 waiting for our fearless leader Matthew Miller. Thank you. Can everybody hear me without a microphone? Awesome. No? Awesome. I will try and speak somewhat loudly. Alright, this is my state of Fedora talk. How many people here were at DevConf in February? Okay, so much of the room. So some of the numbers maybe repeats for you. I've tried to update a lot of the data, but this is essentially the ongoing state of Fedora talk that I do, and some of it may seem very familiar to you. Although now we are in crack-out. I'm also curious, for whom is this the first flock? Wow, awesome. Welcome everybody. Someone this is concerned with getting new contributors into Fedora, so I'm glad to see that you're here. So I like to start out with some press quotes, and although we had good press for Fedora 24, none of them had glowing quotes to pull here, so here are my glowing quotes from Fedora 22 and Fedora 23, and both of these are of no, because especially the register here is a very snarky trade magazine that Tenz didn't have almost always find something negative to say, but in their review here of us that basically is glowing, so I am going to keep highlighting this. I think it's pretty exciting. And then the main part of my presentation here is basically a bunch of numbers, and a lot of these are pulled together from our mirroring data. We don't do any intrusive monitoring of Fedora users at all. There's no tracking cookie or anything that follows you around and makes sure that you're counted, which means that we're kind of doing like a wildlife sampling and trying to figure out what's going on from looking at binoculars from a distance. One of the ways we do that is by looking at connections to our mirror server, and so I promised Steve Smudgeon that when I talk about these numbers, I preface them with the scary dinosaur. It doesn't look so scary, but a whole herd of them will just get you. The important things to know are that, again, because we're not tracking, there's a lot of differences in networks that will really influence how these results look like. Network address translation means that a lot of different systems can be counted only as one system, and changing IP addresses. It happens that a lot of ISPs change your IP address multiple times a day. That might be over counted, and there's a lot of other things that could cause confusion, but I think the general trends can be basically counted on. This is a slightly confusing chart, but this is basically those connections for the last five releases here, going from the green Fedora 20 this left over to Fedora 24 rising there. I guess the important thing to see is we've had a general upward trend of each release being more popular than the last, and we've got a very nice quick start to Fedora 24 with it already being more popular than 22, and I think by the end of the summer it's going to be the release that most Fedora users are running. I think that is very nice. When we look at the absolute numbers, which I don't have a slide for, there's actually a leveling off, which I'm a little concerned with, but I'm hoping it's a summer downturn. We'll see what happens at this talk next year when I have a better handle on what those numbers look like, but right now things are very positive looking. So everybody who worked on this release, the last couple of releases, you should give yourself a hand. Good job. So this is the other metric that people often ask about, and that's ISO download. So this is basically just a raw count over time of every day, how many times people are downloading the release. And I think the interesting thing about this is after a big peak, when we have a first release, it's basically just a constant amount over there. So the number of downloads happens to basically correspond to how long that release lasts. So 21 and 22, there's a total of about 800,000 downloads, and it's like a million for Ecuador 23 because it was a longer cycle there. And if you look back at the previous two releases, which were about a year cycle each, those were like 1.2 million each. So it's kind of a constant thing. Although there's a spike, we don't necessarily have a big uptick just because we have a new release. So I think that's an interesting thing to think about as we're doing marketing, and a little bit more marketing Fedora as a whole rather than marketing just what's in the latest release. Another thing you're probably noticing here is that there's a definite downward trend here, and that actually also comes off to the left. That's higher as well where it's cut off. I believe that this is because we've made upgrades so much easier. So there's no need to download an ISO in order to go to the next release. On the other hand, I would love to see this going up despite that. So there it is. And this is a breakdown of the different Fedora editions and what percentage of the downloads are represented there. So this is downloads, again, not connections. And I have also a slide with all the different spins included here, so which gets a little bit crazy. So I'll look first at this one with just the main editions that we promote. And obviously you can see Workstation is about 70% of the downloads there, but we've got a pretty solid download set for server for blah, blah, blah reasons. This actually ends in January and has not been updated for the last half a year, but it looks to me like the trend basically continues like this. One of the interesting things is this gray bit there. We didn't have a network installer for Workstation earlier. And when we introduced that, it instantly became very popular. And you can see it kind of takes a