 Good. Well, thank you for coming and yeah Hopefully what's some interesting things to say so so really my my talk has two parts There's a sort of a history I think of where we got today and Some thoughts from me on it and then hopefully we can talk about options and discuss things Fair enough Well It's probably on my slide actually, you know, I probably just pasted it in the middle No, okay fine, let's let's let's try and Play with the settings. Why not? It's not it's okay Yeah, then we probably missed the full stop on the end, but it it's it's probably okay eski perfect Anyway, you can be assured that the left-hand side of the logo looked a lot better when we could see all of it So there we are perfect. So you see this is some sort of history of online and How we got here, but so you see there's quite a bit of prehistory sort of Before history began so say a long time ago I went to see these people in Hamburg that had a thing called starter vision and And I can't remember quite when but quail on probably knows all of the answers here And so I'm relying on him to tell me Anyway, they'd created this beautiful active X control that ran in the browser and talked to your son Workstation and they called it star portal and it seems to have died as if it never was on the internet You can't find pictures of it at least I can't But I remember seeing it. Did you see a quail on you? Yeah? Yeah? It was it did exist and Yeah, quite almost anyone who's been interviewed so it must have existed and It talked to a Sun server which was great because you could sell lots of Sun servers and ultimately Sun was a server company So that was kind of useful and There's lots of UNO nurse to go so they used UNO as the remote protocol for that And they did a whole lot of async stuff to make one-way methods work And try make that perform without too many round trips and they brought it up and it yeah, it worked You could you could edit in your browser. You did need an active X control running under Windows, which is a slight downer But anyway, the technology is as it was was quite impressive and some of the underlying things there Actually within subsequently contributed to LibreOffice as they understand it so Yeah, so this vigra and supporting libraries and this base BMP thing sort of all landed in in LibreOffice some some years later and our very own Torsten Behrens was Working on it. I guess back in back in the day and and it lasted quite a long time So base BMP allowed you to do your rendering and and do your VCL rendering at the bottom of LibreOffice In a way that looked the same as X or anything else But without needing a graphic display server so you could run it on the server side a better and Yeah, I mean like I guess Craylon and I why I started made a mess of it And then Craylon finished it beautifully, but the GTK 3 port ultimately brought Cairo and In the end we could get rid of base BMP, which was just a massive celebration very helpful and Yeah, and sort of and some of this vigra and some of this other other staff that was supposedly Sufficient but with templates but turned out not to be and but it's still there in Apache open office If you want to see relics of software engineering, you know and how it used to be you can see it there now So and then there was an SVP plug-in, which I guess sat on top of that so this was the headless plug-in that used this this infrastructure graphics infrastructure underneath and Craylon sort of this quite small commit frankly to to the headless headless module going oh, I'm gonna do all this stuff and In response, I guess unopened sourced there VCL plug-in at the back end the SVP code that today is the basis of the GTK 3 and 4 and you know the headless Work that we use and so the open sourced out and contributed a huge amount of code to doing that So thank you son. Absolutely excellent Yeah So then of course well a little bit later on we announced the idea of the document foundation and you may or may not recall the these times but One of the one of the things that was happening at the time is the Oracle was trying to make an online Collaborative office suite and So that's pretty sticky, you know, it was proprietary wasn't open source But it was it was CSS and JavaScript and so on and so it was really important that we At Lea broth is had a good story. If not if not a Product we at least you know there was a vision there was a direction and you know We live in a tech industry where direction is very important, you know having having Being perceived to have a plan and a future is arguably more important than having it so Yeah, so for the Paris conference. There was quite a large amount of hackery from me The time at Suzer to get the first GTK 3 port working So all sorts of Kyrie rendering damage tracking clipping handling the mess there and Yes GTK extensive exorcisms. I cut and pasting coding is terrible But but we had to take something and we we did that but this is my commit You can go and see it at the bottom of the stack And and one of the things there was to have something to demo to a potential government investor in Paris who could have helped fund You know all sorts of online staff and Libra office and so on and so on Unfortunately, like so many customer pitches never happened But at least there it was and Suzer helped support me doing that which is good And so here actually might my slides from the first first Libra office conference and you discover there's actually still a bug here This is quite annoying where I don't know and I've tried to find this bug many times It won't come back if I don't do it without projecting Anyway, we use GTK 3 and Broadway which is built in and It was amazing You could run Libra office in the browser and for the first time first time ever and that code is still there You can still run. I I expect you can still run a Libra office in the browser Using GTK Broadway and we got screenshots like this which may look familiar for anyone who's played with WebAssembly There may be a feeling of deja vu But rest assured that this was only 10 years ago and we've got something really product ready now so anyway, but so that was that was quite fun and You know here are the here are the conclusions, you know, it's a prototype It's all in public gear to just do this and you know, obviously Alex Larson a red hat did the GTK Broadway piece that made that actually possible So that's all good. So, um, hmm So so but the problem is also at the time. There was a company called. I think Andrew. I don't know They made an Android Apache open office thing. I can't recall what it's what it's what it's called Andropeen and Andropeen office says candy. Yeah, and so again, you know perception is really important that we are the ones with the cool Cool, we are the cool kids that have the cutting-edge stuff in this case. So