Krzysztof Siewicz, an Open Future 2023 fellow, looks into the regulatory requirements and potential challenges of establishing the European Public Digital Infrastructure Fund. Krzysztof’s work was supported by funding from Omidyar Network.
This release is the first public draft of the paper for comments and feedback. The deadline for comments is 12 October 2023, and we plan to release the final version of the paper shortly thereafter.
For at least a few decades already, practically every human activity utilizes or depends on digital infrastructures, especially when “infrastructure” is understood not just as hardware components of IT systems and networks but also and primarily as services and platforms located at higher levels of the whole digital stack and implemented in software1. In particular, many activities involving interaction or cooperation with other individuals, groups, or institutions have been increasingly performed with the aid of software. Hence, digital infrastructures have a broad social impact clearly exceeding any particular individual’s interest and self-ordering capacity. Even more so, given the susceptibility of the infrastructures to power imbalances, including user lock-ins2 that may be (and have been) easily designed into their particular implementations and that can be removed only with a lot of effort and coordination.
Therefore, issues such as the availability of software, its adaptability to individual needs, degree of interoperability, or reusability of data are of utmost importance for the operation of modern society. All these issues can be adequately addressed only if they are carefully included in the design, programming, testing, and deployment processes. Software availability, adaptability, interoperability, and reusability of data are much harder to manage if a given infrastructure was developed without such consideration, especially if it has been operational for an extended period. So, the earlier they are identified and included in the development, the better.
Indeed, it is also necessary to clearly understand the needs and perspectives of all involved parties. The infrastructures should be developed to include as many of these perspectives and to provide for as many of these needs as possible. This means that the whole process has to be flexible, inclusive, provide for good communication, and be well coordinated, especially to reconcile and mediate between some of the contradicting needs. Such an approach serves individual private interests and long-term societal goals3.
The term “Public Digital Infrastructures” has been proposed to denote such infrastructures that are developed, maintained, and governed by their respective communities with non-exclusive membership, instead of the rules being set by a single actor “owning” the tools that allow to unilaterally decide on the issues named above (P. Keller, op. cit.). Such an approach to digital infrastructures is instrumental for developing “Digital Public Spaces,” an interoperable ecosystem of services available to all users and maintained publicly under democratic values with a focus on the needs of the whole digital society (P. Keller, op. cit.).
Currently, a big part of the digital landscape is occupied by platforms whose primary purpose is profit, not any of the individual and social values mentioned above. There are areas, such as social media, where there is a clear dominance of a few platforms controlled exclusively by commercial entities. And while there is nothing wrong in operating for profit, the problematic issue is that the businesses are not designed to consider other values, not to mention prioritizing them. On the contrary, user lock-in, a threat to the proper operation of digital infrastructures from the point of view of long-term individual and social interests, is economically desirable for a corporate-owned platform, as it helps it to operate independently of the competition.
Noteworthy, numerous users of such platforms are entities pursuing broader social goals, such as various communities, non-governmental institutions, or public institutions. It means, in particular, that at least some individual users (when interacting with them) also use the platforms outside of their private interests. This further implies that the lock-in practices do not affect fair competition or consumer relations. Such platforms can exercise exclusive control over a big part of socially and publicly relevant interactions unless a viable, working alternative exists.
Shifting power to entities driven primarily by profit and not other values is hard to reverse. It constitutes an important factor to consider when proposing any regulatory interventions. In particular, this means that such interventions have to address the fact that “owners” of such platforms can and do impact the availability, adaptability, and interoperability of software and data reusability in a way that makes the satisfaction of user needs subject to their discretion. The more network effects are generated by the dominant platforms’ exclusively controlled infrastructures, the more difficult it is for individual users or competitors to contest their power4. By exercising this power to lock users in their products, proprietary platforms may prevent or slow down the adoption and usability of any alternatives, including Digital Public Spaces.
Network effects are an inherent feature of digital infrastructures, given their very nature, which is to facilitate communication between people. However, using network effects to lock users in a particular proprietary platform is not unavoidable. It is only a particular strategy serving particular interests. In other words, one could imagine a society benefiting from the network effects of various interoperable products but not at the expense of dependence on dominant platforms that do not take all important social interests into consideration in a democratic way. Actually, there are numerous examples of cooperation, not competition, for network effects within the Free and Open Source community that utilizes open standards5.