little bit of a chunk out of the server installs because people were using the server in that ISO in order to get a Workstation install and once it was available, they stopped. Can I interject for a second? Yeah. It doesn't look like it actually took a bite out of the server in that install. Do you actually have that as a separate band? We'll say it's a separate bite out of the server, but not in that install. Right. Yeah, people were using the server. Could you repeat the question? Yeah, it didn't take a bite out. People didn't stop using the server in that install very much. It's true. But yeah, that's a good point. I would love to know why that happened. Yeah, it looks like the net, I attributed it to people switching from the server net install, but the point is that actually that doesn't decrease. It's just that the net install increases. I don't know. I don't have a good answer to that. The other thing that I think is very interesting here, as you can see, Atomic is this tiny green band around there. And those are very small numbers. But again, these are downloads of basically the install media, or in the case of that, it's the launchable media. And because these are cloud instances, it's really hard to tell the difference between one person downloaded and ran it one time, or they downloaded it and didn't run it at all, or they downloaded it and ran it a million times. And it happens I had a conversation with a large internet company that you have heard of at Red Hat Summit. And they came up and were asking me some questions about Fedora Atomic. And then they said that they've been actually running it in production for about six months. And I said, oh, that's interesting at what kind of scale. And they said, oh, at all the scale. And this is a company where all the scale is, it may be the most popular Fedora running at all right now. So the little line doesn't necessarily tell the whole story. I'm actually surprised that we have that many cloud downloads. Because it seems like most people getting the cloud edition would actually get it directly from Amazon or Digital Ocean. So one of the reasons we have a lot of cloud downloads, first of all, there's people running it in OpenStack. And the other thing is that this is our only vagrant image we currently produce. So people who are looking for a vagrant image are directed to the cloud image. So that is the bulk of those cloud download images. And actually, I think we should make those for the other editions as well, but we haven't currently. So that's what's going on there. You probably can't hear me so well if I turn this way. So I'm going to try and stop doing that. And I'll try and look at my slides down here. But some of these little lines are very hard to see on the small screen. So this is all of the different spins there. You can see although the individual other spins are never hugely popular, although with KDE being the largest at about a 5% band across there, the aggregate, the net results of all these different things does add up to a lot of various interests. So I think that's kind of a cool thing that we have. I think the breakdown here was 4% KDE and then LXDE and XFCE are about 2.5% each. But I know that those are a very passionate 2.5%, so that's a good thing. Okay, so I also have some external numbers which I think are very interesting as well. So one of the main target for Fedora Workstation is software developers and it happens that the site Stack Exchange, who's familiar with Stack Overflow? All right, awesome, everybody. Right, who's not familiar with this? Yes, awesome. So QA site targeted at developers. They do an annual survey and one of the questions they ask is, what is your preferred desktop operating system? And so this year their survey results came out and it's pretty cool. It's good for Linux at 21 something percent and that's actually a gain from the upper percent from last year. Windows is slowly creeping down to no longer be the majority, just a plurality. OSX is also growing there. So this is cool for Linux. We break that down. This is from a year ago, the breakdown by which Linux distributions are there. Fedora is in the top four, that is to say we are the fourth of the top four, but we get a name showing at least. Now, and there's some reasons that this, it's not a huge surprise. Fedora has had a reputation of being harder to use, fast moving and you can't get your media there, but we would really like to see that growing. And so this year's survey, I would love to say, wow, look, we doubled it. We are up a little bit, but it's kind of noise. So I feel like this is a part where we really have some work to do in the marketing and outreach to software developers if we're going to keep pursuing the strategy of trying to grow in software developer space. And it's not necessarily that we want to compete with the other Linuxes there. This isn't percentage of Linux. This is a percentage of everything. If Ubuntu has 12 percent and we have 20 percent, I'll be perfectly happy. So, but I would really like to see us see that number going up in the future. And it's actually, there's a lot of questions in the survey and they have a full data anonymized data dump of it. So if there's interesting questions we can have about what other technologies people like, what areas we can grow in, I think analyzing that data might be a fun project for somebody. I was starting to do it, but then I realized that would be like a PhD thesis for me and I have other things I should be doing instead. So, but if anybody else wants to do a project like that, come talk to me. All right. So those were numbers basically