So we spent a lot of very heavy lifting work. I mean tour did a lot of work here underneath I was there very very late with candy. I think at a FOSDEM Hacking on this the night before trying to get something that was actually demoable and worked in the Android The other we could show to the assembled masses of hackers at FOSDEM And so yeah, I mean so we managed to get not only broff is in a browser, but something on Android that the predates TDF Hmm and and we showed it to people. I think I have a I think I have a picture of it somewhere Okay, and a minute or two I'll get to that Subsequently February late after FOSDEM TDF was was founded. I guess but interestingly before even TDF existed We have had an online a version and And an Android thing And we look at that in a second So at the same time we had a great programming story about reusing the code and how easy it was to use You know and you know was great and you could remote-control LeBron fist and so on But I was pretty convinced that this was really not easy to do and you had to look lots of libraries and things and link stuff There wasn't really installed in the system and and and it was a nightmare. So The theory was great. Shoot. Yeah. Yeah, you heckling a notification from whom? That's the thing. What did it say? Gokai, oh, right. Is he telling me all sorts of interesting things that I shouldn't know Yeah, um one second one second. I might need to just disconnect and turn off my notifications You didn't see Thank you Perfect Anyway, LeBron fist kid. I will get there in the end. It'll probably take a second to re re re synchronize So the thought was then that we should make it much much easier to reuse LeBron fist And so I created a thing called LeBron fist. Well at the time it was called Lib LeBron office You can see my marketing skills were you know Absolute innovation there in in marketing Luckily in the end there were so many kits. There was package kit and System kit and food kit whatever that even I decided that a better name would be LeBron fist kit So that we renamed it eventually But the idea here was that you have a header only see library So you didn't need to include anything it was all in the box And you just you would link this thing and then you could suck goodness out of LeBron fist And initially we use this for LO com so it's all the first customer wanted to convert files And they didn't want to know anything about LeBron fist They just wanted a small binary that could convert files and do that in process and manage it rather than managing an s-office Shell script that ran an s-office binary and something and something and so you and they got some really good speed wins from that Which is cool, and we have this just this this hook LeBron fist kit hook and we DL sim and load it and all was good So here we go. Aha. So another blue slide. I it's bug really irritates me. Anyway, one day. I'll get it fixed Yes, so this is a FOSDM I was talking about the future and you know all of this, you know, the future is you got to worry about the future a lot I'm so this is 2012 This is a quote, you know directly from in the future cloud computing will be the only choice, you know So so there you know Successful businesses may soon have no chief executive no headquarters and no IT infrastructure. It's amazing, isn't it? I don't know what they're gonna do But anyway, this kind of thing was, you know, I mean like yeah cloud hype before that grid computing Etc. I'm just wondering if anyone here today has presented using an online office suite from their laptop No, ah life. You're you know brilliant brilliant, but at least I think you know like the that one out of you know 30 my my expectation is that we have plenty of time left before this vision occurs, you know, and maybe it's never So yeah, and this this is a quote really that's probably quite useful to bear in mind But if you look at how Microsoft's strategy, you know built their product strategy around online office They essentially duplicated. Well, they rewrote the whole product created something totally different that runs on the in the browser and Started rewriting stuff left and right and this is just a recipe for institutional failure, you know actually Well, anyway, and you can see many examples of companies that have failed anyway I talked about how we're reusing code and the web UI and the traditional UI and the Android UI 95% of code shared Numbers still pretty much the same today Shipping out of the box in LibreOffice 3.5. You see that was the Broadway stuff So and this was oh, yeah, this is our prototype that we were up there late working on late at night You know swapping buffers and GL stuff and you know, this is as good as it gets on Android You know, this is this is these are great times Anyhow eventually other people inside Suzer got wind of the fact that there was one department in Suzer that was having all the fun and You know writing Android apps and doing stuff in the browser. There's nothing to do with kernels and operating systems and so on and so on and When the inevitable You know, I mean and Nevelle went through a lot of rounds of layoffs You know, we would regularly fire You know 5% of the company every year and the real trick was to not be in the 5% and of course See the criteria there I think well and they got different management were more competent arguably and I realized that we weren't making enough money And so they spun us out which was which was great. So Yeah, so finally free, you know, whoo-hoo to do exactly what our customers want, you know so there you go that's that's the life of Being free and but the good news was that cloud on signed up So cloud on had a real problem that they had brought Microsoft Office to the iPad and this was awesome But they did it by running Microsoft Office in virtual machines on Windows servers in the cloud and then streaming it as an H 2 6 1 7 5 video or whatever To the iPad which actually worked and it did it did it really quite well And then they did a bit of you know stuff on the top But the problem was this relied on an interpretation of Microsoft's office licensing That was susceptible to question and also pricing and you know, you know, anyway, they were a massive Amazon user resource user and so that went south are pretty pretty aggressively for them and When that happened they needed a product and they were a startup with money and so anyway the good news was They picked on LibreOffice to build their iPad thing So I guess this is probably another rather prominent LibreOffice technology user And they engaged Collabra the new new company and Miklosh did amazing in-drop work for them So we're still enjoying and tour and and Andre Hunt and various other people did did work with me to To get tiles rendering and they built a very very nice tile rendering app on On