Ref is for sign posted reference numbers, not for things like making markers for old highways appear on the map. Are there signs posted for "OLD M-76"?

Also, you continue to change business routes away from pretty well established, widely used tagging. Please stop it.

In this case it was from a multipolygon relation with amenity+shop tags over both building ways (with building tags on the ways) to two nodes, one for the amenity and one for the shop.

I think the nodes, which are essentially an abstract statement "the thing is here" are a better representation of the situation because they are less specific than the relation. The convenience store building is not part of the OSM amenity, just closely associated with it. And then the other way around for the fuel. The multipolygon is too strong a specification for the modeled features.

I've been changing some stations form way to node in the area also, though I believe in most of those cases the tags on the way were very outdated (so some vague justification based on practice changing since the oldish previous edits and the node making the shop/amenity more visible in editors).

I'm aware that it isn't well settled modeling and am fine adjusting the tagging of the existing modeling if you'd prefer.

There are more than 'few or no intersecting roadways' (like quite a few per mile...) and there are multiple traffic control devices and there is no restriction on non-motorized traffic. By your own definition it doesn't qualify as an expressway.

And then there is the part where length and actual importance should be given more consideration than being divided. A short stretch connecting nothing is just overbuilt, it isn't more important or different than the adjacent portions of US 2.

There's 2 sensible definitions for trunk. For one of them, US 2 is a trunk from Saint Ignace until at least Duluth.

In the other one, an 8 mile section with a small village on one end is not a trunk.

Every 8 months someone sees that it is divided and makes it a trunk. But there's stop light and grade intersections and it isn't connecting significant populations.

Would it be reasonable to express the heights of osm.org/way/445920015 and osm.org/way/445920014 to a centimeter (2 digits after the decimal)?

Likes like some 'ref' values ended up in 'height' tags here. http://overpass-turbo.eu/s/199z

A significant additional issue is that there are hundreds of ways that were not deleted along with the relation. They are not of any value without the relation and should have been deleted along with it.

I've started cleaning them up a bit.

I'm not overly concerned about it, but I used a node for the fuel because in the abstract, I think it's a more sensible model for the situation than tagging the canopy. The pumps aren't up in the air, and they aren't really inside it either, they are just under it. And it is a bit tidier to use a single node when there are two canopies (vs tagging each canopy as fuel, or tagging one as fuel and the other not).

The pieces should be tagged building:part and not building.

building:part=school is a valid tag.

(I was looking as I replied to your comment).

I don't understand why Tower Hill Road would be motorway here. It's not got a high speed limit and it is such a short section that it doesn't really stand on its own, it's conceptually part of the trunk much more than it's conceptually part of the RI 138 that heads east (most of the connectors are pretty severe departures from direction of travel).

This was just a single case that popped out of an analysis I ran https://github.com/maxerickson/powwordchecker

(the analysis is pretty straightforward, just looks for uncommon word use in name vs the value in the religion tag)

Another batch of minor roads upgraded to secondary here. I don't understand the reasoning. It looks like it was done simply because they are state routes, which is not a good reason.

Why did you upgrade these minor roads to secondary?

Hello!

How come you are changing service roads to highway=residential?

Also, Bing is not really the best aligned imagery in Michigan. It's good to either adjust the alignment to the existing data or check if one of the other imagery sources has the elements you want to map and better alignment with the existing data.

Also, it would be preferable to explain where the information inside of '.json data file' came from. People often put 'gpx file' as a source, but that implies that they are using a gps recording, whereas .json data file is a general purpose container.

Max

The Maxar imagery is obviously not well aligned here, please evaluate alignment before making lots of changes based on a given imagery layer. It looks like other Amazon Logistics editors in the area have done the same thing.

Just a heads up, this edit broke the boundary relation for Battle Creek.

The children's hospital is located on the Hartford Hospital campus, but I think including it inside the area for the hospital takes it a step further and implies that it is part of the hospital.

This sort of thing isn't well represented in OSM yet (the hospital and campus are distinct entities in reality and the same object in OSM).

The sounding is for osm.org/changeset/83093063#map=17/44.39898/-84.29277 and doesn't apply to this lake. Sorry.

If you look at the layer USGS Topographic Maps and the Grass Lake sounding on https://www.michigan.gov/dnr/0,4570,7-350-79119_79146_81198_85509_85510-67603--,00.html it isn't particularly obvious that Grass Lake exists anymore. That GNIS has an entry for it and that Wikidata created an entry for the GNIS entry isn't terribly compelling if it is the only reference.

The locals don't mark the lower body any differently: http://www.lakeogemaw.net/about_us.htm

Is there more information that it is an independent feature?

Thanks,

Max

I disagree with this. US 2 is either trunk from the bridge to somewhere west of the UP or it's not trunk anywhere in the UP.

The better use of trunk is to mark the most significant routes through regions.