about the operating system itself, but Fedora is obviously you people more than it is just the downloads. So I tried to answer some questions about this. And if you saw the DevCom talk, you saw some of these things about this. These numbers are not updated from there, but some of the other numbers coming up will be here. So as people probably know, we have a message bus where basically a lot of the activity you do in Fedora causes something to be generated that can be responded to by other applications or other people. And basically most of our online applications are hooked up to this. And so when I wanted to count like what are the active contributors in Fedora, I said, okay, we'll take a look at that and see what we can see. And I particularly looked at certain areas which are easy to count, ones where an actual human being caused the event and rather than it wasn't a bot that just mentioned that person or some sort of event related to a person that wasn't due to their action. And so the three areas that I kind of narrowed down from that being possible are first editing the Wiki page, which is straightforward. And, but it's important, everybody here knows our Wiki is not like the Arch Wiki where it's dedicated just to documentation. This doesn't mean, this section here doesn't represent docs writers. It represents people using it as a workspace in all sorts of different areas, QA and ambassadors and everything else where the Wiki is kind of our workspace. The other one is Bode, which is basically somebody gave feedback on an update. And here I'm only counting logged in users with your fast name, so there's anonymous updates too, but those aren't getting counted. And then finally, in this littler circle here, that's packages who basically did at least one commit to just get made an update to a package there. And so I think it's interesting that there's a lot of non-overlap between those things. And I think this means probably that because these are three random things I could sample, I know there's a lot of areas like translations, which aren't counted here and a lot of ambassador activity. So when I say there's 2,000 plus contributors here, that's 2,000 plus in the easily measured area. So I think we probably have, I'd probably twice that in terms of active contributors in the year to Fedora. And I happen to do this on a year basis, so I'll probably update these numbers for the next DevConf and we'll see how we've got changes here. I would like to measure more things. I know Sonata, the translation thing, recently got hooked up, but it does not generate messages with usernames that I can easily measure here, so that might be some Fed message hacking to get more things hooked up for me to count. If you have an area you think I should be counting, that's the place to start. And I also took basically the top 10% here as a different presentation, and it kind of has the same overall picture ratio of people there. And I'll talk a little more about that top 10% contributors in a little bit. But one thing I wanted to mention here, if you have not seen this slide before, this is one of my favorite ones, is I keep getting asked by people basically inside and outside Red Hat, how much of the work on Fedora is being done just by Red Hat? Is this really a Red Hat thing worth one or two external contributors? Or is this, yeah. So I analyzed that top 10%, which basically is the people who do about two-thirds of the work. And I did not do this with a script because I could knock some of them off easily, but some of them I basically went through name by name and identified if that person is a Red Hatter or not. Because it turns out that there's this blue area here, which I have to explain myself. It's not actually sneaky. If you look there, you'll find that I am actually Matt Diem at Matt Diem.org, at FedoraProject.org. And that's because I had all these filters set up for my email before I started working for Red Hat. And they filtered out my Fedora email very nicely, and I didn't want to redo that. So there's a lot of people who are Red Headers now, but were visible in the community before it came to Red Hat. And that's generally what this 9% here represents. Not really people trying to, not proud of being a Red Hatter. But the basic number here is really we've got two external contributors in the core of people doing most of the work in Fedora for every one Red Hatter, which I think is a pretty awesome area. Now I have geeky breakdowns of where this came from. So this is basically a chart of the messages of this. This is the Bodhi feedback here. Over time for the past three years here, two and a half years. And basically this is the number of people per week. And this is the 1%, and next 10%, and so on there. Basically, and I put these on the same scale here, so you can kind of see. So there are a lot more people who actually make changes per week, even though the number of actual people involved overall in Bodhi was higher. There's just a lot of changes to packages all the time, as we all probably know. And there's the Wiki edits as well. And here's the part where I got to the two thirds number. And this basically holds true across all of the different things here. Where basically of the people who are in like a rolling window of the people who did, who are responsible for the top percentage of the messages. You can basically again see that the top 1% here does somewhere around 20% of the work. So there are a lot of people who are very, very busy. And I'm sure you will be able to guess a lot of those names. If you think of somebody you see all the time doing everything, it's one of