 I'm going to talk very quickly because that means we get to lunch quicker. OK, perfect. So I just want to talk to you a bit about Collabra and particularly Collabra productivity. We're quite an unusual company. Instead of being a boring company for years and then writing a mission statement, we actually believe our mission statement. And it's quite simple. Make open source the de facto standard in the industry. Or make LibreOfficeRock when we apply it to LibreOffice. And it's quite easy to understand. If we do this, the shareholders are happy, which is me and Philippe and increasingly some of our staff as we move forward. And that's cool. So that's what we're about. When you use Collabra, that's what we're doing. I just like to actually talk to you a little bit about latency and remote controls. No. This is the parent company. So Collabra does all sorts of awesome things outside the escape of LibreOffice that might be interesting to you. So maybe you've driven a car. Well, we get Linux into cars. So we're doing a whole load of work to provide services and support, to provide automotive infrastructure, so that hopefully there will be open source, properly run projects out there so you can hack your car and have fun. We work on semiconductor vendors. So we work with all of the big semiconductor vendors that you can imagine. Enabling their hardware, writing Linux kernel drivers for them, doing interesting bespoke development. OEM stuff, so helping people make embedded systems that do crazy stuff. Hopefully not spying on you, but you never know. What they do with it. So we also do digital TV stuff. So actually, we're embedded inside whole loads of televisions. If you buy a smart television, loads of Collabra code has gone into these things behind the scenes to make these beautiful set-top boxes left and right. Digital signage is going to swallow the world. If you go to McDonald's, you can see it happening already, the sort of the cold stuff that you're about to throw away on the rack is being heavily advertised in the signs above, so you buy it and they don't waste anything. Medical devices, so as you reach the end of your life, you can be confident that Collabra has debugged the Linux kernel in the pacemaker inside your body and the device on the side. So it's a whole load of different things. But the mission there, maybe you forgot our mission. Let me go back there, is to make open-source the de facto standard in the industry. We want everything to be open-source. There shouldn't be an excuse for a lame piece of hardware in that device by your bed that isn't open and beautiful and secure and clean and nicely maintained and so on. So that's what Collabra does. And we have a LibreOffice division which is called Productivity and you can see purple people everywhere that are, you know, I see blue people, but I also see purple people, so that's good. How do we fund this? Ah, yes, so fulfilling our mission is made slightly more difficult by the fact that programmers require salaries. This is a problem. Okay, not all programmers require salaries. I see just the most wonderful volunteers around here that do just the most fantastic things in their spare time and we're deeply grateful. It's just amazing, you know, to be in a community where people have that passion and most of the people we hire had that passion too. You know, they worked as students or they just did it for the love of it. So it's great to have that. But at some stage, you know, reality bites and people need to eat. Hopefully, before that point, you know, we've sold something and we've delighted a customer and we've actually got paid and then we can help people eat, which is good, whilst working on LibreOffice. And so one of the things here is there is no money ferry, you know, and this is something that I've worked in big companies and little companies. I've talked to a lot of people in the community and what they perhaps don't realize is that every euro, every dollar, everything that we spend on eating came ultimately from a customer, right? So we had to delight some customers, we had to make their life beautiful and this is really important. This underlies quite a lot of what I spend my time doing and so, you know, just sponsor the conference, it's great, we love to do that, we love to help people but somehow we have to earn all that. It seems obvious but many people never grasp it. Amazing, isn't it? So we have to have happy customers. Luckily we have quite a lot of them. The problem is that as a consultancy, there's two ways of doing consultancy, you can do it badly and create yourself an endless job, you know, like you make a job for life for yourself and you can retire rich. Or you can do a really good job and then you need new customers all the time because you've solved their problem, you know, you're on to the next one. So at least from a consultancy perspective, you know, we have to do a whole load of customer acquisition and product perspective too. So, you know, we do some heavy duty marketing, we have to work, Lenny up there is even now thinking of how to acquire the next lead, you know. We have sales people going out, reaching out and telling people how good free software is and how to use it and how they can get involved