iPad proprietary about that there it was and it was a good time back in 2014 similar time smooths was like Yeah, yeah, but what about Android we can do something good on Android and so this this Smooses is probably I think was run by a sugar daddy with large amounts of money You know, it's like one of those we should do something good And so he did it did do something good and they funded essentially the LibreOffice viewer application for Android So again, they came to to Collabra and we basically took the iPad work We done for cloud on and we reapplied that a LibreOffice kit work to on top of Mozilla's Fennec library Which was a tiled renderer for the web browser to make The first LibreOffice viewer application for Android and that's the foundation of the code you still see in the Android Directory there and of course we created lots of people like Suzer had done a whole load of things in cross-platform support Google Summer of Code people have done things blah blah blah blah blah It Galea did some stuff. I think with with TDF ultimately and of course Mozilla provided a Fennec So and there you can see it looks better than the previous demo. I think we can agree And ultimately TDF then funded the Android editing work So this was to take the the smooth Suzer whatever thing and add to the LibreOffice kit API the ability to edit Which involves adding a whole load of events and driving driving the core like that This was a good time. We were on a roll So I downer is that cloud on were acquired by Dropbox and exited LibreOffice hard Really hard and one could if one was a paranoid speculator note that they also had a large agreement with Microsoft to bundle Dropbox into Windows and so on But there you are so One of the things that we we also and this is I guess another another reuse of it at the time was to try and get LibreOffice Used in GNOME more We like to encourage people to use our technology everywhere Say for viewing arbitrary Documents we started to add these viewer widgets so that you could then embed LibreOffice inside a much simpler UI You know just for viewing a doc like a PDF viewer, but one that actually renders and loads the document So again, I think Calabra did that I think out of the goodness of our hearts and because we thought it was good to help to help You know entrench the technology let's say in the in the mindset of of GNOME people who who I mean everyone Thinks it's easy to write an office suite But until they until they actually get going and they see what Microsoft did and they have to interoperate with So anyway, we pushed that underneath Parts of GNOME. I didn't if it's still there. Is it still there quite long can aim documents the capacity is still there If you install the package from the deep dark behind the tiger, whatever anyway good for enough So this time we we'd managed with our little demos and our totally broken stuff to find another investor Which is which is even more exciting in the Czech Republic and they they had real problems trying to get Microsoft answering their emails and you know being interesting at all In helping them and so they couldn't integrate with office 365 or office online and they made a male client that was there primarily the thing and Yeah, anyway, so we worked together to to do some great work there and bring LibreOffice online into the browser and that work really was yeah initially funded by them Although they they left I guess in in 2017 and then Calabra took that and really you know carried it on hugely And then what happened? I mean these are just some dates own cloud got involved So we signed a partnership to embed in and cloud and have them promote Collaboral line as soon as next cloud was founded and split out in 2016 We had a partnership there and then lots of versions, you know new features collaborative editing, I mean, you know it took quite a long time to get collaborative editing working but Yeah, and yeah. Well, it does work Very nicely in fact 2017 code. Yeah pixel pixel based dialogues So so one of the problems was that it really was just a very simple text editor, you know You could bold things and italic things. There was a toolbar of options but getting the the rich functionality out of the client and into the Browser was was tricky and our first cut there was to bring all of the all the dialogues as pixels you know like teleported dialogues into the into the client and That was quick and easy but brought problems Of course all the dialogues need to become asynchronous Otherwise when you launch one dialogue and then you have another dialogue if you exit this one You can't actually exit it until the stack unwinds back again, which is which is a bit of a nightmare So you so you couldn't really bring all dialogues, but you could bring those that you had tweaked at some length so we're still still on a on a mission to make more things asynchronous and we're thrilled to have Anatropia helping us with our mission because in the world of web assembly If you block the main loop you die, so and you have only one threat Which is which is sexy, but it does mean that everything has to be asynchronous You know what there can be no states that can be no functions calling others that last a long time So it's a nice nice sort of overlap there and in that and the code has been slowly, you know moving in this way I suppose Perhaps C++ closures getting standardized at a sensible time helpful as well So then you're trying to make the rendering prettier adapting to high DPI You know, there's just an unbelievable amount of engineering effort required here You know, you can do doing and trying to get high DPI rendering and tiles to join and not having pixels that are offset and Misaligned and just the designer that's really really something of an absolute nightmare and getting basic mobile support So at some stage, I don't think I mentioned our rather large customer That helped bankroll or not bankroll demand. Let's say the the mobile version but at some some point Yeah, we got a customer on iOS and Yeah, so this is done with Affinis. So Affinis did did a great job of selling a collaborate online And they sold to a company as I may have said before that wanted an online office suite And that's what was in the spec, you know, like when you're online, you have the office suite Me do laughter delivering it. They said, yeah, but we wanted to work offline as well Nice So and of course there wasn't there wasn't anything in that box. There was a sort of by this time dead proprietary cloud on thing But nothing and there was the TDF Android Fennec Mozilla thing But nothing which is written in Java nothing for iOS So anyway, so we crunched to try and get a first version of that We already had a slightly responsive UI that would respond to different sized screens So we spent time making that work