 between 1st and 3rd. So, the most important information, if you haven't figured it out, the Wi-Fi password is block 15, all lowercase. Most important info. So, after Matt's talk this morning, what we're going to do is everybody needs to get out of here because they're going to divide this room so that this will be the main talk room. This is what's red hat or the main room on your schedule. And then we have rooms 1, 2, 3 there and 4, 5 are in the hallway. So, some of you, when you register this morning, or if you haven't registered you should go do that later, got a Jaguar board in your bag. If you didn't, and the guy next to you did, very sorry, Jaguar sponsored 30 boards for us. And so, 30 of you are the lucky winners of Jaguar boards today. And I see somebody like, yay, that was me. This morning at, what? Yes, if you don't want it, I am sure that you'll have no trouble finding a new friend today who would be happy to take it off your hands. We will have coffee at 10.30 this morning, lunch this afternoon, and then coffee sponsored by our late this afternoon, cupcakes sponsored by Rackspace. Tonight there's no sponsored evening event. You're on your own to go get dinner, but we will have games with beer and wine in the same place that we have lunch. You'll have drink tickets. So, come play games. We have a bunch of board games. I know other people brought some more games. We may try to run a trivia thing. It'll be awesome. And then the last thing I want to do is thank our lovely sponsors who helped make all of this happen a lot. And that's Wallspot, Arm, Bluehost, Google, O'Reilly. If you look in your booklets, there's a free O'Reilly book to download, open NMS, Rackspace, and Snow. And with that, I will hand it over to Matt, who will take it away. Hi, everybody. Thank you, Ruth. I'm not used to yelling, except for my children. So, if you can't hear me, raise your hand, and I will try and yell more. If you can't hear me, that's very nice. Okay, so, yeah. This is the State of Fedora talk. That's a state picture. There will not actually be a lot of funds here. The actual State of Fedora is awesome. Thanks to Ruth for the success of Panda picture. We are doing very well as a project. And it's thanks to all of you, so you should give yourselves a round of applause, I guess. I think we're very healthy as a project, and I've got to spend some time going through some numbers here to kind of show some of the exciting things happening recently. So, Smooch put together some stats for me, and as always, I promise to warn that these stats are actually dangerous and misinterpreted, because a lot of them are pulled from server logs in ways that don't necessarily support a lot of the conclusions you can easily draw from them. And so, I'll try and give a little bit of detail to some of the caveats as they go along. So, the first slide. This is exciting, like, really extra hard to see on the board here, especially because it's great yellow over there. So, it's basically completely lost information. So, let's see if I've got this. So, this is basically a graph of young connections for per release. So, basically, every, it's a triad of every IP address checking in per day. So, they're not counted more than once per day to try and catch systems that are checking in multiple times for some reason. It also means that if a thousand people are checking in from behind one proxy server, it will get counted once. So, there's a lot of slump in there. And, of course, some of these genes don't even check in with their updates. So, this would be counted at all for this statistic. But what you can't see over here on the right side is that the most recent releases, 20 and 21 and 22, are actually basically backed up to the peaks over here on the right side after kind of a period of disconcerting down releases where it should be arranged and maybe, hopefully, growing a little bit as well. This slide is a little more readable. This is a slide of, as of this last month, basically the last 31 days of the number of people who are using a certain release. So, I think this is kind of just an interesting thing to see that most people are on current or latest release, the majority of people, and then good sides of people in the last release. And the number of people who are on legacy releases forever are decreasing. So, that's kind of nice to see because we had people who were kind of clinging on to Fedora 15 for a long time. It's actually back to this one. There's an interesting drop here at Fedora 14. Sorry, it's one of the people we're clinging on to. There's an interesting drop about a year ago where a huge number of Fedora 14 systems went away. We don't know why. Maybe they all decided to upgrade to REL7. Maybe they switched to some other thing. It's a whole number of large numbers once. I think it was probably a host in part or something that was running Fedora. But really, we don't have a lot of information on this. You can also see here that tiny little light sliver there. That's the brave people who are running RAHIDE. So, thanks to everybody who's running RAHIDE. There's even a tiny little line of red that just doesn't show up at all. That's Fedora 23 people who are running Fedora. Again, this is before the alpha release. So, I assume that's huge with the bunch of people who installed the alpha. I came out yesterday. So here, this is another, I would focus more on a lot of these statistics if I can see the graph, but I'll go faster past them. So, sorry. This is another way of looking at the data here. And you can see here the red and green lines here. The red line is the number of people