I made up the chart just based on numbers. 02:59.920 --> 03:03.920 It's not particularly accurate because I'm doing unicode art here. 03:03.920 --> 03:09.920 But yeah, like any given week was 1 to 200 commits across a dozen or so specializations. 03:09.920 --> 03:13.920 And they get very, very, very in-depth, very quickly, 03:13.920 --> 03:16.920 to the point where there's a couple of people in the world, 03:16.920 --> 03:19.920 which could reason about what is happening here. 03:19.920 --> 03:20.920 You know, we're lucky. 03:20.920 --> 03:21.920 We have a big tree. 03:21.920 --> 03:27.920 We've got a lot of consumers and free BSD commiters are automatically subscribed to the free BSD source commits list. 03:27.920 --> 03:31.920 So, some people find out they have a commit bit because they get 300 emails overnight. 03:31.920 --> 03:33.920 And they're like, what the hell is going on? 03:33.920 --> 03:37.920 But other than, you know, a very special set of people on the internet, 03:37.920 --> 03:40.920 no one, one person can reason about this. 03:40.920 --> 03:45.920 Yeah, there's some people in free BSD, but anyone reading the commit mail 03:45.920 --> 03:48.920 is already not the audience that needs the commit mail explain to them. 03:48.920 --> 03:50.920 They've already stepped aside of this. 03:50.920 --> 03:56.920 And so, few people can take this on, like, very directly. 03:56.920 --> 04:01.920 I think it would be great if there was regular news about free BSD development. 04:01.920 --> 04:05.920 As a project, we do quite well in documentation, 04:05.920 --> 04:08.920 our managers, I think are better than anywhere else. 04:08.920 --> 04:10.920 I've struggled to think of a man-page average, 04:10.920 --> 04:12.920 wasn't the free BSD one that I really liked. 04:12.920 --> 04:14.920 But, you know, I'd like the man-pages. 04:15.920 --> 04:16.920 The handbook's pretty good. 04:16.920 --> 04:19.920 We've got reasonable tutorial stuff online, 04:19.920 --> 04:21.920 but we don't talk about what we do. 04:21.920 --> 04:23.920 We're really poor at communicating what's going on. 04:25.920 --> 04:27.920 Yeah, so in the last few years, 04:27.920 --> 04:29.920 there's been some efforts towards this. 04:29.920 --> 04:32.920 Mike Carles was helping with release notes, 04:32.920 --> 04:34.920 and he was assembling more stuff. 04:34.920 --> 04:35.920 And it was doing well. 04:35.920 --> 04:37.920 Very slightly passed away last year. 04:37.920 --> 04:39.920 We've been doing status reports, 04:39.920 --> 04:43.920 but I've read all of the status reports and they're very hard to digest. 04:43.920 --> 04:45.920 And they're not immediately tied to things. 04:45.920 --> 04:47.920 And I don't see coverage of them. 04:47.920 --> 04:49.920 I don't see anyone care. 04:49.920 --> 04:52.920 And for other things you do see people care. 04:52.920 --> 04:54.920 And I had this idea a long time ago that, 04:54.920 --> 04:56.920 you know, LWN exists. 04:56.920 --> 04:58.920 LWN is an amazing thing. 04:58.920 --> 05:02.920 And if you want to talk about advocacy for an open source project, 05:02.920 --> 05:05.920 LWN is just above. 05:05.920 --> 05:07.920 They write really high quality technical documentation 05:07.920 --> 05:09.920 based on active development. 05:09.920 --> 05:11.920 It's, for a long time when I was an academic, 05:12.920 --> 05:14.920 it was a first stop before reading a paper 05:14.920 --> 05:17.920 because there'd be a paper a conference for some change. 05:17.920 --> 05:21.920 LWN would give you an approachable explainer for what was going on. 05:21.920 --> 05:24.920 And they're really great for documentation like that. 05:24.920 --> 05:26.920 And they can do this because they have money. 