 So thanks for inviting me and I will talk about commercial open source. So we get started here, a key thing to understand and it's still not widely really well understood is of course that open source by definition and license is not competitively differentiated. Everyone can have it, everyone can use it, any value created stays with those who use it and you can't really commercialize it, which is why saying there's open source business models by themselves is a misnomer because directly of open source you can't really make any money. The key insight to make our way to a concept now of commercial open source is to understand that there must be something else, close complement I call it here, that a company which commercializes open source sells. So identifying what it is that you don't make available for free, developing it, not sharing it and ultimately selling it, that is how a company in relation to some open source software can make money. The idea here is ultimately that the open source available to everyone may create value for users, but then comes a company and tries to capture as much of that value, bring some to it to the close complement, but possibly acquiring more of the total value, basically any customers. This complement is of course the question, what is that? So here are a couple of examples. TensorFlow is a machine learning library, open source software developed by Google, made available for free, fabulous software, many people use it. So Google makes no money off that library, but Google has a service called tensor processing units running in their cloud, which then makes TensorFlow run much faster than if you would run it on your own hardware. So the close complement here in this first example is hardware for software and then also the embedding and a cloud service. And this close complement, of course, is where Google wants users, the free loading users of the TensorFlow library ultimately to go. So we already see a tension here in a sense that Google's actual desire is less that to do good to the world by a good library, but really to get users into that cloud. And this idea of a close complement really goes through all the different aspects of where you might find something complementary. So if you look at some open source software like Drupal, the content management system, well, that's the company, Acreodrupal with its branded Acreodrupal distribution. So the close complement is how all the plugins get configured and then some plugins even open source. So there's complementary software in this case that is creating value, additional value. But then customers go from and users become customers and go from the free open source version to a close source version. That's how a company captures most of that value. Perhaps the most common version these days is operating a piece of open source software in the cloud for customers. So Gradle is a build system that's freely available, widely used, say Java development. And the complement, the close complement is running Gradle at scale in the cloud. So you don't have to wait for yourself to build for half a day, you get it in five minutes. That's a great value proposition, but it's not freely available because it runs in the cloud. Consulting is another option. So PostgreSQL is a fabulous community open source software, but the complement, knowledge, how to deploy it, how to configure it, how to operate it, that's of course labor by people in the consulting business. Similarly, supporting existing software like the GCC, their small scale consultancy suit. So the key insight is, yes, there may be some piece of open source software of different shades and then find what's complementary that you're not giving away for free and possibly even nurturing and building for the purpose of not giving it away for free. That makes sense. So the close complement is the core idea. If you go beyond that, you have to recognize, though, that there are core-screen business models that put things to give put different diverse factors together in a different way. And in my opinion, there are these three core patterns of business model, which are core-service and support firms, single vendor open source firms and open source distributors. The first one is service and support firms. It's basically human labor somehow. So the business model scales by hiring more people, hiring higher skilled people, etc. What you do is you service an open source software for customers who want that service like help with configuring it, help with running it, maybe building applications on top of it. Here, an important distinction is necessary to make between community open source and commercial open source. So I'm talking about businesses right now who are in a living, but in a second, we will really talk about commercial open source software and what that is. Service and support firms, companies who only provide services, human labor for open source software, usually only do that for community open source software, which is open source software that is widely shared in its ownership in its copyright. So many different people contribute it. There's actually a diverse community of contributors so that no single commercial entity can control it. For example, with clever solutions, the GCC, the new compiler figures in because nobody owns the GCC. It's a wide set of people who share in its copyright. Service and support firms, if you take the pure economic perspective are relevant, but the actual value of the company is not that high. It's really hard to build barriers to entry to that market if it's community open source. They will not receive venture capital funding because the return on investment is comparatively low. It's consultancies. We have many more consultancies than product companies. The next two business models are true product businesses. So project or consultancies, product business. And so the first is what I call sender open source firms. And these are companies that develop an open source software, could be an application, could be an infrastructure component for commercial exploitation. And they do have, they can to be the only company who actually services or makes the transfer software into a product. And so there is high barriers of entry to providing services around that open source component. How to do that is a deeper discussion. I'll get to some of the second, but the key really is, and the examples would be MySQL or MongoDB these days, the key really is that this single vendor is and wants to stay the only sole single vendor who can actually make money of that open source software, even though that piece of software is available as open source. They make money off by the way of compliments, but they also make sure through various means that competition stays away. It's