who are checking with young statistics, maybe the green line. Awesome. One of the red and green lines, they correlate very closely. It's the number of people checking in with young, checking for updates. And the other one is the number of people who are hitting the hotspot detection in network manager. Basically, when you connect to the network network manager who tries to fetch this file, if it says okay, it knows that you are on the open internet. If it comes back with some sort of portal page, you know that you're in captive. So that's how if you see network manager put a little question mark up in the bar, that's what's going on there. And so we can have server logs from that. So there's a very good correlation between the young stats and the hotspot stats. And you see that number up there is about 100,000 a day checking into there. So, Smooch says we're about to denial service ourselves with the hotspot file. So we might have to find another solution there. But I thought that was an interesting thing. The blue line across the top is all connections, not just the recent releases. And this goes up like that because that hotspot feature was just added in recent releases. This goes back to 2014, the previous slide back there is on a different scale. It goes back to 2009. Okay. So another way to look at things is ISO downloads. And I really wish I could see the legends here. I couldn't even see it on my screen properly without zooming in. What a pain. This is basically the downloads of the different editions. Matter starts in 19. So this is the release here 19, 20, 21. And again, one of the things I think that interesting takeaway from here is that people basically downloaded the latest release and very, very, very, very, very few people at the bottom go digging up ISOs from the older releases. So basically people are looking for what's current, which I guess makes sense, but it would be useful to know if it were the other way. And it is. So that's good. This is also, this again has one IP per day. There is another version that's much higher. There are some Red Hat servers that download these like hundreds of thousands of times per day or something like that, which was inflating the statistics quite a bit. And it was a little bit disappointing to see them collapse by words of magnitude when you filter those out. But anyways, close the download numbers there. Another thing that is interesting is that the latest releases are kind of smaller than you expected them to be people running. So I think the fact that Fedora 22, when it's small, kind of implies we have a lot of people upgraded online rather than downloading the ISO from the previous release, which I guess is, again, to be expected. So this one has basically here over in the green is the previous Fedora desktop release, and the blue is the DVD. Over here on in the recent side of things, we have Fedora workstation as the purple and Fedora server as the blue and cloud is the red. There's just the additions without the spins. And this is the spins without the additions. I'm going to look at them and this is not a very useful visualization if you can't see it very well. So I have a pie chart, which is hopefully easier to read here. And so this one, I had one of these at DevCon earlier this year. This one I think is a little more rigorous. So last time we didn't put numbers on it. I just had a pie chart that you guessed about this one. I actually feel a little more confident. It still has all these caveats and sketchy IP counting rather than anything else, and particularly downloads don't really show usage in any way. Someone could download it and throw it away, or they could download it and install 10 systems in a lab, or they could download the cloud image and run one million of them. We have no way of knowing from that what those are. But basically the breakdown, Fedora workstation 68% of the downloads and server 14%. And this one, I also, we also included the net installs here. And the breakdown for the net install, which is about 10% of the overall downloads is net install. And that has basically the same ratios for the workstation and server there. So the cloud image comes in at 4%, KDE at 5%, which is actually quite a bit up from the previous numbers. So I think that the new SPINs website is working. People who are interested in trying out the alternate desktops are able to find that and explore that well. So that's kind of, that's cool. LXDE and XFCE about 2.5% each. And sugar and mate, like 1%. And then there's a tiny little group of other things there, which I guess are probably the labs. It seems very, very small. So I don't know if that possibly, they weren't counted right, but I think they're just not very many downloads of those. Those are going to be very special purpose SPINs. Okay. Here is CPU architectures. So we have X8664 and 32 bit going kind of across and decreasing there. You see down in the tiny little corner there, that little bit of red, that's ARM. And actually this is back to young connections, not downloads. So there might have been, again, a lot of ports out there that are not actually connecting for regular updates and things like that. So I think the ARM thing actually reflects probably people using ARM on the desktop rather than using ARM for huge little embedded hack kind of things. And because we're talking about ditching 32 bit Intel architecture, I broke this down a little bit more into a pie visualization for you here. This is basically, you can see the gigantic blue 32 bit six years ago, getting smaller and now down to a much smaller little wedge now. And I don't know if