05:26.920 --> 05:29.920 LWN costs money to read every month, 05:29.920 --> 05:32.920 but some big companies pay for open source maintainers to get it. 05:32.920 --> 05:33.920 But they pay their staff. 05:33.920 --> 05:34.920 They're paying writers. 05:34.920 --> 05:37.920 They're paying probably a staff pool of 20 to 30 people. 05:37.920 --> 05:40.920 And they may have 8 to 10 contributors for each issue of LWN. 05:40.920 --> 05:42.920 And so there's money here. 05:42.920 --> 05:44.920 And you know, we're a smaller project. 05:44.920 --> 05:46.920 We don't have this big kishae. 05:46.920 --> 05:47.920 We don't have as much going on. 05:47.920 --> 05:50.920 We're definitely not enough activity to keep 10 people 05:50.920 --> 05:52.920 contributing freelance articles. 05:52.920 --> 05:55.920 And I would see if we could make a semi automated approach 05:55.920 --> 05:57.920 to summarizing development and frivolousity. 05:57.920 --> 06:00.920 Can we break down the problem a bit 06:00.920 --> 06:02.920 and get a technical writer? 06:02.920 --> 06:03.920 So a technical person, 06:03.920 --> 06:05.920 but not a subject matter expert. 06:06.920 --> 06:10.920 To ingest this fire hose commit log, 06:10.920 --> 06:14.920 speak to people and give us something useful on the other side. 06:14.920 --> 06:17.920 Can we reduce the expert burden on the knowledge? 06:17.920 --> 06:21.920 So you can communicate what we're debating on our projects better. 06:21.920 --> 06:25.920 And I want there to be more writing about frivolousity. 06:25.920 --> 06:27.920 I've been doing this for a very long time now, 06:27.920 --> 06:28.920 which is kind of sad to say. 06:28.920 --> 06:31.920 I think I came to close them for 10 nine years ago. 06:32.920 --> 06:35.920 Since then, I've written a ton of several papers 06:35.920 --> 06:38.920 where I never managed to anything really focused on frivolousity 06:38.920 --> 06:40.920 because I worked in networking. 06:40.920 --> 06:43.920 But you know, we very frequently use frivolousity as a platform. 06:43.920 --> 06:44.920 And we get to, you know, in the paper, 06:44.920 --> 06:46.920 you get to use a wonderful justification of why you pick this. 06:46.920 --> 06:47.920 And you just get to say, 06:47.920 --> 06:48.920 the authors knew it. 06:48.920 --> 06:49.920 And that's why you did it. 06:49.920 --> 06:52.920 So I got to push frivolousity into more places. 06:52.920 --> 06:54.920 I did more frivolousity to the ITF. 06:54.920 --> 06:56.920 I've written far too many presentations. 06:56.920 --> 06:58.920 That's why I've just wrote a shell script 06:59.920 --> 07:01.920 because of the sick of keynote. 07:01.920 --> 07:03.920 And I put our writing in other places. 07:03.920 --> 07:05.920 So I've kept the blog as too scared to see 07:05.920 --> 07:07.920 when I first started writing about frivolousity 07:07.920 --> 07:08.920 but it's probably a long time ago. 07:08.920 --> 07:12.920 And I put out commercial articles, places you've read, 07:12.920 --> 07:14.920 big magazines. 07:14.920 --> 07:16.920 And I'm an editor in the frivolousity journal, 07:16.920 --> 07:18.920 which is a magazine where the word journal 07:18.920 --> 07:20.920 and the title to confuse academics into submitting, 07:20.920 --> 07:22.920 which works really well. 07:22.920 --> 07:25.920 And so I wanted to be more about frivolousity. 07:25.920 --> 07:27.920 And twice a month for some reason, 07:27.920 --> 07:28.920 it was Benedict Washington. 07:28.920 --> 07:29.920 And we record BST now. 07:29.920 --> 07:31.920 And we talk about BST. 07:31.920 --> 07:35.920 And most of what we talk about is based on what other people have written. 07:35.920 --> 07:36.920 And if you stop writing, 07:36.920 --> 07:38.920 we don't have a podcast anymore. 