Certainly, to constitute a viable alternative to the existing dominant platforms, any measure adopted to develop Digital Public Spaces should address the interoperability issue with these platforms. Speaking more broadly, it should utilize every possible means of empowering existing platform users by leveling the power imbalance concerning the abovementioned issues. This means, in particular, that such measures should not be confined to developing public digital infrastructures only but should take further steps necessary to safeguard their truly democratic governance and practical capability of being used in parallel or as an alternative to proprietary platforms. Additional care should be taken so that while providing for the communication needs of society, Digital Public Spaces do not substitute the current dominance with a new one. In particular, they should not require anybody to use a specific application or to transfer data to any particular storage.
Therefore, although Digital Public Spaces can readily draw from the successful practices of Free and Open Source communities, they still need careful design, development, and maintenance. Additional care and effort must be directed to empowering users of existing dominant platforms, particularly by providing interoperability between the platforms and the Digital Public Spaces. Satisfaction of all conditions necessary for this should be linked to financing the Digital Public Spaces. It is thus instrumental to analyze the existing legal framework for possibilities of such financing and identify specific modes in which it could be provided.
In this paper, I analyze the proposal for the European Public Digital Infrastructure Fund (hereinafter “EPDIF”), as laid out by Paul Keller (op. cit.). I look into the regulatory requirements and potential challenges related to establishing such a fund, assuming that it would fund the development of Digital Public Spaces to address, in particular, the consequences of the existing dominance and reliance on proprietary platforms described above. Thus, I assume that it would provide funding to projects that (apart from satisfying the definition of a Digital Public Space) prevent or at least minimize the risk of new lock-ins and attempt to remedy the existing ones.
The EPDIF should, in particular, safeguard that a robust and democratic community develops the projects related to the development of Digital Public Spaces, that they adhere to open standards (and participate in an open-standard setting), and that the projects result in Free and Open Source Software implementations available for participating in the Digital Public Spaces.
Properly organized communities can provide ongoing support for the inclusion of many diverse perspectives and needs of different stakeholders, so the fund should be able to require that project governance and maintenance satisfy criteria relevant from this point of view. The maintenance has to be flexible, inclusive, provide for good communication, be well-coordinated, and oriented towards providing for diverse needs.
Compliance with open standards and availability of FOSS implementations may require additional explanations, provided in detail below.
Interoperability is necessary for Digital Public Spaces to succeed, not only in providing a viable alternative to currently dominant proprietary platforms but also in serving various stakeholders in the long run. Interoperability is provided for, in particular, through common interfaces, protocols, and data formats. To provide for the interoperability of a given system (or its component) with another system (or a component), one has to have access to interoperability information, such as the specification of a given interface, protocol, or data format.
Some specifications implemented in currently used infrastructures are publicly available, while others are privately held. Some may be used freely, while others require obtaining a license. Some are developed and controlled exclusively by specific individuals or groups, while others are maintained in a communitarian approach. Some have undergone a formal standardization procedure. Some are more popular than others or even become de facto “standards.” Using their availability for implementation as a criterion, one could sort the specifications from open standards (one extreme of the whole spectrum) to proprietary specifications (the other extreme).
Providing for interoperability with infrastructures designed according to private, exclusively controlled specifications without their holder’s cooperation requires legal measures such as, e.g., following decompilation rules and other exceptions to exclusive rights in the Software Directive or the essential facilities doctrine (especially as applied in the T-201/04 Microsoft). However, the application of such measures is risky and cumbersome. They work on a case-by-case and ex-post basis, so even if successful, their application often comes too late (after dominance and lock-in have already done their damage). In any case, the possible degree of interoperability is much smaller than possible if all stakeholders had cooperated on designing an open standard from the beginning6.