a constant struggle of how they do that. But here's some idea. Let me take a step back. I already introduced community open source as software that is owned and developed by a community of diverse participants. So the first key thing to understand about single vendor open source is, as I said, there's a sole vendor who develops it and they will actually not accept outside contributions. So no community contributions unless the rights usually get signed over to the company, so that the company as a single entity remains the only sole owner of the software, which gives them a fair amount of flexibility in keeping competitors away and making sure they are the only provider. That is a shortcoming in some people's opinion in the definition of open source, which is purely legal based on at least what the open source initiative website says. It defines open source software as an artifact, the license of which has to fulfill certain criteria. And people often the open source enthusiasts will say that this is not enough, but rather that true open source is only community open source, meaning there's not only that diverse number of stakeholders, but really an open collaborative process that allows everyone to participate, which is exactly what the single vendor open source firms are preventing. So open source traditionally is just by license. You can add an open process to it, but then commercial open source is out of the question. It's not open source in that respect any longer, which is why we have a fair bit of struggle in the industry as to how to define open source. Commercial or single vendor open source companies have another struggle, which is again, how do you prevent to see your margins eroded from your products by anyone who comes in, takes your open source software and builds a product on top of it. After all, it's open source. So I hear today and elsewhere how open source fundamentally is a public good, and that's true, but all these single vendor open source firms try to make it open source with strings attached. Questionable whether it's in the spirit of open source, the whole point is they still try to restrict how other people can use the software using heart measures, trademarks, using soft measures, processes that are not open and forth. From my perspective, sometimes the graph is feathers, that's business. So I don't value, don't make a value judgment on that. The current struggle if you followed it because I put the national question to ask here is what about these funny licensing things going on? So we saw a lot of companies in a more mature stage who are the single vendor open source firms like Radis Labs and MongoDB change the licenses and the new license is a so-called source available license, which is not even open source anymore. And it's actually very nice because these source available licenses say it's like open source. We give you all the benefits of open source unless you want to compete with us and then you're not allowed to use the software, which is like spot on on what the goal of single vendor open source firms is, which is again to prevent competition from their own work. So why would you even open source if actually you have to do the development or yourself or yourself? I have a whole 15 lecture class on the benefits of commercial open source strategies. It's a lot of things, some innovation you might get. I think the strongest argument is usually in the so-called frictionless distribution. You don't have to go through complex sales cycles to get the foot in the door of potential customer companies. You really get a great start to sell your product if it's already in use in the free world with customers and so their business processes of sales behind that, that's a bit too much for this short talk here. Single vendor open source is what is and let me actually show that to you perhaps here. Can I do that? Yes, nicely. So single vendor open source is the dominant enterprise software business model in many spaces possible. Silicon Valley I would argue 15 years ago they said taking that disruptive strategy of open sourcing first to unseat incumbents. Here's a kind of market research attract on the side. You can see there are two different waves really. Sorry, Dirk, to interrupt. I don't think we can see your presentation because I realize I think you're trying to show one from what you're saying but we do not see it. Oh, there's no screen sharing working. Why do I not? Okay, that took you 15 minutes to tell me. Okay, that is funny. Well, here we go. Is that good? I can't screen share. Okay, now it's coming. All right, that is a bummer. Sorry, everyone. It looks like I was too energized to get going. That is silly. Anyway, so what shall I say now? Okay, so I talked about these two models, service providers, which is really just labor and from say a market capitalization, not so important because, well, they are not even in the market. They're not companies. Then afterwards, single vendor open source firms, which are companies that are providing an enterprise software are usually under an open source license, and then they have the close compliments that they sell. Usually it's the cloud service. And finally, I'm talking, given that I have maybe a minute left about open source distributors, in my opinion, a third model that should not be mixed in the single vendor open source firms. So the distributor firms are Ratat, Suzer, Canonical. And the key here to understand is that they don't necessarily own the open source components, but they create a product from combining a large set of otherwise possibly disparate components into one. They own the IP of what's in between components. They are, in my opinion, as possibly rabidly commercial as anyone else. And the IP, though, that they built the business on is not the open source software itself, but rather, again, what's in between. So the IP is expressed as configuration databases, the test speeds, how they set up the build processes, and so forth. Now, the distributor firms are worth mentioning, in addition to the single vendor open source firms, because they are yet one level up. So service and support firms, you can live off it. Single vendor open source firms, now you're in Venture, Silicon Valley, Venture, Capital Land, and you can become a unicorn and can get very rich. But it's still just regular enterprise software. But open source distributors are a level up, as you can see with the evaluation of Red Hat before it was basically acquired by IBM. They do not only do single applications or components, they do whole layers of the stack. So that's why they are much more expensive, much higher valued in the market, and also much less common because they're not that many layers of the stack as their enterprise applications. So here are some