Open science and commoning beyond licensing: Deteriorating digital commons in the absence of collective governance

This post is an adapted/expanded version of an abstract that I have submitted for a track onOpen, collaborative and participatory science: rethinking knowledge legitimacy, policy and science futures at the EU-SPRI Annual Conference

Over the past decades, the open science movement – as well as the many wider, nonacademic, open knowledge movements – have left their marks by building large, open knowledge commons: In the world of academic publishing, the open access movement has helped make the scientific literature a lot more accessible.1 Similarly, the use of open licenses for the publishing and openly sharing data and software has become the norm across increasing numbers of scientific disciplines, creating rich forms of scientific commons too.

Similar to the successful digital commons that can be seen outside the narrower realm of academia – be that Wikipedia, OpenStreetMap, or the many Free/Libre & Open Source Software (FLOSS) projects – these academic commons also rely on – and thrive – thanks to countless contributions made by individuals, be that as part of their actual job or as yet another (semi)-expected academic community service. And of course, we have seen a push towards more participatory research models in the different forms of citizen science, where people who are not working scientists are asked to engage in knowledge production to contribute to commons in different ways.

But, despite these moves towards more participation that are typically touted as “democratizing”, one of the main focuses when looking at these digital commons has been around the types of open licensing that are applied to the content that is created within them. This is true in academic settings around the open data and open access, as well as in the FLOSS spaces, where lobby organizations like the Open Source Initiative or the Free Software Foundation engage in policing of license uses. And while there is a (maybe growing) recognition that just adding an open license does not make or sustain a commons, interventions have often centered around community building for growing the number of commons contributors or community-management that seem to follow a model of pastoral care of emotional support from the top-down. That is to say, without seriously challenging or addressing the issues around sharing power and control that are inherent to commons that are only possible thanks to its contributors but who often might have limited to no decision making power beyond individually deciding whether to contribute or not.

Such disconnects between those who govern a commons (e.g. academic institutions, for-profit companies, closed-off organizations, or individuals) and those who are helping to build it can - and frequently does - lead to conflicts in these open knowledge spaces. Frequently, the use of open licenses is held up as offering a way out of these conflicts, attempting to justify the focus on licensing: The argument being that open licenses allow – at least in theory – “forking” a commons from one regime of control into another one, for example where the existing “powers that be” are seen as illegitimate. But, the reality of taking this “nuclear” option is vastly more complex. Due to a number of interrelated factors, and precisely because openly licensed content is only a small part of what makes for a healthy & dynamic commons.

To sustain, let alone further grow, a digital commons, copying the contents of the commons is not enough. It also takes the community of contributors itself, thus generating a collective action problem to organize the “forking” of not just the content but also the existing network of relationships. Organizing such a move is made more complex by the fact that the existing bodies that govern such a digital commons are often those with ultimate control over both technical infrastructures for organizing as well as financial resources to maintain it. Collectively, these factors can present substantial barriers to “forking” a knowledge commons when the contributor-base find themselves confronted without adequate mechanisms for taking an active role in decision making.2

Of course, the idea that participatory decision making is important for a commons is not new. It is one of the key principles that Elinor Ostrom identified for organizing successful, long-enduring commons. But despite this, in academic as well as non-academic digital commons, governance models most often reproduce the same admin/user dichotomy that is used in the underlying software infrastructures in an inversion of Conway’s law, leaving contributors little if any formal control about the institution that is in charge of the commons.

Among FLOSS projects, one common example of this model is the Benevolent Dictator(s) for Life (BDFL) model for organizing projects. Here, projects are either through single individuals (or small groups of dictators) with full control. While rooted in the way software projects tend to grow beyond their initial creator(s), the limitations of this model regularly rear their head: Forgejo – the open source software forge that is used by Codeberg among many others – is a fork of the Gitea project. This fork came to be after the Gitea lead maintainers moved the Gitea operations into a for-profit company, and towards an open core model. And already Gitea itself was a fork, from the earlier Gogs project, which (somewhat ironically) was started to avoid having a single BDFL. Such a chain of forks is not unique: Last year, the CoMaps project started as a fork from the Organic Maps project as existing volunteer contributors disagreed with the leadership that organized itself in a for-profit company.3 Organic Maps in turn started out as a fork too, after the earlier Maps.me was sold off and made closed source4.

It might look like forking works as advertised to resolve disputes, but the reality is that forking often is a good “dispute resolution” in the same way as “move to a different country if you don’t like it here” is a good way to solve national political issues. For one, it ignores the cost to both the contributor communities as well as the users of the software. Between the fragmentation of contributors between projects, it also requires substantial overheads for the forks to establish their own technical and social infrastructures. And while forking is often presented as a last resort that should make those costs acceptable, the BDFL model means that typically there are no mechanisms to implement graduated sanctions against the BDFL as power flows only in one direction. Despite these shortcomings, moving from BDFL models to more collective models remains an exception. Python’s Guido van Rossum stepping down from the BDFL post is a famous example. And a more recent one is Mastodon-creator Eugen Rochko leaving the BDFL model in favor of giving control to a non-profit organization.