nicely on iOS and we had a first version of it but this was really built completely from scratch with online and the idea essentially was to use a web Window, you know, like all browsers or all phones now have browsers which are used for all sorts of things It's that you know quite a popular API, I guess and so we were just in bed to collaborate online into a into a browser on the device And that worked well enough, although there are all sorts of interesting corner cases around Keyboards and how you pop them up and how windows resize and copy and paste and you know Lots of things at the edge of that are really tricky to get right, but anyway, we did some work So all was well, we had an Android Well, we had an Android version from TDF Repository was made by smooths is a TDF and various other people collaborate obviously and we had an iOS version, which is great Somewhere around then it became clear that the thread heavy model we had so initially we started out with synchronous IO inside collaborate online Which means essentially you do a right and if you can't write because the socket is blocked or it's You know some problem flow control problem on your internet It just sits there and it hangs and that's an easy way to write socket code most people do that start start with but the problem is that if you want to do writes and reads at the same time and You know adapt to different bandwidth situations. You have to have multiple threads so one thread writing one thread reading another thread managing what's going on and You know and that's per user that's connected to the document and any computer science Expert will tell you that threading is hard with a capital H and even even good computer scientists screw it up royally That's probably what it should be software engineers. That's probably why but anyway so it became clear as We started to see some of the bugs caused by race conditions between different different users and things that this was just totally unsustainable and This was on the eve of releasing our product as well. So there was this crisis of Should we meet the roadmap or should we rewrite? the whole way we do threading and IO and Yeah, so one of the exciting decisions I made was that I have to go away and rewrite all of the threading and IO and Yeah, and so we basically Rerote this all so there's one thread per document We had this astonishing realization, which maybe you ought to realize to you That at the end of the day, there is only one network cable coming out of most servers, right? Now it doesn't matter how many threads you have writing to sockets at once There's only one packet gonna come out on the network at once and actually selling a packet It's mostly someone else's problem You know you hand it to the colonel and it can go and do something efficient And so wouldn't it be nice if we had a lot less complexity and you know and we could actually reason about what was going on In the cat. So anyway, so I spent ages and Lots of overnight nightmares and now we have something that's actually much easier to program on so so users can get in and Yeah, it's pretty simple one thread per document. I argue. It's simple. It's simpler a lot simpler. Yeah, I'll say so that was fun What else Android? Why does it say Android at the top? Aha, okay So so we also did code 4.2 at this point. So we we tried to do a six-monthly release cadence of some kind And this added a whole load of sidebar things We're getting there and and because we could use a sidebar in in various ways on the client that we couldn't do dialogues with And because we could use it on mobile by wrapping it into a new UI We added a whole load of functionality to the sidebar So one of the great great drivers for improving the sidebar With like chart wizards and formula wizards and soon-to-be math math formulae you can go and see Mike's Mike's talk on that Improvements and table editing is on is driven by the mobile phone that we want a one-hand UI So that you can wrap these things into the phone and then just use the functionality easily So yeah, so lots of new features then in in 4.2 both for mobile on iOS and also on Desktop and then and then we really start our collaborative office for Android So this is essentially taking that iOS code and the collaborative office Collaborate online call and then just putting it inside a GTK And then what's it called a web view? I don't know. There's some horrible Java class. Can't you probably knows the name? Yeah, and so you say that was essentially based on a completely refreshed approach. There's no fennec There's much much less Java. It's really a web view to a local cool. WSD Wrapping that up just like iOS so entirely entirely new chunk of code that now lives in online for slash Android So after a while and and actually there's quite a big debate at this point. Um, so my my developers Who shall remain nameless were very much of the view that the TDF fennec thing was the future and The hard control of tiles and management of memory and this this was going to give you a better result And it would perform better and he would look better and it should be better And it's it's a idiots game to try and run a web browser on a mobile phone and debug it and you know Well, yeah, it's it's relatively tricky But you can actually connect the Chrome debugger into the mobile phone and debug that thing there and it works actually quite well Surprisingly well for if you're used to GDP. It's Really surprisingly well So yeah, but anyway, I said look sharing code makes more sense than some small performance improvement so we did that and ultimately, I think I was indicated, you know in the end the TDF Android app was dropped from Google Play by TDF nothing to do with me and Yeah, it's really it was done quietly which created some aggro I think in the community and its reviews were really bad and no one was doing anything of course since then Michael Weghorn has done actually a fair bit of work on it and then quite some maintenance So something like it. I don't know a magic 128 commits have been done since December 2020 and most of those by Michael I think for Munich you had some you know strategic interest in that So I think it's a lot better state now than it was then But at the end of the day, you know, it's not not sharing development effort with lots of those other users So so then we come to the contentious political goodness here So one of the things that was happening in parallel here was that there was an increasing openness in the board like this increasing transparency, which was said earlier is probably a good thing and Then there was a vote that was proposed to essentially for TDF to ship online by our bipolar here And I'll get to the text of that in a second. No, no we can talk about it later The problem with that if there is a problem with that