Mikel Maron reports on an interesting development in the OpenStreetMap community of volunteer mappers — the Russian OSM community is debating whether Russian military installations should be removed from the OSM dataset, on security grounds.

It is illegal to do mapping without a permit while in Russia, so it is understandable if Russian OSM contributors refrain from mapping sensitive sites, such as military bases. But the current debate goes further — should the local OSM mapping community in Russia have the final say on what is and isn’t included on the OSM map of Russia? Should it be able to remove content contributed by others to the Russian map in the global OSM dataset if such content is deemed a threat to Russia’s national security?

More broadly, does the OSM’s operational principle of subsidiarity — where decisions are made at the most local level possible — trump the right of anyone to contribute information to the map?

There should be no surprise as to which side of the debate I find myself on. But the arguments are not as clear-cut as they might appear at first glance. Mikel posits an interesting edge case for censorship:

Legitimate cases to me is only information that puts dis-empowered people at risk, such as refugee routes along the Burmese-Indian border. But does this mean I would remove those tracks if someone added to OpenStreetMap? This and other questions remain to be answered.

Here are some arguments against the idea of letting local communities “own” their maps, and against the notion of any censorship in OSM. In no particular order:

Maps of bases are not the vulnerable points in the defense of the motherland:

Russia has indeed seen attacks by militants from the Caucasus region, but on civilian targets, such as the Moscow subway. Yet nobody argues that the proper response is to remove Moscow’s subway stations from the map. Meanwhile, all of the world’s militaries with an interest in Russia know exactly where Russia’s bases are.

OSM’s competitiveness depends on its completeness:

Censorship would make OSM’s dataset less competitive. Google Earth Community contains a exhaustive dataset of military bases, including Russian ones, and these are in no danger of ever being deleted at the behest of local opposition. The only recourse for those favoring censorship would be to try to introduce false information into the GEC dataset, but since such military data is curated by dedicated enthusiasts, this would be hard to pull off.

OSM’s local communities are small, and hence vulnerable to infiltration:

In China, an army of paid astroturfers ensures that the government’s position gets prime place in any online debate it designates as important. This is a effective strategy in small communities, which can be infiltrated by nationalist-sympathizers whose main priority is getting outcomes amenable to the government.

Maps are not owned by the regions they depict:

I mention this at the risk of stating the obvious. Russia may have sovereignty over its territory, but does not have sovereignty over maps of its territory, if these maps are made and kept outside Russia. If this were not the case, via the principle of reciprocity other countries would be free to claim jurisdiction over maps of their territories kept in Russia, and I am sure this is not something Russia would be willing to accede to. It is strange, therefore, that a disparate band of local Russian mappers would try to posit jurisdictional powers that not even the Russian government would pursue.

As for Mikel’s edge case: I see where he is coming from, but would argue that if somebody does want to publish such information online, then OSM is but one of many possible channels. In this case, it is better to know that the information is public and what it contains, so that it can be acted upon defensively (for example by changing the actual Burmese refugee routes). My own response would be to contact the people at risk (if I sympathize with their movement) to alert them that this information is in the public domain.

A course of action I would not condone is the adding of false routes, to make the data less reliable. Injecting false information into a dataset pollutes all the information, and undercuts the reputation of the dataset as a whole.

One edge case that I think is even more difficult than Burmese refugee routes is the mapping of vulnerable historical landmarks, prone to looting. Unlike a route, such historical landmarks cannot easily be moved, so once the information is out in the open, it will long remain relevant.

This debate has been held before, notably by Sweden’s Office of Antiquities (RAA), which four years ago decided that the benefits of a comprehensive public geodatabase of all historical monuments in Sweden outweighed the risks of looting at unprotected sites, above all because crowdsourced labeling of these sites will occur anyway — so you might as well make the best possible geodatabase available.

In sum: if an archaeologist is in possession of the only geodatabase for a certain excavation site prone to looting, then perhaps it is best to keep it secret, but if the information ever enters the public domain, then the right thing to do is to maximise the benefits of having this data be publicly available by making sure the dataset is as good as possible.