Certainly, legal measures aimed at disclosing closed specifications are essential and should not be abandoned. They should be utilized, especially if there is a dominant platform with which a Digital Public Space has to interoperate, and no cooperation with the platform's owner is possible. However, as a rule, Digital Public Spaces should use specifications developed publicly under the auspices of an institution with a non-exclusive membership and made available to everyone under non-discriminating conditions. This would allow everyone to have their perspective included in the process, and it would also allow everyone to implement these interfaces, protocols, or data formats in the software of their choice. Moreover, this would help to increase interoperability of the implementations. At the same time, it would increase the substitutability of general-purpose implementations and help implementers adapt their products to respond to users with specific needs without breaking interoperability.
In other words, Digital Public Spaces should be based on open standards. Although there is no single definition, open standards are usually understood as technical specifications adopted by an independent, transparent, and inclusive standard-setting organization (SSO), publicly available for implementation with no or minimal restrictions (cf., e.g., Free Software Foundation Europe, Open Standards, https://fsfe.org/freesoftware/standards/def.en.html)7.
Many SSOs develop standards important for digital communications. The Internet Standards Almanac published by Article 19, an international human rights organization focusing on freedom of expression and freedom of information worldwide, lists the following SSOs that are developing standards that have human rights implications (https://almanac.article19.org/):
World Wide Web Consortium (W3C),
Internet Research Task Force (IRTF),
Internet Engineering Task Force (IETF),
Institute of Electrical and Electronics Engineers (IEEE),
International Organization for Standardization (ISO, especially through its joint committee with the International Electrotechnical Commission (IEC),
International Telecommunication Union (ITU),
European Telecommunications Standards Institute (ETSI),
3rd Generation Partnership Project (3GPP),
Alliance for Telecommunications Industry Solutions (ATIS),
European Committee for Standardization (CEN), and
European Committee for Electrotechnical Standardization (CENELEC)8.
For example, SSOs differ considerably in their legal status and membership rules, which affect their independence, transparency, and inclusiveness. For example, membership can be subject to a fee (which may constitute a significant barrier for non-entrepreneurs), additional approval from a chair of a particular subgroup, or it can be limited to entities with specific statuses (such as only national standards bodies in ISO, or only governmental representatives in ITU, a UN agency, although in such cases sometimes other entities may apply for a position of an observer or similar). Undoubtedly, transparency throughout the process is crucial for the resulting standard specification to be designed to make it available for implementation under many different circumstances9.
Further differences between SSOs relate to the availability of their standard specifications. The documentation is not always publicly available, and some of the SSOs explicitly claim copyrights and even require a fee payment before granting access. Current case law confirms that standard specifications are protected by copyright law and thus denies that they must be treated as freely available public information (Public.Resource.Org (T-185/19)10. However, this judgment is pending under appeal (C‑588/21). In his opinion, in the appeals case, Advocate General Medina has recommended setting aside the judgment and granting free access to the standards in question.
An additional important source of restrictions to standard implementations may come not from an SSO but from the holders of patents that cover products manufactured according to the standard or processes necessary to implement or comply with the standard (standard-essential patents, SEPs). SEPs have been a known tool to game the standardization process for particular interests not specific to the IT industry11. Some SSOs attempt to address this issue by requiring members through their IPR policies to disclose their SEPs or publicly commit them for licensing to any interested party. Unfortunately, such policies can only affect participants in the standardization process. Also, even though they sometimes set a framework for patent licenses, the exact wording of such licenses must be negotiated between the individual parties.
Ideally, for a genuinely open standard, SSOs should not have any restrictions on membership to facilitate the standardization process inclusively and transparently, as well as making specifications publicly available gratis and under no copyright restrictions. There should also be no SEPs, or they should be available under a royalty-free license that does not restrict implementations in any way. However, in practice, many standards provide only a certain level of openness. Even according to some definitions of open standards, it suffices, for example, that any standard-essential patents have to be available under fair, reasonable, and non-discriminatory (FRAND) terms. For example, the European Interoperability Framework 2.0 states, “intellectual property rights to the specification are licensed on FRAND terms, in a way that allows implementation in both proprietary and open source software, and preferably on a royalty-free basis,” which appears as an attempt to balance the two approaches, by preferring the “royalty-free” but not ruling out proprietary patent licenses.