you can see it, but again, you can kind of see a tiny little sliver of ARM there up there at point 15. But basically, I think we are probably right in saying, you know, the trend is 32 bit Intel is going away. And if we want to move that to secondary, we're getting to the time where that's a good thing to do. Yeah, that is a really good question. I have no idea. Yeah, it's interesting. There's no gigantic jump at any point that says, oh, look, we switched the default download release for the for the assault systems. But I don't know what the case is. I think it's 20. Really? It was pretty recent. So for 20, yeah, 20 would be somewhere near the possibly it's, you can see here on the 32 bit, there's actually kind of a chunk down and another chunk down at the end there. Those those chunk downs do correlate releases there. So if Intel architecture didn't exist, this is just the secondary architectures here. Again, the big red spike there is is ARM and there's all of our other other things. I think PowerPC is the other slightly significant one. And I I cannot see the light. So that's RPC. Which one is I am? RPC and which one is the orange? Spark. Yeah, spark. A primitive is still a pretty good spark. So the last thing on my connection stats slides here is this is basically all the young connections that are Fedora operating system. Fedora updates the green line and the red line is Apple. So not only is Apple, so we've known for a while, Apple has exceeded our number of connections sometime in 2012. But it continues to grow quite a bit in popularity at Apple 7 is taking off. So this is an area where we as Fedora have a big impact beyond the Fedora operating system itself. And this is something that people really appreciate and something that is very, very well used. So I guess that's all I have to say about that right now, but it's worth talking about more over the course of Flock here. All right, back to floss wrap here soon to be getting. Like I said, these statistics tell us a lot of interesting things, but a lot of it is kind of conjecture and guesswork. I would really like to have a more accurate census of the number of Fedora machines out there. And I know we talked two years ago at Flock about running census system automatically. And there was kind of not just kind of there was a lot of resistance to that for privacy concerns. And people wanted to have an opt in system. I know we tried an opt in system before was smolt and it was not particularly successful. I would like to suggest that we have an out system that just sends a randomly generated machine ID and the Fedora version and what the issue spin it is with a lot of care taken getting the privacy concerns, right? Maybe not logging IP addresses even like here, here we're logging IP addresses. We can actually reduce the amount of private information we store by getting rid of that and just having a unique ID that we stay there. And those unique IDs can be used for just counting not for tracking so we can rotate them once a month or something like that. So you can't be tracked over time. I don't really care about tracking. It's fun to track the same system over time but not particularly useful from an aggregate point of view. So I would like to over the next release or so figure out a way to do something like that and then also have opt into more elaborate statistics because it really helps us figure out who our audience is and what we're targeting and what we can sort of count when people are using. I know it's sensitive so I'll try and approach that in a really careful way but I want to point that out that we can at least put some of the dinosaurs in the ground. Okay. So switching to some other statistics. This is from Fedora Magazine which Brian Lurch and Chris Roberts and some of the people who have worked on it very hard over the year. This is a statistics over the past year for visitors in the lower inside part of the bar graph and just page views for the top. So the peak month here in May when we had a release we had 100,000 visitors. That's unique visitors not just people reloading every day. 100,000 unique visitors to the magazine that month which I think is a pretty good audience. It pretty much corresponds to the number of people who are making young connections every day as well so I don't think everybody who uses Fedora looked at the magazine that month. I'm not sure. One of the things, talking to Ryan, one of the things that's interesting is that the articles I get the most views are ones that are very user-focused not contributor-focused so we get a lot of attention when we talk about desktop applications and desktop technologies. A lot of love for GNOME on the magazine. We get a couple of comments saying that GNOME is terrible. We used to switch to whatever title window manager or instead but our actual page views and it are very high. People are very interested. I think that's an interesting thing. We are also in the future working on having a separate Fedora contributors blog and the Fedora hubs thing that will have the contributor-focused content there and make the magazine more and more user-focused because that really seems to work and resonate with people. This is more magazine statistics and this is basically all-time views sorted by, I think this is all-time, sorted or maybe just last year or something like that, sorted by the country that people come from and it's an order of magnitude United States people here. I think we know we've got a lot of users in other countries as well so this is one of the ways to look into that. It's one of the things actually back to caveats for the YUM