Working through some stuff for chapter 3
So a few weeks ago I complained about trying to write about the knotty dynamics of open source and the open source/free software schism in the dissertation, and how I wasn’t sure how important it was. My advisor apparently thought it was very important and her feedback made me kind of go back to the drawing board to think through the chapter and what exactly I’m trying to say—because, as another committee member pointed out, what I had was mostly just “this happened, then this happened” without a clear analysis of the meaning or stakes. Maybe the fix here is just making the chapter shorter, since it’s currently mostly functioning as a buildup to the “2010s geo is ZIRP” chapter, but I don’t want to oversimplify this as merely a “2000s open source geo good, when they get VC money it’s bad and corrupted” story.
The boring question out of the way
One framing question that comes up but that I don’t really find that interesting is why hasn’t open source beaten Esri? And the answers there are kind of the same as “why hasn’t LibreOffice beaten Microsoft”, i.e. liability, procurement, and inertia. Also I don’t think beating Esri is really the point—like maybe some FOSS4G people explicitly want that, but I think they’re also kind of fine with their tools finding the niche they fit into. Also, Esri uses open source stuff plenty because government vendors ask them to.
Does geo have fewer jerks or do I just not tend to meet the jerks
This dissertation is not a deep ethnography of the open source geo world as a whole so I am wary of making broad statements, but from having been in and adjacent to this world through a lot of the 2010s, interviewing people, and reading mailing lists, my impression has been that the culture on project-to-project level is largely pretty good—e.g., people less likely to lose their shit over the premise of adding a code of conduct, taking diversity discussions seriously, welcoming new people and generally not being dicks on listservs. OpenStreetMap was the exception here vibes-wise for a long time, though it seems like it’s improved. So if there’s a stakes question here, one might be what makes open source geospatial less inclined to elevate jerks and/or reasonably functional in ways that other open source projects are sometimes not?
I think one possible reason for this is that a lot of the software people who get into geo stuff understand that the stakes and impacts of geo aren’t theoretical, and this makes them less likely to think about things like access to information or software in abstract absolute terms or make that their entire fucking personality. Schuyler Erle wanted geospatial data so he could start a community ISP. One of the first conversations I had with Paul Ramsey was about how much it frustrated him when he first did contract work for the British Columbia office of Indigenous Affairs that First Nations communities were expected to pay the government money for access to the geospatial data central to land treaty negotiations. GeoServer started as a weird long game towards “how do we do traffic studies to work towards a car-free New York City?” There’s a lot of open source software projects where the stakes back to real life can be easily abstracted away, but geospatial is always referencing back to the world. And a lot of the people in the 2000s scene doing open source geo software stuff were doing it in small to medium scale government or community settings. You can still be a dick about code minutiae when it’s geo code, but you have enough perspective about the broader picture of the world your code exists in that maybe you can get over yourself a little bit.
This isn’t to say that all of open source geo is on the side of justice and doing good in the physical world. Defense and intelligence absolutely have a stake in this stuff. The fact that this year’s FOSS4G North America conference is happening in Reston, Virginia (a major zone of defense contractor and intelligence agency offices) kind of gives the game away, no? And this is something that in general open source isn’t great with: the idea of restricting access to software on moral grounds is decried as a slippery slope and denying freedom (which: IDK man the intelligence agency’s capacity to deny freedom to people seems like a much bigger deal than not letting them use a geocoder).
Again, I don’t like making blanket statements on something that’s very not a monolith but I wonder if open source geospatial is maybe better able to make peace with the amoral applications of their software simply because it has always been thus with geospatial. The intertwining of maps and the state is just too deep to really expect otherwise. In my interviews there was generally an ambivalence to this setup but a sense of both “what can you do” and “if we can use this to do good stuff for people, the means justify the ends.” Selling the Open Geo Suite to the National Geospatial Intelligence Agency kept Open Geo financially sustainable, which meant that they could keep providing the software to other consumers. MetaCarta’s In-Q-Tel and Chevron VC money paid for OpenLayers. Maybe it’s easier to smooth away the friction around morally questionable uses of open source geospatial software when all geospatial work is tangled up with the morally questionable uses. (To be clear, I’m not saying this from some righteous moral high ground—I am literally a geographer, geographers are the most complicit bad guys of social sciences, even anthropologists have more backbone!)
But “maybe open source geo has less jerks because they can make peace with their tools being used for state murder” doesn’t feel like a coherent argument or something I could empirically demonstrate, and is a bit of a monolith sentiment.
