Case Study of OS2borgerPC and MedborgarPC

OS2borgerPC[1] and MedborgarPC[2] are open source operating system solutions implemented in public spaces, notably libraries and citizen service centers. Adoption has been strongest in Denmark, where approximately 50 local governments use OS2borgerPC, while MedborgarPC – recently launched in Sweden – still has only a few local governments as users. While most local governments organise through OS2 and Sambruk, a Danish and a Swedish municipal association respectively, hosting and operational contracts are contracted directly with the service suppliers, where Magenta stands out as the main supplier.

Magenta serves as the primary developer and maintainer of the system on behalf of the local governments in Denmark and Sweden, who mostly collaborate through OS2 or Sambruk, respectively[3]. Some development is also procured directly by local governments from Magenta, which today has two developers more or less dedicated to the project. The development model has been a source of tension in Denmark, where disagreements between OS2 and Magenta on a potential lock-in situation and transparency in development processes have resulted in a temporary code freeze from OS2 while Magenta continues to develop on a fork with direct relationships with a number of local governments. The Swedish local municipalities have, through Sambruk, continued their development of specific features outside of the ongoing tensions, in collaboration with Magenta.

The governance of OS2borgerPC (the Danish version) is formally centred on OS2, yet with some parts de facto centred on Magenta. It is, in essence, a packaged version of the open source operating system Ubuntu[4], enabling visitors to use public library PCs in a safe yet easy way through a custom and simplified interface. MedborgarPC (the Swedish version) is supported by Sambruk, in more direct collaboration with Magenta. MedborgarPC provides the same core software system[5], with some key operational and organisational differences in how they are managed and deployed in Sweden, as well as some necessary localisation efforts[6].

Funding for OS2borgerPC or MedborgarPC is largely centralised through OS2 or Sambruk for the Danish and Swedish local governments, respectively. Local governments can also procure directly from a service supplier like Magenta, which has reportedly become more common since tensions emerged between OS2 and Magenta on the development and maintenance efforts of the project[7]. Tensions centre around a wish from OS2 for multiple suppliers, and separation of services (development, maintenance, hosting, support) in contracts, while Magenta is calling for increased bundling of services and a limit on the number of suppliers involved. Vendor independence and the need for sustainable business models are driving the two sides.

OS2: OS2 is a Danish municipal association focused on enabling its members to initiate and collaborate on common software solutions, mainly in the form of open source. The association today manages 28 projects, with a secretariat providing support[8]. In the OS2borgerPC project, they hold the copyright and governance but have faced challenges in exercising the latter due to a variable focus of the involved local governments, leading to increased bilateral agreements between Magenta and local governments[9]. They have recently pushed for more service supplier independence and transparency in development processes.

Sambruk: Sambruk is a Swedish collaboration of about 150 members, primarily local governments, that co-create and co-maintain digital resources[10]. Similar to OS2, they operate on a membership model that leverages fees paid by local governments, based on their population size per project[11]. In the MedborgarPC project, they work directly with Magenta as a service supplier while ensuring open source principles are maintained. They have contributed to the project through Swedish translation and funding new features like SMS authentication and appointment booking integration[12].

Magenta: Founded in 1999, Magenta is the primary service supplier for both OS2borgerPC and MedborgarPC. Magenta was part of founding OS2 on the service supplier side and developed the original system based on a request from Aarhus Municipality. They provide the solution as a SaaS offering, handling development, security, and maintenance. They currently serve a large number of the Danish local governments and have expanded to serve Swedish local governments through collaboration with Sambruk. Recent tensions have emerged with OS2 over governance and development control.

Borås Public Library: Borås City Library[13] – located in Borås Municipality, Västra Götaland County, Sweden – transitioned to MedborgarPC after years of using Netloan, an expensive and outdated library computer management system[14]. The library connected with Sambruk through another software project and discovered MedborgarPC as a cost-effective open source alternative[15]. The implementation has been largely successful, with MedborgarPC proving significantly faster. Borås covers about 50% of the shared costs and sees the shift to open source solutions as financially advantageous despite the initial investment and the need for enhanced technical skills among library staff[16].

OS2borgerPC, the seed of which was created over 15 years ago, originated in the City of Aarhus. The city laid the foundation for the system through the largesse of an IT system, which had an outsized mandate to innovate and do what it wanted[17]. Since then, the system has grown and expanded as an open source collaboration between local governments, in coordination with the OS2 association and service suppliers like Magenta[18].