is that we didn't really Lots of people in the community weren't aware of many things the board was aware of so in terms of the The number of developers we had how sustainable the project was where we were going and this sort of thing and and and the Problems of trying to attract developers and and make the ecosystem work So at the same meeting Torsten and I and you can go and read the minutes of that Tried to present a whole load of loads of things to starting with this You know the board is a literal quote from what we said and the minutes wants to be more transparent decisions That's good We need to publish the data because it's no good making a decision in in the white community If the white community doesn't really have a deep understanding of the background I think that's intrinsic to doing things more openly. It just makes life. It makes marketing hard and of course it was subsequently picked up by the register in a not in a quite an unfortunate way and But I would argue that that is a Inevitable consequence of sharing the facts of people or perhaps having an overly positive marketing I don't know. I don't know who to blame, but you can blame me. I'm good at being blamed And then there was this vote on the proposal of Paolo 15 15 minute discussion Huh, so we'll look at that in a minute But I first I want to just look at why you might want to have bought to collaborate online in 2025 So these are basically the reasons why you might want to abide. Let me go around this way And the reasons were simple So obviously you got support and SLA and product management. There's a lot of value in these I don't want to diminish them in any way But it's hard to it's harder to explain the value to people versus a free free alternative like, you know, these are good But the cash is king, right? So we had various other things to Apply moral persuasion. So LibreOffice online and code would remind users This is unsupported. You need support, right? Like you should you should care about this piece This is actually important and say to to avoid the impression. It's suitable deployments enterprises when you had something that looked like a big surfer This this popped up didn't stop you using the document It didn't stop you editing it, but it told you politely that you might want to consider, you know Yeah, you want to consider doing something else and also so you could get rid of that by buying Calabra online And you also had access to stable binaries. So one of the one of the key Features for example of red hat is that rel and you know is scarce You can't get rel binaries unless you pay, right? I mean you get the source code Of course you get the source code. It's open source and you can get fedora. That's absolutely fine That's out there and people use it But in terms of getting the binaries of rel you don't get them signed whatever binaries unless you pay I think that's a very traditional and typical thing for distro. Suzer does exactly the same. You'll be pleased to know so Yeah, so the world labor office online development snapshot images, but they weren't stable. They were there you could get them But they they're not not the bees knees And so you'd also get docker images and you get build instructions and things to help you build your own docker images And you'd also get our documentation on installation set up and scaling which was We published parts of but was substantially closed at that time. Okay, so then we look at the vote as proposed And this is just a copy and paste of the what was proposed. So this was the vote 72 hours notice plus I'm sure You know even the infrastructure team to deliver the following packages with TDF and LibreOffice branding docker images own cloud connector So a branded branded version of that of course, this is was written by clabber and own cloud But rebranding that and providing that with the LibreOffice brand Asking next cloud the opportunity to do the same putting it in the univention marketplace Maintaining and updated these on a monthly basis the latest stable version there for free and Yeah I think there's another page there one second. I probably mashed the okay. Yeah, I've mashed the the ordering, okay So LibreOffice branded documentation and help files so to make it really easy to install and to be free and supported only by the community and So so so perhaps you can see the impact of the top pieces. There's some positive pieces as well Like, you know, let me let me bad be balanced, you know, right? So so there were some attempts then to solve some of the obvious problems here So supported only by the community no paid support options from TDF Okay, so and the download page make clear if it's use Lule is used in enterprise environment support options available through the ecosystem So multiple people might be able to provide thing and at a later stage Limitations actually there weren't any limitations at this stage, but but the pop-up I assume is what's meant Concurrent users and documents will be discussed and the vote was pulled at the last minute Okay, and I think Paolo will tell you that you only did it for promoting discussion and something right? Yeah, yeah, but but from my perspective it looked like you couldn't you didn't have the votes to carry it that week and It's hard to you know, no, you didn't didn't care. So So so that's fine, but at least as a business owner when when this level of Change, let's call it is proposed. Um, then then it's it's difficult to to know where you go and So let's just get back a bit to some of the advice from the people who have done this and succeeded So here's Bob young He he's I you know, he's a great guy I think you know, I don't think any longer with red hat in the other thing But here's what he said in his his book open sources, and I hate to read slides But I'm actually gonna read this to you, you know, how do you make money in free software? And I think this is a question that almost everybody asks me, you know I say we write software and we give it away, you know Now what okay, and they say how do you make money doing that? And I'm like that is a really good question, you know, like I mean honestly That's the problem, right? No one expects it to be easy to make money in free software And if I was in church, I would say I'm into that brother, you know, and then you'd all go Yeah, but anyway while making money with free software as a challenge challenge is not necessarily greater than a proprietary software In fact, you make money in free software exactly the same way as in proprietary software You build a great product. You market it with skill and imagination. You build a brand that stands for quality and customer service, right? And I'm broadly with him Problem is that we've already built a great brand. It's LibreOffice and if the product you you are doing all the work