connections. In well-to-be Western nations we have broadband IP and our systems are on all the time and our connections aren't metered hopefully so we tend to have a lot of in the YUM connection statistics are pretty good from the United States and with broadband. Other countries where it's more sketchy connectivity those connections it's more of a mystery what's going on there so this is a little bit now obviously we're kind of English slanted as well there are magazines but in addition to the United States we have a pretty pretty solid viewership from around the world. Okay so this is another kind of social metric that is thanks to Remi for a couple of these slides here. This is basically the number of people who have provided feedback in Bode, the updates system which is basically if a update the throwback is coming out I'm sure everybody knows this record and up things coming out you can go to Bode and say yes this worked for me no this didn't work for me and so this is by year here basically crap looks off from what we're saying before. Anyways I don't know the scale I was expecting but basically we had over the last year about a thousand different people give feedback on packages and that number is growing so I think that's a kind of a nice other metric of the community and so it's we've got a kind of core group of people who are in this room or in you know flock and fun conferences everywhere else but it's nice to see that we've got a really broad base of people we're doing a kind of small level of testing which really helps out a lot and I hope in future this will have some more metrics along with measuring the community there. This is one that I like this is the meetings per month here for the last three years here in IRC so this is IRC meetings not in-person meetings or video meetings or anything else and so we've had last year a 1,066 IRC meetings which is right and that's just that's not people like hanging in IRC chatting that's you know an actual formalized meeting of some sort and had almost 800 so far this year so on track beat last year for the number of IRC meetings so that's a huge amount of activity that's going on in Fedora and I think that's cool for one thing because it shows how vibrant we are but it's also buried in IRC and I think probably again most people here are pretty comfortable with that we live in IRC as a project but this is something when there's a lot of Fedora activity where if you go to the Fedora website you don't see how much is going on every day in making Fedora so I want to look at ways we can bring that to be more visible and again the Fedora hubs with IRC integration is interesting there we've got a new site that the apps team made that presents the meeting logs and kind of a nice visualization there and Bremi has a thing that makes word clouds from these meetings so it's some of the presentations to sort of bring this activity to be more visible I think it's going to be an interesting thing because I think when people you know when you look at Fedora when you look at any sort of open source project it's nice to see how active that project is what's going on is this something where I can where if I you know if I make a bug report yeah I wouldn't say bug report but if I if I want to interact with the project are they going to be there is there somebody around and the answer Fedora is yeah we are definitely around all the time and doing things but if you go to the website in the wiki we've got millions of dead pages of the wiki so a lot bringing a lot of these things to be more visible I think would be a useful thing to do okay uh switching entirely to a different metric here again um this is the copers versus Koji for last year and this year so number of builds and coji number builds and copers and so the point here basically is that copers is we're sure doing a lot more in Koji as well but copers is also even more taking off and Gary Eichman I'm sorry I cannot say that check our thing even a little bit so I'm just going to pretend it's an American are just terrible of me but I had an interesting blog post about how copers is taking off and being very successful and kind of also at the same time that the number of packages in the main store collection has leveled off I was talking to Peter Robinson about that the other night and he made the point that just because it's leveling off doesn't mean that we're not adding new packages we're actually still adding new packages at a big rate we're also retiring older ones so as we get up to about the 20,000 package level in the main collection I think it's reasonable to think about do we actually want that main collection to grow to 40,000 or 60,000 or is it better if we have since other space which has sort of the growing up to 100,000 and then try keep that core to be maybe maybe even smaller but keep the quality higher in that level so that's you know Fedora rings stuff that I've been talking about for a while and so they're kind of ongoing things there okay this ends the random statistics section of this talk does anybody have any random statistics questions answered real quick yeah yeah I think I think it's IRC the question is whether IRC is a measure of engagement and I think it definitely is I guess my concern is that it is also pretty high barrier to entering for people who are we a human with that sorry yeah while that's complicated and while that's still around it's yeah yeah those are those are big concerns and it was really kind of our sort of the heartbeat of Fedora is on IRC and if you're not an IRC you're missing out on a lot of Fedora and that really does