I don’t want to write about the angry OpenStreetMap men in my dissertation so I’m blogging a little instead
There’s a version of this dissertation that could just be about things that were happening in geospatial software over the last 25 years, focused almost exclusively on the web context. I ended up splitting what was supposed to be one chapter into two when I realized I had 20,000 words and wasn’t even close to done explaining the timeline.
And that’s with leaving a lot of stuff out. Leaving out all the people I didn’t interview because I ran out of time, even though they seemed great and game to chat. Leaving out all the stories from the oral history interviews that were absolutely hilarious or extremely telling and insightful but just didn’t flow in the text.
A topic I couldn’t exactly avoid but also am worried about getting mired in and keep kind of putting off really writing about at length: the history of conflicts between what Migurski calls the “craft mappers” and the companies who want to Do Stuff with/on OpenStreetMap. For me, thinking about this stuff is annoying because it means reading things like all the back-and-forth in the comments of Migurski’s post that I linked to, and it all takes itself so fucking seriously. I am sympathetic to the position of “automated map traces made with machine learning will be a coarser capture than ground truth” and at the same time: why are these guys in Mike’s comments talking about this like it’s an existential affront to them personally and not a disagreement in an open source project? Why were those guys so mean to Saman when they launched the iD editor (a story I didn’t even hear from Saman, two separate interviewees told me about that!)?
But I was able to do some good interviews with people involved in Overture, and Overture is kind of the culmination of two decades of norms changing in big corporate tech around open source, the consolidation of the mapping sector in ways nobody could have anticipated, and lots of people trying to figure out what, exactly, OpenStreetMap was or could be. I think it’s worth including in the dissertation, but I can’t really write about it without writing about OpenStreetMap.
I should caveat this with two things: first, I really don’t consider myself part of OpenStreetMap and part of my hesitation with writing too much about it in the dissertation is that I just don’t think I have the relevant community expertise. I interviewed a lot more OSGeo people than OSM people, which has some overlap but not 100%.
Second: I’m making pretty broad generalizations about OSM in focusing on that particular antagonistic pocket of the community, which seems to have chilled out a bit in recent years. Doing a little reading of recent discussions on both the OSM listservs and the Community discussion boards I feel much more confident in identifying the kind of guy that antagonizes Migurski is a minority of the overall “stuff going on in OSM” discourse. There’s plenty of other stuff to attend to that’s not whatever some Dutch guy is made about or whatever, and the question of whether or not big corporations will play a role in the project has basically been settled with baseline terms of engagement for “organized mappers” and the Foundation introducing corporate sponsorships.
Maybe this is really what annoys me about writing about this stuff in the dissertation. That particular kind of antagonistic guy is central to the story that Overture people tell about why they’ve basically forked OSM (well, not Prioleau, he’s very diplomatic in the way I think you have to be when you do business development for a very long time). But those guys who are such a point of contention are not the majority of OSM’s community or culture! So it’s a small number of guys trying to wield power over an open source project mad at another small number of guys (who work for giant corporations) trying to wield power over an open source project. This is absolutely my least favorite kind of thing to spend my time on (this is maybe why I’m such a baby about having to do theory shit because it loves to go into hair-splitting minutiae).
OpenStreetMap is sometimes described as “successful” in the sense that it has a lot of registered mappers with accounts—almost 10 million as of me writing this post—and its data quality is perceived to be pretty comprehensive, albeit not perfect. But there isn’t really consensus on what constitutes “success” for OpenStreetMap, insofar as different priorities and values produce different ideas about what the main goal of OpenStreetMap actually is.
The OpenStreetMap Foundation has a “non-exhaustive” list of its core values, which are as follows:
- The Best Map: We want to make the best map data set of the world
- Free and Open: Our data is available under a Free and Open licence to everyone
- Community: OSM is powered by its Community. Engage positively, be a good and respectful neighbour and assume good intent.
- Useful: We want OSM data to be used as widely as possible.
- Ground Truth: OSM favours objective personal knowledge and “Ground Truth” over all other sources
- Self motivated: OSM wants you to map the things you care about and will ensure that you have the freedom to do so.
