 And so, how can I ask customers and community? Thank you. Thank you, Sophie. That's a very kind of you. This is what I'm going to say when everyone's quiet. Tim, I was, you know, my employee sales crew. Anyhow, this is about collaboration. So you can see lots of things about collaboration. So we have lots of certified developers, we're great, et cetera. But I think, to me, the more interesting thing is our mission, right? So many companies write a mission statement years after they're created because they've completely lost their direction. And then there's a committee of lots of people sit down and they're great and they come up with some immensely long statements. And can I have a guess the other way around? We actually, we just like open source, basically. We want to make it rock. And so that's what we're doing. There is a longer version of it, you know, if you like longer versions. But applying it to LibreOffice, this is our goal, right? We want to make LibreOffice the dominant productivity suite on all platforms. That is what we're focused on. That is the mission that makes our shareholders happy, right? Our shareholders are not there for the money. They're there for the mission, okay? So that makes us, I think, slightly different to your average company. And it makes it fun, I think. So obviously we're not dominant yet, right? So yet. We've got some very encouraging numbers. We've got some good metrics. We've got companies around the ecosystem and partners. So what is Collabra's strategy for becoming dominant? So I've worked with a lot of engineers in my time. And I think there is a startling and important recognition that you need to go through when you're an engineer. And so many of you may have prepped in. Yeah, there is no money fairy, okay? It sounds obvious. But actually, if some projects do have a money fairy. So Mark Shuffleworth, for example, is the Ubuntu money fairy. They don't have to make money, right? It's just a, you know, like, just imagine a lot of stuff and do it in the hand, right? Sadly, that's not our domain. Well, the Mozilla story, you know, this sort of huge chunk of money comes from Google the money fairy. You know, the lands in Mozilla's lab to do awesome things and so on. But as Collabra and as the document foundation, we don't really have that. We have a number of people that donate to TDF. But one of the sad things is some employees never grasp this idea that you're here to serve the customer, right? And it's important to me and it's important to Collabra that actually everything we do is serving our customers. Every dollar, every euro we spend actually comes from a customer, right? So they need to be happy. We need to make them excessively pleased with us. So everything we spend paying developers, hiring new ones, sponsoring conferences, evangelising, all of it comes ultimately from satisfying customers. So we do a lot of this. We dedicate some like 10% of our time, one and a half people going to acquire new customers, talking to them, telling them about LibreOffice, telling them how they can use it, thinking of creative ways to apply LibreOffice in new places where it's not being used. Trying to reach people this message that LibreOffice can solve lots of the problems. And that means hours of conference calls, emails in person meetings, pitching those benefits. Because listening to customer problems, you know, I'm trying to help people by fixed stuff. And so just like to give you some examples of what we actually do, consulting is a pretty obvious business. You know, you find your new customer with a problem. That should be easy, right? Yeah. And then you design, estimate, give them a price and iterate with that to make it fit their budget or their needs. They sign and state the work we execute and they pay, right? So it's pretty simple, right? Yeah. So I'll just give some examples of where we've done that. So here's one example I particularly like. So the customer comes and says, you know, we want LibreOffice to load large XLXX sheets faster than Excel. Okay? It's a small problem. Excel has had, you know, decades of performance work go into it, right? So after some many days of profiling, designing, optimizing, tweaking, mantush, co-haying myself, all sorts of people working on this, four sheets with, well, for workbooks with eight sheets, 100,000 rows with numbers in them. Our previous load time was around 40 seconds. We dated down to four. The reference is seven. So, you know, we did it. Of course, that's just one sheet to remember of other sheets. We don't always win. This is the original. By the way, it's the log scale. You can't really show the performance improvements without, because these two's don't look any pretty small, right? So what we're doing pretty well in orange. So lower is better. Across various large documents. So one example of a customer with a problem, they come to us and we solve the problem for them. Hopefully, well. Hardware name. We want smart spreadsheets to calculate using OpenCL and GPU. This is a huge, huge project. The largest refactoring of calc internals ever makes it run much faster on software. Even if you have no fast hardware, we can do it. A huge team of engineers across a multicore, like Andy and Calabra, delivering a completely new design inside LibreOffice. And tons of unit tests to go with it too, so that we can actually check that it works. So you don't need the hardware to get this speed up. Some of the speed up is just there from the new design. About 30 to 500 times. It depends on the shape of your large sheet, but a very significant performance win. And again, starting to kick the competition pretty vigorously, which is good. Using your standard SUSE, another customer here would like Calada files. It's important. 