How is “Core Software” governance managed in OpenStreetMap, and who decides which projects are included? - General talk - OpenSt

How is “Core Software” governance managed in OpenStreetMap, and who decides which projects are included? - General talk - OpenStreetMap Community Forum
How is “Core Software” governance managed in OpenStreetMap, and who decides which projects are included?
General talk
website
software
osmf
github
core-software
AngocA
(Andres Gomez Casanova)
June 19, 2025, 11:48pm
We are currently witnessing the emergence of new initiatives aiming to complement or replace long-established software components in OSM. This evolution is natural and even necessary—but it raises important questions about governance, continuity, and inclusion.
OSM’s software stack has grown organically over the years. Some components remain critical but rely on just one or two maintainers, with little transparency around succession or decision-making processes. To move forward sustainably, we should address a few key areas:
Decision-making and Core Software inclusion
Who decides what becomes part of the “Core Software” of OpenStreetMap?
What are the inclusion criteria? Are they documented anywhere?
What is the role of the OSMF, the Operations Working Group, or the Software Dispute Resolution Panel?
Is there a process for proposing, evaluating, and incubating new projects, or for replacing outdated ones?
Many other parts of OSM have developed well-defined governance (e.g., board elections or tag approvals). But software governance remains largely implicit. Without clear guidelines, it’s difficult for new contributors to engage meaningfully or for innovation to scale beyond individual effort.
Succession and generational transition
Many key tools have been maintained by the same individuals for over a decade. Their contributions are invaluable, but what happens when they step away? The project risks losing knowledge and momentum.
We should:
Broaden the pool of maintainers for each critical project.
Encourage knowledge transfer and onboarding.
Develop transparent pathways for others to take on responsibility—
before
burnout or disengagement occurs.
Technological sustainability
Some critical software projects are written in languages that, while powerful, are no longer among the most widely used. For example:
Ruby was once cutting-edge, but its popularity has waned considerably.
Meanwhile, Python is one of the most widely used languages globally but has limited presence in OSM Core Software.
We should ask:
Are we building on languages and technologies that will still attract volunteer developers 10 or 20 years from now?
How do we avoid accumulating unmaintainable technical debt due to decisions made decades ago?
Reference language popularity trends:
TIOBE Index (2025)
Statista Developer Survey (2024)
The role of the OSMF and its limits
The OSM Foundation supports but does not govern the project’s technical direction. Its
Strategic Plan
outlines priorities but says little about long-term software infrastructure.
Some reflections:
Software funding is available through OSMF, but there’s no clear process for evaluating competing technical proposals.
The OSMF Board chooses members of the Software Dispute Resolution Panel, but community input is limited.
For a democratic project, there is surprisingly little democratic structure for making decisions about the software that powers it.
Transparency in maintainership and GitHub governance
Who decides whether a pull request is accepted or ignored?
Should “Core Software” projects be under the
openstreetmap
GitHub organization by default?
Should there be written expectations for the maintainers of these projects—including community responsiveness and documentation?
Forking is always possible, but for many projects, being part of the “core” carries recognition, visibility, and potential access to financial support. If the inclusion process isn’t transparent, we risk discouraging innovation.
Community voice and inclusion
We need better paths for the community to express opinions:
Not everyone can evaluate legal or architectural decisions—but everyone should be able to leave feedback on competing approaches.
For multilingual participation, all Core Software should include robust translation workflows and avoid English-only blockers (e.g., Welcome Mat shortcomings).
There should be safe, constructive ways to challenge the status quo—
without being dismissed as “troublemakers.”
Conflicts of interest
Some members of the Operations Working Group are also maintainers of Core Software. This makes it hard for the group as a whole to make impartial decisions about alternative or successor projects.
Could a more neutral steering committee or selection process help mitigate this?
A suggestion
Perhaps it’s time to explore:
A transparent proposal and evaluation framework for new Core Software tools.
A clear set of criteria for inclusion, covering technical soundness, community adoption, maintainership, translation readiness, and long-term sustainability.
A process for community feedback—at least advisory—in decisions about future direction.
This is not about criticizing what’s been done. The past 20 years have brought amazing growth in both the OSM map and its technical ecosystem. But to move forward, we need to imagine OSM in 2045—when few of today’s maintainers may still be active—and plant seeds accordingly.
What are your thoughts? Is there appetite to build a more structured but still community-driven approach to Core Software governance?
12 Likes
Why are we creating a design system for OSM?