to make the product And someone else is marketing it at zero price with the skill and imagination and talent that TDF's marketing brings to it Your brand is not gonna build any brand any kind of brand that gets around that and That's pretty serious in terms of the ability to fund the fund thing and Ultimately marketing is a science that is quite useful. It generates leads if you don't have leads your salespeople can't sell to anyone, you know, and When they sell things you get money and with that money if you're a collaborator you make free and open source software So the lead flow is ultimately fundamental to our business Now there are other ways to generate leads than than You know with with branding the software and downloading it and offering free downloads and so on But they're very expensive and they don't necessarily work well and you can read I write for many years I've written long streets that no one reads with pictures on on on why this is tough But conferences to read mail there are many there are many way I mean like spamming people can kind of work hunting, you know targeting people But it's it's extremely expensive and it's very easy when your software is as cheap as our software is For the cost of sales and marketing to overwhelm anything that you can get back I think this is quite a fundamental problem in in much of the software free software that we have is that we can't afford to market it and You know, you see this with Oracle DBAs you talk to an Oracle DBA and you say why don't you just use my sequel or post Press it's awesome, and they're like mmm, but it's really cheap, right? so So that will you know diminish the whole value of everything that we do No, really, you know, like I can't be if I'm a professional postgres DBA like the whole value of the licensing and the service stuff is is so low that my large salary and you know position as a DBA is Sufficiently, you know is diminished and so they choose to use Oracle which is incredibly expensive like you can't believe And this happens again and again, you know, so ecosystems are interesting and and yeah, anyway, so Yeah marketing marketing against our code that we wrote and like our is kind of odd, but if we wrote it We can claim perhaps it's ours You know as free as in beer is incredibly hard and particularly against a Libra office community brand That's very known enough by all of us So so the votes didn't go through for one reason or another. We decided instead to come up with a marketing plan right So that's a reasonable idea if TDF is going to crush the lead flow that generates the software then Perhaps we should do something else to replace that and that's not you know, it's not a foolish idea, right? And so as I understand it one of the the ideas was you know We take the leads from here that currently go to Calabra and instead we replace all those leads going to TDF and Then well and then what that's the question So there are a lot of people who think that all of TDF's problems would be solved if only it was more diverse You know if only we had more companies You know then then life would be great and there are a number of ways that that shows and I'll show you one in a minute but then Assuming this is even legally possible which was Interesting in many ways and we'll look at Where do you send the leads right? So this is one of the one of the questions if if if leads are ultimately money. What do you do? So when we were lost in Italy, perhaps in Rome October 2017 Here's what was Capone's flyer said what will Capone contribute to LibreOffice online and the punchline are improved user interaction package access and sharing sharing experience, you know So we'll we'll tell you how you can design your UI and we'll package it available for free and we'll shower experience and Calabra had spent quite a lot of time recruiting Capone and visiting them and improving things for them and doing lots of work for them And ultimately they built these packages and for a time. They were listed ahead of Calabra in the download page in the LibreOffice online page So if you went there, you might casual you know as a casual user You might think that Capone who contributed, you know something or they were important, you know, there's a listing and ordering there and To me that was that's pretty that's pretty staggering and astounding because when you look at it actually there was Very little code, let's say that went into online from Capone Yeah Which is odd so so one of the parts of the marketing plan is you know, where do the leads go like how do you make that fair? Like so do you give them to Capone or not or is there some threshold or do you give people a chance when they start as You steer business left and right. How do you how do you do it and on what basis and for what goals? and And you know, how do you make that stick and in terms of legal risk? I'm not going to talk about legal risks in detail for TDF But having a non-profit steering substantial business to one company or another company or anything is Yeah, it's you know, it's it's really it's really not easy, right? so Yeah, see but these are sticking points that maybe could have been solved I'm like I go into these things with an open mind, you know But then the question is when to start doing this should we should we immediately ship? Free binaries at TDF or very rapidly ship them and then just Busk the marketing plan. I hope it works when you can see the the negative impacts kind of obvious or should we? Try it first for the Libra office and see if we can make the moral situation work Well That was very uncertain Destroying business confidence is really it's really easy and creating it It's really hard and perhaps if you if you feel attached to your pension or something like that You can understand why that might be so sorry, you know, I'm making a mess of it anyway So ultimately we lost confidence completely in the being a sensible outcome here and we moved So you can argue why but my my claim is this you know, it's really a Opportunity to unpick lots of badness And so it also allowed us to solve a number of the problems that people have complained about so You know support SLA product management fine, that's all we we should have We made our binaries unlimited binaries Publicly available in a fedora model. So, you know, you could get the development edition And that development edition would have no limits. It would not pop up anything. It would not tell you you should get support at scale And we made all of our documentation public and how we can do our docker images So we chose to use brand awareness then to do that lead lead generation and build a brand So you can argue about forks But essentially everybody Came with us So, you know, generally in a fork you you know, you go along and then it splits into two prongs, right? I