with it. As a sales team, you know, which I enjoy getting involved with myself and Kendi has many gray hairs and you know, there are all sorts of people out there calling people, telling them about LibreOffice, pitching the benefits, investing in pilots, going to places, showing them that it works, you know, helping them see that they can save money and being friendly and there is support and there are people there. Of course, listening to customers, explaining how their problems can be fixed, helping them understand and then, you know, actually selling them things, hopefully. So, the average watch thing at the bottom here is that I'm very interested in money but primarily as a means to fulfill our mission, right? I didn't care about money. I care about LibreOffice and free software and making it rock. Unfortunately, money is the enabling grease that makes this happen. Hence, a corporate plug here. Let me give you some examples of what we've done in the last year. So, here's one of the things that I was involved with and it's quite fun. If you like 3D GL transitions and finding holes in your drivers, it's great. There are all these nice new 3D blingy transitions that are quite nice and interoperable, improved user interface for that. Nice to find a customer that wants it, right? We really thrilled to have that. I'm working with Core, again, Nuanoff, doing this Dutch military project so that you can be confident, I hope, that your data is not leaking outside your organization, that the classification of it is right and it doesn't accidentally get emailed to the press in its current form. OpenGL acceleration, so improving the VCL code structure, polishing that, speeding up. This is things we've done in the last year, I guess. Adapting our code to the very latest GL hardware, so making shaders bigger, strangely, and lots and lots of fun stuff here. The hope is that we can get both quality and performance. So previously you had this dichotomy of you could have very high quality image previews and you could wait for them. Or you could have very interactive editing and you couldn't see the image. And hopefully we can have both. That's the goal and it even works, so yeah. Amazing. PDF, document insert and crop. So this is actually a very small company, came along and said, we have this problem, we want to be able to crop PDFs. We want to be able to embed them like other images and we can't do this stuff. And we pointed out that you can actually load it in drawer and you can paste a GDI meta file and then you can da, da, da, da. That's too difficult, so anyway. So we did this for them and it's beautiful. And of course we have the PDF then preserved inside the meta file so in the future when we can render it better, we'll render it better, et cetera. So that's pretty nice. One of the things we also like to do, this is one of the privileges of employing smart people and just the joys of keeping them in the project. The Miklos in his evenings likes to fix RTF bugs. So actually we fixed, I think 30 plus bug fixes in the last year of the RTF filter. Here's a great example, before, bad, after. You need your LibreOffice logo there. It's very important. I mean it's actually very important for governments. Of course lots of people are still using RTF as the open standard of only two decades ago. So yeah, that's pretty encouraging and loads of unit tests there as well. One of the things we've been investing very heavily in is Colabra Online, which is also known as LibreOffice Online. We've contributed it all upstream, of course. And it looks a bit like this. I think there'll be wall to wall talks about it and demos and stuff. You can go and see purple people and ask for them. This is a new laptop. I'm not gonna demo it now. However, it's pretty awesome. You can get all of the joys and high fidelity of LibreOffice rendering in your browser and it works really well. Particularly when you compare it with other office suites. Let me talk that through. So in terms of why you want Colabra Online, you want it because you can control your data. You have it on your premises. It's in your box, it's in the corner and you know who's touching it. You can see the traffic coming in. Now, some people promise that they'll put your data inside, say the European Union, which is great. So all of your data stays there. But in their cloud, obviously. The only problem is that they don't really control the routing infrastructure. So there was a bug recently where all of Europe's traffic was routed via Hong Kong or something that caused a great chunk of the internet to fail in Europe. Pretty nice, isn't it? But if you can control your network in your buildings and you know where your data's going and you know who else is running stuff on the same CPU, that's kind of nice. There isn't sort of a North Korean running on the same thing as your North Koreans are lovely people, no doubt. But let's say some random hacker with a, you know, trying to look at your confidential business document. So yeah, I think there's some bonus features about online that make it very attractive to us because actually I think it's a place where we can win. People are used to there being a very small feature set in the online products. Google Docs has a very