examples. Denox is the obvious one, that's an operating system. Kubernetes is basically the cloud operating system, if you will. So these are large complex aggregates of software that come in commercial, are commercial distributions, and they are much more valuable than any single vendor open source. But the IP, again, the skills you need to have to build such a business are quite different from the single vendor open source firms. So I think that's it from my side. I specialize in commercial open source. I'm regularly in the Silicon Valley, do my research there, and if you have any questions, I'm here for you. Great, fantastic. So for now, Joshua will come in to speak about open source and making open source default in academia, and then we'll have the opportunity to answer some questions for all three after. So let's see if Joshua comes online now. So make sure that you can share the presentation too. Can hear Joshua? Yeah, give me one second. Can you hear me? Yes, sure, you can see you in a second too. You can hear me and see me? We don't see your face yet, but I think your presentation is coming up now. Give me one more second. Goodness. It always takes a second. There we go. We still don't see your face, but maybe it's fine. I know it's coming. I'm wrong. Now you will see you. Perfect. Take it away. All right, everyone. So what I would like to talk about is a little bit following off of the last group of presentations about things that academia could do to make open source the default. And that goes for open source software and open source hardware. So I want to start off with just the good news. Our team is winning. If we look at how software is being developed free and open source software is becoming the default within industry, many of the top Fortune 500 companies already use it. Open educational resources are everywhere, being integrated into more and more classes if not being completely run off of them. Open access within academia is also becoming the default, particularly in North America are instead of purchasing open access rights with publishers of what we're doing in North America is we're setting up our own repository. So almost every university has their own version of open source scholarship and more and more of the funders are encouraging or demanding that researchers start to use it. And then last but not least is open source hardware. A few years ago it was basically unheard of, but today there's dedicated journals to it. There's an explosion of resources. I try to keep track of all, or I used to try to keep track of all the open source hardware within academia, and at this point it's simply not possible. So what I want to talk about today is how we can make it the default so that rather than open source being sort of the afterthought that sort of tagged onto projects, it becomes the way that we do everything. So there is a lot of momentum in our favor. Anybody that's very serious about computing already knows this. Supercomputers, every single one of them on the planet at this point is running Linux. At the same time, it's starting to take over applications that we all have in our backpockets or smartphones as open source software begins to dominate in that sphere as well. And if we track this out into the future, we've never had good inroads into the desktop market, but mobile and tablets are certainly doing well and the Internet of Things is also already dominated by open source. And we heard from the last presenter talking a little bit about how to commercialize open source. And so on the hardware side, Tesla did a very interesting example where they purposely open sourced their hardware for the purpose of becoming a platform. So if I want to start up any kind of company associated with electric cars right now, it's very likely that I'm going to use one of the open source platforms in Tesla out there first. And this absolutely can lead to money as we saw, you know, I don't know if Elon Musk is still the richest person in the world, but he's very wealthy no matter how you look at it. Within academia, we're starting to take open source quite seriously as well. So this is a log plot of looking up the terms open source software and open source hardware on Google Scholar. And as you can see, the top line is for software and it has had exponential growth and it's now, I don't want to say it's saturated, but there's tens of thousands of articles that come out on open source software, use open source software every year. You know, most of the people that are like writing about code are publishing open source software. The interesting fact is that open source hardware is roughly 15 years behind and on that same exponential growth curve. And so, you know, in the olden days, the thought was you couldn't even have open hardware because, you know, you had to actually physically build things, but now with distributed manufacturing technologies, software and hardware really becoming synonymous with each other. And so how open source software, open source hardware is kind of playing out in academia is through distributed manufacturing. So you have one scientist design something for herself. And rather than just manufacture it and use it in her own lab, if she shares it with the rest of the community with an open source license so that anybody that makes a derivative from it is obligated to reshare it with the community, then we get more and more complex things. So this kind of cartoon that I did in this somewhat now dated book is actually happened in real life. And so it started off with a simple little tube. And now there are dozens of types of centrifuges built around that tube. So if the original person that put the tube up or any of the original test tubes needs a centrifuge, it's freely available. The one in the upper left is the Dremelfuge. This was one of my favorites. It's just a chuck that you attach to a Dremel drill that cost, you know, maybe 50 or 60 bucks. And now you can build yourself an ultra centrifuge on the cheek in the middle of nowhere. Some of the ones on the right are, you know, could be integrated into any lab. And the one in the middle shows how far we can go with distributed technologies for fabricating things. This centrifuge is 100% 3D printed and you can get the quality control with open source software that you can run on your cell phone. So if you have a cell phone and a 3D printer in the middle of nowhere, as long as you have access to electricity, you can manufacture yourself a lab-grade centrifuge and ensure that, you know, when you're spinning down your blood samples or whatever you're doing that you hit the right forces. Open source hardware has infiltrated many of the types of labs. My lab primarily did solar photovoltaic research, so we needed a lot of optics equipment. And a lot