SimonPoole
(Simon Poole)
June 20, 2025, 5:40am
AngocA:
We are currently witnessing the emergence of new initiatives aiming to complement or replace long-established software components in OSM.
Are we actually?
With the exception of OSM-NG I haven’t seen any and that is a non-starter as an actual replacement for numerous reasons that don’t need repeating here (but could have been avoided).
For a democratic project, there is surprisingly little democratic structure for making decisions about the software that powers it.
That seems to be a misunderstanding, while the OSMF does have some democratic decision making processes, the major part of OSM does not. In particular none of the software projects associated with OSM have that outside of for the people actually involved with them, and that is a good thing.
I would note that I don’t think that the current situation is entirely unproblematic (though for example the most active contributor to the website is somebody new, contrary to your characterization). Mainly due to there not being a mechanism for the OSMF to control what is deployed on the website it is legally responsible for, the numerous current compliance issues have been noted elsewhere*.
* ah yes this is a good example of why you don’t want unconstrained voting on features.
3 Likes
Richard
(Richard Fairhurst)
June 20, 2025, 6:54am
AngocA:
For a democratic project, there is surprisingly little democratic structure for making decisions about the software that powers it.
The (non-developer) community only gets to instruct the developers if the community is paying the developers. Otherwise the developers will do what they want. That’s how volunteering works.
(Of course, the trick is to ensure the developers want the same thing as the non-developers at which point this becomes moot. In theory that should be what OSMF does, but ¯\_(ツ)_/¯ )
16 Likes
amapanda_ᚐᚋᚐᚅᚇᚐ
(Amanda)
June 20, 2025, 9:19am
AngocA:
Many other parts of OSM have developed well-defined governance … tag approvals).
17 Likes
ZeLonewolf
(Brian M Sperlongano)
June 20, 2025, 10:19am
While there may be some good points in here, I stopped reading and gave the OP a
solely for it being an obvious AI-generated post, which I find so incredibly lazy and disrespectful to the time of community members in reading and responding to it. Please, present your own ideas. If language barrier is the issue, feel free to do so in your native language.
Well-defined governance… tag approvals… if there was any doubt that this was the work of AI and not genuine, human thought… well, there it is.
17 Likes
Proposal to add :robot: as a reaction to AI content
matheusgomesms
(Matheus Gomes)
June 20, 2025, 11:33am
A pity you gave up on reading the text, just because you feel AI-generated (which clearly is not, since no AI would randomly point out specific topics like Welcome Mat, which is something Andrés started working on that, or that members of OWG are also members of Core Software).
Then, you can argue that perhaps this was AI-improved, but again, what is the problem of using a tool to polish your writing?
Much better using
as the symbol of human intelligence for about anything.
3 Likes
SomeoneElse
(Andy Townsend)
June 20, 2025, 11:50am
matheusgomesms:
Then, you can argue that perhaps this was AI-improved, but again, what is the problem of using a tool to polish your writing?
Things to tidy up grammar and presentation have been around since basically forever, but whatever you used made it read like spam.
Use whatever you like, but the point of writing something is to communicate with people, and if what you are communicating is “I sound like a spammer” your communication will not have the desired effect.
12 Likes
matheusgomesms
(Matheus Gomes)
June 20, 2025, 11:54am
Not sure actually if this is intentional, but we have clear problems with OSM-tech:
outdated stack
few core people
lack of strategy
lack of governance
What if, out of sudden, all the “core” maintainers die (I hope not, just to be clear!)? What happens to OSM? Documentation is quite poor, the already mentioned outdated stack that just few people are still using makes it harder to get new contributors.
And even if no-one dies (hopefully!), how we can project the future of the project? It seems very few people are still risking contributing to the software, despite we having quite a geek community. Why is that? Why people feel more compelled to create an entire new system/piece/plugin/whatever than contributing to the OSM core?
Then, impossible to not mention the Foundation, that for some reason, has as a core vision to not directly deal with OSM tech, in a very techy project. How does that work?
I know it’s very hard to drive a volunteer-based project, but at the same time the lack of future/governance makes it hard to spend any time contributing to OSM-software (not to mention the gatekeeping word that the gatekeepers don’t like to hear about).
SimonPoole:
With the exception of OSM-NG I haven’t seen any and that is a non-starter as an actual replacement for numerous reasons that don’t need repeating here (but could have been avoided).
