WEBVTT 00:00.000 --> 00:07.000 Okay, are you good? 00:07.000 --> 00:16.320 Okay, my name is Yonli Noka, I'm a senior researcher at the National Research Institutes 00:16.320 --> 00:18.000 in Sweden and London University. 00:18.000 --> 00:26.520 I'm going to give some insights and examples of how development and collaboration can be happening 00:26.520 --> 00:30.000 in public sector, open source products. 00:30.000 --> 00:36.680 It's a study I did together with forensic colleagues from Spain and London, or Sweden as well, 00:36.680 --> 00:41.520 called the University and other places, Yonli Noka from Cisco Yonli Noka, that's going to 00:41.520 --> 00:43.440 go with you. 00:43.440 --> 00:51.320 We looked at the called seven different public sector, open source catalogs, and then 00:51.320 --> 00:57.520 we find a quite narrow sample of what we identified as quite mature cases. 00:57.520 --> 01:04.840 So we did a lot of interviews, we did a lot of mining, etc., to really find some insights 01:04.840 --> 01:11.560 and examples to provide others in how you can go about and with doing open source collaboration 01:11.560 --> 01:16.920 development in the public context and how that may differ towards the more general 01:16.920 --> 01:32.120 it may be informal practice that you can find in more products in the world. 01:32.120 --> 01:38.800 So the cases we looked at was one example was the energy plus project, it's a simulation 01:38.800 --> 01:45.320 program for energy consumption in the houses from the US, we looked at those two forms 01:45.320 --> 01:55.600 from our reference, and we looked at Oskary map application platform, powering the geoportal 01:55.600 --> 02:03.880 in Finland, we looked at geotrek platform for managing and enable exploration of national 02:03.880 --> 02:11.520 parks in France, we looked at the Mars simplifier, could call it an ESR platform by the 02:11.520 --> 02:21.800 GNOME in France, and we also looked at the IOAP, so it's an entry app in Italy for accessing 02:21.800 --> 02:28.400 these little public services. 02:28.400 --> 02:34.880 So looking at the development practices, what we saw was that quite a line with open source 02:34.880 --> 02:41.320 in a lot of context is that the development was very centered to a very few set of people, 02:41.320 --> 02:49.720 a core team typically between two to 15 in the most cases, for doing most amounts of development 02:49.720 --> 02:56.360 as we can see in a lot of other contexts as well, generally an open development happening 02:56.360 --> 03:05.560 on GitHub and in these cases, and they used very formal, very structured ALDL, ALDL processes, 03:05.560 --> 03:13.200 very detailed quality assurance processes, and this aligns quite well, because all the 03:13.200 --> 03:18.000 development, there wasn't really a lot of development happening, like the cross organization, 03:18.000 --> 03:24.800 most of it was happening within the main public entity or defender doing that development, 03:24.800 --> 03:32.080 so there wasn't a lot of community development in these cases that we looked at, stuff 03:32.080 --> 03:36.240 was generally reported high-quality, high-level user build in some of which could refer 03:36.240 --> 03:37.240 to that. 03:37.240 --> 03:44.420 We looked at the different types of sponsorships, and one kind of sort of how who develops 03:44.420 --> 03:50.400 or who funds the development and who drives the development basically, and one major part 03:50.400 --> 03:59.200 was where it's really centered to a certain few or one or a very resourceful organizations 03:59.200 --> 04:10.440 to be at a public service company or a ministry or something, it's typically a very business 04:10.440 --> 04:15.160 critical, this application for them, so it's really warrants a major investment for them 04:15.160 --> 04:23.200 in these applications, and then the more decentralized model where you have a lot of different 04:23.200 --> 04:27.200 media, public entities collaborating, co-funding development, driving the development 04:27.200 --> 04:35.280 more in a community way I would call it, together with service providers, in the cases 04:35.280 --> 04:43.320 we looked at, in terms of development resources, again it was typically performed within 04:43.320 --> 04:50.000 one organization, one way or the other, either by the public institution in the self or 04:50.080 --> 05:00.720 through a vendor, one case was the Pagropa, where in the other cases was the resources 05:00.720 --> 05:08.720 was procured in one way or the other, typically consultants, in some cases, really highlighting 05:08.720 --> 05:17.640 this general notion of outsourcing and trusting yourself on the external capabilities, 05:17.640 --> 05:22.600 especially on a municipal level, so Pagropa is quite interesting because they are the 05:22.600 --> 05:28.200 output of an experiment really where Italy or their government wants to build their own 05:28.200 --> 05:35.840 capabilities, build their own development teams and grow vendor independence long term, so 05:35.840 --> 05:44.800 that was a unique in this case, but a very interesting approach, and again all across suppliers 05:44.880 --> 05:51.280 was really highlighted as key for the sustainability of these projects, there needs to be suppliers 05:51.280 --> 05:56.640 out there that can support the development, because again in most cases, as I said, Pagropa, 05:56.640 --> 06:03.120 you were a client on either doing consultants or having vendors to develop the software, 06:03.120 --> 06:09.240 so for the projects to stay sustainable you need to have suppliers with a sustainable business 06:09.240 --> 06:13.400 out there that can support you in this. 06:13.400 --> 06:20.040 So the planning and decision making was typically done top down by the public entity spending 06:20.040 --> 06:25.480 the development, either through some kind of technical steering committee, with those involved, 06:25.480 --> 06:31.920 or through direct communication