"Why OpenStreetMap is in Serious Trouble"

“Why OpenStreetMap is in Serious Trouble” blog post by free software/culture (including OSM) enthusiast/contributor Serge Wroclawski. The why includes both technical and business/organizational reasons. Please read the full post, but an extremely inadequate summary:

  • OSM Foundation very weak, few resources, technical decisions controlled by two sysadmin/developers
  • Project seen as producing dataset, not services; using (end user or API) https://www.openstreetmap.org not encouraged
  • Limited, unchanging API
  • Poor geocoding (place lookup)
  • Messy data/schema (no layers, ids can be resused)
  • Inadequate moderation/review
  • Inability to import government-produced datasets
  • 3D data is now the cutting edge and currently out of scope for OSM
  • Commercial entities selling access to services based on OSM data benefit from some OSM limitations above
  • Some important OSM contributors, perhaps including sysadmin/devs above, want to keep OSM hobbyist-scale

Note “alignment” of last two. Unfortunately Wroclawski doesn’t know how to get out of this trap, presumably not for lack of trying or experience (he has contributed for years as a mapper, developer, and advocate). I imagine the post was made out of extreme frustration. The post made the top of Hacker News. There Wroclawski (emacsen) responded to someone asking what to do with “Honestly, I have no idea” and on twitter:

I’ve been asked this a lot. I think the question is really “Can the culture of OSM change and address the issues I’ve talked about?” If it can, then these problems will solve themselves. If not, then maybe the lessons we learned in OSM need to be applied to a new project. No idea

I’m mentioning all this here because OSM would seem to have much more potential to be an exemplar of commons-based production, destroyer of proprietary rents, a concentrated interest for commons-favoring policy.

On production, it has both wildly succeeded, and as the post above discusses, is severely limited. I suspect that it has had some impact on proprietary geodata rents, but probably much smaller than possible, given its limitations. Has it destroyed any proprietary companies? I’d love to know more. As far as I know OSM hasn’t been active in policy simply because the foundation isn’t very active period, but some OSM aligned entities and OSM activists likely make an impact with less coordination. One of the commercial entities presumably alluded to, MapBox, does some policy advocacy.

Reading the post above I’m wondering whether OSM is much like other commons-based peer production projects that do not have a wealthy/strong institutional center: amazing production, still hard for some to believe, but with plenty of easy to identify limitations with solutions that are not only hard to execute, but hard to identify, meaning distribution is left to reusing companies, and both market and policy disruption relatively limited. For example, Debian.


Wow, this is a shock, I’d never given much thought to how the project was organized, and I thought it was similar to WMF, and that it was so big we has a national level chapter.

Scary to think how a commercial ecosystem could have a hold on a dataset like that. @mlinksva, do you know of any other mapping projects? I’ve been poking around OSM lately since not really being satisfied with how our local localwiki.org was working. I don’t mean to jump ship or anything, but maybe there are interesting projects that could train new mappers right now.

There probably are lots of similarities to the Wikimedia movement, including some strong chapters, but WMF is a strong central entity, OSMF is definitely not.

I doubt there is a better mapping project to contribute to. There have been OSM forks, at least one around licensing, but I don’t think they’ve gotten far. The commercial ecosystem (I note that https://localwiki.org/oakland/map/ uses MapBox) is probably 99% positive, except to the extent it demotivates fixing the problems with OSM (and to be fair, emacsen also cited HOT, or Humanitarian OpenStreetMap Team, a nonprofit, also as possibly a barrier to fixing bigger problems so it can get grants to fix smaller ones).

The parts about layers really concerned me. I thought the project was “healthier”, technically. And I thought importing data was par for the course, so those limitations are big flags for me:

  • Personally: I need geo data
  • Mapping in general: I really overestimated how well we were doing with OSM, and we really need to kick it into gear

I don’t need a map of Earth, but I would like to know about projects that encourage folks to do something like map their city, and allow it to be imported into different projects/systems. Everyone, drop some links if you find something like that. :slight_smile:


Mentions http://fosm.org/ and http://wikimapia.org/ (neither of which use TLS encryption…).

Yes, FOSM is the fork around licensing I was referring to. Though OSM’s ODbL + contributor terms is confusing and makes importing non-public domain data difficult (in theory FOSM is more flexible) I doubt that is what is holding OSM back – if it were a powerful enough attractor, presumably because it solved some or all of the problems emacsen noted, people would deal with the licensing. AFAICT FOSM isn’t all that active. https://wiki.openstreetmap.org/wiki/Fosm.org is probably the best place to click links and look for activity.

WikiMapia isn’t making a complete map, only describing objects on Google’s map, IIUC.

I imagine either somehow OSM’s limitations will be addressed, or someone will start a new commons map project structured technically and organizationally to take over the world, or we’ll be stuck with proprietary maps dominating for a very long time.