07:38.920 --> 07:40.920 And I'll get those hours of my life back. 07:40.920 --> 07:41.920 And nobody wants that. 07:41.920 --> 07:43.920 They want to keep the podcast going. 07:43.920 --> 07:47.920 There's no place right now for what's going on in frivolousity. 07:47.920 --> 07:49.920 There is a great place for what's going on in open BST. 07:49.920 --> 07:51.920 Undeadly the org is an amazing resource. 07:51.920 --> 07:55.920 It was a big draw to me as a nerdy teenager 07:55.920 --> 07:58.920 wanting to just read more and not necessarily having 07:58.920 --> 08:00.920 the technical tropes to do development. 08:00.920 --> 08:02.920 I could read on Devly.org and I could see, 08:02.920 --> 08:04.920 you know, the technical articles we'd go by. 08:04.920 --> 08:06.920 And I didn't know how to pronounce WX or X, 08:06.920 --> 08:08.920 let alone know what it meant. 08:08.920 --> 08:10.920 But I could read about hackathons and how they 08:10.920 --> 08:11.920 progressed the project. 08:11.920 --> 08:12.920 And I could see ways in. 08:12.920 --> 08:15.920 And that came from relatable writing about the projects. 08:15.920 --> 08:16.920 And the things that are happening. 08:16.920 --> 08:18.920 So we need more of this. 08:18.920 --> 08:20.920 And we need some way through. 08:20.920 --> 08:23.920 I did this off my own back. 08:23.920 --> 08:27.920 I'm a staff software engineer at FreeBestie Foundation 08:27.920 --> 08:29.920 to give me latitude just to do stuff. 08:29.920 --> 08:31.920 And so I just did stuff. 08:31.920 --> 08:33.920 And I thought I'd be fun to write some shell scripts 08:33.920 --> 08:35.920 because shell scripts are just the best thing in the world. 08:35.920 --> 08:39.920 Put out some reports because if you ask for opinions, 08:39.920 --> 08:42.920 you're just going to get mired and locked down in 08:42.920 --> 08:44.920 what other people want from something. 08:44.920 --> 08:46.920 They will dream the biggest thing in the world. 08:46.920 --> 08:50.920 And that will slow down the momentum here and I will stop things. 08:51.920 --> 08:53.920 And we'll push this out somewhere really easy. 08:53.920 --> 08:55.920 So I put it on my blog, which is the easiest place I could 08:55.920 --> 08:58.920 put anything, which is also a tangle of shell scripts, 08:58.920 --> 08:59.920 which is just perfect. 08:59.920 --> 09:02.920 And then through some growth hacking techniques, 09:02.920 --> 09:05.920 solicit feedback from the community about what's 09:05.920 --> 09:08.920 good, what's bad, where things should go. 09:08.920 --> 09:10.920 So I wrote a shell script. 09:10.920 --> 09:11.920 You've now seen the shell script. 09:11.920 --> 09:13.920 It's very simple. 09:13.920 --> 09:15.920 It generates a list of commits to review. 09:15.920 --> 09:19.920 It uses the magic of Gitlog to do commits from 09:19.920 --> 09:23.920 last week until now, which anyone who's seen 09:23.920 --> 09:25.920 enough by one hour is like, you're going to miss stuff there. 09:25.920 --> 09:27.920 If you don't do this exactly the same time, it doesn't matter. 09:27.920 --> 09:29.920 If I miss stuff, that's fine. 09:29.920 --> 09:31.920 It shouldn't slow down the momentum. 09:31.920 --> 09:34.920 We want to get stuff out and we don't want to get locked behind 09:34.920 --> 09:36.920 the barrier of perfection because that will just stop 09:36.920 --> 09:37.920 as doing things. 09:37.920 --> 09:40.920 And it's much better to have anything in the world than 09:40.920 --> 09:42.920 nothing in the world, but it'd be really, really good 09:42.920 --> 09:44.920 it's just on your computer. 