The dev who created better-osm script spoke in a very clear manner:
Better-osm-org: a script that adds useful little things to osm.org - #42 by TrickyFoxy
. He is basically saying: “I wanted to improve OSM website, I have good ideas (which seems a lot of people agree with him), but I can’t spend a single second of my life trying to convince maintainers to accept things that everyone wants, except them”.
SomeoneElse:
Use whatever you like, but the point of writing something is to communicate with people, and if what you are communicating is “I sound like a spammer” your communication will not have the desired effect.
Again the same behavior as in the Gustavo’s topic: the problem is how the message is conveyed, not the message itself.
Funny because in the nowadays life where everybody is too lazy to read, the LinkedIn lazy, bulleted-points, emoji-sectioned writing style, really fits this forum, where we have people all around the world, with different English and cognitive capacities to understand complex topics. I found a valid attempt to mitigate these issues. (the translation button is awful, I always have to go copy/paste to Google to understand what is written)
2 Likes
SomeoneElse
(Andy Townsend)
June 20, 2025, 12:09pm
matheusgomesms:
Again the same behavior as in the Gustavo’s topic: the problem is how the message is conveyed, not the message itself.
For the avoidance of doubt - I am trying to help you better communicate your message here. You can choose to ignore that and “shoot the messenger” (as you have done so far), but that will have the opposite effect to the one you desire.
Please, take a step back and look at what you have written. What were you trying to say? How could you have said it differently to communicate better and avoided the immediate negative reaction? If the way that you are saying something is preventing people from engaging with what you are saying, the way you are saying it is wrong.
10 Likes
Minh_Nguyen
(Minh Nguyễn)
June 20, 2025, 12:09pm
10
@AngocA
, you managed to encapsulate in a single post what the UN just devoted
an entire week
to exploring. Governance and maintainership issues are not unique to OSM. Acknowledging that can help us understand the issue better without making people get defensive about it. Unfortunately, I’m sleep-deprived and busy at a conference, or else I’d respond more fully to your line of questioning, but I’ll start with some of the more basic questions you asked to get those out of the way.
AngocA:
Who decides what becomes part of the “Core Software” of OpenStreetMap?
What are the inclusion criteria? Are they documented anywhere?
The important thing to understand is that “core software” has always been a very nebulous term, just like in many organizations and projects. After I took the Core Software Development Facilitator role, I realized the wiki didn’t even define this term, so I belatedly wrote up
a stub
that highlights some common perspectives about what to consider “core”.
So far, in this CSDF role, I’ve taken the approach of prioritizing my attention on some projects over others based on their proximity to day-to-day operations and some other factors like compliance. This has more to do with the mechanism through which the role is being funded than any single right definition of “core”. On occasion, people have asked me for help with more peripheral projects. My response is to try to connect them to people who are in a better position to help – as this community ought to do anyways.
Of course this really isn’t about me – what you’re talking about is about how we move forward as a whole community – but hopefully that gives you an additional data point for something that doesn’t have a clear answer.
AngocA:
What is the role of the OSMF, the Operations Working Group, or the Software Dispute Resolution Panel?
The OWG runs instances of several core software projects, among other things. As the name implies, they’re more focused on keeping things running smoothly than blue-sky feature development, which would more naturally involve the Engineering Working Group. The SDRP is charged with arbitrating disputes in specific software projects, not necessarily core projects. (Only iD is within its purview currently.)
The OSMF is the organization that convenes these working groups and panels. Or did you mean to ask about the OSMF board’s responsibilities?
Anyways, that’s how things are organized on paper. But we’re a big loose collective of projects, so sometimes things get done because people have ignored this structure and made their own arrangements.
AngocA:
Is there a process for proposing, evaluating, and incubating new projects, or for replacing outdated ones?
If you mean this in the context of defining “core software”, then no, the only process is to convince enough of the community to start referring to it as core software.
But it sounds like you’re asking more broadly about the software lifecycle for OSM software. Various parts of the OSM tech stack have gone through replacement cycles in the past, like Osmarender giving way to the Mapnik style, OSM Carto, and vector styles, or the old Gazetteer giving way to Nominatim. I’d love to hear from the folks involved with those transitions about the challenges they had to overcome both technically and socially, and any lessons they took away from that experience.
In the meantime, since website design has been top of mind in recent days, here’s
one perspective
from the people at Mapbox who worked on the last big redesign:
7 Likes
Jarek
June 20, 2025, 1:57pm
11
I do agree that we don’t usually get to tell volunteers what to do or how to run their projects.