Today, in Denmark, approximately 50 local governments are customers of the solution, either by participating through OS2 or operating outside the association’s framework. Each of the local governments participates in the OS2 in different ways, working with a service supplier (e.g. Magenta) directly or working in a more coordinated way through OS2[19].

In Sweden, the adoption is more recent, with two local governments actively using the system and several others showing interest[20]. To onboard local governments to the system more quickly, Sambruk worked directly with Magenta, the same service supplier who was working with local governments in Denmark[21]. One example was the Borås City Library[22]. Speaking to the experience of adopting the solution, one interviewee noted: ‘When we just looked at the numbers and everything like that, we saw this might be an opportunity to [adopt a solution] much cheaper than having the [proprietary system] NetLoan. [...] That's the strength of open source.’ [23]

Denmark has a longer history of open source at the municipal level than in Sweden. Sambruk was more of a traditional municipal association, and its engagement with the world of open source IT has been a relatively recent development[24]. ‘Since I've been working in the Swedish municipality for a number of years, we never used open source. That was more or less banned there [...] It [just] wasn't a big deal in Sambruk. They didn't talk a lot about open source stuff.’ [25]

Magenta serves as the primary developer and maintainer of the system on behalf of the local governments in Denmark and Sweden, who mostly collaborate through OS2 or Sambruk, respectively[26]. Some development is also procured directly by local governments from Magenta, which today has two developers more or less dedicated to the project.

The development model has been a source of tension in Denmark, where OS2 has pushed for more service supplier independence and transparency in development processes[27]. OS2 describes how they: ‘... have tried to make sure that the project was actually reusable by anyone and not only by contacting Magenta [the current vendor]”. They further noted how: ‘... it's basically minor technical stuff that you would expect from any open source project.” OS2 further adds “our vision has always been [...] that the project is ‘real open source’, so any vendor or anyone can download it, install it, start using it, and start contributing to the project. And that wasn't the case.”  Magenta, on the other hand, disagrees, noting that reuse is possible but that certain technical capabilities are required, which some local governments may not necessarily possess[28]. In these cases, suppliers such as Magenta can provide support.

Recently, there has been a freeze on the main codebase while these issues are being resolved by a second service supplier, with Magenta continuing development in their own fork of the project[29]. That said, the development of the Swedish version, MedborgarPC, has contributed new features such as SMS authentication and appointment booking integration. ‘Magenta would gladly cooperate, I think, around new features. And we have had conversations with them [to] develop SMS and booking facilities, but no municipality in Denmark has shown an interest, as far as I know, to use that functionality.’ [30]

Funding for OS2borgerPC or MedborgarPC is to a large degree centralised through OS2 or Sambruk for the Danish and Swedish local governments, respectively. Local governments can also procure directly from a service supplier like Magenta, which has reportedly become more common since the tensions increased between OS2 and Magenta on the development and maintenance efforts of the project[31].

The association model works similarly for OS2 and Sambruk, though there are some small differences. For local governments organised through OS2 and Sambruk, a base membership fee is provided to the respective associations[32]. An additional fee is provided per solution that a member's local government chooses to adopt. In OS2, the fee is based on the number of inhabitants, while in Sambruk, the fee is based on the number of instances of MedborgarPC used per local government. Those local governments in Denmark that are not organised via the OS2 fund development directly through Magenta[33].

On the OS2 side, the joint funding model for OS2borgerPC is focused on development and maintenance costs, while subscription services are managed directly between the local governments and the preferred service supplier[34]. There is a preference from the association for a transparent pricing structure of how funds are allocated between development, maintenance, and services[35]. The project achieved financial sustainability after about three years, when it had enough clients to support two full-time developers[36]. Current challenges include balancing service supplier sustainability with municipal cost expectations and ensuring transparent allocation of funds between different aspects of the service[37].

For Sambruk, the relationships remain stronger with Magenta. Sambruk supports local governments working with Magenta on feature development, maintenance, and customisation[38]. ‘It was the easy way to go, and they had a quite fair pricing model. It was a no-brainer for me anyway to go [with Magenta]. But if they had been expensive or so, then we might have considered doing a procurement of another vendor instead.’ [39]

On the service supplier side, Magenta prefers a funding model from the local governments that incorporates continuous development and maintenance costs with the SaaS subscription. A lag in procurement requests from local governments on development and maintenance implies a risk of the number of resources being dedicated to the project. The sustainability and quality of the project, by extension, come at risk due to the lack of continuous maintenance and oversight, and competing solutions taking root[40].