09:44.920 --> 09:47.920 I made an interface to go through the commit so I could 09:47.920 --> 09:48.920 take them and reject them. 09:48.920 --> 09:49.920 It's one shot. 09:49.920 --> 09:51.920 So if I hit enter too quickly, I get a commit, which is 09:51.920 --> 09:52.920 fine. 09:52.920 --> 09:53.920 I'm a human. 09:53.920 --> 09:54.920 I'll just remove it later. 09:54.920 --> 09:55.920 And if I hit no, it's gone. 09:55.920 --> 09:56.920 I don't cover it. 09:56.920 --> 09:59.920 Hopefully, it was in a series that I will catch. 09:59.920 --> 10:00.920 Or someone will tell me I missed it. 10:00.920 --> 10:02.920 And I can go back and talk about it. 10:02.920 --> 10:04.920 It keeps the log of the items that we're going to be 10:04.920 --> 10:05.920 to buck it. 10:05.920 --> 10:07.920 But that's just so that I could make a plot like I did 10:07.920 --> 10:08.920 at the start here. 10:08.920 --> 10:11.920 So I can think about trends in the future with 10:11.920 --> 10:12.920 some supporting numbers. 10:12.920 --> 10:14.920 I didn't know before starting with how many commits 10:14.920 --> 10:16.920 there would be in a week or three BSD. 10:16.920 --> 10:17.920 I'll let alone a year. 10:17.920 --> 10:18.920 And so that was good. 10:18.920 --> 10:20.920 I templated the final report. 10:20.920 --> 10:24.920 Templating the report gives you something repeatable. 10:24.920 --> 10:28.920 And all I mean, English gives you like a nicer, consistent document. 10:28.920 --> 10:31.920 And consistency is a real benefit for anything because it 10:31.920 --> 10:32.920 helps the reader. 10:32.920 --> 10:35.920 For a technical material of the reader knows what to expect 10:35.920 --> 10:36.920 next. 10:36.920 --> 10:37.920 They won't be surprised. 10:37.920 --> 10:40.920 They'll know how to skip through the boring stuff they don't want. 10:40.920 --> 10:44.920 But they'll also know the cues in the writing that will help them read. 10:44.920 --> 10:49.920 It's a really good way to get better understanding from something. 10:49.920 --> 10:53.920 So each Friday, I was streaming this, but it proved too hard. 10:53.920 --> 10:56.920 Did you get pool and I run a shell script. 10:56.920 --> 10:59.920 And then do some other stuff. 10:59.920 --> 11:03.920 I was streaming this, but it conflicted with having a six month old. 11:03.920 --> 11:07.920 And while I can write a report, I can't write a report and talk about it 11:07.920 --> 11:10.920 without sounding like an absolutely crazy person. 11:10.920 --> 11:13.920 When I've had no more than two hours sleep at a time for three weeks. 11:13.920 --> 11:14.920 So I had to put that away. 11:14.920 --> 11:17.920 Maybe we'll try again in the future. 11:17.920 --> 11:19.920 The report generates a file with a header. 11:19.920 --> 11:24.920 The header consists of going on in FreeBSD because I like the English turner phrase. 11:24.920 --> 11:27.920 I don't know how well this translates outside of English. 11:27.920 --> 11:31.920 It has a file I keep just in the director for this called Notes, 11:31.920 --> 11:36.920 where I've assembled stuff that isn't going to appear in the commit log. 11:36.920 --> 11:38.920 So good example of this is the stab. 11:38.920 --> 11:43.920 The stabilization we remind us that the lab has been running, 11:43.920 --> 11:48.920 where they pool part of the FreeBSD tree and they try and just test it in production 11:48.920 --> 11:50.920 for Netflix and other places. 11:50.920 --> 11:53.920 And they land in the FreeBSD source mailing list, but they're not commit, 11:53.920 --> 11:55.920 so they don't appear and get. 11:55.920 --> 11:57.920 So they just sort of disappear. 11:57.920 --> 11:59.920 I've put the section headers. 11:59.920 --> 12:01.920 So I remember what the section headers are called. 12:01.920 --> 12:04.920 So when you read this, you can look in the section you looked in before. 