of the mechanical pieces for optics equipment get very pricey, you know, that starts off for you for around $50 and you need dozens of them to make an actual optics setup. Optical rail runs $380 a meter, but you can get open beam for $12 a meter and then you could 3D print all your fixtures for pennies rather than say buy them for $25 or $50. The overall sort of take-home message is that if you use this completely parametric open source optics library, you'll save between 97 and 99% on any kind of optics that you're putting together. You can also use the distributed manufacturing tools like 3D printers as scientific automation tools themselves. So here's an example of using an upside down Delta 3D printer for both mixing of chemicals and for fluid distribution. So this is like a 96 well plate that you normally have to pay a lot like tens of thousands of dollars to have that automated and now that's it's just a couple thousand. And so when we're seeing these large amounts of savings, particularly in open hardware for scientists, where you use distributed manufacturing, generally you save between 90 and 99% off of commercial proprietary tools. When I did a Fulbright All-Till University Distinguished Chair, one of the tasks I was set with was looking at all of Finland to say, you know, should we strategically start to invest in open hardware development for tools that we're buying all the time or tools that we're importing from other countries or outside of Europe, you know, would it make any sense what kind of money would we make? And so the kind of an exhausted study of what purchases finished researchers were making and which ones would make the most sense to open source, we found that if they were, you know, even if they just did the bare minimum, they'd save almost 3 million euros a year. And if they got aggressive about it, it'd be closer to over like 28 million euros. Recently, you know, there were some thinking that a lot of the easy open source stuff like the typehead racks and stuff like that had already been done and were the technologies that we're developing now still seeing the same amount of savings. So last year I did a review of all the open hardware for science and found that the savings were 94% for free and open source tools in general over a commercial equivalence. And if you used a substantial amount of 3D printing and open source electronics like an Arduino, the savings could be even higher. And so we're getting what we're finding is that the tools are becoming more and more sophisticated. And as they're becoming more sophisticated, they're going after the higher price point items. And from a national perspective, those are the ones that you would make the most sense to go after from funding. So you want to fund open hardware development rather than fund people just to purchase those. So to give an example of how powerful this can be, consider the open source syringe pump library that we developed. And so this, again, was done in OpenSCAD. And so you can put any motor and any syringe combination together that you'd like in order to get the performance that you're looking for for your particular application. And we ran it off of Raspberry Pi so you can also do this off of Wi-Fi off the web. And more and more sophisticated versions like the one that Vila Clar did with force feedback. So you can start to use this, not just for kind of base medical applications or electrospinning and chemistry, but also to do some real material science with it. And so because we developed this in my lab, I have all the economic numbers for it. And so it took about 30,000 U.S. and with a 52% overhead to develop this open source syringe pump library, it replaces syringe pumps on the absolute lowest end at $150, the double pump like the one I'm showing here is cost over $2,400 commercially. In the first month, it was downloaded 1,000 times. So the downloaded substitution value on the very low end after only a month was $168,000, which is far more than we invested in putting it for the overall scientific community. And if we go on the high end, over $2.5 million, since that time, this particular design has been downloaded more than 10,000 times. And so I can conservatively say, we've saved the scientific community $20 million. Now, if you are a funder of science, this is absolutely one of the highest return on investments you're going to get. And the numbers get silly. You're going to be well over a thousand percent return on investment. If your goal is to push science forward as fast as possible, you want to make low cost open source tools that other people can replicate and start to kind of, you know, stand on the shoulder of giants, take the thing that they just read about in the latest journal and start to be able to do it themselves. So to give you a feel for how this actually, what this actually looks like, I have this cartoon. And so this is using U.S. numbers as an example. So in the U.S., the two biggest funders are the National Science Foundation and the NIH, both of whom have kind of success rates under 10%. The last time I looked at the NSF one, it was like 7%. So if you submit a grant, your chances of winning it are only 7%. So let's say you're a science funder, big one, and you're putting a million dollars into some sort of technology and you have two choices of the way to go. You can use the standard model where you would pay, say $100,000 a year for 10 years to buy scientists' equipment. Or you could use the open source model. So in the first year, the both models look about the same. Only one scientist out of 10 is funded. And that scientist has to be a pretty, you know, an older gentleman usually that has substantial experience in the field. It has built up a large number of papers in order to have an H&X high enough that you're even going to consider funding them. And so you make one person happy, but everybody else loses. Where open source really starts to shine is in year two through 10. And so the first year, the open source person that won the contract did a good job and made the open source hardware available to everyone. The second year, because we know at the very most it's going to be 10% for the cost to actually, you know, buy out of materials to build the hardware, you could fund everybody that applies for the grant. And so no matter what we're talking about, no matter how sophisticated the tool, after you funded the open hardware development of it, you could start to provide it to all of your grant, everybody that's applying for a grant. Whereas in the proprietary model, every year you're only funding one person. And so in the open hardware model, because you could fund them and maybe fund them to improve upon the design, that's what the little stars represent that you're getting better and better equipment over time. And so if we add this all up in the proprietary model, after 10 years, you've only funded 10 scientists and most of them are have now outdated, unusual equipment. If you think about how fast science is moving now to have a tool that's, you know, eight years old is probably unacceptable. 