SotM 2022 and the OSM data model

It’s two weeks after the State of the Map 2022 in Florence and its high time I write this blog post. I had a great time and can only agree with what Ilya said in the closing session, how great it was to see everyone again (or for the first time) after such a long dry spell (Ilya’s words were much better though).

I gave a talk about the OSM data model (slides) and how we could change it based on the on the study commissioned by the EWG. This generated a lot of interest and I spent many hours during the conference (and also afterwards at the FOSS4G) discussing many of the technical details and also a lot about the organisational issues.

I think that almost everybody liked the idea to move forward with something, nobody likes the stagnation we feel in the project. Not everybody agrees on what that something should be or what the priorities are, but most people seem to like the ideas I put into the paper in general. There are still lots of details to discuss, of course. The paper is intended to help start an informed discussion of those details, not to end it.

Most concerns I got were from people worried about the size of possible changes and how to implement them with minimal disruption for the wider community. I agree that this is the biggest problem. I believe that the changes an ordinary mapper will see are not that big, but the changes are, without question, large and intrusive for implementors, especially editor developers, but also for some “power users” of OSM data. And I don’t see a way to “cushion” the switch for everybody. For most data users who only consume planet files or extracts, I believe we can provide forwards and backwards compatible solutions. But for editors and users of the change feeds I don’t see how to do this in a technically feasible way. If there were, we wouldn’t need the switch in the first place. I am convinced that it is needed, and the sooner we do it the easier for everybody, but I do understand it if some people have a different view on this.

Many discussions at SotM were less about the technical details and more about the process. How best to implement a change process in the community? How do we decide what we want? Shouldn’t we start with some smaller changes to get things unstuck, learn how to do this and get some quick benefits to build upon? How do we pay for all of this? How do we make sure we get everybodies input? I am glad that we are having these discussion now and that EWG is stepping up to lead them.

The EWG will need a while to digest the paper and decide on the next steps. If you want to be part of the more organizational aspects, I encourage you to become a member of the EWG. All further questions and discussion about the data model and proposed (or not yet proposed) changes can happen on the osm-data-model repository for the time being. This is a good place to capture any ideas and concerns until the EWG decides to move forward with this project.