12:04.920 --> 12:06.920 If you want to know what happened in wireless. 12:06.920 --> 12:07.920 And it puts a fruit in. 12:07.920 --> 12:08.920 And the fruit has two things. 12:08.920 --> 12:10.920 It says, please send me an email now. 12:10.920 --> 12:12.920 I want to know this is good. 12:12.920 --> 12:18.920 And it thanks the FreeBSD Foundation for giving me the time to do this. 12:18.920 --> 12:26.920 I get this log and I get a list of 20 to a couple hundred commits for this one time. 12:26.920 --> 12:30.920 And I have to rearrange it into logical groups that would make sense to discuss 12:30.920 --> 12:33.920 and make sense for a reader to understand. 12:33.920 --> 12:37.920 I've not been trimming out some of the stuff which is just going to fall in the bucket of some boxes. 12:37.920 --> 12:43.920 Big, some bug fixes updates and improvements because it's a bit tedious, but it's also good to see. 12:43.920 --> 12:48.920 By rearranging these, the order gets thrown away. 12:48.920 --> 12:50.920 But if you want to know the order things happen, you have to get logged. 12:50.920 --> 12:52.920 You can just go pull this so I just don't worry about it. 12:52.920 --> 12:59.920 And I had commentary as best as I can with the desire to do things quickly first and accurately second. 12:59.920 --> 13:04.920 Because there's just nothing better in the world than saying something wrong to get the right answer. 13:04.920 --> 13:07.920 I published the easiest place I can. 13:07.920 --> 13:10.920 I published somewhere easy because I could publish them quickly. 13:10.920 --> 13:12.920 And so it goes on my blog. 13:12.920 --> 13:16.920 A part of it going on my blog means that nobody else has to check over this because it just goes out. 13:16.920 --> 13:18.920 And you can't stop me blogging. 13:18.920 --> 13:21.920 If it went out through the FreeBSD Foundation, 13:21.920 --> 13:24.920 I think things would get a bit more mired. 13:24.920 --> 13:26.920 They would probably want to have a run over. 13:26.920 --> 13:30.920 And they would then, no, give me the freedom just to do this whenever I wanted to do it. 13:30.920 --> 13:33.920 This just gives me the speed to get the right thing out there. 13:33.920 --> 13:34.920 And then I solicit feedback. 13:34.920 --> 13:36.920 So I sent emails to FreeBSD Net for BST Current. 13:36.920 --> 13:38.920 Net because that's where net work and things happen. 13:38.920 --> 13:42.920 Current because that's where people should be reading if they care about what's going on in current. 13:42.920 --> 13:44.920 I put on the Fediverse. 13:44.920 --> 13:48.920 I drop it into places that will not complain if I put my blog frequently. 13:48.920 --> 13:53.920 I think I've syndicated on some planets, but they're on now just all say FreeBSD network staffs report on them. 13:54.920 --> 13:56.920 Which helps no one at all. 13:56.920 --> 13:59.920 And yeah, I just try and get comments. 13:59.920 --> 14:02.920 So I started doing this. 14:02.920 --> 14:06.920 I think in September, 2024, it was week 39. 14:06.920 --> 14:11.920 The week numbers are dictated by the date command with the V capital V flag. 14:11.920 --> 14:14.920 If you have an argument, you can argue with ISO rather than me. 14:14.920 --> 14:15.920 I don't mind. 14:15.920 --> 14:17.920 I did 12 through the end of 2024. 14:17.920 --> 14:22.920 I stopped on week 50 because we were traveling for the winter break. 14:22.920 --> 14:25.920 And so I wasn't going to have time to sit down and read a ton of stuff. 14:25.920 --> 14:26.920 I picked them up again. 14:26.920 --> 14:28.920 The third week of January. 14:28.920 --> 14:32.920 And I did a mega catch up report in between which covered just under 800 commits. 14:32.920 --> 14:33.920 And it was horrible. 