concern me a lot because you know it's very functional it works very well I don't want to get rid of it and get rid of what we have there but it is kind of increasingly something that it's hard to bring in new contributors when we have that whole level and it's not just using the technology it's it's a whole IRC culture that goes with it that can also be confusing and it's very itself yeah I was going to comment on that it's also hard to surface that the activity is going on yeah and it's also hard to measure the extent of it yeah there's a Zod bot is our the bot Fedora and it has a thing where Adam can fire people in IRC I had not been counting the number of firings but that is something we could do yeah if you're interested in yeah this is the excellent place up Mo are you leading that talk when is it okay Mo and Megan are leading a talk on Fedora Hums sometime if you're interested in barriers to entry in general IRC and those kind of things go to that talk I know I protected on my schedule 180,000 yeah so and we go above the top number so it's like why was the number jumped up so much I don't know I think it actually might be due to the scratch bills there being no okay so does anybody from release engineering have an idea why that number jumped up so much okay so mass rebuild this year didn't do it last year that might be yeah one more step yeah yeah right so yeah there's a whole bunch of speculation about why we had a big drop in the Fedora 15 numbers and so one thing one possibilities that people were holding on for 14 Intel the monthly desktop came available I'm not sure what the number of downloads we have for the Montes explains the amount of jump there but it possibly could some of those things are things that if we had that number that you know the thing said what edition you're using we could get some better statistics on that kind of thing okay so current current state of things that are 23 alpha came out on time yesterday with a little bit of don't look at the sausage yeah definitely that definitely is deserving of a pause so we are well underway and not only unscheduled for our Halloween release this year so that's that's very exciting I actually I think that everybody here is probably a lot of people here are kind of in the midst of working on it so I'm not going to dig too deeply into the details of what the alpha is except for I want to don't read the actual text here it's blah blah blah one of the things that as someone who talks to the press about Fedora and some of the Fedora marketing team it is nice to have a story for what this release is about so for Fedora 21 we had the first release with Fedora next look at our shiny new editions for Fedora 22 we had a little bit of yeah and then we put some polish on that so it would be nice to have a kind of a story that the press or anybody could kind of grab into you know what what what should I talk about with this release so if you're interested in helping us figure that out join the Fedora marketing group you're just basically saying you know we kind of updated Python to the aversion and there's a new GCC and all this is mildly interesting but it's kind of expected of Fedora so it's nice to have things that can be headline grabbing I'm hoping that some of the work we're doing with Fedora Atomic will generate some interest doing a two-week cycle for the Fedora Atomic release but we'll have to see where that goes but please join the marketing group now Joe if you would like to get involved with that my talk on marketing is at 430 tomorrow okay there's a talk on marketing Joe is leading at 430 tomorrow so yeah come to that okay Fedora Governance as well last flock we had an awesome meeting run by Heiko and Toshio about how we were going to reform the Fedora couple governance and then after that meeting we over the course of months did it so Thursday at 10 there is a full session with the current Fedora Council so again I'm not going to go into a huge amount of details here but we have this new governance model in place and actually we try and make it a leadership model as well the distinction governance is making sure resources are allocated responding to requests and leadership is kind of helping to set direction for the project and I think I'll talk about that a little bit next up the slides here actually one of the things that I think is an exciting thing is that the council is consensus decision-based and a lot of times lazy consensus which means we don't need to wait for a full vote of everything in order to move that lets us kind of cross things off the list without making sure everybody's around into the forum as long as nobody's objecting we actually go ahead and do things and the consensus based model also means that we have to worry a little bit less about the actual right proportion of people from different parts of the project because if even if there's only one person who represents QA or some some area that one person can say wait stop we're not going to move ahead and tell the concerns of my constituency or satisfy and we've got some the blah blah blah governance details about how that all works but I think that's an interesting and exciting model for project like Fedora and we also have a lot of a lot of this is set up so that me it's all for me but our project leader I didn't get burned out as much so a lot of it's been sharing the load between the rest of the council members so we have Remy Cosmaker on board from OSAS at Red Hat and he is not just helping put together this awesome conference but it's the action and impact community action and impact believe and so it's working a lot