Given the vagueness of the term “open standard” and the current status quo, in which standards developed by even the most open SSOs feature a different level of openness, the proposed European Public Digital Infrastructure Fund should diligently analyze the circumstances of every particular project and either:
indicate a particular standard for the beneficiaries to implement after verifying its openness,
specify detailed openness criteria for selecting a standard by project owners themselves or
decide to direct the funding to the standardization process if no sufficiently open standards are available.
A widely recognized open standard could be mandated for projects implementing a well-known, general-purpose feature. In more complex projects, such as social networking services, there are likely multiple solutions at different stages of development and maturity12. In such a case, a standard, or a set of standards, could be chosen after a careful technical analysis that should be performed within the project itself, but according to the fund’s specified openness criteria and under the fund’s supervision (which may include an independent audit). However, such standards should not be treated outside of the project. The fund should give project maintainers resources to engage in the standardization process, including the perspectives and needs of the community the project is expected to serve13.
Finally, in the case of projects aiming to interoperate with existing dominant platforms, there might be no open standards available. Such platforms frequently use interfaces, protocols, or data formats developed internally, with no complete specification available to outsiders. So, it might be prudent to direct the funding to the beneficiaries’ efforts to obtain interoperability information from the platform. Then, funding should be provided to facilitate the standardization process so that the resulting standard is open. One could imagine many different forms of sponsoring open standardization processes by the EPDIF. For example, it may provide financing for elaborating or implementing an SSO’s IPR policy. It may provide financing for implementers for patent investigations, patent revocation proceedings, or patent licensing. Or, financing could be directed at a beneficiary’s involvement in elaborating interoperability measures that gatekeepers must offer under the Digital Markets Act. Such financing would aid smaller stakeholders in safeguarding that a gatekeeper’s API is well documented and adjusted to their needs, not just to the needs of other gatekeepers or others who already hold resources necessary to implement a measure designed in a probably not too inclusive manner.
The requirement to adhere to open standards (or as open a specification as possible under particular circumstances) introduced as a condition of the proposed funding is necessary but insufficient for developing Digital Public Spaces. Open standards can be implemented in closed, proprietary software that is not readily available to anyone and cannot be easily adapted without additional cooperation from the software producer. There is a need for implementations in software developed in democratic communities. Thus, Digital Public Spaces require working Free and Open Source Software (FOSS) implementations to prevent software from being used as a leverage to control users and communities because the public availability of software source code without constraints is an inherent feature of FOSS, making it possible to develop it in a communitarian way.
From the formal point of view, FOSS is software distributed under terms compliant with the Free Software Definition14 and the Open Source Definition15, which require the availability of source codes and enable anyone to run, modify, and distribute programs with or without modifications, in any form and for any purpose. The mere availability of any software (free or proprietary) is insufficient for it to be useful, interoperable, and adequately maintained. The additional necessary condition is that it is developed and maintained considering the diverse needs of all users, with a particular focus on interoperability. In practice, this condition is met in the case of FOSS projects developed by robust and well-organized communities (see, e.g., J. Bacon op. cit.). Therefore, the EPDIF should direct its funding to all possible beneficiaries: entrepreneurs, non-governmental organizations, public institutions, or even individuals, as all these types of entities are involved in developing Free and Open Source Software in such communities. At the same time, the funding should be directed to democratically governed projects with transparent and inclusive development processes.
In the European Union, public funding can be distributed in several ways. Funds allocated in the EU budget may be managed directly by the European Commission, managed together by the Commission and national authorities, or their management may be entrusted to a selected organization. Member states may establish their funding programs individually or in interstate cooperation after following state aid rules. Additionally, specific products or services can be ordered under public procurement rules at the EU and the member state levels.
A relevant example of EU funding directly managed by the Commission is the “Digital Europe Programme” (https://digital-strategy.ec.europa.eu/en/activities/digital-programme). It has been established under Regulation 2021/694 to “support and accelerate the digital transformation of the European economy, industry and society,” it is subject to the general rules of the Financial Regulation (Regulation 2018/1046). Implementing the Program is entrusted to European Digital Innovation Hubs that stimulate the adoption of digital technologies by serving as access points, promoting open innovation, and providing support, for example, by coordinating educational activities.