But I also think that some parts of OSM software stack, potentially including parts maintained by volunteers, need to be subject to some oversight (“community gets to instruct the developers”). To give a purposefully absurd example, what if volunteers maintaining the OSM API decide to disallow changesets coming from a range of IP addresses, or from an IP address geolocated into a particular region - like, say, the entire United States of America?
Although this doesn’t appear to be the definition of “Core Software” currently in use, IMO this kind of oversight should extend to anything hosted on
openstreetmap.org
or loaded by default on
openstreetmap.org
, and to be explicit, yes, currently this includes iD and carto. This level of exposure on
openstreetmap.org
means that projects greatly influence the OpenStreetMap database and the experience that everyone has with OpenStreetMap, and thus IMO they are a core part of OpenStreetMap.
8 Likes
SimonPoole
(Simon Poole)
June 20, 2025, 4:10pm
12
matheusgomesms:
outdated stack
Constantly repeating a many times debunked trope doesn’t make it truer (and as I’ve pointed out many times too, I’m not even a particular fan of rails).
matheusgomesms:
few core people
True, but three, plus a number of people that could fix things in a sticky situation, is a lot better than many many other projects that do not have permanently employed developers.
The problem is that
osm.org
is not your run of the mill hobby website, it is used by a large number of people every day and has operational, legal, financial and support constraints that can’t just be ignored. For example that is the reason that as far as possible core functions shouldn’t depend on third party services or that are real compliance issues that will not simply go away if you ignore them.
matheusgomesms:
lack of strategy
Yes there is no strategy but then is that really a loss? In the end that is a thing that the current maintainers need to decide on. On the other hand I’ve long suggested that there should be a prominent statement of purpose for the website by the OSMF board for the OSMF instance to alleviate the burden and blame the devs get because it isn’t intended as a replacement for google maps.
matheusgomesms:
lack of governance
Again as long as they are not being paid by anybody it is obvious that the maintainers can organize themselves as they seem fit.
Yes the OSMF is a complete fail at governing what gets deployed on its website which is a very different question, but tell me news. Not to mention that you wouldn’t like the results if it was more effective, see below.
matheusgomesms:
He is basically saying: “I wanted to improve OSM website, I have good ideas (which seems a lot of people agree with him), but I can’t spend a single second of my life trying to convince maintainers to accept things that everyone wants, except them”.
He actually explains in that posting why he didn’t expect to have success with getting his ideas supported, and they are essentially those that I mentioned above. But if you avoid violating those constraints it is quite easy to get things merged.
BTW a more effective governance by the OSMF would be much more restrictive with what is deployed in its responsibility, think WMF.
4 Likes
Mateusz_Konieczny
(Mateusz Konieczny)
June 20, 2025, 9:24pm
13
AngocA:
For a democratic project, there is surprisingly little democratic structure for making decisions about the software that powers it.
why you claim that OSM is a democratic project? If there is a single word to describe overall system it is a doocracy.
Can you give examples of projects where “democratic structure for making decisions about the software that powers it” worked well?
AngocA:
Who decides whether a pull request is accepted or ignored?
On whether someone bothers to review it. See “doocracy” part.
AngocA:
Should “Core Software” projects be under the
openstreetmap
GitHub organization by default?
no, there is no reason to force Github on people
(there are some reason why migrating away at all costs from Github and ignoring tradeoffs is not a good strategy, but forcing people to use Github is not a good idea)
AngocA:
Should there be written expectations for the maintainers of these projects—including community responsiveness and documentation?
Only if you pay them appropriately or you want to kill this projects.
AngocA:
Community voice and inclusion
Structure of this post looks exactly like LLM-generated text using default stylistic settings. If it is not LLM generated I would suggest you to change styling in future.
If it is your text heavily rewritten by LLM I would advise you to ask LLM for suggestions, not for text rewrite.
(feel free to ignore this suggestion, but the feel is so strong that I feel like I am reading autogenerated text, and decided to stop reading further)
PS
How can the following text be improved? List changes, with a specific listing of what was changed and why.
is incantation that I sometimes use - and edit text manually, ignoring about 80% of what LLM generated.
If you just tell LLM to rewrite/reformat/improve text it will end with the same standard blob that people will easily recognize as LLM output, at least for typical use of LLMs like chatgpt.
11 Likes
aighes
(Henning)
June 20, 2025, 11:21pm
14
AngocA:
For a democratic project, there is surprisingly little democratic structure for making decisions about the software that powers it.
Where do you think OSM as a project is democratic? We map what’s there in reality and if there is a mapper who actually wants to add it, not what 51% of the mapper want.