The governance structure of OS2borgerPC and MedborgarPC is both complex and multifaceted, incorporating OS2, Sambruk and Magenta from different perspectives. In Denmark, a de facto dual governance structure is emerging with OS2 and Magenta facilitating and orchestrating two parts of the community of local governments, which overlap to a certain extent [41].

From OS2’s side, the OS2borgerPC project follows OS2’s standardised governance model that they have refined through the years and applies to all of their open source projects[42]. A technical steering committee performs and facilitates the collaborative requirements engineering process and procures the development and maintenance activities accordingly from the service suppliers[43]. Amounts are generally below thresholds, which is why direct procurement can be applied, hence[44]. From Magenta’s side, they serve as a focal point for its customers, coordinating requests and development efforts.

A transfer of local governments from the auspices of OS2 to Magenta is attributed to a period of inattention from the local governments involved in OS2’s governance structure. Today, the governance for OS2borgerPC has reportedly been revitalised, including a dedicated project coordinator helping to facilitate the collaboration.

Between the two factions, differences relate not just to financial matters, but also to issues of choice and autonomy in response to the governance of the solution itself. There are differing views over how contracts are procured and the existence, or lack thereof, of exit strategies for local governments that would like to change how the solution is leveraged, developed, or maintained[45]. Sambruk, by comparison, has a more straightforward relationship with Magenta, operating primarily through a service supplier-client model while ensuring open source principles are maintained[46]. In Sweden, governance is facilitated through Sambruk and its technical coordinators, who maintain communication with concerned members. The engineering requirements were coordinated with Magenta, and the necessary activities were procured accordingly.

The cross-border collaboration lacks formal governance structures, relying mainly on service supplier-mediated coordination. The unique tripartite relationship between OS2, Magenta, and Sambruk creates some complicated dynamics for the cross-border collaboration, given that there is no formal organisation or mechanism for facilitating it. One interviewee noted: ‘My intention [has been] that we should have a couple of meetings a year in order to synchronise what we are doing and what they are doing. That hasn't actually happened to a very large extent because [...] Magenta has been doing what OS2 has been saying to them to do. And Magenta has also been doing what we have said to them to do. And they have merged the code in there.’ [47]

In effect, in this relationship, Magenta becomes the hub for developing and maintaining OS2borgerPC, while the governance is managed in clusters of OS2 and Sambruk, as well as directly by single local governments through bilateral direct contacts with Magenta. This seems to be working for now, but there are questions about the long-term sustainability of the model, especially if local governments in other countries or jurisdictions become interested in developing their own business models for customising and maintaining the solution at the local government level.

1. Local governments should continuously check and review relevant projects to ensure compliance with open source principles and avoid potential lock-in effects.

The case of OS2borgerPC and MedborgarPC illustrates the real and sometimes challenging realities PSOs face when maintaining and improving open source code over time. In the case of OS2borgerPC, OS2 identified concerns at a late stage, prompting them to contract a second supplier for a code review[48]. This move led to longer-term tensions between Magenta and OS2, the main service supplier, highlighting the importance of transparent and recurring reviews to build mutual trust and prevent misunderstandings. Such practices also support a more balanced competitive landscape by encouraging accountability and ensuring software sustainability, while addressing risks of dependency on a single service supplier.

Local governments should generally strive to retain some technical expertise (internally or through collaborations such as OS2 and Sambruk) and introduce regular checks and reviews of open source projects to ensure adherence to open source principles and good practices in open source software engineering. Over time, this may help to improve code quality and prevent service supplier lock-in.

2. Procurement should consider market size and turnover in relation to the number of suppliers and service bundling, enabling sustainable business models and open source project maintenance.

A mutual understanding between local governments and service suppliers is critical to, on the one hand, ensure competitive prices and digital sovereignty, while on the other, ensuring profitable business models for service suppliers, and by extension, a healthy maintenance of the open source projects.

One factor to consider in this balance is the number of service suppliers that are actively procured by the local governments, in relation to the market turnover these local governments make up. After the use of a second service supplier to review the code base of OS2borgerPC and address a number of concerns, the supplier started providing hosting services for the software. From Magenta’s point of view, competition is welcome, yet it limits their margins, making it difficult to keep their engineers dedicated to developing and maintaining the project. A second factor to consider regards the bundling of services in the procurement from service suppliers. A separation between development and maintenance activities and operation, support and implementation-related services, as in the case of OS2, enables service suppliers to potentially focus on the latter only.