The Commission allocates the funding to specific work programs launched every two years. Actually, in the Program, provisions are already aimed at ensuring interoperability and encouraging FOSS. For example, the Program’s “Specific Objective 5 – Deployment and Best Use of Digital Capacities and Interoperability” includes support “for the integration and use of trans-European digital service infrastructures and of agreed European digital standards in the public sector and in areas of public interest to facilitate cost-efficient implementation and interoperability,” as well as “the development, update, and use of solutions and frameworks by public administrations, businesses and citizens, including of open-source solutions and the re-use of interoperability solutions and frameworks.”
The funding is provided in the forms specified in the Financial Regulation, particularly through procurement, grants, and prizes (art. 14 of the Digital Europe Programme Regulation). The list of the funded projects is available at https://ec.europa.eu/info/funding-tenders/opportunities/portal/screen/opportunities/projects-results;programCode=DIGITAL.
The provisions of the Digital Europe Programme explicitly require that its activities are coordinated with similar activities under Horizon Europe. EPDIF may be introduced similarly. It could start as a specific work program under the Digital Europe Programme Regulation. The Commission prepares multiannual work programs (typically two-year, although annual programs are also admissible). The scope of the programs is framed by the technical annex to the Regulation and the specific objectives set out in it (Rec. 65-66). Work programs are adopted in the decisions of the Commission under art. 110 of the Financial Regulation, art. 24 and 30 of the Digital Europe Programme Regulation. Before adopting such a decision, the Commission is required to consult experts delegated by each Member State. After adoption, decisions must be notified to the Parliament and the Council and enter into force if these institutions do not object within two months. The Digital Europe Program Coordination Committee assists the Commission in the process.
The following work programs are currently underway:
DIGITAL Europe Work Programme 2023-2024
DIGITAL Europe - EDIH Work Programme 2021-2023, specifically dedicated to the European Digital Innovation Hubs,
DIGITAL Europe - Cybersecurity Work Programme 2023-2024,
DIGITAL Europe - High Performance Computing (managed by the EuroHPC Joint Undertaking)16
The “Digital Europe Programme” is scheduled until 2027.
For other existing programs in the digital area relevant from the point of view of the proposed EPFID, see, e.g.,
Horizon Europe17. Its 2022-2023 work program includes the Next Generation Initiative, with one of the topics being coordination and support action in the NGI Commons Policy. The NGI was launched in 2018 to increase interoperability on the Internet while observing the fundamental European values of openness, inclusivity, transparency, privacy, cooperation, and data protection. A few hundred FOSS projects have been funded by the NGI so far.
the European High Performance Computing Joint Undertaking18. Established in 2018 and currently operating under Regulation 2021/1173, it is a public-private partnership for implementing parts of such EU programs as Horizon Europe and Digital Europe. Its members are the EU, member states, and third countries associated with EU programs and associations of entities active in the research and innovation, HPC, and quantum computing areas. Its mission includes the operation of a world-leading federated, secure, and hyper-connected supercomputing, quantum computing, service, and data infrastructure ecosystem. While the subject matter of the EHPC JU is not directly related to the operation of the proposed EPDIF, experience and good practices of its coordination mechanism might be useful.
the proposal for the Regulation establishing the Strategic Technologies for Europe Platform (‘STEP’) (https://commission.europa.eu/system/files/2023-06/COM_2023_335_1_EN_ACT_part1_v11.pdf), as an example of leveraging existing funding programs (including, e.g., cohesion policy instruments, Innovation Fund, European Innovation Council) by providing for their bigger flexibility, strength and creating synergies. STEP would focus on deep and digital technologies, clean technologies, and biotechnologies. STEP’s experience in funding selected technologies with a mixture of existing funds and excellent practices in coordination and creating synergies should be used to operate the proposed EPDIF.
Member states that wish to fund undertakings on a selective basis (i.e., outside of general measures open to all entrepreneurs but directed to specific companies, sectors, or regions) should only do so in a manner compatible with the internal market. Under TFEU art. 107.1, state aid which distorts or threatens to distort competition by favoring certain undertakings or producing certain goods is incompatible with the internal market (if it additionally affects trade between member states). Art. 107.3 contains a list of types of state aid that may be considered compatible with the internal market, which includes “aid to promote the execution of an important project of common European interest” or “other categories of aid as may be specified by decision of the Council on a proposal from the Commission.”