90% of your scientists remain unfunded. Where in the open hardware model, you get a huge return on investment. Over 809,000%, 91% of your scientists have equipment that you funded them for. And all of those research tools are easily upgradable only for the cost of materials. And so they're much more likely to be the latest and greatest, the state of the art. And so if we start thinking about how can we encourage open source, both on the software and the hardware side within academia. And so an idea I had was to start thinking about making open source endowed chairs. And so the typical terms for an endowed chair is you just have to demonstrate excellence in your field. You become a really big, fancy person and they give you an endowed chair. But an open source endowed chair would have two more requirements. That requirement would be that everything that you do must be made open access. So all of your writing, you write a paper, you make an open access version available, or you publish it in an open access journal. And the second one is that you release all your intellectual contributions into the public domain or under open source licenses. And for some academics, that might sound extreme, but the kind of the beauty of this is you're getting the endowed chair. So it comes with a lot of prestige. And it also may come with something interesting. So I ran two surveys, one in Canada and one in the US with this idea of open source and down chairs and sent electronic surveys to many of my colleagues. And so the result also and so within Canadian faculty, 81% of the people that responded were willing to accept an open source endowed professorship. 34% more than a third of Canadian faculty members would take the endowed chair with nothing else. They would be willing to publish all their work open access and put all of their technical work in the public domain or with open source licenses without giving them a single penny. The majority of them though wanted the normal things that come with an endowed chair. And so they would want about a third wanted graduate assistance to be paid for and about half wanted the discretionary budget. And so the results that we can get from this, this kind of little mini study is that there's a huge widespread sentiment in favor of knowledge sharing within academia, within Canada, and that an open adult professorship might be one effective way to catalyze and pre-sharing in this community that's kind of already prepped for it. And you might be thinking to yourself, sure, this is Canada, they're known for being nice people. Maybe, you know, they're already sort of already pre-prime for open source. Well, what about, you know, the Americans, which maybe don't have that same kind of national characteristic of what they're known for? But interestingly enough, the numbers come out even slightly better in the American case. So there's a supermajority of almost 87% would be willing to accept an open source endowed professorship. And only 13% were unwilling under any circumstances. And so you have 13% of these hardcore people that maybe think that they're going to become rich by getting patents and selling their technology to a VC. This generally doesn't happen. Of all the people that I interviewed and everyone that I've ever talked to, I don't know a single professor that funds his or her own research with money that comes from patents that they've developed at the university, not a single one. You might know one or two. If you are, please send them my way. I'm very interested to talk to them to see how they did it. In general, that is simply not an economic way to make money as a professor. What the results of this US study show is that there's a willing, even in American academia, to expand open access to science. They want to have science progress at a faster rate. That's kind of use it falling along the open source developmental platform. And from a funder's perspective, there is a large opportunity to move towards open science by funding open source endowed chairs. Something that we just have started at Western University in the engineering department is to start to also fund students to develop open source hardware projects. And so the idea behind this fund is that if an engineering student has a good idea and they would like to take it to the next step, but often engineers or often students of all kinds don't have a lot of excess capital. They're spending a lot on tuition. They don't really have time to work with their studies. And so what this does is spread a small amount of money that gets them started so that they can make a prototype and to see if the idea has any legs. And so this just started so I don't have any data yet on whether or not this can be successful, but of the students that I've shared it with so far, they are certainly energetic and enthusiastic about it. And so to summarize the opportunities to maximize the return on investment for research funding is obviously all publicly funded science should be open access. There should be open access mandates, and that can happen at the national and even at the university level. And we're starting to see more and more funders start to move in this direction that now I think this is becoming the default. And so we've kind of won on the open access front. A lot of faculty members still aren't actively sharing their open access articles, even when they're required to. You can see this kind of in the the shaming that Google Scholar does on your personal profile. The other kind of obvious thing to do the next step would be if we're funding science with our tax dollars, not only should we be able to read the science, but we should be able to use it. And that means that all software that's developed using public funds should be freely available to everyone. And that all hardware developed should be freely available to everyone. I think something that that not everybody appreciates is that especially for patents, there is no clause that allows you to use it unless you've paid for it. So if I patent something, you can't use it. You can't do science on it. You can't do research on it. You have to get a license for me even to make it. And so the that kind of flies in the face of, you know, dumping millions of dollars into trying to move science forward. And yet we're allowing individual researchers to patent or the universities to patent the ideas and not allow anyone else to get to it. And so this is at both the national and the funder