between the municipality or public entity and the vendor, 06:31.920 --> 06:38.480 so and this could be run in parallel, so they could be like overall coordination happening 06:38.560 --> 06:45.360 while still municipality is going directly to the vendor to fast-track some things, and in some 06:45.360 --> 06:55.280 cases also than for the internal teams, and this is something to maybe consider is that by how 06:55.280 --> 07:01.440 you include vendors or the service suppliers in your collaborations, excluding them fully and just 07:01.520 --> 07:08.960 doing with the public entities in collaboration, that may lose insights from the vendor 07:08.960 --> 07:15.760 service supplier which could help you actually in the development, and also by fully relying 07:15.760 --> 07:22.400 on the vendor or the service supplier in coordinating, that can lead to soft lockings, so you 07:22.400 --> 07:28.080 kind of lose track of the project, so it's really finding this balance in how you include 07:28.160 --> 07:32.400 the service suppliers and vendors in your projects, and this becomes then a matter of how you 07:32.400 --> 07:37.520 how you do public procurement, for example, if you're doing it below the fresh whole delimits and 07:37.520 --> 07:44.400 you can do direct procurement, or if you're frame frame agreements as well, so that needs to be 07:44.400 --> 07:52.560 considering in context as well. And there was also an interesting note in the interviews that 07:52.560 --> 07:59.360 was a large preference for local service suppliers and vendors to have some kind of closer collaboration 07:59.360 --> 08:06.000 and relationship with them. And then in terms of communication, all the all the projects have some 08:06.000 --> 08:13.600 kind of formal communication, but a lot of the communications was happily closed between 08:14.640 --> 08:18.640 inside the organization, like in within the vendor or the public institution doing the 08:19.600 --> 08:26.720 development, but also directly between the public entities and vendors. So this is also like a risk 08:26.720 --> 08:32.720 it risks creating these clicks or isolated communication, really losing a lot of the benefits with 08:32.720 --> 08:39.760 the openness and transparency of open source, really creating a big barrier for others to come in 08:39.760 --> 08:45.360 and start collaborating and see what's done and so on, so this is also something to think about 08:45.360 --> 08:53.040 how communication is done. Community engagement, the community say we're very user oriented, 08:54.560 --> 09:00.320 but we can still talk about some kinds of contributions in terms of funding, sustaining the 09:00.320 --> 09:05.840 development, of course, providing subject matter expertise to their requirements in generic process, 09:05.840 --> 09:13.360 but also in terms of the general shaping and knowledge sharing and identifying a bugs and so on. 09:14.320 --> 09:20.000 So not not not much about like the more technical code contributions here, so more about the higher 09:20.000 --> 09:29.520 level in sustaining the projects. The user basis, so the is often quite limited, is typically 09:29.520 --> 09:38.640 public sector entities, but there's somewhat higher user communities, so not these developer 09:38.640 --> 09:47.120 communities as we find in the world basically or mostly, but the number of end users are much 09:47.120 --> 09:53.280 much greater. So depending on how we draw, the borders of the user community, it will be much 09:53.280 --> 09:59.280 much greater because these projects that we look at are really core in different kinds of digital 09:59.280 --> 10:03.680 public infrastructure or public services, so they are still of great importance. 10:04.320 --> 10:10.720 In terms of sustainability, we have touched upon it, but in the central model, there's a lot of 10:10.720 --> 10:18.720 reliance on this core entity like the land survey in Finland, for example, if they were to pivot 10:19.440 --> 10:26.480 or if any other case, then a lot of other ministries or agencies relying on this project would 10:26.560 --> 10:34.720 have to reconsider, and if you're in the heavily invested, it can be costly, so it's important 10:34.720 --> 10:41.040 that's open source projects in general to think about the robustness and the bus factor or 10:41.040 --> 10:46.720 whatever we call it, how development they concentrate and what you're doing to address this. 10:48.320 --> 10:52.560 And on the other way around in a more decentralized sponsorship model where you have a lot of 10:52.560 --> 10:59.920 different public entities collaborating, co-funding, co-pooling money. You often need to grow up 10:59.920 --> 11:06.320 until a certain level, a number of municipalities, for example, for it to help sustain the development 11:06.320 --> 11:13.040 and so on. Otherwise you need some a certain few larger ones with more capability to fund 11:13.040 --> 11:18.640 the development, but that's typically not sustainable in the long term because again, it comes 11:18.720 --> 11:26.160 to reliance on someone, be it like the capital city in the country. They want to move on, 11:26.160 --> 11:32.560 so there needs to be a sustainable funding model for public sector collaborations. 11:32.560 --> 11:37.840 Well, I talked in the funding dev room this morning and we talked more about these community 11:37.840 --> 11:43.680 rooms, but it's also an issue here. Some further recommendations, 11:44.000 --> 11:53.040 really you need to consider growing an open collaboration involving enabling more public entities 11:53.040 --> 11:58.800 to come in and help co-fund the development, grow capabilities, consider these suppliers, 12:00.080 --> 12:06.160 business models that they are profitable, enabling that you don't end up in a lock in having 12:06.160 --> 12:13.360 transparent continuous check-ins on code and so on to and also consider involving the 12:13.360 --> 12:17.840 divenders and the suppliers. Okay, that was a...