Generally, any new state aid measure must be notified by a member state to the Commission before it is put into operation. Notification is not necessary only for state aid that meets the criteria of the General Block Exemption Regulation (651/2014), de minimis state aid as specified in Regulation 1407/2013 (which requires that the aid granted to a single undertaking does not exceed EUR 200 000 over any period of three fiscal years), or aid provided under an already authorized scheme. Certainly, there would be projects related to the development of Digital Public Spaces that fall, e.g., into the research and development category (GBER art. 25) or innovation (GBER art. 28, art. 29), or that do not require funding in the amount exceeding the de minimis threshold. The member states could fund these without notification.
As an example of the notification track, Regulation 2021/1060 sets a framework for the member states to notify various programs authorized by the Commission in a decision and then managed by relevant state agencies. It was used, for example, by the Polish government to establish the “European Funds for Smart Economy” program (https://www.nowoczesnagospodarka.gov.pl/media/113927/FENG_zaakceptowany_przez_KE_ang.pdf), which provides funding for, among others, innovation development by increasing the level of cooperation between enterprises, technology transfer, digital transition. Hence, a similar program could be notified by a member state that would like to provide funding for Digital Public Spaces, as discussed in this paper, at the national level. Conversely, if EPDIF is set up at the EU level (e.g., as a work program under the Digital Europe Programme Regulation, or as an EDIC – see below), its activities should be coordinated with the funding provided under such national programs as the Polish one mentioned above.
There are also guidelines that help member states design their programs, such as the Communication “Framework for State Aid for research and development and innovation (2014/C 198/01).” These particular guidelines invoke TFEU art. 179 that envisages a European research area “in which researchers, scientific knowledge and technology circulate freely…” and aims to address market failures of too low R&D&I due to market actors not considering positive externalities (items 2 and 4 of the guidelines). The guidelines continue to lay down conditions of compatibility with the internal market of, among others, “aid for innovation activities” (which includes “obtaining, validating and defending patents”) and “aid for innovation clusters” (which includes “supporting the investment in open and shared infrastructures”). The criteria are specified in item 36 of the guidelines as follows:
contribution to a well-defined objective of common interest,
need for state intervention,
appropriateness of the aid measure,
proportionality of the aid,
avoidance of undue negative effects on competition and trade between Member States, and
transparency of aid.
The exact interpretation of these criteria is provided in item 40 et seq. of the guidelines. Nothing there would prima facie prevent providing state aid to projects that contribute to delivering FOSS implementations of Digital Public Spaces components compliant with open standards developed by democratic communities. The practical application of these guidelines might perhaps require some more exploration, as they are driven mainly by the paradigm of “innovation” perceived as a breakthrough, individual activity instead of an incremental, communitarian process as is more the case with the development of FOSS (or Digital Public Spaces in general). From this point of view, for example, “innovation activities” (“obtaining, validating, and defending patents”) should include challenging standard-essential patents (or obtaining patent licenses under open standard-compliant terms). Also, “open and shared infrastructure” should encompass the application layer developed transparently by democratically governed communities19.
The German Federal Ministry of Economics and Climate Protection has recently started an interesting scheme for financing the development of FOSS. It has set up the Sovereign Tech Fund, formally operated by its wholly-owned limited liability company SPRING GmbH. (the project’s website: https://sovereigntechfund.de/de/)20. It has been incorporated after a feasibility study, which found out that the Open Source ecosystem supports decentralization and independence in software development, prepares a “breeding ground” for innovations, stimulates further training and quality standards in the community, and increases the attractiveness of Germany as a strong player in the digital realm. It envisaged broad funding availability for individuals, organizations and communities, companies, and service providers to develop FOSS-licensed, basic infrastructures of the internet and software development and technologies that increase decentralization and independence. After considering different models for the funding, the “one-stop-shop” was selected, which combines various funding instruments admissible within the EU state aid law and provided for in Sections 23 and 44 of the Federal Budget Code21. This means, in particular, that each specific funding provided by the Sovereign Tech Fund has to either fall under such exceptions as the GBER or de minimis or undergo notification procedures. At the moment of this writing, the fund is already in its post-launch phase, implementing experience gathered during the pilot round22.