those people. But this next 10% ends up doing basically about two thirds of the work, which I actually think is a pretty good ratio. When I started to look at this, I was afraid we'd see something more like 9010 or even 8020. So I think that's pretty good. I mean, it'd be kind of nice to see if we could expand the long tail and have more of the work done by thousands of other people. But it's still the 300 people doing that. There is still a pretty big number, so that's a good core group. So there's some interesting spikes here. These big spikes here are packaged, the mass rebuilds, where basically somebody in release engineering rebuilt everything. So they get credit for that. That is being done by a relinch user now. This data set counts the relinch user as a human, even though I didn't count it in that 300, but I actually put it in my blacklist of bots for the next time I run this and I tried to rerun it and then I broke data grabber. So that's not updated to show that, but that's what's going on there. And there's one other one that has a funny spike here, these dropouts here. What you're seeing here is Christmas vacation, Christmas vacation, Christmas vacation there every year, where people tend to do a lot less work across all of the board there. You can see in the packages here, particularly the casual contributors drop out during Christmas, which makes sense. If you're making an update to a package once a year, the odds of you doing it on your vacation, yeah, it's smaller. Right, exactly, but people, I was going to note, the people down here, Peter is working 24 seven and Dennis and a lot of other people update things no matter what and so this one is basically a broken down percentage by the colors are different here. This is basically the green are actions by people who are new in this week. The yellow is people who are new in this month, but not the week. And the green or the red is people who are new this year at least. And then the blue is people who have been active for more than a year. So we've got a constant influx of new people here all the time. This is in the Bode feedback here, but a pretty large majority of this is done by people who have been around for at least a year. And in the package, see, this is the release engineering user, by the way, when it was created that caused that spike. In general though, the packaging in particular like 90% of the packaging is done by people who have been around for more of a year, which that's kind of a hard technical skill. So it's not completely surprising. And I should mention that been around for means mentioned in this data set. So people could have been Fedora contributors for five years, but it's the activity by someone who started that year that is counted here. So it would be kind of nice to get some of them, get more new users into the packaging group in particular. And Wiki, as you might expect to see, has a lower barrier to entry and more newer users there. And we should have a special, shout out to Patrick. Where's Patrick here? I can't pronounce his last name because I think that nobody, I don't think he can pronounce his own last name. Yeah, there, good. That's right, because as close as I can get, and I know it's horribly wrong. So we've been under attack by spammers in the Wiki. And so you may know that if you are not in at least one additional Fedora group of some sort, you cannot edit the Wiki right now, which is kind of unfortunate. Although, because it is sort of a group workspace, it doesn't really impede new contributors to Fedora as much as it might if it were really meant to be an easy to get to resource. But I think it's kind of a problem, and I don't know a good way around it. But we've done a good job of at least keeping that down to zero for the past couple weeks, or at least zero that are visible to me. OK, and I'm going to try and not use all my time on numbers here. But I also wanted to mention the Fedora magazine. This is also something that's been very successful. And I think that these are just basically page youth for months. And the important thing is we've got a nice upward trend here. The red dots are when there was a Fedora release that month. So you can see those are obviously going to be peaks. But this is a general trend in growth as well. And there's been, if you've been following the content, a lot of growth in really good solid articles that are really useful to people there. So I'm really excited for and proud of and thankful to you. Everybody's been working on Fedora magazine. So that's the end of my numbers section. I'll have some room for questions at the end, I think. But we'll go ahead and see how this goes. Last time at DEF CONF, I had about three minutes to rush through the big goals section. So I'll try and give myself a little more time here. The important thing is that I do not set the goals for Fedora. I have a name leader in my title. But really what I try to do, and what we as Fedora leadership in the Fedora Council try to do, is discern the collective goals of basically everyone here in this room and everyone who couldn't be here today in the Fedora contributor community. So, yeah, it's you guys who set the goals. And actually as I was going through the things, I was like, okay, yeah, that's important. I should mention that. I realized that most of these correspond to talks that are happening at this conference, which means I think things are going exactly as they should be. That's what this conference is for. And it means that these aren't just ideas that I pulled out of my head, but they're ideas that we're all working together on. The first one is