Are you calling it democratic because 20 mappers vote on a proposal in the wiki? 20 out of a huge number of contributors.
The only thing I would consider democratic is the OSMF itself. But OSMF is not the project
If you have a cool idea for whatever, do it, showcase it, maintain it and it might be considered as a core software at some point.
8 Likes
Wynndale
(Wynndale)
June 21, 2025, 4:46pm
15
There have been multiple cases where software development has got interesting over the years.
Xapi was written in
Mumps
, which made finding a pool of Ruby developers easy by comparison. When its developer had differences with OSM someone hastily rewrote it in Java, leading to the Overpass API we enjoy today.
The one-time developers of iD became so notorious for their attitude on the issue tracker there was a page about it on the wiki. The Foundation eventually stepped in to pay a developer long term.
OSM-Carto has had a long history of declining issues and even offers of code. This has often led to double tagging; in cases such as busways where this isnʼt possible this has led to ugly rows. Because vector tiles are considered the long term solution it has been allowed to sit there with planning blight.
When the wiki implemented a Wikidata-style database to manage descriptions of attributes of tags across all the different languages that they are documented in, Taginfo declined to use it. This means that if wiki pages are written the most maintainable way Taginfo doesnʼt see anything. It also means that even if descriptions are duplicated, taglist tables, which could otherwise be an effective way to manage the persistent problem of generating complex wiki pages, are much less useful.
8 Likes
Jarek
June 21, 2025, 7:11pm
16
Mateusz_Konieczny:
why you claim that OSM is a democratic project?
Personally I understand OSMF board member elections to form a kind of representative democracy in OSM
OSMF website says it “supports but does not control” OSM, but ultimately the people who pay the server bills control OSM as it exists today
Mateusz_Konieczny
(Mateusz Konieczny)
June 21, 2025, 7:53pm
17
yes, elections to OSMF are democratic
but OSMF is only part of OSM, it does not make OSM a democratic project
And in case of OSMF board having enough time/courage/energy/weird hobby preferences to become candidate has typically far greater influence - we typically have only few candidates. I would say that it is still more doocracy (who even tries to be on board) than democracy (that selects from that small pool).
(not saying that it is good or bad, just that is the current situation)
3 Likes
osmuser63783
(Osmuser63783)
June 23, 2025, 5:39am
18
There’s an interesting question here. If the Board is formally in charge of the website, of convening the working groups and of funding some software development work, then it is through electing the Board that OSMF members can influence the direction of the website, the working groups, and software development.
Questions of software governance played next to no role in last year’s Board elections, so maybe they should in this year’s? Maybe we should be asking the candidates questions like:
Where do you see the Board’s role in the governance of OSM software projects, especially those hosted on OSMF infrastructure or featured on
OpenStreetMap.org
Do you agree with the OSMF Strategic Plan that “strengthening maintenance and feature development of core software” should be a core priority of the Board? What should the Board in your view be doing to achieve this?
What role, if any, do you think the Board should take in setting the strategic direction for the development of OSM software projects?
What is your vision for the future of the
openstreetmap.org
website and the iD editor?
6 Likes
SimonPoole
(Simon Poole)
June 23, 2025, 6:46am
19
osmuser63783:
… then it is through electing the Board that OSMF members can influence the direction of the website, the working groups, and software development.
That is jumping to conclusions, that is very much the same as saying, because the OSMF uses the WMFs mediawiki software for the OSM wiki, it can “influence the … software development” at the WMF.
For all practical purposes the rails port is an independent project that the OSMF has now and then supported the development of certain features for and just happens to operate its largest deployment, but it is nothing that the OSMF “owns” even in a wide sense of the word.
Any “control” would have to involve money, or at least control over what gets deployed on the OSM website.
2 Likes
aighes
(Henning)
June 23, 2025, 10:19pm
20
osmuser63783:
then it is through electing the Board that OSMF members can influence the direction of the website, the working groups, and software development.
That would be true for the small amount of paid software development the OSMF does. Which is quite limited to keep everything somehow running and develop vector tiles. Most of the software is only operated by OSMF but not developed/maintained by the OSMF. Like OSMcarto, Nominatim…
You can have 100% of the OSMF member vote for rendering busways, still OSMcarto devs might say
We don't care, fork it if you want to.
OSMF is funded by donations which might also be limited to certain goals. Like if you donate 50k for operating the server, OSMF members (and the board) can’t decide to take the money and beautify the website.
4 Likes
next page →