mean my forks do or maybe more for them If it only has one I'm arguing it's a move, but I went I went I went hassle you about that Ultimately, there were no commits between the move being announced for two months There's a split vote as to whether to freeze it the chairman's vote was a casting vote and Subsequent we had this atticization process approved for putting things in the attic And and the fact that it could be moved back out again was an integral part of that with rules for that The SC ultimately voted in June They were given a week's notice a heads up for this there will be a formal vote next week and Links to the proposals and so on it was then discussed There was a vote which is very rare the SC almost never votes Maybe five times in its existence All of the calabra and allotropia people abstained from that and it was passed effectively unanimously Then there was a board decision, which was much less unanimous And it resolved to postpone formally at the kizing online until Well soon I guess and that's the result of the vote But at least gives a process for bringing it back Some things sprung up at that point. I guess so Andreas manka was stimulated. I'd say by the SC Decision to work on an updated LibreOffice online So you can see that in the free office online repository on github Yeah, I mean he took the calabra online repository from 2022 all of the work we've done in the last I don't know how many years 18 months 20 months plus there and his first commit added this file is part of the LibreOffice project to all of the headers Now shoot Paolo Yes, so let me reflect on Sure, so Paolo said that's the first thing we removed when when it was no longer part of the LibreOffice project absolutely right and When you reflect on the trademark Nonsense that we've had recently and the problems of claiming that something is LibreOffice when it isn't I think it's it's a little bit rich Simultaneously to attack people for using LibreOffice brand and also to encourage other people to call something that isn't LibreOffice LibreOffice I find that hard to understand but yes fair enough There was also a renaming of number of files changing of some strings and he continued to take commits from calabra Tweak and then merge them some people talk about stealing code from I don't know who you're stealing it from but Um, I think it's interesting to have a balanced view on those things There were some silliness here where Andreas changed a string in his his fork I mean this clearly is a fork although how much work is going on here really is unclear That broke broke it and then there was a big long thread on a board discuss which you can read but the punchline is It didn't break anything except Android and iOS Releases of which there were none From that side. So hopefully this has all been resolved in the obvious way just by having a sensible string There's only one string that read cool in the LibreOffice core and it now rens. Okay, which stands for LibreOffice kit There's another alternative here, which is OX office. So OX office OSSI I Sorry, I primarily funded by a Taiwanese. Well, I understand there's quite a lot of tiny Taiwanese government money there I've long had of their own fork of both LibreOffice and online and It's great that they're building their own brand around that. I think that's that's kind of cool But it's hard to see what they're doing and the base of the code is is rather different there's a lot of Chinese or Traditional Chinese commit messages in there and there's a lot of this sort of stuff going on You know, look at this activity all of this adding caching of tunnel dialogues Which generally is thought to be poor style to take other people's code and put your name on it So, yeah, and I don't see a lot happening there the base is a very old base So that that really does start from I think even before the move They they forked they also checked the thing out and checked it in They copied it and created a new git repository having changed it and checked in so it's very hard to diff One side to the other because you don't know what git commit was the base and you don't know what was changed And that's quite hard to do I mean you have to actually try to obscure, you know, the origin of it and we'll do that but that's been done both for their LibreOffice fork and also their Their online fork and they do contribute some things back. There are a few patches a year that come back So I think this framing is probably maximally unfair and deeply polarizing so I'd encourage you not to use that one I don't think it's reasonable We've put a fortune into writing floss and we contributed, you know with lots of other people and when pushed into a corner In our view we left TDF as a community project and if people show up and do good things it goes well. So, you know, I think we should You know, not not try to use the community brand to destroy the people doing the work So here's another thing I hear that we we should have done something useful during COVID Well, what we did during COVID was the move So we had unlimited easy to install binary packages docker images anyone can use We had public set up an installation documentation removed this warning and At the same time we helped many schools and universities and companies to deploy it and they fund development today And of course TDF can still hope, you know promote what it wishes and this is our fun I really I guess LibreOffice does rock, you know, it is it is cool It's an outstanding piece of software and we love it and we recommend and commend it to you It all came up with as part of the marketing plan this very useful LibreOffice technology unifying way of Showing this thing I think it's excessively helpful way of winning together and we worked to credit LibreOffice technology And put this in there it even shows up as I said earlier in in thing and we do community crediting and we take that quite seriously So we have good lists of translators and contributors and so on And we carried on releasing new features and doing good good new things Next cloud decided as a voter confidence, which is Brilliant actually given the competitive situation in the market to base on LibreOffice technology by default Which is cool and so next-cloud office are using that and of course they use their brand and that's just, you know, great and She'd Next cloud So half how big are they how big a next cloud? So if you look at the docker image downloads, I think They have half a billion downloads of next cloud on the docker. Oh, that's the docker number. You can see I think Now not all of those have collaborate online in them or next-cloud office But it's it's certainly a great way to get our work out there and get people using the LibreOffice technology for sure So the marketing plans the question is we did do many