of things time to actually grow the community Fedora community and grow the Fedora user base and provide support to that growth so that's exciting we are in the process of choosing a diversity advisor to add to the council as well and that's kind of an ongoing thing right now I kind of hope to have that by now but we're still still working out the details and that will be a basically somebody on the council who can give input to any sort of decisions we're making that sort of impact community and contributor diversity and I think that's a really important thing for any open source project these days and not open source as well but somewhere where we can show leadership and then the other thing we have on the council is this idea of objective leads I'm going to talk a little bit about objectives here basically if you don't know where you're going how do you know when you got there the idea with Fedora is that we've had this kind of very high mission statement of leading the open source and open content community and I should be able to quote it on hand but it's a it's a very ambitious and it's a good mission statement but it's very high up and so a lot of the times what we've done is make a distribution and here's our mission about leading leading the world open source and the connection between those is kind of lost and I think this is the thing I'm having I'm going to plug my talk at Linuxcon I'm going to have a full talk about this sort of method of strategic planning this is something from this is an asking our chart major we do you see I'm kind of proud of it so I put it in this version I think from my talk at Linuxcon I will have a shinier version but here it is basically this is a way of connecting that high mission statement at the end to something that is actually what is the impact of the actual things that you're doing and the basic idea is you plan from the right side the high level and then your actions come across from the left side and so anyways without going way way too much on that the point of this is on the council we have objectives which the council selects between two or four two to four of these my goodness I'm running out of time already geez okay so we select some objectives and we have tried to move towards making the objectives happen over the course of an actual defined time frame we'll talk more about those in the council sessions I would like to use these objectives in the future for things like selecting which talks get priority at flock which funding goes where funding goes for all sorts of things priorities in general and so I think it's very important that we as a community make sure that the two to four objectives we've officially selected represent the things we want to focus on overall and having selected that we can actually follow that focus here are our current ones right now the fedora additions Steve Gallagher it's led and that is wrapping up right now so that's going to be an empty empty slot for a new objective so we can talk about at this conference what other things we want to focus on below that we've got a university outreach objective that Remi is leading up and then Langdon is also leading an objective having to do with fedora modularization and fedora rings come to Denise's talk on what redhead wants to hear some more about redhead's perspective on that all idea and then this is the basic idea again we've got these empty objectives the idea is where do we want to go 12 months from now not just where do we want to go in 10 years but what are we going to do what are we going to do over the next year and I don't have a magical answer to that although I think some of those things I talked about are important I want to have more community visibility I want to grow the user base I want to work on the modularization thing I've got some ideas but it really shouldn't just be my ideas it should be all of your ideas so let's take those ideas and come together and make some of these solid objectives for the next next release and then go ahead and do that over of course of it and that's pretty much it for my slides thank you very much again to all of our awesome sponsors for questions are we three three minutes for questions any more questions yes yeah so the question was is internet of things and lower end new intel chips something worth considering keeping 32 of intel alive probably I think part of this is redefining what we mean by secondary architecture as well right now we have kind of an early big split primary and secondary built way off in its own shadow installation and everything so one of the things we might look at doing is building our secondary architectures in the main infrastructure and that way it's less of a hard line and so we can move things back and forth and we can say things like this is secondary from the point of view if it or a workstation but it's primary for internet of things which is a different way of looking with primary secondary mean but that might be useful for that more questions remy more of a comment and question but thank you to some of the local rochester hacker spaces particularly interlock rochester and and dan schneiderman from maker space for helping out with some hardware infrastructure today you folks came through the question it's worth thanking you from the whole audience so thank you very much and thank you also to tom and ruth and josh and remy and the people who put together this awesome hey man yeah as a quick and perfect draft for the good war magazine try to get that during the next session and share it around the world all right i think i'm making a voice out of world