 So, I mean, when LibreOffice started, I think there had been a whole lot of reasons why we couldn't improve the source code. And these all suddenly went away. And so there's this flowering of, well, we should fix that. It's broken, let's fix it. And it's just really, really good then to see this huge, huge cleanup effort get underway and, you know, just so much has improved. You know, I mean, it's all little things, but actually, you know, little things really annoy you off for a while. And they also make the code harder to read. So actually now the code is just so much easier to get into and so much better structured, so much safer to develop in. Because previously, lots of things were coupled together in very non-obvious ways. And now, you know, with proper smart pointers and real reference counseling, it's much harder to go wrong. So I think people can feel confident they can contribute without being too worried. And we can accept people's patches more easily too because you can skim through it and see it. It's going to work. Yeah, so we've done some work there. We've done some presentations on sort of basic structural overviews from top-down and graphs and so on. And again, I think there's a difference there to how originally it was presented. So, you know, originally, I think OpenOffice was presented as something that's impossibly difficult to understand, you know. And the graphs were like, you know, whole walls, you know, we are so awesome. Look at us. But actually, the gimmick isn't that bad. And if you do a few little things, you can make the graph look relatively simple, relatively easy to understand. And then, you know, so you can explain the structure. There's good slide decks around that. There's good slide decks about the basic types and strings and the widgets and so on. So it's actually, you know, I think any moderately competent C++ program can sit down and start to struggle and make real progress, you know, in a day or so, in a particular area. And of course, we have easy hacks then to try and make that easier so that we'll point you at the area. So you don't have to go, oh, I want to fix this problem, but I don't know where it is, you know. So you get to help mentor people and show them where they can go and struggle for a while and make a real difference, actually. One of the challenges we face is trying to get people to use ODF everywhere and open standards and open formats are, I think, very important. But if you have to install software on your machine to be able to read a document you've just been emailed, life is not so great. But if you happen to have a browser, you know, and you can follow a link and you can see that document and you can edit, you know, online in your browser with a, you know, native ODF application on the other end and ideally collaborate and interact with other people. That's a really rich proposition. So that was our vision, really, to bring, you know, sort of interchangeable open standards to everybody in their browser. So you don't have to install LibreOffice. Everyone has access to it. So that was the vision we had, I guess, and so we spent a lot of time, you know, building product around that, obviously, in the marketplace, I think. You know, having an online solution is really vital. We see lots of users with whole, you know, as our customers have a whole, whole. Let me just show you my t-shirt. You know, I have the, you know, the Calabra, you know. Look at this. Look at that. Calabra online. Yes, little Superman underneath. But we have a whole load of customers, different things. So people using PCs, I think, are still, you know, still users out there with individual installs. I think it's eternal servers, quite a lot, which is sort of halfway to remoting your desktop, that's easier to manage. And, of course, online and then, you know, Android device. And we've been involved in all of these domains trying to improve LibreOffice to make them work there. But I think, yes, so online is really a very, very rich proposition. Of course, you can host the data on your own premise. You can be sure where the network traffic goes to and from it from a compliance issue. That's very, very nice. So, say, banking customers or something like that. They're really not very interested in even, you know, like, yes, it's in the EU, but I want it in my building where I can see it. And I want to know who is connected to that network. So, you know, and if someone is misconnected, the alarm bells sound and people run and, you know, it's bad. And so, in these kind of very compliance-oriented industries, pharmaceuticals, you know, banking and so on, then we can do some really great stuff to improve things for those customers. So, they can be sure that, you know, they've got that stuff. In conjunction with all sorts of other people, say, you know, own cloud, next cloud, party O, C file, you know, there's a great long list. I'm going to forget at the top, partner's name is getting in trouble now. But, you know, we've just been overwhelmed with how many people want to partner with us, which is very helpful. One of the things we do is tile rendering. So, for the mobile devices, I mean, it turns out that they're very good at flipping textures. So, your ideal texture is really a square. And, you know, you can move it around and there's a whole load of them. And so, we did some work there for the Document Foundation, actually, funded by our donors, to make it work on Android. And actually, the same stuff is then used for LibreOffice online. So, quite a lot of that stuff is basically bringing a whole load of tiles for your document. So, the Chrome and the menus and things run inside your browser, natively JavaScript. But the actual document itself is then rendered on the server. And so, there's a protocol to do that. I think there have been quite a lot of changes to LibreOffice over time to make that work well. But, yeah, I mean, there's a huge amount of work in LibreOffice online. Don't get me wrong, we've been working on it for two years. The team of five to ten people the whole time. So, it's really a big investment from Flabra. But, yeah, you know, quite a lot of the infrastructure is there. And the great thing is you've got this wonderful LibreOffice rendering engine. You know, 25 years of file format work, interoperability fixing, rendering, improvements and so on. So, every improvement you do in LibreOffice then also works in LibreOffice online and helps improve rendering and fidelity. All sorts of things, you know, more performance, more stability. You know, I'm scaling to huge, huge sets. And that's, I think, one of the things that LibreOffice online can do really well is scale to large, large numbers of users if, you know, for enterprise use. I think richer functionality, you know. So, at the moment we have very basic editing. So, you can do basic styles, basic, you know, commenting, basic this. But making that just much, much richer, I think. There's really no reason why we can't bring almost the whole functionality of this suite to the client. But, you know, we do what our customers like at Collaboron. So far we're, you know, 99% of the connects it. You know, that's what we do. We please, please those guys. I use both VI and Emacs, although Emacs would be my preferred one. And unfortunately, I have to use Windows quite a bit. So, I end up using, well, actually, VI more frequently on Windows. Visual Studio, unfortunately there. Occasionally I use Xcode if I have to because we do some stuff on the Mac. Yeah, I don't know. Emacs is where I feel at home. But that's just because I'm old, you know. I'm sure the young people are all using, you know, whatever is latest. G-Edit, oh dear, no. But it's a real professional tool and there are lots of other professional tools out there. All I'd say to people on this thing is, don't use G-Edit. Don't use Notepad. I've seen people programming in Notepad and that just makes me sad because there's so much richness and, you know, goodness to a real professional tool. So, one of the interesting things about processors is they're getting much, much, much wider. So, there are various limits to, you know, silicon and clock speeds and so on. You see that, you know, the gigahertz race is topping out. Four gigahertz, that sort of speed. But the number of threads is going out massively and I'm very excited. One of our partners, AMD, a collaborator, is shipping them called Ryzen quite shortly and it's, you know, like 16 threads by default. So, nowadays, if you're single-threaded, you're wasting 93% of your CPU. So, how can we improve concurrency and do that in a safe way? And there's a lot of people working on this problem of free software, Rust, new languages, you know. And so, where can we use that in LibreOffice to best effect? I think it's an exciting field and also a risky one. It's at the edge of, you know, making computer science work well on, I think. So, that's what I'm talking about, trying to encourage people to get involved in doing crazy, thread-y stuff with LibreOffice. So, we've already done some good work there but there's much more to do. Yeah, there's lots of exciting things going on in LibreOffice, I think. It's a really encouraging project. It's growing. There's fun people like yourself getting involved and, you know, doing good things around it and it's just a really rewarding place to get involved and to make your blow for freedom. And more than that, there's now so many people that can use it. You know, previously, we had 120 million users using it, we think. But now, with LibreOffice online, we can start to grow that massively. We don't have to get them to download anything, install anything. We just have to provide them access to this thing. And so, in combination with our partners, you know, we're growing that user base. Massively. And so, any small fix or improvement that you make in LibreOffice now, you can be confident is helping vastly more people than it could have done before. So, I think that's fantastic.