14:33.920 --> 14:37.920 It seemed like four hours to digest all of these. 14:37.920 --> 14:40.920 I could feel myself getting tired right in this just going. 14:40.920 --> 14:42.920 It just just fixes. 14:42.920 --> 14:43.920 It doesn't mean anything. 14:43.920 --> 14:46.920 And so I could see what would happen if you tried someone else tried this. 14:46.920 --> 14:51.920 You didn't have any knowledge or intuition about how stuff works in FreeBSD. 14:51.920 --> 14:53.920 I had done two more reports. 14:53.920 --> 14:54.920 I did one yesterday. 14:54.920 --> 14:56.920 I thought it was good. 14:56.920 --> 14:58.920 I've had some great outcomes from this. 14:58.920 --> 15:00.920 You should buy Kristoffa Beer. 15:00.920 --> 15:01.920 He's left. 15:01.920 --> 15:02.920 So. 15:02.920 --> 15:04.920 I'm going to beer. 15:04.920 --> 15:05.920 You should buy him a beer. 15:05.920 --> 15:16.920 Kristoff is consistently committing a series of improvements to PF, which is I find a standing. 15:16.920 --> 15:20.920 It's funded by NACA and you can see this continuous work going into improving this 15:20.920 --> 15:21.920 firewall. 15:21.920 --> 15:24.920 It's really good. 15:24.920 --> 15:26.920 I've gone a lot of positive feedback. 15:26.920 --> 15:27.920 I've asked for feedback. 15:27.920 --> 15:31.920 And this really helped Caleb Sasser the CEO of panic ink. 15:31.920 --> 15:32.920 They make max software. 15:32.920 --> 15:37.920 You have a talk at the last XOXO conference last year where one of those messages from 15:37.920 --> 15:38.920 this was send a nice email. 15:38.920 --> 15:44.920 If you see someone doing something you appreciate or you want to happen again or has helped you, 15:44.920 --> 15:45.920 you should just thank them. 15:45.920 --> 15:47.920 It's, you know, four words. 15:47.920 --> 15:49.920 Like, like, thanks for that post. 15:49.920 --> 15:51.920 And that helps wonders. 15:51.920 --> 15:54.920 Most of us are saying a home by ourselves and we never really speak to anyone. 15:54.920 --> 15:56.920 And we get bugs. 15:56.920 --> 16:00.920 But just putting out some positive feedback to the world makes everything much better. 16:00.920 --> 16:02.920 Because you can't come to a conference every week. 16:02.920 --> 16:05.920 But you can't just get an email from someone who's gotten some benefit from this 16:05.920 --> 16:06.920 stuff you do. 16:06.920 --> 16:10.920 I've gotten a real strong sense of what's going on in the 3BSD project from doing this. 16:10.920 --> 16:14.920 And, you know, I said that 3BSD commuters get source emails. 16:14.920 --> 16:15.920 I read them. 16:15.920 --> 16:17.920 Sometimes I didn't read them for six months at a time. 16:17.920 --> 16:22.920 But I've read them, like, cursorily, like, not particularly in depth. 16:22.920 --> 16:26.920 Not reading them in depth because I have to categorize them and getting a lot of value from this just to know what's happening. 16:26.920 --> 16:28.920 And hopefully I'm just doing that out. 16:28.920 --> 16:31.920 I learned about the internal driver modernization that Kevin Bowling was doing, 16:31.920 --> 16:33.920 which I think I would have missed. 16:33.920 --> 16:35.920 I would have been like, it's just some numbers that skip that. 16:35.920 --> 16:38.920 But now I get to refer people to talk, you know, 16:38.920 --> 16:39.920 people with bugs. 16:39.920 --> 16:42.920 I can say, well, you could try 3BSD current because it has the new things in it. 16:42.920 --> 16:44.920 You should try and do that. 16:44.920 --> 16:49.920 I've got a great sense for the continual work that goes into the 3BSD TCP stack. 16:49.920 --> 16:53.920 And I sit on the 3BSD transport call every two weeks. 