level. Lastly is strategic investment. And so we should start to invest public research funds for science into developing open source technologies, whether software or hardware that are in the best interest of the nation. We should start to fund open source development. This can be at the university, the funders in the national level. And then finally, we should think about funding open source and doubt chairs. And so I'll leave you with this. Everything that my group does is at apropedia.org slash fast. We're the free appropriate sustainable technology research group. If you're interested in anything that I've ever done, you can get it there. And so I've essentially done what what I'm talking about is get a funded endowed chair to do open source research. Speaking of which, if any of you are a little younger and looking for a fully funded PhD to do open source development, particularly on the technology side, please drop me an email. I would be happy to talk to you. Wow, thank you very much. I think I have to echo Jan who wrote a mind blowing. I can only agree. Well, let's let's invite the unique next up for her presentation on IT organizations. If you have a presentation, you can also share it now. We should make you the center. Yes. Perfect. And you're still muted in case I think presentation has come up. We can see you. So I think we can get going. So hello, everyone. I'm happy to share with you the work we have done with Professor Chauvet and Goodarzy on how the communities are linked with formal structures. And so we are interested in commercial open source as Dirk. But in this in this article, we really concentrate on the links with communities. Why? Why is it going to the second? Okay. Okay. This is fine. So why is this subject important? First, we define knowing communities as networks that interact and exchange knowledge in order to foster creation and innovation. And we know that communities play a critical role in the innovation of organizations. And this is particularly true in software industry. The critical role of these communities in the software industry is all the more important as all industries are going digital. And therefore, all companies are linked in one way or another to one or more open source communities. As Dirk has said, interactions between organizations and communities are sometimes difficult. And he spoke about MongoDB and the recent cases of elastic search and MongoDB show that these relations are sometimes very difficult. I would say that use is a common factor for the whole software industry. And if there was no use of open source, there would be no implication in communities. But use comes often without getting involved at all in these communities. And I would say that implication grows along with experience. This sharing innovation with communities has already been theorized or theorized. But it was either studied from community point of view or when it was studied from first point of view, it was with qualitative studies, studying often big companies and mainly studying the interactions with only one community. And here our purpose is to study what it means to be involved with internal knowing communities when there are many knowing communities and how does the use of open source affect this involvement. And software industry for us is an emblematic case of this question. So we use the mixed research design, first quantitative study, then qualitative and then again quantitative. For the quantitative study, our objective was to validate the scales and to test the effect of use and open source on community involvement. Then we tried to explicit our model results. And then we came back to our quantitative data to explore the central role of formalization in community involvement. I will focus a little bit on our data. For the quantitative studies, we used the national survey on open source on 2017 that was done in France with all professional organizations for IT and open source software. They sent themselves the survey to all their members and we did then the analysis together. The respondents are either organizations using and organizations editing or integrating open source software. We had 345 answers but we eliminated a few of them and we kept 307 managers using open source and among which 126 are providing open source offers. As you can see, half of the the respondents are from the market side and half of them are more from the technical side. We used AIR for the scales development and then we did some structural modeling using two approaches in order to make sure our results were effective. For the case studies analysis, we chose three organizations, three French organizations who are heavily involved in open source ecosystems. We participated in numerous meetings with them and we conducted semi-structured interviews and we did some content analysis using open source tools and we coded according to the scale dimensions that I will talk to you a few minutes later. Now I want to speak to you about the contributions to research that we make. There are three of them. The first of them is the creation of scales for open source use and for involvement in open source communities. Use of open source has already been described a lot but this has been described with several facets and usually making a distinction between the final user for everyday software and open source used or incorporated in product. Our study reveals that the use of open source is more than that. In fact, it is the organizational use of open source methods, tools and components. For the community involvement, we find a description of three categories of involvement. The formalized involvement concerns organizations that adapt their employees work contracts that set up rules to organize contributions and that set up a formalized process for recognizing contributors. The result oriented is the second one and it concerns organizations that are searching for K people in the open source communities to recruit them or to influence development decisions. The interconnected organizations are concerns organizations that are members of one or more open source consortia and that have close relationships with them and that actively contribute to the projects. I wanted to make a little poll with you but I think I don't have the opportunities to do it. If you want, you can do it with the plus symbol in the bottom left corner. But if it's a bit complicated, we can also just use it in the discussion section. I wanted to ask you whether you think the link between organizations and the hang on. I wanted to ask you, in your opinion, what is the most frequent the most frequent way of involving in communities? All right, that's a people can vote for I guess for a little bit now but I guess you can continue with the vote. Okay, so this is