3D models, it's a new OpenStandard. Is it with XML files in it? Has anyone heard of that before? Amazing, hey? But all new 3D models. The new OpenStandard for exchanging them. Actually, Google's SketchUp warehouse does that. It is Calada. And we can render these down, many of them down to GLTF and embed 3D models. Now, this one is not going to animate. Sorry, I'm too blame. You can pan around it and hopefully we'll be moving to starting to walk through things. You're an estate agent and you want to show the house, the model of the new house that you're selling to a guy. You can walk through it in your slide deck. This is what happens when you go to the lounge. How about upstairs? So, fun stuff. 3D chart it. People want to have a live data stream. Consider a financial analytics coming in, being crunched on your GPU. So streaming data out of some live feed. Doing funky analytics on it. And then, streaming it into a chart like this. Except, we have 3D charts already, but they're no way near capable of handling this much data. But consider the high frame rate, zooming around and clicking these things and showing points of interest. And getting just a very immersive view of your 3D data. So, pretty cool. Interoperability. Fixed in. Thanks to Calada. So some of these customers have named. We're very pleased with our customers. Like I mentioned, every dollar we can do, they've been doing a huge amount of interoperability work. I have only one slide here, because I thought it might come out of it, but I didn't want to. But this is one nice example of a very old, very horrible problem of embedding all sorts of crazy stuff inside these shapes. And now we can do a great job of that, change tracking, smartart embedding, and so on. It's been there for years, and we've been embarrassed about it. And now it's fixed, which is really great. So thanks to them. Other platforms. So we are very eager to have an Android viewer and Smoothz after a very much discussion has just been brilliant and funded a great part of the development of this. Calabaros is doing some too. To finally get LibreOffice, Calc, Impress, and Brighter onto the Android platform. So you can at least view your documents. So a whole lot of tile rendering work, making it quick and a performance, reusing Mozilla, Fenec, and so on. Encouraging consulting, consulting thing. Measuring your PC. Turns out that people want real-world benchmarks to test how fast their computers are. And the speed of your computer is now measured in LibreOffices. How many LibreOffices does your have? So if you have a fast computer, it will go quicker. This is obviously a very useful, large, real-world workload. And it's shipping in PCMark 8.2. You can download it and you'll even see it pop up. Spreadsheets do fun things while benchmarks running. So that's pretty encouraging. So once the customers come to us and have a problem, they want it solved, we can do something for them that improves their life and deliver it to the format. But we're also interested in producing products so we can get LibreOffice out into places it couldn't go otherwise. So we ship on Windows, Linux, and OSX and we'll support on all those platforms. So people can actually deploy. So we have these three years of security maintenance and support that takes the problem of doing this long-term support away from the community so we can move on collectively quicker while someone goes and does all the boring work in the back room to please the customers of giving a very predictable release version, any version you like, humans with customer fixes, MSP patches, easy to deploy, extremely expensive. Like, you know, you should haggle with us on this one. 21 cents per user per month, that's Euro cents in long, I think. So that's in volume, two Euros 50 a year. And of course that's for the maintenance product for level three bug fixing, we charge a fixed price. So you know, you have five bugs, you want to know they'll be fixed before you can deploy. We can help make sure those bugs are fixed and you can deploy and it will have a predictable price, Euro is in volume. We also provide an insurance like scheme. Maybe you didn't use your entitlement last year but it's comforting to know if you had five problems, they will be fixed. So we provide no claims, kind of a bonus scheme. So it becomes cheaper over time if you don't use it. Average pricing is one key. We are a bit of a pain about this. We do insist that we do