16:53.920 --> 16:55.920 I should know what's going on. 16:55.920 --> 16:59.920 But actually I've got a much better idea of just the small changes that are going through. 16:59.920 --> 17:03.920 I've got a great sense of the small changes which are constantly going into the TCP stack, 17:03.920 --> 17:06.920 because it's working on really complex code. 17:06.920 --> 17:11.920 I learned a much better idea of how many changes to 3BSD are just continuous stuff. 17:11.920 --> 17:16.920 There are a bunch of commitments who are just fixing typos, which seems, you know, 17:16.920 --> 17:18.920 not the highest value work. 17:18.920 --> 17:21.920 But actually, it's increasing the code quality. 17:21.920 --> 17:23.920 They're going through in their reading things. 17:23.920 --> 17:25.920 So we actually get to know the code has been read. 17:25.920 --> 17:28.920 If somebody comes in ads or removes an apostrophe somewhere, 17:28.920 --> 17:30.920 that has been looked at in the last 10 years. 17:30.920 --> 17:34.920 And for a large part of our code base, it's four years old and no one's looked at it. 17:34.920 --> 17:35.920 Sometimes. 17:35.920 --> 17:40.920 But they also find small changes and fixes and it's just this constant cycle. 17:40.920 --> 17:45.920 We don't get, you know, four major releases a year the way the Linux kernel does, 17:45.920 --> 17:47.920 where you have to figure out what all the features are. 17:47.920 --> 17:53.920 We don't see stuff coming as much necessarily because no one is doing media coverage on fabricator 17:53.920 --> 17:54.920 and stuff coming in. 17:54.920 --> 17:58.920 I don't think the Fruity developers would want media coverage on stuff in fabricator 17:58.920 --> 18:00.920 in the way they do in the Linux kernel. 18:00.920 --> 18:03.920 But we get to see the, you know, from looking at this, 18:03.920 --> 18:07.920 you get to see the continuous work that goes through and how much better things get. 18:07.920 --> 18:11.920 You know, BF is dying is a very old meme now. 18:11.920 --> 18:13.920 But it's definitely not dying if you look at the commit log. 18:13.920 --> 18:16.920 It's just continually moving forwards. 18:16.920 --> 18:21.920 Doing the selection process to bucket things into networking or I don't care. 18:21.920 --> 18:22.920 Highlight at a ton of stuff. 18:22.920 --> 18:24.920 I really cared about which wasn't networking. 18:24.920 --> 18:29.920 So that, um, truncated post by Kirk, where, you know, 18:29.920 --> 18:32.920 he's doing a fix for UFS one file systems. 18:32.920 --> 18:36.920 So the date stamps will work in 2038, which is Kirk's, um, 18:36.920 --> 18:39.920 84th birthday, right, a card now. 18:39.920 --> 18:44.920 Um, and I don't know if I would necessarily have seen that, 18:44.920 --> 18:46.920 I would even like that to a big commit message move on. 18:46.920 --> 18:50.920 But instead I read it, and it's really nice just to see that come through. 18:50.920 --> 18:53.920 There's, there's great value there. 18:53.920 --> 18:56.920 So I said I wanted to be more free BSD content. 18:56.920 --> 18:57.920 There's a podcast to fill. 18:57.920 --> 18:59.920 Ari needs release notes. 18:59.920 --> 19:01.920 So people need to write more stuff. 19:01.920 --> 19:04.920 And so I want to ask everyone here to write more. 19:04.920 --> 19:07.920 You should just write like anything. 19:07.920 --> 19:09.920 The scale doesn't matter. 19:09.920 --> 19:11.920 It maybe needs to be more than one sentence to have value, 19:11.920 --> 19:13.920 but it might be a really funny sentence. 19:13.920 --> 19:16.920 Like you might really hit the mark there. 19:16.920 --> 19:18.920 You should write where you can first. 19:18.920 --> 19:21.920 And we can figure out other places for you to write in the future. 