limited feature set. Microsoft offers 365 online. It's even more limited. It's quite extraordinary. And yet there is this great collaborative, light editing use case that we can actually win in so we can have features that they don't. And the other thing about it is that it makes document formats less relevant. Now we support ODF, we ship ODF by default, we have fans of ODF, but the reality is that in the cloud, it's in the cloud, man, who cares, you know, like whatever, you know. And so we can create ODF documents left and right and people don't even know it, you know. They're just there, you know. The default format is that. And hey, it's in my browser, so who cares, right? Which is cool. But more interesting than that, it's actually really quite hard to encourage people to pay for LibreOffice support, even though our support and service is excellent, simply because it's pretty good. And PCs, you know, if it fails, then probably we'll work around it. It's not really a server mission critical use case. So why don't we turn LibreOffice into an online mission critical use case? And then you'd be mad not to have services and support, right? So from a business perspective, hopefully, you know, we can bring money into the ecosystem through this vector and through certification and server revenue and going to market with lots of people who have an existing business model selling very similar things. So that's the thinking behind it. And here's how we're doing it. So we like an indirect sales model, so we partner with people and we strongly prefer them to sell things rather than us selling them because, well, you know, they're much better at it, probably. And, you know, we like to focus on doing the development piece. So these are the people that we've partnered with. They're all signed. They've all committed to distribute collaborative online integrated with their products. That's rolling out, own cloud and next cloud do it now. I think Cfile, Pydio are either shipping right now or coming extremely soon. VNC, again, this video's out there, but the hope is that all these people have existing customer bases and will be distributing and shipping LibreOffice, effectively, to a much larger customer base, which is great. So, what's coming next? Okay, so I'm just gonna blow whoever else's talk. So the next key thing in Collaboral Online is, I guess, collaboration. So some of those things, well, okay, so here's a feature actually from the shipping now in 1.0, which is revision browsing. So, you know, you can look at your revisions and you can see all of your old documents, what they look like, and grab versions, restore to them, which is pretty nice, I don't know if you've seen that, but just a really useful add-in for enterprise file sync and share. So people can quickly browse their versioning and see what documents look like. I mean, it's pretty neat, because of course, you know, you can have all of your files synced to your local device, whether it's an Android device or a Windows PC, but the revisioning thing is harder to interact with, and we can make that super easy. But what's coming next is really collaborative editing. So here are some screenshots of me playing with the creatively named, you know, someone else and another user and all this sort of thing, and, you know, getting nice notifications when people join, multiple cursors, selections, you know, names on top of it. Lots of fun new interactions there, lots of extreme uses. We use this in Calabra actually quite a lot internally. I just have project planning, and I don't know if there are any guilty Google docs developers here, but we'll have some good news for you later in the talk. What else? So my friend, Svanter, at the back here, is always telling me about composable operations and the importance of everything being composable, and I'm sure you're right, but at the end of the day, you get down to these conflicts where you can't do anything, and you have to repair the document. So we have a simple undo redo model that we're adopting, say that people get to do the very easy cases, and beyond that, they get effectively the whole undo stack exposed to them. So what user did what, and then we hope that LibreOffice is undoing it very, very nicely. So a repair mode to complement the versioning. We also also save relatively regularly. So my hope is that with that, we fix all of the real world use cases relatively elegantly, and simply in terms of coding. So one of the issues we have with online is that so far you can see the commit breakdown of who committed it, and there's no prizes for guessing who the blue part of the pie chart is. So all the code is already open. It's all a document foundation, but it can be hard to set up and work on this stuff. It's a mixture of C++, JavaScript, and whichever document management or enterprise files then can share a solution you like. So next cloud, own cloud. You saw the list, right? And so we're really planning to do some more stuff with the document foundation to try and get this out to more people. So we're recommending that LibreOffice 53 coming up in January would be the first source released from the document, foundation from this, and we'll be providing then de-branded, LibreOffice-ized containers so