A consequence is that development efforts are concentrated on one supplier, Magenta, in the case of OS2borgerPC, who will have difficulty in upkeeping their engineers dedicated to the project, when development and maintenance are procured sporadically. Operation, support and implementation-related services provide a higher and continuous margin (based on the business model), enabling the funding of dedicated resources. By requiring service on both development and use from suppliers, local governments will also reduce the risk of lock-in effect further, while strengthening the sustainability by encouraging more investments into the project.

3. Local governments must decide whether to take an active or passive role in the governance and maintenance of their open source projects and engage service suppliers accordingly.

Robust and active governance and coordination are crucial for the sustainable management of open source projects. A period of inactivity from OS2's side related to OS2borgerPC led some local governments to engage directly with Magenta, bypassing community structures of the association[49]. This undermined collaborative governance within the association itself and created divisions within the community, which should have been bound to their participation in a collective.

To prevent this, local governments must maintain clear and active communication channels and uphold consistent governance practices. Quarterly meetings to discuss roadmaps are not sufficient, as maintenance is a continuous effort. An alternative approach is to outsource maintenance and lower-level governance to suppliers, while remaining active on high-level roadmap decisions and the strategic matters of the projects. The latter requires extra care in establishing transparent decision-making processes, including recurring reviews, and involving all stakeholders – local governments, service suppliers, and associations – can enhance trust and mitigate the risks of lock-in.

4. Including suppliers in the governance and planning process from the outset of a project promotes collaboration, awareness, knowledge-sharing, and synergies in development for PSO open source projects.

Irrespective of whether local governments must decide whether to take an active or passive role in the governance and maintenance of their open source projects, involving suppliers from the outset of open source projects has the potential to foster collaboration, awareness, and knowledge-sharing, leading to more efficient development cycles and better project outcomes. By integrating suppliers into governance and planning processes, open communication can be maintained and misunderstandings avoided. In OS2borgerPC, tensions emerged when local governments communicated directly with the service supplier because they were bypassing established governance channels.

The local governments should still ensure a safe and controlled space where suppliers are not involved. Such a space is needed to discuss procurement and higher-level strategies where they, as customers, can talk freely and without risk of influencing or favouring certain suppliers over others.

5. Coordinated co-funding and procurement enable a synchronised and consistent communication between users and service suppliers.

Coordinated co-funding and procurement are essential for aligning communication between users and service suppliers in open source projects. In the case of OS2borgerPC and MedborgarPC, OS2 and Sambruk implemented their respective co-funding models, where fees are determined by the population size of a local government or the number of instances used. This approach helps distribute costs fairly and encourages broader adoption of both solutions across local governments and libraries.

While direct procurements may be necessary to address specific development needs, as some local governments did with Magenta, even though they were part of the OS2 association, such arrangements can also cause silos and disrupt community cohesion if not properly synchronised. There also needs to be consideration of what role such business arrangements play in impacting development and maintenance across local governments. There could also be some debate over whether this approach scales across countries of different sizes, as opposed to Denmark and Sweden, where the populations are quite small.

In any case, what remains clear from the OS2 and Sambruk example is that a coordinated funding approach enables transparent communication and shared development goals, minimising conflicts and ensuring that the different stakeholders involved remain aligned around a common support model. While it will vary case by case, maintaining a balanced ecosystem should mean that local governments should be encouraged to collaborate through structured co-funding arrangements, promoting a unified approach to procurement and sustaining the PSO open source community.

6. Local governments should consider generalisability and localisation beyond local use cases early on to enable reuse, but typically require external funding to make this a reality.

To enable cross-border reuse, local governments sharing solutions and code – even via associations – should generally prioritise generalisability of solutions and the ability to localise them from the start. For MedborgarPC, localisation to Swedish and feature integration were relatively straightforward due to the underlying open source project's maturity (Ubuntu).[50] However, other local governments faced challenges due to high customisation costs, limiting the potential for reuse over time.[51] By way of comparison, a more modular design approach with native localisation support can facilitate broader adoption across local governments with varying cultural and administrative processes.

Conversely, excelling in localisation needs can lead to other challenges for open source development and reuse. For example, while localisation to Swedish from Danish was relatively straightforward, other cross-border use cases might require significant customisation, which might be too expensive for a local open source project, even one supported by an association. This demonstrates the need for additional funding sources beyond local budgets to localise solutions for broader use.

External funding, such as national or European grants, can support the development of modular designs and native localisation features, making the software adaptable to different administrative contexts. A funding approach that encourages the development of more customisable architecture that can cater to different local needs might be important if the use of OS2borgerPC continues to expand into other contexts. If it does, planning for generalisability and securing appropriate funding will allow future potential end-users in local governments to maximise the impact and scalability of their projects.