The timing of open source geo?
Another question I could use to frame the chapter might be _why does open source geospatial thrive in the early 2000s?_I think the answer there is mostly low-key “the same reason lots of open source thrived in the early 2000s, which is the fucking Internet happens, genius.” Esri overtaking the GIS market is a little bit of a part of the story (nobody in the open source world that I interviewed really had a ton of love for Esri), but it doesn’t seem central to it in the way that something like LibreOffice exists in opposition to Microsoft Office. The Esri resentment was at times a second-order thing for the people I talked to—some of them didn’t know Esri or GIS even existed when they started getting into map stuff. There’s maybe something with 2006 being when a bunch of stuff kicks off and the Google Maps of it all, but that also feels part of the ZIRP buildup story.
Planet scaling and capitalist realism
Maybe the interesting thing is to think through how both geospatial tech and open source have a porousness that makes them both potentially really empowering and potentially really oppressive. Disentangling the complications of both to be less oppressive is perceived as really fucking hard, partly because of the dependency on companies and governments that do the oppressive stuff (and go on to sponsor or contribute to or monetize the open source projects).
I’ve been thinking about how the philosophies of free software and open source are a lens on what Mark Fisher called capitalist realism: neither really wants to see a world made otherwise, but they also can’t imagine it because their critique of capitalism is only raised at the level of intellectual property law. Theories of labor in open source exist but they don’t come out of the Open Source Initiative or the Free Software Foundation.
It feels weird to talk about geo as manifesting capitalist realism because geo describes more a set of techniques and methods, some of which predate capitalism. But there is a kind of learned helplessness to wanting to do geo for international development or social justice or whatever and being like, “the only way that I can do this is if Facebook and Amazon Logistics are also part of it” that maybe mirrors the dynamic of “well, The World Bank and free trade agreements are problematic but also simply How Things Are Done.” If the past year has demonstrated anything it’s that How Things Are Done really can be undone at a moment’s notice.
I think some of this has to do with scale. Not that geospatial can’t be used for oppression at lower scales (redlining was municipal), but planet scale generally is something that’s needed by actors who are interested in eating the planet and fucking the world (imperial nation states, extractive industries, car companies, platform capitalism). Think about the once-open source geo companies that steadily switched to closed licenses and more expensive subscription setups as they raised venture money. They also became providers of planet scale systems. (There’s maybe a joke here about how web Mercator really is at its least problematic at low scale; maybe planet-scale is fine if you only have z10 and higher.) This is one of the reasons it low-key stresses me out that the place where the most interesting cool open source innovation work is happening is Cloud Native Geospatial—tech that is, from jump, necessarily planet scale. Like it’s cool that a bunch of people are working on interoperability and performance and doing interesting things with complex remote sensing databases, but it seems pretty obvious who will reap the benefits of this work the most—and it’s not going to be conservationists or climate scientists.
Also: is “the commons” just “infrastructure” for romantics
I’m being pretty obviously flippant here but also, through a supply chain capitalism lens so much of the talk of “the commons” in open source and even like dated remix-culture discourse feels like the atheist technologist version of Wal-Mart’s invocation of Christian service rhetoric to make people feel OK about low wages and erratic scheduling. The digital commons of codebases and community art and personal images being turned into the foundation for Copilot and AI slop isn’t exactly enclosure, it’s somehow something worse? It’s like deliberately poisoning the soil of the commons and then telling people to keep growing their food there, don’t worry about the fact the potatoes have teeth or whatever.
One of my more salacious takes (but one backed up by etymology, kind of!) is that infrastructure is a word people often use when they’re too embarrassed to say “the means of production”, and it seems pretty obvious that the commons is or could be made a means of production—that’s why capital is so keen to enclosure it.
Geospatial data and technology for geospatial analysis are a weird kind of commons and/or infrastructure insofar as they’re a thing a lot of people need and benefit from but that very few everyday consumers want to pay a lot of money for—kind of because they’re accustomed to the upended model that Google Maps normalized around geo data, but also it’s not like Thomas Guides were crazy expensive back in the day. And doing geospatial well and thoroughly is expensive. People don’t like spending lots of tax dollars on a bridge, but they get that a bridge is going to be expensive to do well and want it done well because the stakes are high. They’re less inclined to see those stakes with a map, print or digital.
Other stuff: the deeply American-ness of this entire framing, how crown copyright and the not-public-domain-ness of geo data in other places transforms this whole premise, am I just being a cunt about these revisions because I’m still annoyed at my advisor (a little).