other people can sell it or take credit for it or whatever. And we'll provide Docker containers that are suitable for home use of the latest master builds so people can easily interact with it, develop on it. We want to get translators involved too, so you have nightly builds so that you can see what's changed. Maybe we can do it quicker, let's see, but let's not commit. Now it would be nice if you could translate the string and almost immediately see it in the user interface, but we'll get there. Luckily, there aren't that many strings yet. It's very, very small. And we also want to give counts to people so they can demo it, play with it, use it, and tell us what's wrong with it, and file bugs and stuff. So we'll be providing accounts for everyone that's here and anyone that's interested and plausibly related to the document foundation on Collaborar Online Instance. That also provides some enterprise files like a Dropbox-like functionality. Please don't upload your DVDs there. Storage costs money. So we have other products. So we have PC products. So we sell maintenance and support. So we have Collaborar Office, which has three years of support and releases twice a year. And we have Gov Office, which has five-year support and releases annually. And this makes life good for different kinds of people. And of course, you get security maintenance support. You pick a version, blah, blah, blah. I'm not gonna read the slide to you. It costs money. It's good, isn't it? We invest all that money in making it better. Fantastic. We also provide level three bug fixing. So if you really have a bug and you really know what it is and you can describe it clearly to us, then you're at the level three support level, which is great. So we'll provide an SLA for fixing your bug or at least starting work on it because a bug takes about a man week to fix, on average. And we'll provide fixed builds for you so that if something is blocking your business, you can be confident that we will unblock it, again, rapidly. As rapidly as we humanly can. And we fix any bug for a flat cost, which is quite an interesting problem because though the average is one week, some bugs take a month at least and some take 10 minutes. So we insist on capturing the 10 minute ones as well as the month ones. So we have to fix all of your bugs. Otherwise, the law of averages is mean to us. So there you go. We also have dedicated engineers. So you can buy a hybrid engineer who has a leg of a writer, developer, an arm of a calc hacker, a head of a VCL hack. So there is sort of a synthesis of our engineers. And you know, because this person doesn't exist that understands the whole code base and can fix anything, but we can luckily assemble him, like Frankenstein's monster. And you can buy this machine, you know, for only, and it will come and fix all your problems. Anyway, but it's easy to buy on budget for so you can provide, you know, an engineer and we do all that flexibly. So in terms of taking that to the market, we very strongly prefer to sell via partners. So there are many things that we don't do and we don't want to do. We don't do training, we don't do migration, we don't do level one and two support, and we don't speak Japanese, Finnish, Dutch, you know, like that sort of thing. Maybe some of us do. You know, we have a very diverse, globally distributed team, but that's basically not what we want to do. We want to specialize on LibreOffice, developing it and being excellent at that. So interestingly, now 41 countries we have commercial support in, which is plus 20 from last year. So that's good, isn't it? Last year was only 21. And those I've seen here so far, so Studio Storti, we had Marina in the keynote and Italo somewhere, and Nuanoff core is sitting here as a representative. iCraft, we have Shinji somewhere around, over here, perfect. We have EDX for Olivia, I guess, wherever Olivia, oh, there you go, see, you see. And Yak-May, we have Moritzio. So there you go. If I've forgotten any partners that happen to be here, no, never mind, good. So these people are experts in their domain. They can go into your business. They can wear a pretty suit. They can sell things to you. They can talk to you in your language. They can write reports and explain things to you in words you can understand. And we can focus on the coding and doing the deep, interesting stuff. And that's our really preferred way of doing this. So what does this look like? So say we have a large American company that's a customer, prints all of their 10,000 invoices a day through LibreOffice, which is cool, right? Sitting on a Linux server somewhere in a basement and they use us primarily for PDF generation and then I guess print. The problem is these invoices are OCR'd by some of their customers, which shows you how broken the world is, doesn't it? You should print the invoice, post it, someone else OCR's in, it's electronic again. But the world is broken. Anyhow, they're running on a to be anonymous enterprise Linux distribution and they need to upgrade versions from one version to the next and they've got the whole product, it's just about to deploy and suddenly they discover that it's not rendering the same as it used to. But luckily they have a support contract. So they