There are some unspoken assumptions in here that make these values a little at odds with each other. What does “best” mean here? Is the “best” map the map made by a community or the map with the most precise data? If precision is the goal, wouldn’t a data import from a municipal GIS be the preferable option? And a map that has what the OSM wiki derisively calls “armchair mapped” data is pretty obviously more useful than having no map at all, which is what happens if someone can’t convince people to go do the ground truthing or maybe the area that needs ground truthing is a pedestrian-hostile zone.
Recent analyses of corporate mapping (what the OSMF calls “organized mapping”, which implies ground truthing is a chaos mode maybe) indicate that corporate mappers still makes up a relatively small fraction of total OSM edits worldwide, but their presence has grown a lot over a relatively short period of time. However, from a country/regional scale analysis in some places corporate mappers significantly outnumber local contributions, and even as corporate mapping remains the minority of total contributions, those contributions have grown a lot in a relatively short period of time. (This also assumes that all corporate mappers are identifying themselves at all times, which is a whole other thing—Amazon Logistics, for instance, has a page on the OSM Wiki but basically no users listed!)
So on a practical, technocratic level the issue with corporate mappers is they don’t bother to learn best practices (as seen in this osm-talk discussion about edits made by Amazon Logistics mappers). On a values level, the issue with corporate mappers would be their lack of community-level engagement with space.
My partner is not a Geo Person and explaining this stuff to him is always very illuminating. “So OSM is Wikipedia for maps and a cartel of private companies are…copying Wikipedia for maps and making their own,” was his summary of Overture, and yeah when you put it like that it sounds very bad. But where a corporate-controlled Wikipedia might revise common knowledge to suit some business agenda (eg removing the “Controversies” sections of various pages), corporations that use geo data generally really need it to be a reasonably accurate reflection of physical reality because like, they have to use it to deliver packages or decide where to put data centers or help some random consumer drive to a place. (That doesn’t mean weird shit isn’t likely to happen—remember the urban legends of Google removing its data centers from aerial imagery layers on Google Maps?—but I think that most of the too big to fail firms have reached the point where they have enough power to not need to have secret buildings). This is sort of what hinders GeoAI stuff, too: you really can’t have a system with a Wrong Mode labeling the one-way streets and getting it wrong. I’m not really worried about Overture yet, is what I’m saying, but I am worried by the state of open source and open data stuff being increasingly captured and dominated by big tech.
The argument I’m increasingly hovering around with trying to write in OSM and Overture is that a lot of the shitty internecine conflict that makes me not want to engage with this stuff fundamentally comes back to the political immaturity of FOSS in which “anyone can use/participate” really means anyone, including oligarchs and monopoly vendors. Basically, FOSS isn’t a very good framework for talking about power and sort of just assumes the “rising tide lifts all boats” strategy even though some of the boats are megayachts that are going to probably crush the little flotillas.
To be fair: OpenStreetMap slightly predates Google Maps, and it really doesn’t seem like it set out to be the only competing alternative to limited monopoly vendors—I think it just kind of wanted to compete with the Ordnance Survey. It’s weird to have your hobby turn into load-bearing infrastructure for global behemoths!
But FOSS also doesn’t really provide good mechanisms for dealing with those tensions. The craft mapper contingent could have pushed for a change to the license or to contribution guidelines that basically said “no imports or remote mapping”—and I don’t know, maybe someone did and they lost that fight, but if they had it would have sparked a whole debate over whether OSM was actually “open” and all the ways it would hold the project back. In the absence of a culture of setting restrictions on who can use a so-called “open” tool, the primary means available to people in an open source project to deter certain actors is basically being a total dick on the listserv. I’m not saying that’s the thought process any particular antagonistic person went through, I’m saying that people tend to lash out in situations where they want to control an outcome but can’t.
I don’t know, I might just leave this stuff out entirely in the interest of “a good dissertation is a done dissertation” and work out some of this stuff in a paper or something later. I feel bad that all the writing has meant I’ve been neglecting the more public sharing of stuff—I will try to put some post-mortem stuff here as I go/turn in chapters.