all of your bug fixing so that you don't give us the vilest bugs in the world and you then pay much less money from all the cheap ones. Because this is an average price, right? We need the easy ones as well as the hard ones. Of course, luckily, it's very hard to tell looking at a bug if it's easy or hard. But even so, that's some important part of what we do now. We also want to brand LibreOffice. We want to tell people about LibreOffice primarily. We like to tell them about Collaborate as well. We want to leave with LibreOffice. So, you know, our brand is I guess inspired by, you know, Red Hat and Susan LibreOffice. It's sort of iterating in that tradition. It's very purple. Hopefully, it's distinctly purple. And, you know, we have a license from TDF for the trademark to do that. And, of course, LibreOffice can come from many people. It can come from the document foundation. You'll see the document foundation strap on there. And it comes from Red Hat. It comes from many people. I'd love to see other people using that model so that we're always talking. When we talk about Collaborate, and we leave with the LibreOffice brand. We have a branding gap. You may have noticed or may not have noticed. So, the more we can talk about it, the better. I'm trying to open. Everything we do is in public git, I think. Yes. And it's a series of git branches so you can just check it out. Nothing is closed. All the bug fixes are contributed back. But we have proprietary licensed binaries. So, you can compile it yourself if you wish. But we got a bit fed up with this and we'll buy one copy and deploy it on 100,000 machines type thing. That kind of irritated me. Other vendors who don't do this have weird termination clauses in their service contracts to achieve the same end. But we just go for the basically a stock viewer for the binaries. But otherwise everything is upstream. And we can't really sell. We have a small piece of the puzzle. A customer really wants native language so they can pick up the phone to German in Switzerland or French or Belarusian, this kind of thing. And we also didn't do level one or level two support. We're really doing any debug fixing in the back end. We don't do migration or training. So we love to sell through partners. Our primary goal is to sell through partners. And so we encourage and market our partners as well. Let me show you some of them. So that you can actually find and get a LibreOffice deployment in Europe. We have great friends, perhaps you know, people in EDX in Brazil, Elian, Alivia and so on and lots of other companies around the place. In Europe we have a number more. Hopefully some familiar names here. Even so far as I'm carrying on the list. Look at this. Wonderful. And Anthony's group here, I guess. I mean, Core, Malz, Atomic. Maybe you can do some great customers here. And friends as well. StudioSport what can you say? Brilliant people. I can't emphasize more that you just do so much better to go to a local partner to us. Not only is the price the same. We're not going to give you a discount your partners don't get. But you just get a charming, competent, smiling local person to help with your problems and to provide a great interface to our level 3 services. So what does a level 3 bug fix look like? Well, this is fixing a pretty nasty bug. Before and after I hope you can tell the difference. So we're good to see that. Similarly, the not having anything there to having something there is a nice transition. And we also support people in strange situations. So I can't talk about all of our customers. The huge stationary company, for example, in America. And all of their invoicing goes through LibreOffice. So if you want to print an invoice out, you get a PDF and an XLXX or whatever and you can print it out on paper. But LibreOffice is doing all the layouts and so we'll support that. You don't want your invoicing machine to break down. It's bad for your cash flow. So we help with that. We help with document migration. So if you use SharePoint, for example, SharePoint scribbles on the document. So as you check it out of SharePoint, trying to migrate to something else, it is scribbling on the metadata. So you can't see this document is the same as all these others. So you need something clever and LibreOffice can provide that clever. These two documents are really the same. Apart from some SharePoint scribble. And again, you need support if you're doing this kind of thing. So, of course, our mission is to try and make LibreOffice available to everyone, to make it dominant, to smooth the wheels, to allow us to sell to people who have to have support. There are some companies who just can't deploy unless they can say, look, we have support. We have checked this box. Oh, and by the way, we've done intrusion detection on the person that is going to give support and we've done this and we've validated this huge form that has been filled in by some of the 57 Shareboxes. So we help do that. But we also think that we contribute quite a lot in terms of actually giving back to the community. These are really just a few product