call us up and instantly we're in a morass of phone calls, font debugging, infrastructure stuff and we fix the problem for them. Actually one of the interesting problems we have with commercial support is that people are usually so relieved that their bug is fixed. We just never hear from them again. They just wander away happily. They don't even close the ticket. They don't, you know, you just, so you have to proactively go and say, look, you know, you're happy. Oh yeah, yeah, yeah, very happy, no problem. It's done. They have another crisis somewhere else. You know, there's fires burning anywhere. Of course in this case, it's not a collaborative office problem at all. It's the underlying operating system, which has broken in a interesting way. But hey, we fix that for them too because we're nice like that. I particularly like this one because it's very graphical and you need pictures to stay awake. But I think you can see the general gist. How about a large Irish bank, which we were working with. They had a beautiful solution on the Linux desktop. It was absolutely amazing. It used LibreOffice and inside a browser with the Java bean using JavaScript, talk to the Java bean, and then Firefox on the next desktop. And it worked amazingly. It worked. So that was surprising in itself. However, there was then a strategy change that everything should be windows and terminal servers. And so dutifully, this beast of all of these things was then ported to windows and terminal servers. And even more amazingly, it still worked, which is extraordinary. Unfortunately, under heavy load in production on the terminal server with whatever, it failed at random. And so you end up with these wonderful meetings where you go there and one guy is doing hardware for one company. There's two different consulting companies maintaining the software. There's us, there's Microsoft, there's Oracle. And the typical strategy of everyone else is to go, it's them, like this. So, but instead we fly there, we sit there in the office, we bootstrap the build, we debug it, and of course it's highly secure, ultra lockdown, you can't download things, you can't connect your laptop, you can't. So we defeat the security schemes, we get in there, we bisect the problem, we build the logs, we work around it, and eventually we fix this horrendous horror. We build new binaries and again, they're happy, which is cool. So the branch offices of the bank can now continue to sell exciting financial products. Yeah, excellent. Another thing we do, of course, is to try and keep these things from failing again. So we like to write unit tests to stop regressions escaping into the customer's deployments. That's really bad. And also, having charged someone a fixed price for fixing their bug. You know, I mentioned about bad consultancies, you know, constantly refixing. So it would be bad if you build them again the next year for fixing the same bug. So we actually write quite a lot of regression tests. We may be only a quarter of commits, but we're 36% of commits to the unit test directories. So I think many of these are cleanups as well. So yeah, auto testing where no one has before. So this is pretty much the end of my talk. I'm just gonna do a quick summary of three years. We've been in business as a subsidiary of Collaborative for three years. The parent company is a 10-year-old business, coming out to 11 now, they're doing awesome open source stuff. So 18,000 commits in 25% of the last 12 months of commits to LibreOffice, 15 commits per day every day throughout the year, sustained over those three years. 17, 1,780 bug reference and commit, so a commit with a bug, a bugger number in it. And we had 23 ad collaborations in the last 12 months. And you know, one of the great joys for me, and I have to do a lot of quite boring things in my role. One of the great joys to me is just to look at the list of people and see what awesome guys I've had the privilege to work with and help fund. And many of these people are students or interns or people who are between jobs, they come through Google Summer of Code. Obviously we have a core of full-time people who are there all the time, but it's just fantastic to have been able to provide gainful employment, improving LibreOffice for everybody that has so many good people. And here's the commit statistics, you get a handout of that. As always, I'm hopeful it will continue to improve diversity-wise, and we'll be a smaller part of a bigger pie or something. It's moving in that direction, so that's good. What else? Yeah, so here's a summary. Our mission is to make LibreOffice rock. Of course, along with you guys, without whom we could do nothing. It's all paid for by our customers. So, you know, if you're a customer of ours, you're doing something awesome. You're changing the world for the better and making LibreOffice cool. Thanks so much to our partners who go out there and sell it alongside us and do just the great work. It's real pleasure to see many of you here and to share beers and enjoy time together. It's all done by the staff, who are just awesome, and without the community, as I say, nothing would be possible. So, yeah, it's our pleasure to sponsor the conference. Thanks for your time and patience listening.