Another interesting measure has been developed for the cooperation between member states in the form of EDICs, i.e., the European digital infrastructure consortia has been introduced under Decision 2022/2481, establishing the “Digital Decade Policy Programme” 2030. Their purpose is to achieve the general objectives of the program, which include, for example: “promoting a human-centered, fundamental-rights-based, inclusive, transparent and open digital environment where secure and interoperable digital technologies and services observe and enhance Union principles, rights and values and are accessible to all, everywhere in the Union,” as well as other objectives very much in line with the goals for the EPDIF already discussed above. When implementing the program, Member States draw their national roadmaps, which they then consult with the Commission.
However, the multi-country projects deserve particular attention here. EDICs are one of the envisaged implementation mechanisms for these projects. They are separate legal persons in which the involved Member States are represented by one or more public entities or private entities with a public service mission. Setting an EDIC requires application to the Commission, which includes, in particular, the draft of its statutes and a technical description of the multi-country project. After the positive assessment of the application, an EDIC is set up in the Commission’s decision. At least three member states are required as members of an EDIC and membership with voting rights requires a financial or non-financial contribution. Non-contributing member states can be observers without voting rights. EDICs are open for additional member states under terms specified in the statutes. The statutes may allow third countries, international organizations of European interest, and public or private entities to become members of an EDIC. Still, in such a case, member states have to jointly hold the majority of voting rights. EDICs are governed by an assembly of members. The assembly includes the Commission, but without voting rights save for a right of veto against decisions related to actions financed under centrally managed EU programs. The assembly has full decision-making powers, including on the EDIC’s budget, which means that spending is agreed on by the members in the framework of a given multi-country project (and subject to the rules of the applicable centrally-managed EU program). EDICs also have a director appointed by the assembly as the executive and legal representative. Statutes may provide for additional bodies of an EDIC. Statutes have to provide for, among others, a detailed description of the multi-country project, member tasks and timeline, the EDIC duration, members' voting rights, and rules of ownership of assets (including intangibles). As a legal person, EDIC is liable for its debts, and members’ liability is limited to their contributions, but statutes may extend their liability. EDICs are required to produce annual reports that must be approved by the assembly, transmitted to the Commission, and published. The Commission may provide guidance regarding the matters covered in the report.
Given the close alignment of the goals of the Digital Decade Policy Programme with the goals of the proposed EPDIF, using EDICs as a vehicle to introduce the funding as discussed above is an excellent option for a coordinated activity of member states, supplementary to establishing EU-wide work programs, such as under Digital Europe Programme discussed in the preceding chapter.
As an alternative to, or in parallel to, funding independent projects that develop components of Digital Public Spaces, the EPDIF could procure specific implementations that would be then deployed and maintained under the auspices of the public institution or using separately procured third-party services. The EPDIF would then operate according to the rules of public procurement. These rules have always very clearly required equal treatment of all economic operators, non-discrimination, transparency, and proportionality.
Art. 18.1 of Directive 2014/24 on public procurement prohibits the artificial narrowing of competition through unduly favoring or disadvantaging certain economic operators. This general rule is further specified, in particular in Art. 42 of the Directive laying down rules for drafting technical specifications. Art. 42.1 allows the drafters of the specifications to refer, e.g., to a specific method of production provided it is linked to the subject matter of the contract and proportionate to its value and objectives. It also allows to specify whether the transfer of intellectual property rights will be required. Art. 42.2 requires that technical specifications afford equal access to economic operators or shall not result in creating unjustified obstacles for competition. Moreover, under art. 42.3.b it is clear that procurement may reference standards.