While moving towards a non-profit structure can be an improvement for more community-based governing, this is far from guaranteed. Maybe the most striking example of this comes from the Mozilla Foundation , the non-profit with a complex legal structure between the foundation and fully owned for-profit subsidiaries, that nominally is in charge of the Firefox browser. It consists of a self-perpetuating leadership that continues to be criticized for its overpaid highly-paid leadership that is out of touch with its user community, replacing their volunteer translators with ‘AI’, while also axing features people liked in favor of unpopular “AI” integrations and wanting to get into the advertising market.

Mozilla is far from unique, neither in its lack of understanding for volunteer motivations nor in the lack of giving its contributors any power. In the world of citizen science, iNaturalist ’s partnership with Google for an ‘AI’ project sparked a backlash. As volunteer contributors did not have a way of effectively influencing the self-perpetuating board of the non-profit that runs iNaturalist, some contributors felt that “voting with their feet” was the only option open to them, leading to them stopping to contribute or even deleting their existing contributions of observations. A few months later, iNaturalist co-founder Ken-ichi Ueda published that he had left iNaturalist over disputes in which direction the non-profit should take. Among other things, he specifically outlines his view that the organization should move towards having a board that is community-selected and giving the contributors more power. Similarly, as Bradley Kühn chronicles, the Open Source Initiative ’s “member-based elections” for the board – which are only advisory and subject to approval by the existing board – broke down the moment any community candidate actually came close to challenging an existing dogma, ultimately leading to the organization suspending its elections in 2026.

All of these examples5 demonstrate the relevance of adequate pathways for contributors to have power to influence decision making in digital commons, across a wide range of different types of knowledge commons. To me, this points to the fact that it is not about if a commons will become brittle in the absence of shared decision making, but when. The question is: How can we do better?

Different digital commons have found different answers to that: After moving on from the BDFL model, Python adopted a model of ‘core contributors’ who can induct more core members into their circle and elect a steering committee. Both OpenStreetMap and Codeberg operate as “traditional” member-run non-profits that have an open-enrollment membership that gives voting rights to those members, for both leadership positions and concrete issues. The _Open Bioinformatics Foundation_6 operates on a hybrid model, with a membership that votes on concrete issues, while the board is mostly self-perpetuating. Inspired by Wikimedia’s community seats, the _Open Humans Foundation_7 used to have community-elected board seats when the community was more active.

When it was more active hints at one of the big challenges with implementing such approaches: Organizing more democratic and inclusive decision-making approaches takes work– and often it is labor that does that might not seem urgent or not feel particularly rewarding. This is especially the case when a commons is just starting to form, there are still few contributors and potential conflicts seem still far away in the future. As such, it can become easy to procrastinate the building out of decision making mechanisms while everything still goes well. And by the time there is urgency, due to the emergency of disputes, it can be too late. Especially if there was never any effort to think through decision making, e.g. about what types of decisions need to be made, how to make them and who gets to be involved. To avoid this, learning how to organically grow the governance in an ad hoc fashion alongside with a commons can help train those “decision making muscles” while avoiding over-defining things.

To avoid that our commons become brittle and deteriorate, everyone should take an interest in how they operate and see how/if one can be involved in steering it, beyond just looking at the presence of an open license. In an ideal world, there would not just be many different guides for how to pick an open license for code or ‘content’, but also similar advice for how to pick and implement decision making models for your commons to support projects. In addition to supporting new projects, it could be a lobbying instrument vis à vis organizations that stewards a commons, but still lack participatory decision making models. Otherwise, there will be many more days where the only way for reclaiming a commons is a painful fork.

References

1. Strasser BJ, Tancoigne E, Baudry J, Piguet S, Spiers H, et al. (2023) Quantifying online citizen science: Dynamics and demographics of public participation in science. PLOS ONE 18(11): e0293289. https://doi.org/10.1371/journal.pone.0293289
2. Greshake Tzovaras, B. (2025, June 12). Generative ‘AI’ in citizen science: the iNaturalist backlash. Bastian Greshake Tzovaras. https://doi.org/10.59350/7ykzb-dkx88
3. Greshake Tzovaras, B. (2025, June 18). Power-sharing in online communities. Bastian Greshake Tzovaras. https://doi.org/10.59350/jpevj-h4y27
4. Greshake Tzovaras, B. (2025, March 31). Sunsetting openSNP - a personal retrospective. Bastian Greshake Tzovaras. https://doi.org/10.59350/8se4t-tmx68
5. Herb, U. (2025). Open Access Publishing Patterns 2015–2024: Global Article Output and Modal Shifts [Data set]. Zenodo. https://doi.org/10.5281/zenodo.15423474
6. Ostrom E. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press; 2015.

Footnotes

1. That is, at least more accessible for reading, high and rising article processing charges have made publishing in open access journals that are run by commercial publishers substantially harder for many researchers—Another example of non-community governed commons. ↩

2. Or rather an active and binding role, as participation washing is all-to-common. ↩

3. A brief conflict of interest -note: I’ve started volunteering with CoMaps since after they forked. ↩

4. The timeline and circumstances reported about the Maps.me/OM fork are unclear and vary, depending on who of the involved retells that story. ↩

5. These examples are far from an exhaustive list. The Wordpress drama and OpenTofu would be other examples, involving corporate-control issues. ↩

6. A note for transparency: I have been on the board of it for many years and still am. ↩

7. Another note for transparency: I am also a member on that board. ↩