basically that the release train is gonna keep going. We're gonna have Fedora 25 released in October-ish, November, that kind of thing. And we really want to stick to this idea of having an October release and a May release every year so that you can know that that's gonna be the cadence of what Fedora does. And if we happen to slip in one release, we'll make the next release short to make up for it. This is kind of a new policy in the last couple of releases even though it's sort of not only been the policy, but really trying to stick to it as a new experiment we're doing. I think it's gonna be okay. This particular release is kind of rushed because that's the consequences of doing that. So we'll see how that comes out. But I, and I think probably most people here know what the basic features are going in this release. If this were a talk to a general audience, I would probably dwell on those a bit more. There are some talks about the release process. Dennis Gilmore has a talk about getting things into Fedora and Ralph Bean has a talk about a thing called Factory 2.0 which is basically upgrading this to a diesel engine of some sort or maybe like an electric locomotive. It's gonna be awesome. What's that? Maglev, Maglev awesome. That's why we're going way next generation. I thought it was 2.0. We're going right to 10.0 or something. So one of the really big things here I think is this idea called modularity. And this is something that started at Flock four years ago or maybe a little before that, but I had a talk at Flock there about this idea called Fedora rings. And I had, I think this Lego slide was in that talk and it's been carried forward ever since because everybody loves Lego. And at DevConf I basically said this should be the year where we can do, it stopped talking about this so much and actually get to show and tell. And I'm really excited Langdon White is going to be actually doing a show and tell presentation. So if you're interested in this, how we're gonna put the distribution together in the future, I think you should go to that talk. It's gonna be very exciting. And it's actually not just abstract concepts but actual working demo. It's pretty cool. And a lot of the thing is the idea that a module is something that's not an RPM, it's bigger than an RPM, but you can actually put it together in a workflow that looks very much like what you do to put together a package. So there's a disk it like thing and that kind of same mechanisms that we use for packaging will be used for making a module. And I think the goal is to open it up so it's not just an elite few who make Fedora modules but we have a whole contributor community to making modules just like we have for packages now. Another big thing that I am interested in is a university outreach goal. And this is again going back to new users and maybe attracting the next generation of developers. This is something that has been, we've been talking about for a while but has not quite picked up steam. So I hope that this is the conference where we get that steam going and gets back to the train metaphors of picking up steam. Justin Florey and Yona, is that, hey, print's your name? Are you here? Have a talk about that. And I forget what day that's on. But Justin Florey, today. Today, awesome, there you are. Yeah, talk today about that. So I'm excited to see how that goes as well because I think we really need to make sure we keep getting young people into Fedora or else it's gonna end up being an old people's distribution which is not what we want to be. Another one of our outreach efforts is this Fedora loves Python. Miro Hunchek has a talk about this. This is kind of a marketing idea. Again, we'd said we were gonna try and address the needs of developers. And in talking about that in Fedora marketing earlier this year, we came to the conclusion that developers is a very broad target and that maybe by biting off one smaller bit first and saying, okay, we've got good relationships in the Python community. We've got good Python tooling in Fedora. Let's try and market specifically to Python developers. So we went to PyCon and handed out awesome Fedora Python t-shirts and talked to a lot of people there. So there's an ongoing effort to specifically attract some of those Python developers. Kind of to show if we try to address these developers, here's an area where we can have growth and then we can broaden that out bringing in the Java developers and the .NET developers because I talked to Microsoft at Red Hat Summit and they were very interested about how they could use the Fedora packaging guidelines to correctly get .NET packaged up for Fedora. I said, wow, and they were very serious about it. So let's see how that goes. But if we're now Python. So another big thing I've been talking about for, yes, Adam? Yeah, .NET is an open source thing. We're not gonna include some closed source .NET in Fedora. Don't worry. .NET, Microsoft seems to have a change of heart and at least from the engineering level, they seem really sincere about it. So that's awesome. Langdon? Also called a change of CEO. A change of CEO also helps. Yeah, exactly. Hubs is something I've been talking about for a long time. I didn't have the IRC meeting slide on here. But as everybody here knows, the core of Fedora activity is on IRC and mailing lists which are awesome and useful tools but also completely invisible to people outside of a very nerdy subset of the world. They're a hard barrier to get into. IRC culture, let alone setting up an IRC client and registering a NIC and figuring out what a NIC serve is and a channel and all of these things. Those