the point of my second contribution. I can see from the results that have appeared already that for instance results oriented and interconnected are main ways of involving according to you and our research shows that in fact the main involvement well in fact we show that you was always always comes with community involvement and community involvement comes with experience but what we see and what really surprised us is that the formalized involvement is the most important because of all literature that we read before we know that a lot of organizations are supposed to be very opportunistic so we would have thought result oriented would be more prevalent and also we know that in the different categorizations that exist the interconnected way is considered as the most achieved way of involvement so we would have thought that perhaps interconnected way of involving would have been more important but rather than that we had the formalized way of involving and therefore we went to our case studies to understand why interconnected involvement is so it's not so important and we then studied the specific role of formalization. First our case studies they referred us that anyway for any IT organization there is a need to use code and to rely on communities and to rely on many communities and there is a ratio of one to ten between what you take from the communities and what you produce yourself. Second from this organization that we are deeply involved in communities we could see that okay they use a lot and sometimes they help but only when it brings value to the communities and when they help they do it well. Third they all of them told us that there is a great need for tools in order to be involved in communities which whatever it could be governance rules or guide versions tracking tools and so on so the formalizations seem to have a specific role formalization in their experience is different from bureaucracy for organization in the traditional industry I would say formalizations often means bureaucracy but here it is a different sense for formalization and therefore we came back to our data and we studied the role of formalized involvement as a mediator between the two other involvement and we can see that in this case the two other community involvement the two direct links become become non significant and formalized really acts as a strong mediator with result oriented involvement but even more with interconnected involvement and this is a more okay or in condense was with what our case studies said as a conclusion I can say that here we bring two scales one for the use of open source and one for community involvement we assess the link between use and involvement and we really try to understand what is what could be a symbiosis or interconnected involvement and what the specific role of formalization is for we bring managerial contributions and give customer arguments to organizations first to explain them that using open source is much more than using everyday tools and that involvement comes with experience and then at last what we would study the few years to come study in more details the specific role of formalization as a mediator then try to assess the performance of the three facets of involvement and then we would like also to generalize these scales and these links to other industries for instance robotics or mobility sectors where lots of things happen in open source hardware as Joshua said so that's all for me thanks very much thank you very much very unique that's very nice then I would invite Dirk and Joshua back to the stage so we can maybe answer a question what will have been asked before and the results are in yeah result on it's interesting yeah okay interesting let's just see if Dirk can also join us maybe still tracking the camera maybe let's start with one question sorry well how do I join I'm sorry Sivan how do I join ah you click the camera symbol symbol the camera okay yeah I can also actually share my face again um let's maybe start one question uh for ps because one you have already answered in the chat that's very nice um but there was another one um sorry uh by Carla on um you pursuing responsible innovation um how do you see open source aligning with responsible innovation sorry that's a that's it both of my last questions are really good the other one's much easier to answer so uh the the open source work that I do would all be fall under the upper career technology umbrella and appropriate technology was that kind of 1970s idea that you know we would only do technologies that were in the best interest of the people that were using them and it was super super popular in the 70s and then the 80s it kind of died and you know back then the idea was you know say you needed a water pump in a developing country village you would send away to a nonprofit in britain with a self-addressed stamped envelope and they'd put the plans in the envelope for you and send it back well now with the internet basically being everywhere and distributed technology finally being good enough that you can actually make real things out of it with it I think the kind of the golden age of open source appropriate technology you're oh satisfyingly here and so we're seeing you know a lot of the tools that we develop in my lab and people are developing kind of in um you know wealthy countries can be immediately applied in countries that aren't so wealthy or by individuals that aren't so wealthy and so I think a lot of the more appropriate and responsible technologies particularly the environmental ones are kind of open for business and then to answer your other question and a little bit more detail um why would a business want to fund open source development and dirt touched on some of these where you would have an open source business but a lot of the businesses I work with aren't necessarily open source to begin with but they fund my lab to create open source hardware usually and that the reason is I think of it as a secondary supplier model you want to find open source things that you can make that make then more money if they're shared more widely so you actually want as many people copying as possible and to give you a couple of examples one is a enzyme company which has nothing at all to do with my research but we had made an open source colorimeter for for our stuff and they saw it and they were selling a colorimeter for a couple hundred dollars to use their enzymes to tell if there was nitrate in water or soil or um forage for for animals and that what they were interested in is going after a new market because their several hundred dollar colorimeter was sort of outside the price point of something you'd see in grade school or high school you're wondering if we could make an open source one and so we collaborated