things that you can see. So stopping regression and escaping into our customers' deployments is important to us. So something like 40% of the test commits in the last year since 4.1 came from Calabra. And we love that. Unit tests are a key part of reducing the cost for us and improving quality and improving it incrementally at a time. Systematically testing documents. I don't know if you've seen Marcus' emails with the lists of things that crash or don't crash and the statistics there, but we test 55,000 horrible bug documents that we've had from Bugsilla and we then look at how many crash during load, how many crash during save and these multiple formats are the files valid at the end of that. So there's a significant impact on product quality. Performance testing. So Matt, back here, has created a performance tinderbox for us so we can profile lots of document loads for start, various operations that tend to get problems with them. So we build a whole load of profiles in a very nice controlled environment and we've caught some significant performance problems. You need to have been some 10x slowdowns that we've seen and you see them in the graph so it goes flat and then the time goes massively up and then hopefully it goes down again when we fix it. But previously you just really couldn't see those. You'd have to release it and wait for people to tell us who was banned and now we can catch them payfully as we program. So, my conclusions. I'm sure this is to make me profit rock. That's the punchline. And it's all paid for by our customers so thank you to our customers. Unfortunately many wish to be anonymous but AMD, Cloud Home, a smooth suzerable responsible in large part for funding everything we do. So I'd like to clap then. But of course the staff actually do the typing. So we have an awesome team as perhaps you saw on the earlier slides. Without them nothing would be possible. And without you guys, the whole community producing this great product and working alongside us and having fun and doing all the right things around open source again nothing would be possible. So it's our pleasure to sponsor the conference. In the middle. Thank you so much. Do you have questions? I mean for the question. Do you have questions? Right, yes. I wonder how you did the licensing trick. Did I with this right do the proprietary the license. That's great. So of course we derivative the open source license. Anything that has an open source license that you know we don't want to do anything to trade on an open source license. So if you read our Euler it says that in big text, right? So that's the first thing to say we want to respect the rights of everyone that has contributed the stuff. Having said that the Missoula public license which we use for 98% of our code explicitly allows this proprietary binaries as long as you have contributed the code back. So it's really focused on making sure the code grows with the copy that license so the code has contributed back and everyone can share it and use it. And that's the key. So it's really focused on the source and on the binaries. And again the reason I think that's a win because of the free loading problem. We have seen large numbers of deployments that think it's great to just contribute bug reports and angry noise saying my buggy is still not fixed. And at least from our perspective that's not entirely helpful to our code. We won't be profits to dominate the industry. And every developer that we employ will pay you with money. And there is no money ferry so we have to get that money from a customer. So everything that impedes the flow of cash from customer through to making great and sort of software is something that we're concerned about. So does that help explain it? So if I can say everything was open to you so if I check out you can combine it. That's absolutely fine. And we encourage you to for your LibreOffice deployment on something I would really recommend Labros LibreOffice 4.2 branch because it's the pinnacle of tested hard and beautiful LibreOffice and of course you can combine it with your wish. And if you wish to keep recompiling it when you have security fixes and redeploying it and so on competitors can take it and build on it and perhaps do. But we of course love to do that in public so it's there, it's very easy it's all transparent Cool, good question. No questions? Oh yes, sorry if you like chocolate there are little postcards in the pack and you can post them to your nearest and nearest. In fact stand up Tim you know you want to stand up. Tim is our charming charismatic salespersonage and he would love to put a free stamp on your postcard and send it to your nearest and nearest or even your colleague in your office. They make it the same, you never know. And so you know help them enter themselves in a prize pool for absolutely truckloads of timber which is inspired by made-in-burn food stuff. So yes, file online, wait early on and maybe you'll get some great timber. So I think we can make a short break. Yeah. Yes. 10 minutes 10 minutes 10 minutes break and then the state of the project. Thanks.