are difficult mailing lists or something that basically people don't want to subscribe to anymore. And even as something other than a barrier to getting involved in Fedora, it's something where you look at the Fedora website and you can't tell that anything is going on. When we really have thousands, literally of IRC meetings every year and multiple meetings a day and so much lines of chat and so much activity happening, people can't see it and that makes it easy to dismiss Fedora as a vital ongoing project. So this Hubs idea is basically to bring all that onto the web which is what most people think of when they think of the internet today. So we want to make sure we are visible on the internet. So there is a talk on Hyperkitty and there's a talk about bringing IRC to the hubs and then more important than that, there is actually a hack fast to actually work on this on Friday. So if that's interesting to you, go to that. Next, Fedora Atomic. We've been doing this kind of cool thing where instead of waiting every six months to have a release, we've had a every two weeks give or take a release of the latest version of the Fedora Atomic host with updates. As I said before, large company is relying on us for this, so that's cool. But more than that, I really think that this idea of containerized bits of the operating system are probably the future of the OS from the server to the desktop and we see operating systems like CoreOS and RancherOS in the bleeding edge there and I often talk about Fedora not wanting to be the bleeding edge, but we definitely wanna be the leading edge. We wanna make sure that when those, well it's okay to let other people experiment with the crazy ideas, but when they start looking like, hey, that crazy idea might be onto something, we should be in the forefront of the experimentation as well. And so Atomic is basically our way of getting involved in that. And I think that as time goes on and especially in connection with the modularity work, this is really gonna be core to how we put together Fedora. And so I think this is a really big deal and one of the parts of this particular is looking at OpenShift, which is not just a single host cluster system, but a container system, but a clustered one. And so Josh Burkus here has a hackfest about that. Is that also on Friday? Thursday, thank you. Thursday, a hackfest about that, which is important. And we've got a lot of other talks about Atomic in general and containers. One of them is the Docker layered image build service that Adam Miller put together and with a lot of other people's help as well. And that service basically, like I said, with the modules, you'll be able to sign up to be the maintainer of a Docker file, a Docker container, just like you can sign up to be the maintainer of a package and you can maintain the container right now composed out of RPMs from the RPM set. And it uses the same sort of the exact same tools, builds and co-g, the same way that RPMs built. So I think that's really exciting. 10 minute warning, okay. Yeah, so the next thing, talking about the similar kind of technologies, Flatpak is a method for distributing and running desktop applications with the eventual goal of running them in a secure sandbox way. So you can download random applications from somewhere and trust that they won't take over your system. It's pretty cool technology coming from the free desktop.org people and there's a big push on that in Fedora Workstation. And I don't know if we have any specific talks on that but there is a talk about the future of Fedora Workstation which I'm sure will mention Flatpak as well. And we also have an ongoing conversation in the Fedora Council about allowing people outside of Fedora to make Flatpaks available to Fedora users in the software center in Fedora Workstation and how we will handle that kind of thing because availability of applications is one of the things that comes back to people, comes back to us a lot when we were like, why did you choose something other than Fedora? So that's one of the interesting technologies and that is the end of my highlights there. Although I'm sure that there are lots of other exciting highlights in this conference. I didn't mention everything. Those are just some of the ones that really stuck out to me. I wanna thank the people who helped me put together the statistics at the beginning. Steve Smudgen and Ralph Bean. And also just thanks to everybody for being so awesome. Fedora would not be what it is without you so thank you very much. And I guess we have like eight, no, Joe, go ahead. Yes, we have about eight minutes for questions or between talks. I forgot one very, very important thing earlier. We need help recording and transcribing talks. So, folks who have not yet volunteered, I created a Bitly link to make it easier to get here. We have a wiki page for people to sign up to start the recordings. Time in the talks as somebody has been with here and also we need help transcribing so that you're a past type or into IRC rooms for each talk. You go to bitly slash splot 16 help all over case. That will take you to the wiki page we're gonna sign up. Thanks. Awesome, okay, any questions, anybody? Yes, so I found your statistic about the new contributors who are packaging, pretty interesting but I think it would be even more interesting if Copper were in there because there's a lower barrier to entry for Copper and it would be curious to see that they're the same statistic for both ODE and Copper. Yes, so the question there was, not a question at all but a comment to me which was that Copper, including Copper, would be good and I