they funded us we got ended up getting some federal funding and funding from the company to make an open source nitrate tester ours cost $50 and you can build it yourself it was so good that it was better than the one that the company was was selling and so they started to manufacture them ourselves using distributed you know 3d printing uh so they you can buy one from the company directly and you can make one yourself for for substantially less but the beauty is they don't care if you buy the carameters because that's not what they're selling they're selling enzymes and the more carameters that are out there that need their enzymes in order to run the nitrate test and it's you know it's completely environmentally safe and it helps kind of do citizen science based testing that the environment that everybody wins and so they funded open hardware development on one of their secondary products in order to sell more of their primary product and then with you know solar photovoltaic vendors I do this all the time anything that I do that increases market for them they're all about and that goes from the material suppliers that are you know selling the back sheets to photovoltaic manufacturers to the photovoltaic manufacturers themselves if I can open source something downstream that makes it you know easier for them to sell more modules they're all over that and so I think the the trick for either academics like me that are trying to you know keep students funded or for um businesses is to find those places where you want to have more copying of your design in order to create more wealth uh for everybody um you know and most notably the company that you're working with I don't know Dirk has some other ideas on that because he certainly has touched on this I would love to sit in on all those lectures yeah so maybe a very quick answer to both questions um I think responsible innovation is not an open source it's in the behavior of the innovators so if you try to encode some notion of responsibility an open source you'll probably fail because open source by definition should not have restrictions on fields of use etc if you try to do that you get ethical licenses which I think by community consensus are not open source licenses so I would argue it's in the responsibility of the people to the second question uh yeah I focused on commercial open source where the primary thrust is to make money in a direct way of the open source by a complement the much more common way of why companies invest in open source is indeed in community open source but it clearly has a commercial or an economic goal which is to take away any money out of a particular component that would otherwise be dominated by a monopolist so we've got linux so that microsoft didn't dominate the operating system they are with windows we've got the eclipse so that again microsoft didn't dominate the developer tool space and we got open stack as an attempt now kubernetes as at least a partial attempt against aws and keep them at bay so it is a very prominent strategy in our industry in my observation that as soon as there's one company that is about to dominate as a monopolist or something like it a particular component or even layer in the stack uh the rest of the industry will gank on them and take the money out of that layer by putting community open source against it and now that we understand with the help of open source foundations and good community governance how to do that that really is an important tool we have for the industry have completely moves things away from traditional sustainable positional advantages into really a continued need for innovation because community open source in my book has the potential of eating up every classic monopolist soft monopolist position from the bottom if you will so um folks gang up on monopolist and that's what open source lets us do i hope my microphone was on and my video was on and i'm so happy that why that i'm getting a reaction thank you really what do you think um yeah it's interesting that you know the let's say growing the pie um uh dynamic there that also josh i talked about this is quite interesting um but there was another question conscious to kind of get them all uh get them all in um by simon for dirk on um if you have information on how often a fork has successfully competed with a well adopted project um so there's some examples but i don't know if you are aware of anything i have no quantitative data no quantitative data all right um let me see we have i think another one from yes another one from simon uh which is on well i've actually attempted monetization i think you answered that one in some sense already i was wondering in terms of actually sustainability maybe um question goes to um maybe specifically dirk but in general terms of sustainability are the different business models in terms of sustainability are there some that you have identified that tend to um provide a more sustainable way of financing a project so if you're asking me again i think community open source like linux or kubernetes is sustainable because folks recognize they need it as an important layer in the stack that's not going away anytime soon and i really want to prevent that anyone else monopolizes that layer in the stack and takes all the money basically the commercial open source is really harder because when you look at say the single vendor open source one would be that the unique also mentioned and think elastic search was your other example um as they mature they close down so there's a life cycle to these because at the end open source as a strategy go to market strategy has run its course on the one hand single vendor open source is great because we get open source software funded by venture capitalists but the long term perspective is usually that they will try to close it down as a market matures to basically turn as many of the potential users that could be customers into customers so you either need to fork or do something else and that's always just not so present pleasant so um yeah so that's right all right um i think marco had a question now i would say there's also an ecosystem building strategy that is relevant in business models don't you are you sure i'm sure what the question is to be honest can i can i answer to marco behrlinger i would when he says that i would say that i also an ecosystem building strategy that is relevant in business models i would say that the the ecosystem is the is these communities and really they must they must be they need to be addressed as a specific target as a specific business model in fact it's as if you as an open source company you needed to have business models for your paying customers and a business models to towards the community to in fact a two-sided business model