Thanks for this update! Do you happen to know if there’s still a tornado siren on the roof? osm.org/node/6719056946 I suppose it’s possible it got replaced by a modern, pole-mounted siren nearby.

Any idea which Golden Gate Transit bus route got conflated with U.S. 101? I’m afraid I’m not familiar enough with that system to know what the stops should correspond to.

This changeset somehow replaced the route relations for northbound and southbound U.S. Route 101 between Los Angeles and Hopland (71162 and 108619) with the members of a bus route relation, presumably for Golden Gate Transit. The relations were then deleted as tagging mistakes in changeset 98079317. The relations have been restored in changeset 108175328 and unrepurposed in changeset 108175438.

These relations were somehow broken in changeset 93689318, but the correct fix would’ve been to revert that changeset rather than delete the relations, which represented U.S. Route 101. The relations have been restored in changeset 108175328 and were unrepurposed in changeset 108175438.

Documentation and consensus is the standard for bulk additions of data into OSM, not for backing out undiscussed bulk edits. Large-scale edits have been summarily reverted countless times in the past, not necessarily by anyone in a position of authority. What sets Coke County apart is only that it took a couple years for anyone to get around to it.

This does not need to be an adversarial situation. It could be an opportunity to make improvements to the data processing and redo the import while engaging with the broader mapper community according to the import guidelines. It’s quite rare to have the opportunity to cleanly refine and update an import years after the initial import.

Granted, it’s more work than you may have bargained for, but as someone who recently voluntarily followed the import guidelines for something that wasn’t even technically an import, I’d say it does lead to better results long-term.

Additionally, the surface street that runs parallel to the freeway has a speed limit of 30 mph, not 65 mph, even it also happens to be named Guadalupe Parkway.

Changesets 106625261 and 106625786 correct the speed limits along these two roads.

Changeset 106625514 changes the speed limits back to mph, for consistency with the signs posted along the tracks. I’m unaware of any light rail system in the U.S. that posts speed limits in km/h.

Speed limits measured in miles per hour must have the suffix “mph”; otherwise, they’re interpreted as kilometers per hour.

(Coming here from a brief discussion in OSMUS Slack.) The discussions about COGs on that wiki talk page may not be 100% applicable to this case, since the MPO function is only one of Metro’s functions. Unlike every other COG or MPO, Metro has at the very least an electoral boundary.

That said, I’d argue that Metro still fits neatly within the definition of a special district. Special districts defy boundary=administrative’s rigid admin_level hierarchy. For example, according to https://www.oregonmetro.gov/sites/default/files/2020/02/19/JurisdictionRegional.pdf, the Metro boundary has a complex relationship with the Hillsboro and Tigard city boundaries. Such complex relationships are also common with admin_level=7 townships in some Midwestern states, but they are the rule rather than the exception when it comes to special districts.

You’re right that boundary=administrative isn’t strictly limited to “full governments”, but data consumers have come to expect boundary=administrative relations to represent the principal political subdivisions of the containing boundary=administrative. While one may consider a given location in Portland to lie within the Metro boundary for some purposes, it would be a stretch to say that Portland is subordinate to Metro in general.

There have been some experiments with tagging other kinds of special districts, such as osm.org/relation/9992727 and osm.org/relation/6534979 for school districts, but I’d hope it isn’t common to overload boundary=administrative for this purpose. I agree that boundary=MPO is an inadequate tag for this boundary. Since Metro serves multiple distinct purposes, perhaps it’s time to coin a new boundary=special_district tag. We could probably find enough existing special district experiments in OSM to justify a wiki page about them.

Also, you probably noticed that the access road was replaced by one just a couple yards over to the east. In the future, please consider adding the replacement road. After all, a missing road is worse than a misaligned one.

OK, changeset 105516179 replaces junction=roundabout with junction=circular, which is more general, but keeps the circular way to retain that level of detail.

Hi, please don’t replace properly mapped roundabouts (even minor traffic calming roundabouts) with the traffic_calming=island tag. Oversimplifying the roundabout creates ambiguity about the relationship of the roadway to the objects inside the island, such as osm.org/node/6511629050 .

Hi, these stubs are already built, not just proposed. It’s important to map these stubs as one-ways; otherwise, when a driver hears a voice instruction for this roundabout, the exit count will be off by one. This situation occurs frequently in areas that are being built in phases, so I’d suggest sharing this information with your colleagues. In the meantime, I reverted these changes in changeset 105110930. Thank you for your attention to this matter.

This was based on the SCCProperty2 layer tagged on this changeset. Normally when I encounter a building with individual units in that layer, I add points for each of the units and then tag some of them with the information from the SDP dataset. That way we aren’t left with a false sense of having completed the per-unit mapping of the building.

In this case, it does look pretty messy. I wonder if the layer incorporated a combination of different data sources, none of which accurately positioned the units within the building. Since the data is suspect, we can remove it in favor of a proper survey.

This changeset incorrectly changed San Francisco into a speed limit enforcement relation. I changed San Francisco back into an administrative boundary in changeset 103293987, but if you had intended to add a speed limit enforcement relation for N. Five Points Rd., you may want to go back in and try again with a brand new relation. Let me know if you need help with that.

That was the general sentiment a long time ago, when we were daunted by the challenge of mapping every road and meanwhile found it annoying that imported TIGER roads often misclassified driveways as through roads. (If I’m not mistaken, at the time, the map legend in the right sidebar even said that the building polygons were only for “important” buildings.) These days, OpenStreetMap has evolved to consider any driveway and building important enough for inclusion in the database.

You can use http://revert.osmz.ru/ to revert this changeset before someone comes along and creates a conflicting edit (which most likely would recreate some of the driveways). If there’s already a conflicting edit by the time you get to it, just ask and one of us can help you out with it.

Years-old mailing list threads can still be useful for context on perennial debates in OpenStreetMap, but it’s best to take them with a grain of salt and ask around for second opinions. You can subscribe to that mailing list at https://lists.openstreetmap.org/listinfo/talk-us/ or ask the larger U.S. community in OSMUS Slack (https://slack.openstreetmap.us/ ).

Fixed in changeset 8195579319.

Campos is the owner’s last name, not the Spanish word for “camp site”.

This changeset overwrote intentional differences between Wikidata labels (which are optimized for page titles) and OSM name tags (which are optimized for map labels and, in some cases, place more weight on on-the-ground usage). Some of these changes have since been undone, but the rest of the changes may need review – similar to what happened to the place POI and boundary for Washington, D.C.

Some of the addresses in this batch wound up getting assigned to garages or carports instead of houses.