think Copper actually generates these messages with user type to them so I'm not sure why I didn't include it so I will definitely do that once I've unbroken data gripper. Thank you. Yes, Josh, and not Josh, Jason. Do you have any stats on packages that they can all win for ODE without getting any carbon? Yeah, that would be interesting. Yeah. Basically, never get more than one carbon but that doesn't mean that my package usually has a barrier in there. Yeah, I do not have those stats off hand. The question was, do we have stats on packages that do not ever get that feedback in BODY and make it through without anybody ever looking at them? Does somebody from QA have the answer to that off hand? I know that, I know that. Yeah. That may not be a slide you want to show us. Yeah, I am happy to show the slides that don't look so good. Yeah, those are what we call things to work on slides so that's, yeah, I don't know but that's a good, interesting question. Yeah. Do you have any statistics, the last year in change, we've been pushing on the university and the diversity initiatives. Have you had any kind of statistics on whether or not that has had any effects? Yeah, so the question is, do we have any statistics on the diversity and university initiatives? The answer is no, I think. I would... No, no, no data or effects? No data. I hope that we've been having some effect. The university thing has been kind of slow but I think we've been having, there's a Fedora Women's Day which was, I think, very successful. So hopefully that is having an effect. And I know that Tadaka is our diversity advisor is working on a survey and that once we start having that survey that will give us some sort of measurable things to do but we don't actually have that yet. Five minutes. Yeah. So, regarding the number of contributors from the local organization. For 24, we ran the spring, it was 50 participants. Okay. I think we have already seen the post for Tadaka and the fragmented implications. Okay, so the comment was there were about 50 people who worked on translations over the last, for Fedora 24, is that just in the sprint or was that over the whole release? Whole. Okay, so about 50 people working in the translation. Do you know what percent, did that break down into, like, again, 10% of people doing most of the work who probably follows that same pattern as well. But yeah, I'd definitely be interested in adding that. Yeah. Do you know how many statistics how many groups do it, whether we install it, send it, and might they upgrade to all kinds of things? Do we have statistics about what upgrade method they use? No, I do not. But I can tell that it's, again, because the ISO downloads are going down and the number of people running the new releases are going up, some good chunk of people are choosing to upgrade. I think there's some ways we can pull that out of the mirror data, but it's a little bit hard because you definitely can't tell easily the difference between somebody doing a DNF upgrade that pulled down a lot of packages and an actual update without doing deeper analysis that would involve keeping more logs than ever keeping. I have some ideas. Okay, yeah, let's talk about that later. There are definitely some things we can do and I think that when people do an upgrade using the graphical upgrade, I think that we've got a way to measure that, but that just started, so I haven't been able to do that. Justin. Is there a way that we have, are there any statistics available for measuring activity per continent or per region of the world? Yeah. So the question is activity per region of the world. Do you mean contributor activity or user activity? We have information on where contributors are from, so that could be analyzed, but I have not really done that, but yeah, we actually do, people put in where they're from when they sign up for an account, so we could go and look at that and that would be a good thing to do. We've tried to do some analysis of where people are coming from in the mirror manager stats and that's actually a thing that is, one of those things that ends up being a dinosaur, the stats are heavily skewed by those two users in connect areas where there are reliable always on internet connections. So people in Latin America, people in remote parts of Asia, and definitely people in Africa, are being under counted there. People who have high speed internet in Europe and North America are over counted. So if you look at the breakdown there, it basically tells you that's the case, so it doesn't really, but for contributors we have more granular data that we could look at. One more question. You already asked a question. Is there anything else? Yes. How's it going on on the tech message that's in charge of privacy interaction? Yeah, so the question is, have we thought about privacy in the broadcast with all this stuff going on Fed message? A lot of the stuff, yeah, so a lot of the reasons we don't have more invasive data is less you being careful about privacy. But the Fed message stuff where your name is attached to a thing you did, we have, I know you raised that issue recently, we have really not looked into that very much before. I really find the data valuable, but I also understand the importance of not tracking individuals, like particular location and time of action. So it's possible that we could look at anonymizing older data or something that would make that, like even just knocking things down to actions per day instead of actually per minute kind of thing would probably make a lot of additional privacy. All right, thank you everybody for coming and have a wonderful flock.