The above-mentioned provisions provide a legal framework for public procurement that would not discriminate against any potential supplier. At the same time, they allow to specify conditions for the ordered products or services to fit the public institution’s needs. So, it is possible that these conditions can be met by only particular suppliers as long as they have an objective justification. For the development of Digital Public Spaces, these needs include the delivery of FOSS implementations compliant with open standards and developed by democratic communities that are transparent and inclusive. Requiring compliance with such conditions would not violate the said rules as long as the procured products are technologically neutral. Firstly, because public availability of FOSS with permissions for broad reuse significantly increases the potential base of redistributors and reusers (i.e., releasing procured software as FOSS allows everyone, including entrepreneurs, to reuse it, for example, by offering their own products based on it). Secondly, formulating conditions of technological neutrality (obtained in particular through open standards and by transparent and inclusive development) would further guarantee that nobody is discriminated against in their capacity to compete for procurement and against using and reusing procured products.
Referencing standards in public procurement for developing open infrastructures has been recognized as an essential tool for a long time. In the communication “Against lock-in: building open ICT systems by making better use of standards in public procurement” (COM(2013) 455 final), EU institutions confirm that lock-ins to proprietary systems are responsible for EUR 1.1 billion annual losses for the public sector and call for better use of standards to remedy this situation. It has also been observed that lock-ins lead to referencing brand names in future procurement documents, thus permeating the problem. While the use of standards is recommended in communication as a remedy, it has also been observed that procuring authorities have difficulties in selecting standards that fit their needs. Therefore, the Commission has prepared the “Guide for the procurement of standards-based ICT — Elements of Good Practice” (SWD(2013) 0224 final), which provides specific aid to the authorities. It envisages a centrally available repository of knowledge and agreements, lists of recommended standards, etc. It also envisages public authorities planning their ICT procurement by describing their needs, long-term plans, and budgetary choices. Finally, it turns to drafting particular procurement documents in a way that does not result in future lock-ins. The Guide continues with specific recommendations for each problem area under the “What to do?” and “Who should act?” headings. For example, it includes ready-to-use templates for the most important clauses in procurement documents and contracts.
The Guide mentioned above is not the only source of information that can be used to develop Public Digital Spaces through public procurement. The additional ones include Communication “Setting out the EU approach to Standard Essential Patents” (COM(2017) 712 final) or papers written by Björn Lundell and Jonas Gamalielsson, who analyze work practices of public-sector organizations and develop guidelines for implementation23.
There are also numerous examples of the practical application of the rules of public procurement for commissioning infrastructures based on FOSS, and the Barcelona City Council technological sovereignty guide provides an excellent summary of recommendations and best practices in this area while at the same time citing various national law provisions regulating the details of IT procurement24. Examples of coordinating the use of standards include Nordic Smart Government and Business project, aiming at “making real-time business data accessible and usable for innovation and growth” (https://nordicsmartgovernment.org/), or Slovenian National Interoperability Framework (https://nio.gov.si/nio/cms/page/links?lang=en).
Digital Public Spaces can be developed with the help of public financing. This does not require introducing specific measures but just applying the already working schemes, some of which already provide for using FOSS and open standards. In this paper, I demonstrated that the financing for projects developing components of Digital Public Spaces could be distributed by an EPDIF established within the existing framework of public funding at the level of the EU and the member states. Given the international coverage and importance of Digital Public Spaces, cooperation and coordination of institutions operating at all these levels should be encouraged and performed whenever possible. To this end, EDICs can be established as introduced in Decision 2022/2481.
The funding of Digital Public Spaces deserves a work program within the Digital Europe Programme, which already covers areas close to the development of Digital Public Spaces. After such a work program is adopted in the Commission’s decision, it should certainly be coordinated with other similar funding activities in the EU.
At the level of individual member states, funding can be provided under the rules of state aid, especially for projects falling into GBER or de minimis exceptions. Still, using the notification track for projects meeting the criteria necessary for developing Digital Public Spaces is also possible. Notified state aid schemes should pay particular attention to the Commission’s guidelines and interpreting such concepts as “obtaining, validating, and defending patents” or “shared infrastructures.” Also, we should observe the German Sovereign Tech Fund’s operation closely and pay attention to their good practices and experience in this area. Moreover, any such funding activities should be accompanied by coordinated public procurement, which should require FOSS and open-standards-compliant solutions, per the detailed guidelines in this paper.