changes to improve things the marketing plan And they and maybe they could have worked. I don't know it's possible If you should be on TDF internal there are many interesting things there do if you remember do subscribe but The punchline is that this didn't have any impact at all on our business. So on one stage that's sad on another Angle it's good that we didn't do radical changes expecting this to have an effect when in fact it didn't so There we go. Where do we go next? So there are many things that we could do and I'm sure you have more ideas and do come and talk to me or You know others afterwards There is this hungry school kids in Africa need free binaries concern and that's fair enough Of course, they can use code TDF uses code or We can provide them with free binaries if that's particularly important to them Obviously running collaborate online needs a next-cloud and own cloud of IDOC file file Blah blah blah blah right and it needs hardware to run on and clearly someone needs to pay for the hardware I don't think anyone's arguing. We should give away hardware Although maybe we could perhaps that's an option If there are strategic accounts and people feel that they want to install something in their local school And they want that just talk to me we can we can help you Someone needs to pay something we try and steer people to do that in many ways in in LibreOffice software My hope is the post-atchization. We can do a better job of promoting LibreOffice technology You know, we like LibreOffice and we like to promote it So I think telling people about it more enthusiastically is certainly something we can do As we use the LibreOffice brand to tell people about LibreOffice We need clear legal rules and they have to be credible trustworthy and enduring commitments which unfortunately in the current world is very hard for TDF to say and Commit anything credibly. I think Yeah Binding future boards is really hard and getting boards to agree is hard, but it's possible that we can make some progress here We try and use a fair use description of our software and fairly Credit LibreOffice and we should work more with the TDF marketing team because they're cool And we can do case studies and we're a LibreOffice success story. We need to be You know telling people like that to try and create more LibreOffice success stories. It seems to me This is another idea. We could have a new and more effective marketing plan that Repersuades people that they should pay for LibreOffice and use that to drive LibreOffice development Maybe I think there's be a massive uphill struggle there and that's not something I advocate I don't want to go there again. Um and furthermore, it's not clear to me that TDF is a good foundation to build a business on I mean if you look at the branding that CIB have had to change And and the way that they've been treated. It's not clear to me That anything that brings leads and business into TDF and then TDF distributes it is stable or sensible So another option is to hey just take the hit. It doesn't matter. It's all renaissance We should just give it away free as in beer is the key and we should just give it all away now, you know So there are people that think that So anyway, thank you for listening. I'm afraid I've consumed much more time that I'd hoped but we can discuss outside Five minutes. Okay, so so we can have some questions. I hope you can see well, maybe you can think about this as you sit there It's unfortunately slightly cut off at the bottom. What what do these companies have in common? Sun, Oracle, Smooth, Sousa, Cloud on Ice, Warp, Capone Anyway, so thank you for listening. You've been very patient Thoughts, question. Oh, yeah, you can clap life guy. I've got a clack at the back, you know Comments or thoughts? Yeah, so possibly I slightly so do starving children in Africa need an online version of a productivity suite So so obviously I slightly caricature the perspective, right? So so let's say, you know Like there's a lot of quite wealthy people in third world countries and in cities with infrastructure And maybe we should be compassionate to them and give them them software It's a viewpoint. I hear regularly from people that these are mission to close the digital divide as TDF And so we should do that and that's I don't want to diminish that argument And so the question is then what do we do about that? How do we do that in a way that doesn't destroy? You know the the software itself. I mean, there's no benefit in doing something today That means that we can't do anything tomorrow, you know, like we will deploy software that's then not sustained. Oh, we have a problem Does that make sense? It is and we should definitely think about that later But I think it's a real concern that some people have and and people have different views of what TDF's mission is that we have a lot of goals we have a big manifesto and People like different parts of it and attracted to different parts of that. I think that's reasonable That's why we have lots of you know different strands. The question then is how you reconcile it all Make everyone will make it work Paulo, I'm sure you have a question or a comment or a Am I totally unreasonable? I mean, I know I am but So so Paulo like the beginning of the talk, I think it's the punchline the technical the technical details The the the the first beat is, you know was a very very nice I like, you know, actually the story because you know, I used to do a lot of stuff and similar stuff many many years ago So the struggle to create a new technology just to innovate and that is just fantastic So I as I said many times are really really appreciate your your efforts as a developer. You did actually great things Yeah, so I always said that and I naturally I always said I always said that I appreciate What Collabora is is doing? Let's say that the rest of the story probably needs fine tuning Because there is one side and naturally we can't take hours and hours of Conversation in the discussion we had to reach compromises to find the way to Actually work on liberal official line together, but you know, that is something that maybe We talk about another time I Have 30 seconds so then let me just just say that if you if you read my slides carefully the beginning bit shows That investment is what makes code We had these ideas for a long time But we had no money to do it and that actually the economics are what drive it what the what's in common between all these companies Is that they all once invested in? Online Libra office like this online document thing and now none of them do So the last people left standing are Collabora and well our partner network luckily We've tried to build economics that works. Anyway, thanks Glenn. That's that's it You've been very patient grab me outside if you want to carry on