Towards a dedicated public issue tracking/project management system for OSM

Posted by imagico on 10 August 2018 in English (English).

We have various communication channels in OpenStreetMap being used for different needs in communication. The mailing lists and forum work reasonably well for free and open discourse of the community, changeset discussions allow communicating on specific edits in the map (and we have for example Pascal’s tool to look through these). We have the user diaries for people publishing their thoughts and experiences on the project and discussing them with others. And we have the OSM wiki which is used as a place to document things.

All of these have their issues and room for improvements but they are widely used and accepted as the platforms where communication happens. And they all have relatively low entry barriers as evidenced by the fact that quite a lot of people use them actively.

What we don’t have and where we have in OpenStreetmap a fairly obviously increasing need for is a means for project organization and related communication, task and issue tracking etc. There is a very old trac instance but this is hardly used any more and has a fairly awkward usability, in particular for non-programmers. Safe to say this is not an established communication platform any more.

Because of that people have started widely using external commercial platforms, in particular github, for this kind of work.

Specific examples:

  • corporations doing organized edits have github repositories to track their work - like here, here, here and here
  • import planning is frequently performed on github - like here and here
  • there are attempts to move tagging discussion to github issue trackers here
  • the OSMF and its working groups using github for issue tracking (both publicly and internally), public examples here and here

For OpenStreetMap this is not a good development for various reasons:

  • github is designed for software developers and is practically much less accessible for non-developers. Even if non-developers manage to adapt to this they will always feel less at home there and as a result there is an inherent dominance of the software developers over non-developers on github.
  • the requirement to register on an external platform and accept the terms of service there poses a highly problematic hurdle. It should always be the goal that an OSM community member should be able to participate in all public community discourse without such hurdles.
  • quite a few people have principal ethical concerns regarding platforms like github which are usually financed through either advertisement or sale of personal information about its users.
  • since the github software is not open source use of github is in conflict with the general culture of OpenStreetMap to base itself on open source technology.

Because of these problems i am generally inclined to boycott attempts to move non-development discussions to github. But this is somewhat difficult if you can’t point to a suitable alternative. I would therefore propose we set up an open source project management system that can be used by everyone with an OSM account for use by the OSM community. There are quite a few software products available for this.

There are various questions and arguments that might come up regarding this suggestion:

  • Do we really need this kind of tool in OSM? Yes, the fact that github is used so widely for OSM projects is a clear indicator.
  • But github is so convenient, everyone already knows how to use it while something else you would have to newly learn to use. Yes, for you that might apply - but you are putting the convenience of you and a few other people familiar and comfortable with github over the interests of the vast majority of mappers.
  • Why should the OSMF invest money and work into self hosting something when there are github alternatives based on open source software available for use that might offer affordable service plans for an organization like OSM? Mostly to ensure a low entry barrier for people to participate by requiring nothing more than an OSM account. If this could be achieved with an externally hosted tool and reliability of the service and access to and ownership of the database are ensured, external hosting would IMO also be an option.
  • Won’t this fragment the discourse in the OSM community by creating yet another set of communication channels you need to follow to stay informed? Yes, that is a possibility - but as said this is already happening through the use of github at the moment. I think a dedicated OSM platform would improve the situation on this matter.
  • Should this be a pure project management/issue tracking platform or also a source code repository and version management system? That’s a good question. Many of the free software options available offer both. But most software development projects around OpenStreetMap are independently managed and you can’t force any of them to move. The core arguments for not using github i listed do not necessarily apply to all of these projects. The main use case would at least initially be non-development projects. And therefore usability for non-developers should be a primary concern.
  • Great, but who does the work necessary to set this up? Ideally such a platform would be integrated into the existing OSM website with notifications via the OSM website messaging system, using the configured language settings and possibly connections to changeset discussions etc. That would be a lot of work to set up. But running it separately similar to the OSM Forum would already be a useful first step. This would require some work from operations to set this up and maintain it. But the more difficult steps are probably to come to a decision with wide support what we need in terms of features, what software should be chosen for this and to configure and adjust it for OpenStreetMap’s needs. This post is meant to start the discussion on these questions.

Comment from gileri on 13 August 2018 at 21:23

Nice writeup. I slightly disagree on the usability of Github, which I think is easy enough to pick up, even for non-programmers, but OSM would I believe benefit from using a self-hosted tool instead of depending on Github.

Comment from imagico on 13 August 2018 at 22:17

I know many non-programmer mappers who dislike being forced or urged to use github to participate in OSM related discussions. This might not always have identifiable reasons in usability - it might in parts simply be a psychological effect of visiting a place that is obviously primarily meant and optimized for programmers.

The way github presents and scores the users for example (with contributor activity information, repositories etc.) is clearly favoring developers. Objectively this is not all that meaningful but it still communicates certain priorities and preferences to people.

Comment from alesarrett on 1 September 2018 at 06:45

I would agree on setting up a git-based solution for discussing various issues related to OSM, but I disagree GitHub is the right choice, mainly because it is a proprietary solution (recently acquired by Microsoft OSM can’t risk to build on. In case a git-based solution is proposed, I strongly suggest to use GitLab ( that is an open source, highly developed and supported solution.

