51–59 of 59 posts
Hahaha, do they have logs of who did this?
Yes, here is the changelog: https://www.openstreetmap.org/changeset/61555047?node_page=3
Kind of interesting to see all the other active blocks in effect and the corresponding explanatory notes.
It continues to amaze me that OpenStreetMap is not massively plagued by this (I know that MapBox uses OSM as one of many sources, as mentioned in the article, but the changed name did not appear in the live OSM map, afaik, so I am guessing it came from somewhere else). In the past, the main explanatory argument was always that OSM's complicated editing process was the gatekeeper, but I don't think this is true anymor…
It was spotted and fixed within 24 hours in OSM (about 30 days ago), that's the power of a community driven map. Mapbox don't keep their maps up to date, and probably used a snapshot of the data which happened to have gotten in the few hours before it was reverted.
Earlier quoted context omitted.
It was spotted and fixed within 24 hours in OSM (about 30 days ago), that's the power of a community driven map. Mapbox don't keep their maps up to date, and probably used a snapshot of the data which happened to have gotten in the few hours before it was reverted.
Do you happen to know the OSM relation ID that was affected? The "New York City" relation [1] does not have anything related in its history [2] [1] https://www.openstreetmap.org/relation/175905 [2] https://www.openstreetmap.org/api/0.6/relation/175905/histor...
Earlier quoted context omitted.
Do you happen to know the OSM relation ID that was affected? The "New York City" relation [1] does not have anything related in its history [2] [1] https://www.openstreetmap.org/relation/175905 [2] https://www.openstreetmap.org/api/0.6/relation/175905/histor...
I wouldn't be surprised if it was purged from the history. Otherwise might it imply that New York City was once given that name, historically?
The OSM history of an object shouldn't really be seen as the historical status of the thing.
Correct me if I'm missing something, but what's the usecase for user submitted data on the map? Do city names change so frequently that they need to be dynamically updated? I can see some use for it when you get to the neighbourhood level to account for local nicknames, but not full cities.
The entire map is user submitted! It's a wiki map.
OSM isn't finished (and never will be), so we need to keep adding to it. Even the place node for New York (which was edited here) has dozens of different tags as well as just the "place=city name=New York"
Earlier quoted context omitted.
It sounds like building a system on top of OpenStreetMap that is similar to Wikipedia's "protected pages" could be worthwhile. There are legitimate reasons to edit even the information for major world cities (most of them, for example, don't have translations into all major languages yet), but there's little reason to need those changes to be applied immediately.
It's ironic as this vandalism was spotted and reverted in less than 24 hours in OSM. So any improvements and community fixing of vandalism cannot be taken advantage of by Mapbox if they rely on a monthly snapshot.
Even better, less than 2 hours.
And this is why incorporating user submitted information is a tricky strategy. On one hand, there sure are a lot more users than employees. On the other hand, bad actors are often far more motivated and diligent than good faith actors, and they will try to screw over your system for no good reason.
That is true, but at the same time, Wikipedia wouldn't be possible without it. The CEO confirmed that their absuse-detection system immediately flagged the label, but apparently some human reviewer approved it. Which shows that even if your system is perfect, humans are always the weakest link.
OpenStreetMap wouldn't be possible without it either. :)
For OSM, humans are the strongest link, they are often the best way to fix these kinds of changes. A strong local community is the solution. OSM fixed this vandalism within 2 hours.
Correct me if I'm missing something, but what's the usecase for user submitted data on the map? Do city names change so frequently that they need to be dynamically updated? I can see some use for it when you get to the neighbourhood level to account for local nicknames, but not full cities.
Do city names change so frequently that they need to be dynamically updated? Maybe that's the bug. It didn't occur to me until I read your comment [0] that the system could simply accept user-submitted data only above a certain level of detail. It won't eliminate the damage, but it would keep it more contained. [0] OTOH, GIS isn't my day-to-day like it would be if I worked at, say, MapBox.
(after all, someone has to be able to edit things like the label for New York)
For instance, there's tens of millions of objects in OSM. Step one to locking some of them is deciding which ones. That alone is a big project.
We are just finishing building out our OSM infrastructure, and this morning in the shower I was thinking how nice it is that we can fix the maps by adding new construction and new roads (we do real estate for part of our state), and contribute this directly to OSM and then update tiles from OSM. Then I came in to read this story. Yesterday we were trying to figure out what our update strategy would be, if we would pu…
The vandalism was reverted in 2 hours over 20 days ago. If Mapbox had stayed current with the diffs we publish, the vandalism would have been gone in a blink.