19:21.920 --> 19:24.920 Um, I'm on the editor of the free BSD journal, 19:24.920 --> 19:26.920 which is a magazine about free BSD. 19:26.920 --> 19:31.920 And we, we moved to quarterly, but we're always after 19:31.920 --> 19:32.920 contributors. 19:32.920 --> 19:35.920 And we want new contributors as much as we want continuing contributors. 19:35.920 --> 19:36.920 Because it's quite a burden. 19:36.920 --> 19:40.920 We ask for people to write, you know, reasonably long technical articles. 19:40.920 --> 19:44.920 Um, but yeah, if you want to write more and you don't have a venue, 19:44.920 --> 19:45.920 I'll help you find one. 19:45.920 --> 19:48.920 Sign up at blog easy, but I can also help you with that. 19:48.920 --> 19:49.920 You can also write for the foundation. 19:49.920 --> 19:51.920 I don't think there's any organization in the world, 19:51.920 --> 19:54.920 which should say they've got too many good technical writers. 19:54.920 --> 19:58.920 Most places say it's impossible to find a technical writer who can deliver stuff. 19:58.920 --> 20:01.920 And so you should write stuff to get to the point where you could be a good technical writer. 20:01.920 --> 20:03.920 It'll make everything you do better. 20:03.920 --> 20:04.920 And you should write for other places. 20:04.920 --> 20:07.920 Uh, I spoke to the journalists from a friend to the next magazine today, 20:07.920 --> 20:09.920 which I didn't know existed. 20:09.920 --> 20:11.920 How cool is that? 20:11.920 --> 20:12.920 So you need to write for other places. 20:12.920 --> 20:13.920 And there is. 20:13.920 --> 20:18.920 Um, the VPP work I did, which was funded by Arginance in the free BSD foundation, 20:18.920 --> 20:22.920 we pushed really hard to get into as many venues as possible. 20:22.920 --> 20:25.920 And this led to someone a conference in December saying to me, 20:25.920 --> 20:28.920 I know your name because of the VPP work. 20:28.920 --> 20:30.920 They did not have any context for free BSD, 20:30.920 --> 20:31.920 but we got the word out of there. 20:31.920 --> 20:34.920 And we spread a lot more awareness about the project. 20:34.920 --> 20:35.920 That's all I have. 20:35.920 --> 20:36.920 Thank you very much for listening. 20:36.920 --> 20:38.920 Please go write stuff. 20:38.920 --> 20:40.920 Thank you. 20:44.920 --> 20:45.920 Five minutes. 20:49.920 --> 20:50.920 Just a comment. 20:50.920 --> 20:51.920 I listen to this. 20:51.920 --> 20:56.920 I think we, and in that BSD, we should do a lot more of this. 20:56.920 --> 21:02.920 Because I feel like we're suffering from lower mind shared and say free BSD. 21:02.920 --> 21:04.920 And part of this is that nobody writes. 21:04.920 --> 21:05.920 Yeah. 21:05.920 --> 21:06.920 Yeah. 21:06.920 --> 21:09.920 So the comment was, in that BSD, we should do more of this. 21:09.920 --> 21:12.920 We're suffering from lower mind shared than other projects. 21:12.920 --> 21:13.920 And I think you're right. 21:13.920 --> 21:18.920 On deadly.org is, and the artwork are the best bits of advertising in the BSD world. 21:18.920 --> 21:20.920 And it's got this massive cliché. 21:20.920 --> 21:23.920 Well, outside of his user's base, compared to free BSD. 21:23.920 --> 21:26.920 And it's because they are frank about what they do. 21:26.920 --> 21:29.920 There are bit means people and they have great artwork. 21:29.920 --> 21:32.920 And they do really well just stimulating interest. 21:32.920 --> 21:34.920 Yeah. 21:34.920 --> 21:36.920 Well, the songs have been cringed. 21:36.920 --> 21:49.920 I know your name from your BSD working. 21:49.920 --> 21:50.920 Right. 21:50.920 --> 21:53.920 And I refer to the voice. 21:53.920 --> 21:56.920 Okay. 21:56.920 --> 21:57.920 Thank you. 21:57.920 --> 22:07.920 Thank you.