Comment from SimonPoole on 2 September 2018 at 08:46

I’ve discussed this with imagico before.


  • while it can argued that having a non-github like issue tracker would remove the real or imagined bias towards software developers, it is however unlikely that the devs would want to give up the tight integration of the issue tracker with git that github-like systems provide. Having two -different- public issue tracker systems (plus trac for JOSM and itrs internally) would be a step back from what we do today,
  • in the end gitlab has exactly the same issues as github and what we would want is something thta allows single sign on with osm,org. That would for example be possible with a self-hosted instance of gitea.

Comment from imagico on 2 September 2018 at 09:20

while it can argued that having a non-github like issue tracker would remove the real or imagined bias towards software developers, it is however unlikely that the devs would want to give up the tight integration of the issue tracker with git that github-like systems provide.

Which is why i specifically focused on projects that are not primarily software development projects - just look at the examples i gave.

Should we in the future at some point have such a system established it seems pretty likely to me that software developers would deliberately use it for purposes where they seek input from the broader OSM community and not just developers.

And i think even having one of the various open source github clones set up specifically for OSM use (with as you said single sign on) could already - through a different culture of use and nuances in configuration - be significantly more non-developer friendly.

Comment from kocio on 4 September 2018 at 15:24

I have no clear opinion on that, but this is right question to ask, thanks!

Comment from Władysław Komorek on 4 September 2018 at 16:13

I think it is the right way to reorganize certain issues and create a new approach, how we perceive OSM.

Comment from Mateusz Konieczny on 17 September 2018 at 08:46

The main problem here is an opportunity cost.

So, first of all, we need somebody who will set up something initially (either as a hobby or someone needs to raise funds for that or spend OSMF funds).

And later it would be necessary to maintain, upgrade and in general keep system working.

I am not sure whatever benefits are larger than drawbacks, though I am in a group that is most likely to not notice problems of Github being programmer-oriented.

I am also worried that it will end like trac or - effort was spend to set it up and it is quite clunky, with poor usability. I am not sure how likely is that this new issue tracking platform will be overall better than Github.

Comment from tordans on 22 September 2018 at 16:31

This comic sums it up pretty nicely:

IMO, the OSM community has more important projects that are closer to its core then replacing Github …

Comment from joost schouppe on 10 November 2018 at 10:58

I don’t agree with your assumption that github-style systems are uncomfortable for non-developpers. I can speak from personal experience, since I’m not a developper. The strength of the Github system is exactly the interface between the “code” and the users. But the code can just be a currated document, or even an Excel file! I have been introducing Github to a lot of people in the government sector, because it is the easiest way to introduce them to open source logics for collaborating openly accross organisations. Mind you that we’re not talking about git -at all-, we just use the web interface. With osmbe we just set up our first GitLab account recently, for practical, not philosophical reasons. The web interface lacks some of the sexyness of Github, but it might be good enough to be an alternative to us non-developpers. But is GitLab open enough?

Comment from imagico on 10 November 2018 at 11:29

I don’t agree with your assumption that github-style systems are uncomfortable for non-developpers.

  1. I have never said that, what i have said is that github is primarily designed for software developers so non-developers will therefore inevitably feel less at home there. This is not only a matter of how it formally presents itself to the user, it is also a matter of who else is present there. In a pub where 90 percent of the customers are men women will feel less at home even if the design and presentation of the establishment on its own is equally attractive for men and women.
  2. You should always be careful projecting you own personal experience as representative for the group you consider yourself part of.

Regarding use of ‘github clones’ like gitlab - i listed four reasons why widespread use of github for OSM matters is problematic. You can see for yourself which of these also apply for an alternative. The most likely point that could be solved by a github clone would be the ‘not open source software’ aspect.

Comment from PlaneMad on 30 November 2018 at 07:29

One option is to use phabricator which is the open source project management system used by Wikimedia Foundation and used by the community as well.

Its worth looking into how its used by WMF to get an idea of what could be possible on the platform:

Comment from Mateusz Konieczny on 30 November 2018 at 10:26

Is phabricator less programmer-centric? As far as I can say it is not less programmer centric and it is more complicated.

I think it would not solve the problem of “too software developer focused”.

It may solve “external login” and would solve two other mentioned issues but at cost of maintaining one more service and even more increased complexity of a platform.

Comment from imagico on 30 November 2018 at 10:51

My impression is that while github and github clones are in total fairly programmer centric (though you can argue about to what extent this applies to the issue tracker aspect of them alone) phabricator seems more management centric. Since Wikimedia has a much higher degree of centralized management than OSM you would need to think about to what extent it makes sense to use a tool designed specifically for that. In general i would prefer using software that is being developed for a wide audience and avoid tools that are developed for a specific application outside of OSM that is in the future not unlikely to develop in a very different direction.

But ultimately i think this should be about collecting the options we have, find out the pros and cons and then make an informed decision.

Login to leave a comment